Ziele Untersuchung
mit Columbo Integrität von
Datenbanken Interaktion und
Portierbarkeit Ergonomie der
Schnittstellen

Angebot Produkte Projekt Beratung

Mittel Analytik Modellierung Sprachen Algebra Logik Hardware Denken Kreativität

Zusammenhänge Gesellschaft Wirtschaft Branche Firma


products/Sources/formale Sprachen/GAP/pkg/ctbllib/doc2/ (Algebra von RWTH Aachen Version 4.15.1^©) Datei vom 25.4.2025 mit Größe 597 kB

Quelle chap11.html

Sprache: HTML

products/Sources/formale Sprachen/GAP/pkg/ctbllib/doc2/chap11.html

<?xml version="1.0" encoding="UTF-8"?>

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
 "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">

<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en">
<head>
<title>GAP (CTblLibXpls) - Chapter 11: GAP Computations Concerning Probabilistic Generation of Finite
Simple Groups</title>
<meta http-equiv="content-type" content="text/html; charset=UTF-8" />
<meta name="generator" content="GAPDoc2HTML" />
<link rel="stylesheet" type="text/css" href="manual.css" />
<script src="manual.js" type="text/javascript"></script>
<script type="text/javascript">overwriteStyle();</script>
</head>
<body class="chap11" onload="jscontent()">

<div class="chlinktop">Goto Chapter: <a href="chap0.html">Top</a> <a href="chap1.html">1</a> <a href="chap2.html">2</a> <a href="chap3.html">3</a> <a href="chap4.html">4</a> <a href="chap5.html">5</a> <a href="chap6.html">6</a> <a href="chap7.html">7</a> <a href="chap8.html">8</a> <a href="chap9.html">9</a> <a href="chap10.html">10</a> <a href="chap11.html">11</a> <a href="chapBib.html">Bib</a> <a href="chapInd.html">Ind</a> </div>

<div class="chlinkprevnexttop"> <a href="chap0.html">[Top of Book]</a> <a href="chap0.html#contents">[Contents]</a> <a href="chap10.html">[Previous Chapter]</a> <a href="chapBib.html">[Next Chapter]</a> </div>

<a href="chap11_mj.html">[MathJax on]</a>
<a id="X7BE9906583D0FCEC" name="X7BE9906583D0FCEC"></a>
<div class="ChapSects"><a href="chap11.html#X7BE9906583D0FCEC">11 GAP Computations Concerning Probabilistic Generation of Finite
Simple Groups</a>
<div class="ContSect"> <a href="chap11.html#X8389AD927B74BA4A">11.1 Overview</a>

</div>
<div class="ContSect"> <a href="chap11.html#X7B4649CF7B7CFAA1">11.2 Prerequisites</a>

<div class="ContSSBlock">
 <a href="chap11.html#X7B6AEBDF7B857E2E">11.2-1 Theoretical Background</a>

 <a href="chap11.html#X79D7312484E78274">11.2-2 Computational Criteria</a>

</div></div>
<div class="ContSect"> <a href="chap11.html#X7B56BE5384BAD54E">11.3 GAP Functions for the Computations</a>

<div class="ContSSBlock">
 <a href="chap11.html#X806328747D1D4ECC">11.3-1 General Utilities</a>

 <a href="chap11.html#X7A221012861440E2">11.3-2 Character-Theoretic Computations</a>

 <a href="chap11.html#X83DACCF07EF62FAE">11.3-3 Computations with Groups</a>

</div></div>
<div class="ContSect"> <a href="chap11.html#X7A221012861440E2">11.4 Character-Theoretic Computations</a>

<div class="ContSSBlock">
 <a href="chap11.html#X86CE51E180A3D4ED">11.4-1 Sporadic Simple Groups</a>

 <a href="chap11.html#X821778BC7D665AB4">11.4-2 No longer necessary computations for the Baby Monster and the Monster</a>

 <a href="chap11.html#X84E9D10F80A74A53">11.4-3 Automorphism Groups of Sporadic Simple Groups</a>

 <a href="chap11.html#X80DA58F187CDCF5F">11.4-4 Other Simple Groups – Easy Cases</a>

 <a href="chap11.html#X7B1E26D586337487">11.4-5 Automorphism Groups of other Simple Groups – Easy Cases</a>

 <a href="chap11.html#X78B856907ED13545">11.4-6 O_8^-(3)</a>

 <a href="chap11.html#X84AB334886DCA746">11.4-7 O_10^+(2)</a>

 <a href="chap11.html#X84E3E4837BB93977">11.4-8 O_10^-(2)</a>

 <a href="chap11.html#X8307367E7C7C3BCE">11.4-9 O_12^+(2)</a>

 <a href="chap11.html#X834FE1B58119A5FF">11.4-10 O_12^-(2)</a>

 <a href="chap11.html#X7C5980A385C088FA">11.4-11 S_6(4)</a>

 <a href="chap11.html#X829EDF7F7C0BCB8E">11.4-12 ∗ S_6(5)</a>

 <a href="chap11.html#X85162B297E4B67EB">11.4-13 S_8(3)</a>

 <a href="chap11.html#X8495C2BF7B6EFFEF">11.4-14 U_4(4)</a>

 <a href="chap11.html#X7A3BB5AA83A2BDF3">11.4-15 U_6(2)</a>

</div></div>
<div class="ContSect"> <a href="chap11.html#X8237B8617D6F6027">11.5 Computations using Groups</a>

<div class="ContSSBlock">
 <a href="chap11.html#X815320787B601000">11.5-1 A_2m+1, 2 ≤ m ≤ 11</a>

 <a href="chap11.html#X7B5321337B28100B">11.5-2 A_5</a>

 <a href="chap11.html#X82C3B4287B0C7BEE">11.5-3 A_6</a>

 <a href="chap11.html#X85B3C7217B105D4D">11.5-4 A_7</a>

 <a href="chap11.html#X84EA645A82E2BAFB">11.5-5 L_d(q)</a>

 <a href="chap11.html#X855460BE787188B9">11.5-6 ∗ L_d(q) with prime d</a>

 <a href="chap11.html#X7EA88CEF81962F3F">11.5-7 Automorphic Extensions of L_d(q)</a>

 <a href="chap11.html#X7C8806DB8588BB51">11.5-8 L_3(2)</a>

 <a href="chap11.html#X7B7061917ED3714D">11.5-9 M_11</a>

 <a href="chap11.html#X82E0F48A7FF82BB3">11.5-10 M_12</a>

 <a href="chap11.html#X7FF2E8F27FBEB65C">11.5-11 O_7(3)</a>

 <a href="chap11.html#X7F80F2527C424AA4">11.5-12 O_8^+(2)</a>

 <a href="chap11.html#X78F0815B86253A1F">11.5-13 O_8^+(3)</a>

 <a href="chap11.html#X85BACC4A83F73392">11.5-14 O^+_8(4)</a>

 <a href="chap11.html#X86EC26F78609618E">11.5-15 ∗ O_9(3)</a>

 <a href="chap11.html#X8393978A8773997E">11.5-16 O_10^-(3)</a>

 <a href="chap11.html#X7BBBEEEF834F1002">11.5-17 O_14^-(2)</a>

 <a href="chap11.html#X8477457780B69BC7">11.5-18 O_12^+(3)</a>

 <a href="chap11.html#X854D85F287767342">11.5-19 ∗ S_4(8)</a>

 <a href="chap11.html#X82CFBAF07D3487A0">11.5-20 S_6(2)</a>

 <a href="chap11.html#X826658207D9D6570">11.5-21 S_8(2)</a>

 <a href="chap11.html#X82A6496887F80843">11.5-22 ∗ S_10(2)</a>

 <a href="chap11.html#X7A03F8EC839AF0B5">11.5-23 U_4(2)</a>

 <a href="chap11.html#X7D738BE5804CF22E">11.5-24 U_4(3)</a>

 <a href="chap11.html#X7D4BC6A38074BF68">11.5-25 U_6(3)</a>

 <a href="chap11.html#X7A92577A830B5F23">11.5-26 U_8(2)</a>

</div></div>
</div>

<h3>11 GAP Computations Concerning Probabilistic Generation of Finite
Simple Groups</h3>

Date: March 28th, 2012

This is a collection of examples showing how the GAP system <a href="chapBib.html#biBGAP">[GAP24]</a> can be used to compute information about the probabilistic generation of finite almost simple groups. It includes all examples that were needed for the computational results in <a href="chapBib.html#biBBGK">[BGK08]</a>.

The purpose of this writeup is twofold. On the one hand, the computations are documented this way. On the other hand, the GAP code shown for the examples can be used as test input for automatic checking of the data and the functions used.

A first version of this document, which was based on GAP 4.4.10, had been accessible in the web since April 2006 and is available in the arXiv (no. 0710.3267) since October 2007. The differences between that document and the current version are as follows.

<ul>
<li>The format of the GAP output was adjusted to the changed behaviour of GAP until version 4.10. This affects mainly the way how GAP records are printed.

</li>
<li>Several computations are now easier because more character tables of almost simple groups and maximal subgroups of such groups are available in the GAP Character Table Library. (The more involved computations from the original version have been kept in the file.)

</li>
<li>The computation of all conjugacy classes of a subgroup of PΩ^+(12,3) has been replaced by the computation of the conjugacy classes of elements of prime order in this subgroup.

</li>
<li>The irreducible element chosen in the simple group PΩ^-(10,3) has order 61 not 122.

</li>
</ul>
<a id="X8389AD927B74BA4A" name="X8389AD927B74BA4A"></a>

<h4>11.1 Overview</h4>

The main purpose of this note is to document the GAP computations that were carried out in order to obtain the computational results in <a href="chapBib.html#biBBGK">[BGK08]</a>. Table I lists the simple groups among these examples. The first column gives the group names, the second and third columns contain a plus sign + or a minus sign -, depending on whether the quantities σ(G,s) and P(G,s), respectively, are less than 1/3. The fourth column lists the orders of elements s which either prove the + signs or cover most of the cases for proving these signs. The fifth column lists the sections in this note where the example is treated. The rows of the table are ordered alphabetically w.r.t. the group names.

In order to keep this note self-contained, we first describe the theory needed, in Section <a href="chap11.html#X7B4649CF7B7CFAA1">11.2</a>. The translation of the relevant formulae into GAP functions can be found in Section <a href="chap11.html#X7B56BE5384BAD54E">11.3</a>. Then Section <a href="chap11.html#X7A221012861440E2">11.4</a> describes the computations that only require (ordinary) character tables in the GAP Character Table Library <a href="chapBib.html#biBCTblLib">[Bre25]</a>. Computations using also the groups are shown in Section <a href="chap11.html#X8237B8617D6F6027">11.5</a>. In each of the last two sections, the examples are ordered alphabetically w.r.t. the names of the simple groups.

<div class="pcenter"><table class="GAPDocTable">
<caption class="GAPDocTable">Table: Table I: Computations needed in <a href="chapBib.html#biBBGK">[BGK08]</a></caption>
<tr>
<td class="tdleft">G</td>
<td class="tdcenter">σ < frac13</td>
<td class="tdcenter">P < frac13</td>
<td class="tdright">|s|</td>
<td class="tdright">see</td>
</tr>
<tr>
<td class="tdleft">A_5</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">5</td>
<td class="tdright"><a href="chap11.html#X7B5321337B28100B">11.5-2</a></td>
</tr>
<tr>
<td class="tdleft">A_6</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">4</td>
<td class="tdright"><a href="chap11.html#X82C3B4287B0C7BEE">11.5-3</a></td>
</tr>
<tr>
<td class="tdleft">A_7</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">7</td>
<td class="tdright"><a href="chap11.html#X85B3C7217B105D4D">11.5-4</a></td>
</tr>
<tr>
<td class="tdleft">A_8</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">15</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">A_9</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">9</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X815320787B601000">11.5-1</a></td>
</tr>
<tr>
<td class="tdleft">A_11</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">11</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X815320787B601000">11.5-1</a></td>
</tr>
<tr>
<td class="tdleft">A_13</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">13</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X815320787B601000">11.5-1</a></td>
</tr>
<tr>
<td class="tdleft">A_15</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">15</td>
<td class="tdright"><a href="chap11.html#X815320787B601000">11.5-1</a></td>
</tr>
<tr>
<td class="tdleft">A_17</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">17</td>
<td class="tdright"><a href="chap11.html#X815320787B601000">11.5-1</a></td>
</tr>
<tr>
<td class="tdleft">A_19</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">19</td>
<td class="tdright"><a href="chap11.html#X815320787B601000">11.5-1</a></td>
</tr>
<tr>
<td class="tdleft">A_21</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">21</td>
<td class="tdright"><a href="chap11.html#X815320787B601000">11.5-1</a></td>
</tr>
<tr>
<td class="tdleft">A_23</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">23</td>
<td class="tdright"><a href="chap11.html#X815320787B601000">11.5-1</a></td>
</tr>
<tr>
<td class="tdleft">L_3(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">7</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X7B1E26D586337487">11.4-5</a>,</td>
</tr>
<tr>
<td class="tdleft"> </td>
<td class="tdcenter"> </td>
<td class="tdcenter"> </td>
<td class="tdright"> </td>
<td class="tdright"><a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a>, <a href="chap11.html#X7C8806DB8588BB51">11.5-8</a></td>
</tr>
<tr>
<td class="tdleft">L_3(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">13</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X7B1E26D586337487">11.4-5</a>,</td>
</tr>
<tr>
<td class="tdleft"> </td>
<td class="tdcenter"> </td>
<td class="tdcenter"> </td>
<td class="tdright"> </td>
<td class="tdright"><a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">L_3(4)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">7</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X7B1E26D586337487">11.4-5</a></td>
</tr>
<tr>
<td class="tdleft">L_4(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">20</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">L_4(4)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">85</td>
<td class="tdright"><a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">L_6(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">63</td>
<td class="tdright"><a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">L_6(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">182</td>
<td class="tdright"><a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">L_6(4)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">455</td>
<td class="tdright"><a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">L_6(5)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">1953</td>
<td class="tdright"><a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">L_8(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">255</td>
<td class="tdright"><a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">L_10(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">1023</td>
<td class="tdright"><a href="chap11.html#X84EA645A82E2BAFB">11.5-5</a></td>
</tr>
<tr>
<td class="tdleft">M_11</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">11</td>
<td class="tdright"><a href="chap11.html#X7B7061917ED3714D">11.5-9</a></td>
</tr>
<tr>
<td class="tdleft">M_12</td>
<td class="tdcenter">-</td>
<td class="tdcenter">+</td>
<td class="tdright">10</td>
<td class="tdright"><a href="chap11.html#X84E9D10F80A74A53">11.4-3</a>, <a href="chap11.html#X82E0F48A7FF82BB3">11.5-10</a></td>
</tr>
<tr>
<td class="tdleft">O^+_8(2)</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">15</td>
<td class="tdright"><a href="chap11.html#X7F80F2527C424AA4">11.5-12</a></td>
</tr>
<tr>
<td class="tdleft">O^+_8(3)</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">20</td>
<td class="tdright"><a href="chap11.html#X78F0815B86253A1F">11.5-13</a></td>
</tr>
<tr>
<td class="tdleft">O^+_8(4)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">65</td>
<td class="tdright"><a href="chap11.html#X85BACC4A83F73392">11.5-14</a></td>
</tr>
<tr>
<td class="tdleft">O^+_10(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">45</td>
<td class="tdright"><a href="chap11.html#X84AB334886DCA746">11.4-7</a></td>
</tr>
<tr>
<td class="tdleft">O^+_12(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">85</td>
<td class="tdright"><a href="chap11.html#X8307367E7C7C3BCE">11.4-9</a></td>
</tr>
<tr>
<td class="tdleft">O^+_12(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">205</td>
<td class="tdright"><a href="chap11.html#X8477457780B69BC7">11.5-18</a></td>
</tr>
<tr>
<td class="tdleft">O^-_8(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">17</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a></td>
</tr>
<tr>
<td class="tdleft">O^-_8(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">41</td>
<td class="tdright"><a href="chap11.html#X78B856907ED13545">11.4-6</a></td>
</tr>
<tr>
<td class="tdleft">O^-_10(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">33</td>
<td class="tdright"><a href="chap11.html#X84E3E4837BB93977">11.4-8</a></td>
</tr>
<tr>
<td class="tdleft">O^-_10(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">122</td>
<td class="tdright"><a href="chap11.html#X8393978A8773997E">11.5-16</a></td>
</tr>
<tr>
<td class="tdleft">O^-_12(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">65</td>
<td class="tdright"><a href="chap11.html#X834FE1B58119A5FF">11.4-10</a></td>
</tr>
<tr>
<td class="tdleft">O^-_14(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">129</td>
<td class="tdright"><a href="chap11.html#X7BBBEEEF834F1002">11.5-17</a></td>
</tr>
<tr>
<td class="tdleft">O_7(3)</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">14</td>
<td class="tdright"><a href="chap11.html#X7FF2E8F27FBEB65C">11.5-11</a></td>
</tr>
<tr>
<td class="tdleft">S_4(4)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">17</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X7B1E26D586337487">11.4-5</a></td>
</tr>
<tr>
<td class="tdleft">S_6(2)</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">9</td>
<td class="tdright"><a href="chap11.html#X82CFBAF07D3487A0">11.5-20</a></td>
</tr>
<tr>
<td class="tdleft">S_6(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">14</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X7B1E26D586337487">11.4-5</a></td>
</tr>
<tr>
<td class="tdleft">S_6(4)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">65</td>
<td class="tdright"><a href="chap11.html#X7C5980A385C088FA">11.4-11</a></td>
</tr>
<tr>
<td class="tdleft">S_8(2)</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">17</td>
<td class="tdright"><a href="chap11.html#X826658207D9D6570">11.5-21</a></td>
</tr>
<tr>
<td class="tdleft">S_8(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">41</td>
<td class="tdright"><a href="chap11.html#X85162B297E4B67EB">11.4-13</a></td>
</tr>
<tr>
<td class="tdleft">U_3(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">6</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X7B1E26D586337487">11.4-5</a></td>
</tr>
<tr>
<td class="tdleft">U_3(5)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">10</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, <a href="chap11.html#X7B1E26D586337487">11.4-5</a></td>
</tr>
<tr>
<td class="tdleft">U_4(2)</td>
<td class="tdcenter">-</td>
<td class="tdcenter">-</td>
<td class="tdright">9</td>
<td class="tdright"><a href="chap11.html#X7A03F8EC839AF0B5">11.5-23</a></td>
</tr>
<tr>
<td class="tdleft">U_4(3)</td>
<td class="tdcenter">-</td>
<td class="tdcenter">+</td>
<td class="tdright">7</td>
<td class="tdright"><a href="chap11.html#X7D738BE5804CF22E">11.5-24</a></td>
</tr>
<tr>
<td class="tdleft">U_4(4)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">65</td>
<td class="tdright"><a href="chap11.html#X8495C2BF7B6EFFEF">11.4-14</a></td>
</tr>
<tr>
<td class="tdleft">U_5(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">11</td>
<td class="tdright"><a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a></td>
</tr>
<tr>
<td class="tdleft">U_6(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">11</td>
<td class="tdright"><a href="chap11.html#X7A3BB5AA83A2BDF3">11.4-15</a></td>
</tr>
<tr>
<td class="tdleft">U_6(3)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">122</td>
<td class="tdright"><a href="chap11.html#X7D4BC6A38074BF68">11.5-25</a></td>
</tr>
<tr>
<td class="tdleft">U_8(2)</td>
<td class="tdcenter">+</td>
<td class="tdcenter"> </td>
<td class="tdright">129</td>
<td class="tdright"><a href="chap11.html#X7A92577A830B5F23">11.5-26</a></td>
</tr>
</table> 
</div>

Contrary to <a href="chapBib.html#biBBGK">[BGK08]</a>, Atlas notation is used throughout this note, because the identifiers used for character tables in the GAP Character Table Library follow mainly the Atlas <a href="chapBib.html#biBCCN85">[CCN+85]</a>. For example, we write L_d(q) for PSL(d,q), S_d(q) for PSp(d,q), U_d(q) for PSU(d,q), and O^+_2d(q), O^-_2d(q), O_2d+1(q) for PΩ^+(2d,q), PΩ^-(2d,q), PΩ(2d+1,q), respectively.

Furthermore, in the case of classical groups, the character tables of the (almost) simple groups are considered not the tables of the matrix groups (which are in fact often not available in the GAP Character Table Library). Consequently, also element orders and the description of maximal subgroups refer to the (almost) simple groups not to the matrix groups.

This note contains also several examples that are not needed for the proofs in <a href="chapBib.html#biBBGK">[BGK08]</a>. Besides several small simple groups G whose character table is contained in the GAP Character Table Library and for which enough information is available for computing σ(G), in Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, a few such examples appear in individual sections. In the table of contents, the section headers of the latter kind of examples are marked with an asterisk (∗).

The examples use the GAP Character Table Library, the GAP Library of Tables of Marks, and the GAP interface <a href="chapBib.html#biBAtlasRep">[WPN+22]</a> to the Atlas of Group Representations <a href="chapBib.html#biBAGRv3">[WWT+]</a>, so we first load these three packages in the required versions. The GAP output was adjusted to the versions shown below; in older versions, features necessary for the computations may be missing, and it may happen that with newer versions, the behaviour is different.

<div class="example"><pre>
gap> CompareVersionNumbers( GAPInfo.Version, "4.5.0" );
true
gap> LoadPackage( "ctbllib", "1.2", false );
true
gap> LoadPackage( "tomlib", "1.2", false );
true
gap> LoadPackage( "atlasrep", "1.5", false );
true
</pre></div>

Some of the computations in Section <a href="chap11.html#X8237B8617D6F6027">11.5</a> require about 800 MB of space (on 32 bit machines). Therefore we check whether GAP was started with sufficient maximal memory; the command line option for this is <code class="code">-o 800m</code>.

<div class="example"><pre>
gap> max:= GAPInfo.CommandLineOptions.o;;
gap> if not ( ( IsSubset( max, "m" ) and
> Int( Filtered( max, IsDigitChar ) ) >= 800 ) or
> ( IsSubset( max, "g" ) and
> Int( Filtered( max, IsDigitChar ) ) >= 1 ) ) then
> Print( "the maximal allowed memory might be too small\n" );
> fi;
</pre></div>

Several computations involve calls to the GAP function <code class="func">Random</code> (<a href="../../../doc/ref/chap30.html#X7FF906E57D6936F8">Reference: Random</a>). In order to make the results of individual examples reproducible, independent of the rest of the computations, we reset the relevant random number generators whenever this is appropriate. For that, we store the initial states in the variable <code class="code">staterandom</code>, and provide a function for resetting the random number generators. (The <code class="func">Random</code> (<a href="../../../doc/ref/chap30.html#X7FF906E57D6936F8">Reference: Random</a>) calls in the GAP library use the two random number generators <code class="func">GlobalRandomSource</code> (<a href="../../../doc/ref/chap14.html#X7AC96008820FAF1F">Reference: GlobalRandomSource</a>) and <code class="func">GlobalMersenneTwister</code> (<a href="../../../doc/ref/chap14.html#X7AC96008820FAF1F">Reference: GlobalMersenneTwister</a>).)

<div class="example"><pre>
gap> staterandom:= [ State( GlobalRandomSource ),
> State( GlobalMersenneTwister ) ];;
gap> ResetGlobalRandomNumberGenerators:= function()
> Reset( GlobalRandomSource, staterandom[1] );
> Reset( GlobalMersenneTwister, staterandom[2] );
> end;;
</pre></div>

<a id="X7B4649CF7B7CFAA1" name="X7B4649CF7B7CFAA1"></a>

<h4>11.2 Prerequisites</h4>

<a id="X7B6AEBDF7B857E2E" name="X7B6AEBDF7B857E2E"></a>

<h5>11.2-1 Theoretical Background</h5>

Let G be a finite group, S the socle of G, and denote by G^× the set of nonidentity elements in G. For s, g ∈ G^×, let P( g, s ):= |{ h ∈ G; S ⊈ ⟨ s^h, g ⟩ }| / |G|, the proportion of elements in the class s^G which fail to generate at least S with g; we set P( G, s ):= max{ P( g, s ); g ∈ G^× }. We are interested in finding a class s^G of elements in S such that P( G, s ) < 1/3 holds.

First consider g ∈ S, and let ��(S,s) denote the set of those maximal subgroups of S that contain s. We have

|{ h ∈ S; S ⊈ ⟨ s^h, g ⟩ }| = |{ h ∈ S; ⟨ s, h g h^-1 ⟩ ≠ S }| ≤ ∑_M ∈ ��(S,s) |{ h ∈ S; h g h^-1 ∈ M }|

Since h g h^-1 ∈ M holds if and only if the coset M h is fixed by g under the permutation action of S on the right cosets of M in S, we get that |{ h ∈ S; h g h^-1 ∈ M }| = |C_S(g)| ⋅ |g^S ∩ M| = |M| ⋅ 1_M^S(g), where 1_M^S is the permutation character of this action, of degree |S|/|M|. Thus

|{ h ∈ S; ⟨ s, h g h^-1 ⟩ ≠ S }| / |S| ≤ ∑_M ∈ ��(S,s) 1_M^S(g) / 1_M^S(1) .

We abbreviate the right hand side of this inequality by σ( g, s ), set σ( S, s ):= max{ σ( g, s ); g ∈ S^× }, and choose a transversal T of S in G. Then P( g, s ) ≤ |T|^-1 ⋅ ∑_t ∈ T σ( g^t, s ) and thus P( G, s ) ≤ σ( S, s ) holds.

If S = G and if ��(G,s) consists of a single maximal subgroup M of G then equality holds, i.e., P( g, s ) = σ( g, s ) = 1_M^S(g) / 1_M^S(1).

The quantity 1_M^S(g) / 1_M^S(1) = |g^S ∩ M| / |g^S| is the proportion of fixed points of g in the permutation action of S on the right cosets of its subgroup M. This is called the fixed point ratio of g w. r. t. S/M, and is denoted as μ(g,S/M).

For a subgroup M of S, the number n of S-conjugates of M containing s is equal to |M^S| ⋅ |s^S ∩ M| / |s^S|. To see this, consider the set { (s^h, M^k); h, k ∈ S, s^h ∈ M^k }, the cardinality of which can be counted either as |M^S| ⋅ |s^S ∩ M| or as |s^S| ⋅ n. So we get n = |M| ⋅ 1_M^S(s) / |N_S(M)|.

If S is a finite nonabelian simple group then each maximal subgroup in S is self-normalizing, and we have n = 1_M^S(s) if M is maximal. So we can replace the summation over ��(S,s) by one over a set ��/~(S,s) of representatives of conjugacy classes of maximal subgroups of S, and get that

σ( g, s ) = ∑_M ∈ ��/~(S,s) frac1_M^S(s) ⋅ 1_M^S(g)1_M^S(1).

Furthermore, we have |��(S,s)| = ∑_M ∈ ��/~(S,s) 1_M^S(s).

In the following, we will often deal with the quantities σ(S):= min{ σ( S, s ); s ∈ S^× } and ��/~(S):= ⌈ 1 / σ(S) - 1 ⌉. These values can be computed easily from the primitive permutation characters of S.

Analogously, we set P(S):= min { P( S, s ); s ∈ S^× } and P(S):= ⌈ 1 / P(S) - 1 ⌉. Clearly we have P(S) ≤ σ(S) and P(S) ≥ ��/~(S).

One interpretation of P(S) is that if this value is at least k then it follows that for any g_1, g_2, ..., g_k ∈ S^×, there is some s ∈ S such that S = ⟨ g_i, s ⟩, for 1 ≤ i ≤ k. In this case, S is said to have spread at least k. (Note that the lower bound ��/~(S) for P(S) can be computed from the list of primitive permutation characters of S.)

Moreover, P(S) ≥ k implies that the element s can be chosen uniformly from a fixed conjugacy class of S. This is called uniform spread at least k in <a href="chapBib.html#biBBGK">[BGK08]</a>.

It is proved in <a href="chapBib.html#biBGK">[GK00]</a> that all finite simple groups have uniform spread at least 1, that is, for any element x ∈ S^×, there is an element y in a prescribed class of S such that G = ⟨ x, y ⟩ holds. In <a href="chapBib.html#biBBGK">[BGK08, Corollary 1.3]</a>, it is shown that all finite simple groups have uniform spread at least 2, and the finite simple groups with (uniform) spread exactly 2 are listed.

Concerning the spread, it should be mentioned that the methods used here and in <a href="chapBib.html#biBBGK">[BGK08]</a> are nonconstructive in the sense that they cannot be used for finding an element s that generates G together with each of the k prescribed elements g_1, g_2, ..., g_k.

Now consider g ∈ G ∖ S. Since P( g^k, s ) ≥ P( g, s ) for any positive integer k, we can assume that g has prime order p, say. We set H = ⟨ S, g ⟩ ≤ G, with [H:S] = p, choose a transversal T of H in G, let ��^'(H,s):= ��(H,s) ∖ { S }, and let ��/~^'(H,s) denote a set of representatives of H-conjugacy classes of these groups. As above,

<div class="pcenter"><table class="GAPDocTablenoborder">
<tr>
<td class="tdright">|{ h ∈ H; S ⊈ ⟨ s^h, g ⟩ }| / |H|</td>
<td class="tdcenter">=</td>
<td class="tdleft">|{ h ∈ H; ⟨ s^h, g ⟩ ≠ H }| / |H|</td>
</tr>
<tr>
<td class="tdright"> </td>
<td class="tdcenter">≤</td>
<td class="tdleft">∑_M ∈ ��^'(H,s) |{ h ∈ H; h g h^-1 ∈ M }| / |H|</td>
</tr>
<tr>
<td class="tdright"> </td>
<td class="tdcenter">=</td>
<td class="tdleft">∑_M ∈ ��^'(H,s) 1_M^H(g) / 1_M^H(1)</td>
</tr>
<tr>
<td class="tdright"> </td>
<td class="tdcenter">=</td>
<td class="tdleft">∑_M ∈ ��/~^'(H,s) 1_M^H(g) ⋅ 1_M^H(s) / 1_M^H(1)</td>
</tr>
</table> 
</div>

(Note that no summand for M = S occurs, so each group in ��/~^'(H,s) is self-normalizing.) We abbreviate the right hand side by σ(H,g,s), and set σ^'( H, s ) = max{ σ(H,g,s); g ∈ H ∖ S, |g| = [H:S] }. Then we get P( g, s ) ≤ |T|^-1 ⋅ ∑_t ∈ T σ(H^t,g^t,s) and thus

P( G, s ) ≤ max{ P( S, s ), max{ σ^'( H, s ); S ≤ H ≤ G, [H:S] prime } } .

For convenience, we set P^'(G,s) = max{ P(g,s); g ∈ G ∖ S }.

<a id="X79D7312484E78274" name="X79D7312484E78274"></a>

<h5>11.2-2 Computational Criteria</h5>

The following criteria will be used when we have to show the existence or nonexistence of x_1, x_2, ..., x_k, and s ∈ G with the property ⟨ x_i, s ⟩ = G for 1 ≤ i ≤ k. Note that manipulating lists of integers (representing fixed or moved points) is much more efficient than testing whether certain permutations generate a given group.

Lemma:

Let G be a finite group, s ∈ G^×, and X = ⋃_M ∈ ��(G,s) G/M. For x_1, x_2, ..., x_k ∈ G, the conjugate s^' of s satisfies ⟨ x_i, s^' ⟩ = G for 1 ≤ i ≤ k if and only if Fix_X(s^') ∩ ⋃_i=1^k Fix_X(x_i) = ∅ holds.

Proof. If s^g ∈ U ≤ G for some g ∈ G then Fix_X(U) = ∅ if and only if U = G holds; note that Fix_X(G) = ∅, and Fix_X(U) = ∅ implies that U ⊈ h^-1 M h holds for all h ∈ G and M ∈ ��(G,s), thus U = G. Applied to U = ⟨ x_i, s^' ⟩, we get ⟨ x_i, s^' ⟩ = G if and only if Fix_X(s^') ∩ Fix_X(x_i) = Fix_X(U) = ∅.

Corollary 1:

If ��(G,s) = { M } in the situation of the above Lemma then there is a conjugate s^' of s that satisfies ⟨ x_i, s^' ⟩ = G for 1 ≤ i ≤ k if and only if ⋃_i=1^k Fix_X(x_i) ≠ X.

Corollary 2:

Let G be a finite simple group and let X be a G-set such that each g ∈ G fixes at least one point in X but that Fix_X(G) = ∅ holds. If x_1, x_2, ... x_k are elements in G such that ⋃_i=1^k Fix_X(x_i) = X holds then for each s ∈ G there is at least one i with ⟨ x_i, s ⟩ ≠ G.

<a id="X7B56BE5384BAD54E" name="X7B56BE5384BAD54E"></a>

<h4>11.3 GAP Functions for the Computations</h4>

After the introduction of general utilities in Section <a href="chap11.html#X806328747D1D4ECC">11.3-1</a>, we distinguish two different tasks. Section <a href="chap11.html#X7A221012861440E2">11.3-2</a> introduces functions that will be used in the following to compute σ(g,s) with character-theoretic methods. Functions for computing P(g,s) or an upper bound for this value will be introduced in Section <a href="chap11.html#X83DACCF07EF62FAE">11.3-3</a>.

The GAP functions shown in this section are collected in the file <code class="file">tst/probgen.g</code> that is distributed with the GAP Character Table Library, see <a href="http://www.math.rwth-aachen.de/~Thomas.Breuer/ctbllib">http://www.math.rwth-aachen.de/~Thomas.Breuer/ctbllib</a>.

The functions have been designed for the examples in the later sections, they could be generalized and optimized for other examples. It is not our aim to provide a package for this functionality.

<a id="X806328747D1D4ECC" name="X806328747D1D4ECC"></a>

<h5>11.3-1 General Utilities</h5>

Let <code class="code">list</code> be a dense list and <code class="code">prop</code> be a unary function that returns <code class="keyw">true</code> or <code class="keyw">false</code> when applied to the entries of <code class="code">list</code>. <code class="code">PositionsProperty</code> returns the set of positions in <code class="code">list</code> for which <code class="keyw">true</code> is returned.

<div class="example"><pre>
gap> if not IsBound( PositionsProperty ) then
> PositionsProperty:= function( list, prop )
> return Filtered( [ 1 .. Length( list ) ], i -> prop( list[i] ) );
> end;
> fi;
</pre></div>

The following two functions implement loops over ordered triples (and quadruples, respectively) in a Cartesian product. A prescribed function <code class="code">prop</code> is subsequently applied to the triples (quadruples), and if the result of this call is <code class="keyw">true</code> then this triple (quadruple) is returned immediately; if none of the calls to <code class="code">prop</code> yields <code class="keyw">true</code> then <code class="keyw">fail</code> is returned.

<div class="example"><pre>
gap> BindGlobal( "TripleWithProperty", function( threelists, prop )
> local i, j, k, test;
> 
> for i in threelists[1] do
> for j in threelists[2] do
> for k in threelists[3] do
> test:= [ i, j, k ];
> if prop( test ) then
> return test;
> fi;
> od;
> od;
> od;
> 
> return fail;
> end );
gap> BindGlobal( "QuadrupleWithProperty", function( fourlists, prop )
> local i, j, k, l, test;
> 
> for i in fourlists[1] do
> for j in fourlists[2] do
> for k in fourlists[3] do
> for l in fourlists[4] do
> test:= [ i, j, k, l ];
> if prop( test ) then
> return test;
> fi;
> od;
> od;
> od;
> od;
> 
> return fail;
> end );
</pre></div>

Of course one could do better by considering unordered n-tuples when several of the argument lists are equal, and in practice, backtrack searches would often allow one to prune parts of the search tree in early stages. However, the above loops are not time critical in the examples presented here, so the possible improvements are not worth the effort for our purposes.

The function <code class="code">PrintFormattedArray</code> prints the matrix <code class="code">array</code> in a columnwise formatted way. (The only diference to the GAP library function <code class="func">PrintArray</code> (<a href="../../../doc/ref/chap24.html#X7DEBC9967DFDFC18">Reference: PrintArray</a>) is that <code class="code">PrintFormattedArray</code> chooses each column width according to the entries only in this column not w.r.t. the whole matrix.)

<div class="example"><pre>
gap> BindGlobal( "PrintFormattedArray", function( array )
> local colwidths, n, row;
> array:= List( array, row -> List( row, String ) );
> colwidths:= List( TransposedMat( array ),
> col -> Maximum( List( col, Length ) ) );
> n:= Length( array[1] );
> for row in List( array, row -> List( [ 1 .. n ],
> i -> String( row[i], colwidths[i] ) ) ) do
> Print( " ", JoinStringsWithSeparator( row, " " ), "\n" );
> od;
> end );
</pre></div>

Finally, <code class="code">CleanWorkspace</code> is a utility for reducing the space needed. This is achieved by unbinding those user variables that are not write protected and are not mentioned in the list <code class="code">NeededVariables</code> of variable names that are bound now, and by flushing the caches of tables of marks and character tables.

<div class="example"><pre>
gap> BindGlobal( "NeededVariables", NamesUserGVars() );
gap> BindGlobal( "CleanWorkspace", function()
> local name, record;
> 
> for name in Difference( NamesUserGVars(), NeededVariables ) do
> if not IsReadOnlyGlobal( name ) then
> UnbindGlobal( name );
> fi;
> od;
> for record in [ LIBTOMKNOWN, LIBTABLE ] do
> for name in RecNames( record.LOADSTATUS ) do
> Unbind( record.LOADSTATUS.( name ) );
> Unbind( record.( name ) );
> od;
> od;
> end );
</pre></div>

The function <code class="code">PossiblePermutationCharacters</code> takes two ordinary character tables <code class="code">sub</code> and <code class="code">tbl</code>, computes the possible class fusions from <code class="code">sub</code> to <code class="code">tbl</code>, then induces the trivial character of <code class="code">sub</code> to <code class="code">tbl</code>, w.r.t. these fusions, and returns the set of these class functions. (So if <code class="code">sub</code> and <code class="code">tbl</code> are the character tables of groups H and G, respectively, where H is a subgroup of G, then the result contains the permutation character 1_H^G.)

Note that the columns of the character tables in the GAP Character Table Library are not explicitly associated with particular conjugacy classes of the corresponding groups, so from the character tables, we can compute only possible class fusions, i.e., maps between the columns of two tables that satisfy certain necessary conditions, see the section about the function <code class="code">PossibleClassFusions</code> in the GAP Reference Manual for details. There is no problem if the permutation character is uniquely determined by the character tables, in all other cases we give ad hoc arguments for resolving the ambiguities.

<div class="example"><pre>
gap> if not IsBound( PossiblePermutationCharacters ) then
> BindGlobal( "PossiblePermutationCharacters", function( sub, tbl )
> local fus, triv;
> 
> fus:= PossibleClassFusions( sub, tbl );
> if fus = fail then
> return fail;
> fi;
> triv:= [ TrivialCharacter( sub ) ];
> 
> return Set(
> List( fus, map -> Induced( sub, tbl, triv, map )[1] ) );
> end );
> fi;
</pre></div>

<a id="X7A221012861440E2" name="X7A221012861440E2"></a>

<h5>11.3-2 Character-Theoretic Computations</h5>

We want to use the GAP libraries of character tables and of tables of marks, and proceed in three steps.

First we extract the primitive permutation characters from the library information if this is available; for that, we write the function <code class="code">PrimitivePermutationCharacters</code>. Then the result can be used as the input for the function <code class="code">ApproxP</code>, which computes the values σ( g, s ). Finally, the functions <code class="code">ProbGenInfoSimple</code> and <code class="code">ProbGenInfoAlmostSimple</code> compute ��/~( G ).

For a group G whose character table T is contained in the GAP character table library, the complete set of primitive permutation characters can be easily computed if the character tables of all maximal subgroups and their class fusions into T are known (in this case, we check whether the attribute <code class="func">Maxes</code> (<a href="../doc/chap3.html#X8150E63F7DBDF252">CTblLib: Maxes</a>) of T is bound) or if the table of marks of G and the class fusion from T into this table of marks are known (in this case, we check whether the attribute <code class="func">FusionToTom</code> (<a href="../doc/chap3.html#X7B1AAED68753B1BE">CTblLib: FusionToTom</a>) of T is bound). If the attribute <code class="func">UnderlyingGroup</code> (<a href="../../../doc/ref/chap70.html#X81E41D3880FA6C4C">Reference: UnderlyingGroup for tables of marks</a>) of T is bound then this group can be used to compute the primitive permutation characters. The latter happens if T was computed from the group object in GAP; for tables in the GAP character table library, this is not the case by default.

The GAP function <code class="code">PrimitivePermutationCharacters</code> tries to compute the primitive permutation characters of a group using this information; it returns the required list of characters if this can be computed this way, otherwise <code class="keyw">fail</code> is returned. (For convenience, we use the GAP mechanism of attributes in order to store the permutation characters in the character table object once they have been computed.)

<div class="example"><pre>
gap> DeclareAttribute( "PrimitivePermutationCharacters",
> IsCharacterTable );
gap> InstallOtherMethod( PrimitivePermutationCharacters,
> [ IsCharacterTable ],
> function( tbl )
> local maxes, i, fus, poss, tom, G;
> 
> if HasMaxes( tbl ) then
> maxes:= List( Maxes( tbl ), CharacterTable );
> for i in [ 1 .. Length( maxes ) ] do
> fus:= GetFusionMap( maxes[i], tbl );
> if fus = fail then
> fus:= PossibleClassFusions( maxes[i], tbl );
> poss:= Set( fus,
> map -> InducedClassFunctionsByFusionMap(
> maxes[i], tbl,
> [ TrivialCharacter( maxes[i] ) ], map )[1] );
> if Length( poss ) = 1 then
> maxes[i]:= poss[1];
> else
> return fail;
> fi;
> else
> maxes[i]:= TrivialCharacter( maxes[i] )^tbl;
> fi;
> od;
> return maxes;
> elif HasFusionToTom( tbl ) then
> tom:= TableOfMarks( tbl );
> maxes:= MaximalSubgroupsTom( tom );
> return PermCharsTom( tbl, tom ){ maxes[1] };
> elif HasUnderlyingGroup( tbl ) then
> G:= UnderlyingGroup( tbl );
> return List( MaximalSubgroupClassReps( G ),
> M -> TrivialCharacter( M )^tbl );
> fi;
> 
> return fail;
> end );
</pre></div>

The function <code class="code">ApproxP</code> takes a list <code class="code">primitives</code> of primitive permutation characters of a group G, say, and the position <code class="code">spos</code> of the class s^G in the character table of G.

Assume that the elements in <code class="code">primitives</code> have the form 1_M^G, for suitable maximal subgroups M of G, and let ��/~ be the set of these groups M. <code class="code">ApproxP</code> returns the class function ψ of G that is defined by ψ(1) = 0 and

ψ(g) = ∑_M ∈ ��/~ frac1_M^G(s) ⋅ 1_M^G(g)1_M^G(1)

otherwise.

If <code class="code">primitives</code> contains all those primitive permutation characters 1_M^G of G (with multiplicity according to the number of conjugacy classes of these maximal subgroups) that do not vanish at s, and if all these M are self-normalizing in G –this holds for example if G is a finite simple group– then ψ(g) = σ( g, s ) holds.

<div class="example"><pre>
gap> BindGlobal( "ApproxP", function( primitives, spos )
> local sum;
> 
> sum:= ShallowCopy( Sum( List( primitives,
> pi -> pi[ spos ] * pi / pi[1] ) ) );
> sum[1]:= 0;
> 
> return sum;
> end );
</pre></div>

Note that for computations with permutation characters, it would make the functions more complicated (and not more efficient) if we would consider only elements g of prime order, and only one representative of Galois conjugate classes.

The next functions needed in this context compute σ(S) and ��/~( S ), for a simple group S, and σ^'(G,s) for an almost simple group G with socle S, respectively.

<code class="code">ProbGenInfoSimple</code> takes the character table <code class="code">tbl</code> of S as its argument. If the full list of primitive permutation characters of S cannot be computed with <code class="code">PrimitivePermutationCharacters</code> then the function returns <code class="keyw">fail</code>. Otherwise <code class="code">ProbGenInfoSimple</code> returns a list containing the identifier of the table, the value σ(S), the integer ��/~( S ), a list of Atlas names of representatives of Galois families of those classes of elements s for which σ(S) = σ( S, s ) holds, and the list of the corresponding cardinalities |��(S,s)|.

<div class="example"><pre>
gap> BindGlobal( "ProbGenInfoSimple", function( tbl )
> local prim, max, min, bound, s;
> prim:= PrimitivePermutationCharacters( tbl );
> if prim = fail then
> return fail;
> fi;
> max:= List( [ 1 .. NrConjugacyClasses( tbl ) ],
> i -> Maximum( ApproxP( prim, i ) ) );
> min:= Minimum( max );
> bound:= Inverse( min );
> if IsInt( bound ) then
> bound:= bound - 1;
> else
> bound:= Int( bound );
> fi;
> s:= PositionsProperty( max, x -> x = min );
> s:= List( Set( s, i -> ClassOrbit( tbl, i ) ), i -> i[1] );
> return [ Identifier( tbl ),
> min,
> bound,
> AtlasClassNames( tbl ){ s },
> Sum( List( prim, pi -> pi{ s } ) ) ];
> end );
</pre></div>

<code class="code">ProbGenInfoAlmostSimple</code> takes the character tables <code class="code">tblS</code> and <code class="code">tblG</code> of S and G, and a list <code class="code">sposS</code> of class positions (w.r.t. <code class="code">tblS</code>) as its arguments. It is assumed that S is simple and has prime index in G. If <code class="code">PrimitivePermutationCharacters</code> can compute the full list of primitive permutation characters of G then the function returns a list containing the identifier of <code class="code">tblG</code>, the maximum m of σ^'( G, s ), for s in the classes described by <code class="code">sposS</code>, a list of Atlas names (in G) of the classes of elements s for which this maximum is attained, and the list of the corresponding cardinalities |��^'(G,s)|. When <code class="code">PrimitivePermutationCharacters</code> returns <code class="keyw">fail</code>, also <code class="code">ProbGenInfoAlmostSimple</code> returns <code class="keyw">fail</code>.

<div class="example"><pre>
gap> BindGlobal( "ProbGenInfoAlmostSimple", function( tblS, tblG, sposS )
> local p, fus, inv, prim, sposG, outer, approx, l, max, min,
> s, cards, i, names;
> 
> p:= Size( tblG ) / Size( tblS );
> if not IsPrimeInt( p )
> or Length( ClassPositionsOfNormalSubgroups( tblG ) ) <> 3 then
> return fail;
> fi;
> fus:= GetFusionMap( tblS, tblG );
> if fus = fail then
> return fail;
> fi;
> inv:= InverseMap( fus );
> prim:= PrimitivePermutationCharacters( tblG );
> if prim = fail then
> return fail;
> fi;
> sposG:= Set( fus{ sposS } );
> outer:= Difference( PositionsProperty(
> OrdersClassRepresentatives( tblG ), IsPrimeInt ), fus );
> approx:= List( sposG, i -> ApproxP( prim, i ){ outer } );
> if IsEmpty( outer ) then
> max:= List( approx, x -> 0 );
> else
> max:= List( approx, Maximum );
> fi;
> min:= Minimum( max);
> s:= sposG{ PositionsProperty( max, x -> x = min ) };
> cards:= List( prim, pi -> pi{ s } );
> for i in [ 1 .. Length( prim ) ] do
> # Omit the character that is induced from the simple group.
> if ForAll( prim[i], x -> x = 0 or x = prim[i][1] ) then
> cards[i]:= 0;
> fi;
> od;
> names:= AtlasClassNames( tblG ){ s };
> Perform( names, ConvertToStringRep );
> 
> return [ Identifier( tblG ),
> min,
> names,
> Sum( cards ) ];
> end );
</pre></div>

The next function computes σ(G,s) from the character table <code class="code">tbl</code> of a simple or almost simple group G, the name <code class="code">sname</code> of the class of s in this table, the list <code class="code">maxes</code> of the character tables of all subgroups M with M ∈ ��(G,s), and the list <code class="code">numpermchars</code> of the numbers of possible permutation characters induced from <code class="code">maxes</code>. If the string <code class="code">"outer"</code> is given as an optional argument then G is assumed to be an automorphic extension of a simple group S, with [G:S] a prime, and σ^'(G,s) is returned. In both situations, the result is <code class="keyw">fail</code> if the numbers of possible permutation characters induced from <code class="code">maxes</code> do not coincide with the numbers prescribed in <code class="code">numpermchars</code>.

<div class="example"><pre>
gap> BindGlobal( "SigmaFromMaxes", function( arg )
> local t, sname, maxes, numpermchars, prim, spos, outer;
> 
> t:= arg[1];
> sname:= arg[2];
> maxes:= arg[3];
> numpermchars:= arg[4];
> prim:= List( maxes, s -> PossiblePermutationCharacters( s, t ) );
> spos:= Position( AtlasClassNames( t ), sname );
> if ForAny( [ 1 .. Length( maxes ) ],
> i -> Length( prim[i] ) <> numpermchars[i] ) then
> return fail;
> elif Length( arg ) = 5 and arg[5] = "outer" then
> outer:= Difference(
> PositionsProperty( OrdersClassRepresentatives( t ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t ) );
> return Maximum( ApproxP( Concatenation( prim ), spos ){ outer } );
> else
> return Maximum( ApproxP( Concatenation( prim ), spos ) );
> fi;
> end );
</pre></div>

The following function allows us to extract information about ��(G,s) from the character table <code class="code">tbl</code> of G and a list <code class="code">snames</code> of class positions of s. If <code class="code">Maxes( tbl )</code> is stored then the names of the character tables of the subgroups in ��(G,s) and the number of conjugates are printed, otherwise <code class="keyw">fail</code> is printed.

<div class="example"><pre>
gap> BindGlobal( "DisplayProbGenMaxesInfo", function( tbl, snames )
> local mx, prim, i, spos, nonz, indent, j;
> 
> if not HasMaxes( tbl ) then
> Print( Identifier( tbl ), ": fail\n" );
> return;
> fi;
> 
> # Now we are sure that the order of the characters returned by
> # 'PrimitivePermutationCharacters' is compatible with 'Maxes( tbl )'.
> mx:= List( Maxes( tbl ), CharacterTable );
> prim:= List( PrimitivePermutationCharacters( tbl ), ShallowCopy );
> for i in [ 1 .. Length( prim ) ] do
> # Deal with the case that the subgroup is normal.
> if ForAll( prim[i], x -> x = 0 or x = prim[i][1] ) then
> prim[i]:= prim[i] / prim[i][1];
> fi;
> od;
> 
> spos:= List( snames,
> nam -> Position( AtlasClassNames( tbl ), nam ) );
> nonz:= List( spos, x -> PositionsProperty( prim, pi -> pi[x] <> 0 ) );
> for i in [ 1 .. Length( spos ) ] do
> Print( Identifier( tbl ), ", ", snames[i], ": " );
> indent:= ListWithIdenticalEntries(
> Length( Identifier( tbl ) ) + Length( snames[i] ) + 4, ' ' );
> if not IsEmpty( nonz[i] ) then
> Print( Identifier( mx[ nonz[i][1] ] ), " (",
> prim[ nonz[i][1] ][ spos[i] ], ")\n" );
> for j in [ 2 .. Length( nonz[i] ) ] do
> Print( indent, Identifier( mx[ nonz[i][j] ] ), " (",
> prim[ nonz[i][j] ][ spos[i] ], ")\n" );
> od;
> else
> Print( "\n" );
> fi;
> od;
> end );
</pre></div>

<a id="X83DACCF07EF62FAE" name="X83DACCF07EF62FAE"></a>

<h5>11.3-3 Computations with Groups</h5>

Here, the task is to compute P(g,s) or P(G,s) using explicit computations with G, where the character-theoretic bounds are not sufficient.

We start with small utilities that make the examples shorter.

For a finite solvable group <code class="code">G</code>, the function <code class="code">PcConjugacyClassReps</code> returns a list of representatives of the conjugacy classes of <code class="code">G</code>, which are computed using a polycyclic presentation for <code class="code">G</code>.

<div class="example"><pre>
gap> BindGlobal( "PcConjugacyClassReps", function( G )
> local iso;
> 
> iso:= IsomorphismPcGroup( G );
> return List( ConjugacyClasses( Image( iso ) ),
> c -> PreImagesRepresentative( iso, Representative( c ) ) );
> end );
</pre></div>

For a finite group <code class="code">G</code>, a list <code class="code">primes</code> of prime integers, and a normal subgroup <code class="code">N</code> of <code class="code">G</code>, the function <code class="code">ClassesOfPrimeOrder</code> returns a list of those conjugacy classes of <code class="code">G</code> that are not contained in <code class="code">N</code> and whose elements' orders occur in <code class="code">primes</code>.

For each prime p in <code class="code">primes</code>, first class representatives of order p in a Sylow p subgroup of <code class="code">G</code> are computed, then the representatives in <code class="code">N</code> are discarded, and then representatives w. r. t. conjugacy in <code class="code">G</code> are computed.

(Note that this approach may be inappropriate for example if a large elementary abelian Sylow p subgroup occurs, and if the conjugacy tests in <code class="code">G</code> are expensive, see Section <a href="chap11.html#X85BACC4A83F73392">11.5-14</a>.)

<div class="example"><pre>
gap> BindGlobal( "ClassesOfPrimeOrder", function( G, primes, N )
> local ccl, p, syl, Greps, reps, r, cr;
> 
> ccl:= [];
> for p in primes do
> syl:= SylowSubgroup( G, p );
> Greps:= [];
> reps:= Filtered( PcConjugacyClassReps( syl ),
> r -> Order( r ) = p and not r in N );
> for r in reps do
> cr:= ConjugacyClass( G, r );
> if ForAll( Greps, c -> c <> cr ) then
> Add( Greps, cr );
> fi;
> od;
> Append( ccl, Greps );
> od;
> 
> return ccl;
> end );
</pre></div>

The function <code class="code">IsGeneratorsOfTransPermGroup</code> takes a transitive permutation group <code class="code">G</code> and a list <code class="code">list</code> of elements in <code class="code">G</code>, and returns <code class="keyw">true</code> if the elements in <code class="code">list</code> generate <code class="code">G</code>, and <code class="keyw">false</code> otherwise. The main point is that the return value <code class="keyw">true</code> requires the group generated by <code class="code">list</code> to be transitive, and the check for transitivity is much cheaper than the test whether this group is equal to <code class="code">G</code>.

<div class="example"><pre>
gap> if not IsBound( IsGeneratorsOfTransPermGroup) then
> BindGlobal( "IsGeneratorsOfTransPermGroup", function( G, list )
> local S;
> 
> if not IsTransitive( G ) then
> Error( "<G> must be transitive on its moved points" );
> fi;
> S:= SubgroupNC( G, list );
> 
> return IsTransitive( S, MovedPoints( G ) ) and
> Size( S ) = Size( G );
> end );
> fi;
</pre></div>

<code class="code">RatioOfNongenerationTransPermGroup</code> takes a transitive permutation group <code class="code">G</code> and two elements <code class="code">g</code> and <code class="code">s</code> of <code class="code">G</code>, and returns the proportion P(g,s). (The function tests the (non)generation only for representatives of C_G(g)-C_G(s)-double cosets. Note that for c_1 ∈ C_G(g), c_2 ∈ C_G(s), and a representative r ∈ G, we have ⟨ g^c_1 r c_2, s ⟩ = ⟨ g^r, s ⟩^c_2.)

<div class="example"><pre>
gap> BindGlobal( "RatioOfNongenerationTransPermGroup", function( G, g, s )
> local nongen, pair;
> 
> if not IsTransitive( G ) then
> Error( "<G> must be transitive on its moved points" );
> fi;
> nongen:= 0;
> for pair in DoubleCosetRepsAndSizes( G, Centralizer( G, g ),
> Centralizer( G, s ) ) do
> if not IsGeneratorsOfTransPermGroup( G, [ s, g^pair[1] ] ) then
> nongen:= nongen + pair[2];
> fi;
> od;
> 
> return nongen / Size( G );
> end );
</pre></div>

Let G be a group, and let <code class="code">groups</code> be a list [ G_1, G_2, ..., G_n ] of permutation groups such that G_i describes the action of G on a set Ω_i, say. Moreover, we require that for 1 ≤ i, j ≤ n, mapping the <code class="code">GeneratorsOfGroup</code> list of G_i to that of G_j defines an isomorphism. <code class="code">DiagonalProductOfPermGroups</code> takes <code class="code">groups</code> as its argument, and returns the action of G on the disjoint union of Ω_1, Ω_2, ..., Ω_n.

<div class="example"><pre>
gap> BindGlobal( "DiagonalProductOfPermGroups", function( groups )
> local prodgens, deg, i, gens, D, pi;
> 
> prodgens:= GeneratorsOfGroup( groups[1] );
> deg:= NrMovedPoints( prodgens );
> for i in [ 2 .. Length( groups ) ] do
> gens:= GeneratorsOfGroup( groups[i] );
> D:= MovedPoints( gens );
> pi:= MappingPermListList( D, [ deg+1 .. deg+Length( D ) ] );
> deg:= deg + Length( D );
> prodgens:= List( [ 1 .. Length( prodgens ) ],
> i -> prodgens[i] * gens[i]^pi );
> od;
> 
> return Group( prodgens );
> end );
</pre></div>

The following two functions are used to reduce checks of generation to class representatives of maximal order. Note that if ⟨ s, g ⟩ is a proper subgroup of G then also ⟨ s^k, g ⟩ is a proper subgroup of G, so we need not check powers s^k different from s in this situation.

For an ordinary character table <code class="code">tbl</code>, the function <code class="code">RepresentativesMaximallyCyclicSubgroups</code> returns a list of class positions, containing one class of generators for each class of maximally cyclic subgroups.

<div class="example"><pre>
gap> BindGlobal( "RepresentativesMaximallyCyclicSubgroups", function( tbl )
> local n, result, orders, p, pmap, i, j;
> 
> # Initialize.
> n:= NrConjugacyClasses( tbl );
> result:= BlistList( [ 1 .. n ], [ 1 .. n ] );
> 
> # Omit powers of smaller order.
> orders:= OrdersClassRepresentatives( tbl );
> for p in PrimeDivisors( Size( tbl ) ) do
> pmap:= PowerMap( tbl, p );
> for i in [ 1 .. n ] do
> if orders[ pmap[i] ] < orders[i] then
> result[ pmap[i] ]:= false;
> fi;
> od;
> od;
> 
> # Omit Galois conjugates.
> for i in [ 1 .. n ] do
> if result[i] then
> for j in ClassOrbit( tbl, i ) do
> if i <> j then
> result[j]:= false;
> fi;
> od;
> fi;
> od;
> 
> # Return the result.
> return ListBlist( [ 1 .. n ], result );
> end );
</pre></div>

Let <code class="code">G</code> be a finite group, <code class="code">tbl</code> be the ordinary character table of <code class="code">G</code>, and <code class="code">cols</code> be a list of class positions in <code class="code">tbl</code>, for example the list returned by <code class="code">RepresentativesMaximallyCyclicSubgroups</code>. The function <code class="code">ClassesPerhapsCorrespondingToTableColumns</code> returns the sublist of those conjugacy classes of <code class="code">G</code> for which the corresponding column in <code class="code">tbl</code> can be contained in <code class="code">cols</code>, according to element order and class size.

<div class="example"><pre>
gap> BindGlobal( "ClassesPerhapsCorrespondingToTableColumns",
> function( G, tbl, cols )
> local orders, classes, invariants;
> 
> orders:= OrdersClassRepresentatives( tbl );
> classes:= SizesConjugacyClasses( tbl );
> invariants:= List( cols, i -> [ orders[i], classes[i] ] );
> 
> return Filtered( ConjugacyClasses( G ),
> c -> [ Order( Representative( c ) ), Size(c) ] in invariants );
> end );
</pre></div>

The next function computes, for a finite group G and subgroups M_1, M_2, ..., M_n of G, an upper bound for max { ∑_i=1^n μ(g,G/M_i); g ∈ G ∖ Z(G) }. So if the M_i are the groups in ��(G,s), for some s ∈ G^×, then we get an upper bound for σ(G,s).

The idea is that for M ≤ G and g ∈ G of order p, we have

μ(g,G/M) = |g^G ∩ M| / |g^G| ≤ ∑_h ∈ C |h^M| / |g^G| = ∑_h ∈ C |h^M| ⋅ |C_G(g)| / |G| ,

where C is a set of class representatives h ∈ M of all those classes that satisfy |h| = p and |C_G(h)| = |C_G(g)|, and in the case that G is a permutation group additionally that h and g move the same number of points. (Note that it is enough to consider elements of prime order.)

For computing the maximum of the rightmost term in this inequality, for g ∈ G ∖ Z(G), we need not determine the G-conjugacy of class representatives in M. Of course we pay the price that the result may be larger than the leftmost term. However, if the maximal sum is in fact taken only over a single class representative, we are sure that equality holds. Thus we return a list of length two, containing the maximum of the right hand side of the above inequality and a Boolean value indicating whether this is equal to max { μ(g,G/M); g ∈ G ∖ Z(G) } or just an upper bound.

The arguments for <code class="code">UpperBoundFixedPointRatios</code> are the group <code class="code">G</code>, a list <code class="code">maxesclasses</code> such that the i-th entry is a list of conjugacy classes of M_i, which covers all classes of prime element order in M_i, and either <code class="keyw">true</code> or <code class="keyw">false</code>, where <code class="keyw">true</code> means that the exact value of σ(G,s) is computed, not just an upper bound; this can be much more expensive because of the conjugacy tests in G that may be necessary. (We try to reduce the number of conjugacy tests in this case, the second half of the code is not completely straightforward. The special treatment of conjugacy checks for elements with the same sets of fixed points is essential in the computation of σ^'(G,s) for G = PGL(6,4); the critical input line is <code class="code">ApproxPForOuterClassesInGL( 6, 4 )</code>, see Section <a href="chap11.html#X7EA88CEF81962F3F">11.5-7</a>. Currently the standard GAP conjugacy test for an element of order three and its inverse in G ∖ G^' requires hours of CPU time, whereas the check for existence of a conjugating element in the stabilizer of the common set of fixed points of the two elements is almost free of charge.)

<code class="code">UpperBoundFixedPointRatios</code> can be used to compute σ^'(G,s) in the case that G is an automorphic extension of a simple group S, with [G:S] = p a prime; if ��^'(G,s) = { M_1, M_2, ..., M_n } then the i-th entry of <code class="code">maxesclasses</code> must contain only the classes of element order p in M_i ∖ (M_i ∩ S).

<div class="example"><pre>
gap> BindGlobal( "UpperBoundFixedPointRatios",
> function( G, maxesclasses, truetest )
> local myIsConjugate, invs, info, c, r, o, inv, pos, sums, max, maxpos,
> maxlen, reps, split, i, found, j;
> 
> myIsConjugate:= function( G, x, y )
> local movx, movy;
> 
> movx:= MovedPoints( x );
> movy:= MovedPoints( y );
> if movx = movy then
> G:= Stabilizer( G, movx, OnSets );
> fi;
> return IsConjugate( G, x, y );
> end;
> 
> invs:= [];
> info:= [];
> 
> # First distribute the classes according to invariants.
> for c in Concatenation( maxesclasses ) do
> r:= Representative( c );
> o:= Order( r );
> # Take only prime order representatives.
> if IsPrimeInt( o ) then
> inv:= [ o, Size( Centralizer( G, r ) ) ];
> # Omit classes that are central in `G'.
> if inv[2] <> Size( G ) then
> if IsPerm( r ) then
> Add( inv, NrMovedPoints( r ) );
> fi;
> pos:= First( [ 1 .. Length( invs ) ], i -> inv = invs[i] );
> if pos = fail then
> # This class is not `G'-conjugate to any of the previous ones.
> Add( invs, inv );
> Add( info, [ [ r, Size( c ) * inv[2] ] ] );
> else
> # This class may be conjugate to an earlier one.
> Add( info[ pos ], [ r, Size( c ) * inv[2] ] );
> fi;
> fi;
> fi;
> od;
> 
> if info = [] then
> return [ 0, true ];
> fi;
> 
> repeat
> # Compute the contributions of the classes with the same invariants.
> sums:= List( info, x -> Sum( List( x, y -> y[2] ) ) );
> max:= Maximum( sums );
> maxpos:= Filtered( [ 1 .. Length( info ) ], i -> sums[i] = max );
> maxlen:= List( maxpos, i -> Length( info[i] ) );
> 
> # Split the sets with the same invariants if necessary
> # and if we want to compute the exact value.
> if truetest and not 1 in maxlen then
> # Make one conjugacy test.
> pos:= Position( maxlen, Minimum( maxlen ) );
> reps:= info[ maxpos[ pos ] ];
> if myIsConjugate( G, reps[1][1], reps[2][1] ) then
> # Fuse the two classes.
> reps[1][2]:= reps[1][2] + reps[2][2];
> reps[2]:= reps[ Length( reps ) ];
> Unbind( reps[ Length( reps ) ] );
> else
> # Split the list. This may require additional conjugacy tests.
> Unbind( info[ maxpos[ pos ] ] );
> split:= [ reps[1], reps[2] ];
> for i in [ 3 .. Length( reps ) ] do
> found:= false;
> for j in split do
> if myIsConjugate( G, reps[i][1], j[1] ) then
> j[2]:= reps[i][2] + j[2];
> found:= true;
> break;
> fi;
> od;
> if not found then
> Add( split, reps[i] );
> fi;
> od;
> 
> info:= Compacted( Concatenation( info,
> List( split, x -> [ x ] ) ) );
> fi;
> fi;
> until 1 in maxlen or not truetest;
> 
> return [ max / Size( G ), 1 in maxlen ];
> end );
</pre></div>

Suppose that C_1, C_2, C_3 are conjugacy classes in G, and that we have to prove, for each (x_1, x_2, x_3) ∈ C_1 × C_2 × C_3, the existence of an element s in a prescribed class C of G such that ⟨ x_1, s ⟩ = ⟨ x_2, s ⟩ = ⟨ x_2, s ⟩ = G holds.

We have to check only representatives under the conjugation action of G on C_1 × C_2 × C_3. For each representative, we try a prescribed number of random elements in C. If this is successful then we are done. The following two functions implement this idea.

For a group G and a list [ g_1, g_2, ..., g_n ] of elements in G, <code class="code">OrbitRepresentativesProductOfClasses</code> returns a list R(G, g_1, g_2, ..., g_n) of representatives of G-orbits on the Cartesian product g_1^G × g_2^G × ⋯ × g_n^G.

The idea behind this function is to choose R(G, g_1) = { ( g_1 ) } in the case n = 1, and, for n > 1,

R(G, g_1, g_2, ..., g_n) = { (h_1, h_2, ..., h_n) ∣ (h_1, h_2, ..., h_n-1) ∈ R(G, g_1, g_2, ..., g_n-1), h_n = g_n^d, for d ∈ D } ,

where D is a set of representatives of double cosets C_G(g_n) ∖ G / ∩_i=1^n-1 C_G(h_i).

<div class="example"><pre>
gap> BindGlobal( "OrbitRepresentativesProductOfClasses",
> function( G, classreps )
> local cents, n, orbreps;
> 
> cents:= List( classreps, x -> Centralizer( G, x ) );
> n:= Length( classreps );
> 
> orbreps:= function( reps, intersect, pos )
> if pos > n then
> return [ reps ];
> fi;
> return Concatenation( List(
> DoubleCosetRepsAndSizes( G, cents[ pos ], intersect ),
> r -> orbreps( Concatenation( reps, [ classreps[ pos ]^r[1] ] ),
> Intersection( intersect, cents[ pos ]^r[1] ), pos+1 ) ) );
> end;
> 
> return orbreps( [ classreps[1] ], cents[1], 2 );
> end );
</pre></div>

The function <code class="code">RandomCheckUniformSpread</code> takes a transitive permutation group G, a list of class representatives g_i ∈ G, an element s ∈ G, and a positive integer N. The return value is <code class="keyw">true</code> if for each representative of G-orbits on the product of the classes g_i^G, a good conjugate of s is found in at most N random tests.

<div class="example"><pre>
gap> BindGlobal( "RandomCheckUniformSpread", function( G, classreps, s, try )
> local elms, found, i, conj;
> 
> if not IsTransitive( G, MovedPoints( G ) ) then
> Error( "<G> must be transitive on its moved points" );
> fi;
> 
> # Compute orbit representatives of G on the direct product,
> # and try to find a good conjugate of s for each representative.
> for elms in OrbitRepresentativesProductOfClasses( G, classreps ) do
> found:= false;
> for i in [ 1 .. try ] do
> conj:= s^Random( G );
> if ForAll( elms,
> x -> IsGeneratorsOfTransPermGroup( G, [ x, conj ] ) ) then
> found:= true;
> break;
> fi;
> od;
> if not found then
> return elms;
> fi;
> od;
> 
> return true;
> end );
</pre></div>

Of course this approach is not suitable for disproving the existence of s, but it is much cheaper than an exhaustive search in the class C. (Typically, |C| is large whereas the |C_i| are small.)

The following function can be used to verify that a given n-tuple (x_1, x_2, ..., x_n) of elements in a group G has the property that for all elements g ∈ G, at least one x_i satisfies ⟨ x_i, g ⟩. The arguments are a transitive permutation group G, a list of class representatives in G, and the n-tuple in question. The return value is a conjugate g of the given representatives that has the property if such an element exists, and <code class="keyw">fail</code> otherwise.

<div class="example"><pre>
gap> BindGlobal( "CommonGeneratorWithGivenElements",
> function( G, classreps, tuple )
> local inter, rep, repcen, pair;
> 
> if not IsTransitive( G, MovedPoints( G ) ) then
> Error( "<G> must be transitive on its moved points" );
> fi;
> 
> inter:= Intersection( List( tuple, x -> Centralizer( G, x ) ) );
> for rep in classreps do
> repcen:= Centralizer( G, rep );
> for pair in DoubleCosetRepsAndSizes( G, repcen, inter ) do
> if ForAll( tuple,
> x -> IsGeneratorsOfTransPermGroup( G, [ x, rep^pair[1] ] ) ) then
> return rep;
> fi;
> od;
> od;
> 
> return fail;
> end );
</pre></div>

<a id="X7A221012861440E2" name="X7A221012861440E2"></a>

<h4>11.4 Character-Theoretic Computations</h4>

In this section, we apply the functions introduced in Section <a href="chap11.html#X7A221012861440E2">11.3-2</a> to the character tables of simple groups that are available in the GAP Character Table Library.

Our first examples are the sporadic simple groups, in Section <a href="chap11.html#X86CE51E180A3D4ED">11.4-1</a>, then their automorphism groups are considered in Section <a href="chap11.html#X84E9D10F80A74A53">11.4-3</a>.

Then we consider those other simple groups for which GAP provides enough information for automatically computing an upper bound on σ(G,s) –see Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>– and their automorphic extensions –see Section <a href="chap11.html#X7B1E26D586337487">11.4-5</a>.

After that, individual groups are considered.

<a id="X86CE51E180A3D4ED" name="X86CE51E180A3D4ED"></a>

<h5>11.4-1 Sporadic Simple Groups</h5>

The GAP Character Table Library contains the tables of maximal subgroups of all sporadic simple groups, so the primitive permutation characters of all sporadic simple groups can be computed via the function <code class="code">PrimitivePermutationCharacters</code>.

<div class="example"><pre>
gap> sporinfo:= [];;
gap> spornames:= AllCharacterTableNames( IsSporadicSimple, true,
> IsDuplicateTable, false );;
gap> for tbl in List( spornames, CharacterTable ) do
> info:= ProbGenInfoSimple( tbl );
> if info <> fail then
> # keep the table columns narrow
> if info[2] <= 10^-13 then
> info[2]:= "<= 10^-13";
> fi;
> if info[3] >= 10^13 then
> info[3]:= ">= 10^13";
> fi;
> Add( sporinfo, info );
> fi;
> od;
</pre></div>

We show the result as a formatted table.

<div class="example"><pre>
gap> PrintFormattedArray( sporinfo );
 B <= 10^-13 >= 10^13 [ "47A" ] [ 1 ]
 Co1 421/1545600 3671 [ "35A" ] [ 4 ]
 Co2 1/270 269 [ "23A" ] [ 1 ]
 Co3 64/6325 98 [ "21A" ] [ 4 ]
 F3+ 1/269631216855 269631216854 [ "29A" ] [ 1 ]
 Fi22 43/585 13 [ "16A" ] [ 7 ]
 Fi23 2651/2416635 911 [ "23A" ] [ 2 ]
 HN 4/34375 8593 [ "19A" ] [ 1 ]
 HS 64/1155 18 [ "15A" ] [ 2 ]
 He 3/595 198 [ "14C" ] [ 3 ]
 J1 1/77 76 [ "19A" ] [ 1 ]
 J2 5/28 5 [ "10C" ] [ 3 ]
 J3 2/153 76 [ "19A" ] [ 2 ]
 J4 1/1647124116 1647124115 [ "29A" ] [ 1 ]
 Ly 1/35049375 35049374 [ "37A" ] [ 1 ]
 M <= 10^-13 >= 10^13 [ "59A" ] [ 1 ]
 M11 1/3 2 [ "11A" ] [ 1 ]
 M12 1/3 2 [ "10A" ] [ 3 ]
 M22 1/21 20 [ "11A" ] [ 1 ]
 M23 1/8064 8063 [ "23A" ] [ 1 ]
 M24 108/1265 11 [ "21A" ] [ 2 ]
 McL 317/22275 70 [ "15A", "30A" ] [ 3, 3 ]
 ON 10/30723 3072 [ "31A" ] [ 2 ]
 Ru 1/2880 2879 [ "29A" ] [ 1 ]
 Suz 141/5720 40 [ "14A" ] [ 3 ]
 Th 2/267995 133997 [ "27A", "27B" ] [ 2, 2 ]
</pre></div>

Here are the exact values for the Baby Monster and the Monster.

<div class="example"><pre>
gap> ProbGenInfoSimple( CharacterTable( "B" ) );
[ "B", 1/174702778623598780219392000000,
 174702778623598780219391999999, [ "47A" ], [ 1 ] ]
gap> ProbGenInfoSimple( CharacterTable( "M" ) );
[ "M", 1/5622007631255133978225347923531983224832000000000,
 5622007631255133978225347923531983224831999999999, [ "59A" ], [ 1 ]
]
</pre></div>

We see that in all these cases, σ(G) < 1/2 and thus P( G ) ≥ 2, and all sporadic simple groups G except G = M_11 and G = M_12 satisfy σ(G) < 1/3. See <a href="chap11.html#X7B7061917ED3714D">11.5-9</a> and <a href="chap11.html#X82E0F48A7FF82BB3">11.5-10</a> for a proof that also these two groups have uniform spread at least three.

The structures and multiplicities of the maximal subgroups containing s are as follows.

<div class="example"><pre>
gap> for entry in sporinfo do
> DisplayProbGenMaxesInfo( CharacterTable( entry[1] ), entry[4] );
> od;
B, 47A: 47:23 (1)
Co1, 35A: (A5xJ2):2 (1)
 (A6xU3(3)):2 (2)
 (A7xL2(7)):2 (1)
Co2, 23A: M23 (1)
Co3, 21A: U3(5).3.2 (2)
 L3(4).D12 (1)
 s3xpsl(2,8).3 (1)
F3+, 29A: 29:14 (1)
Fi22, 16A: 2^10:m22 (1)
 (2x2^(1+8)):U4(2):2 (1)
 2F4(2)' (4)
 2^(5+8):(S3xA6) (1)
Fi23, 23A: 2..11.m23 (1)
 L2(23) (1)
HN, 19A: U3(8).3_1 (1)
HS, 15A: A8.2 (1)
 5:4xa5 (1)
He, 14C: 2^1+6.psl(3,2) (1)
 7^2:2psl(2,7) (1)
 7^(1+2):(S3x3) (1)
J1, 19A: 19:6 (1)
J2, 10C: 2^1+4b:a5 (1)
 a5xd10 (1)
 5^2:D12 (1)
J3, 19A: L2(19) (1)
 J3M3 (1)
J4, 29A: frob (1)
Ly, 37A: 37:18 (1)
M, 59A: 59:29 (1)
M11, 11A: L2(11) (1)
M12, 10A: A6.2^2 (1)
 M12M4 (1)
 2xS5 (1)
M22, 11A: L2(11) (1)
M23, 23A: 23:11 (1)
M24, 21A: L3(4).3.2_2 (1)
 2^6:(psl(3,2)xs3) (1)
McL, 15A: 3^(1+4):2S5 (1)
 2.A8 (1)
 5^(1+2):3:8 (1)
McL, 30A: 3^(1+4):2S5 (1)
 2.A8 (1)
 5^(1+2):3:8 (1)
ON, 31A: L2(31) (1)
 ONM8 (1)
Ru, 29A: L2(29) (1)
Suz, 14A: J2.2 (2)
 (a4xpsl(3,4)):2 (1)
Th, 27A: ThN3B (1)
 ThM7 (1)
Th, 27B: ThN3B (1)
 ThM7 (1)
</pre></div>

Essentially the same approach is taken in <a href="chapBib.html#biBGM01">[GM01]</a>. However, there s is restricted to classes of prime order. Thus the results in the above table are better for J_2, HS, M_24, McL, He, Suz, Co_3, Fi_22, Ly, Th, Co_1, and J_4. Besides that, the value 10999 claimed in <a href="chapBib.html#biBGM01">[GM01]</a> for ��/~( HN ) is not correct.

<a id="X821778BC7D665AB4" name="X821778BC7D665AB4"></a>

<h5>11.4-2 No longer necessary computations for the Baby Monster and the Monster</h5>

At the time when <a href="chapBib.html#biBBGK">[BGK08]</a> was written, not all character tables of maximal subgroups of the Baby Monster and the Monster were available, hence it was not straightforward to compute the primitive permutation characters of these two groups.

In this section, we show the original computations for B and a modification of the computations for M.

For the sporadic simple groups B and M, we choose suitable elements s. If G = B and s ∈ G is of order 47 then, by <a href="chapBib.html#biBWil99">[Wil99]</a>, ��(G,s) = { 47:23 }.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "B" ), "47A",
> [ CharacterTable( "47:23" ) ], [ 1 ] );
1/174702778623598780219392000000
</pre></div>

If G = M and s ∈ G is of order 59 then, by <a href="chapBib.html#biBDLPP24">[DLPP24]</a>, ��(G,s) = { 59:29 }.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "M" ), "59A",
> [ CharacterTable( "59:29" ) ], [ 1 ] );
1/5622007631255133978225347923531983224832000000000
</pre></div>

Before the publication of <a href="chapBib.html#biBDLPP24">[DLPP24]</a>, it had been claimed that ��(G,s) = { L_2(59) }, see <a href="chapBib.html#biBHW04">[HW04]</a>. Under that assumption, we get a bound that is worse. In this situation, the <q>permutation character</q> is not uniquely determined by the character tables, but all possibilities lead to the same value for σ(G).

<div class="example"><pre>
gap> t:= CharacterTable( "M" );;
gap> s:= CharacterTable( "L2(59)" );;
gap> pi:= PossiblePermutationCharacters( s, t );;
gap> Length( pi );
5
gap> spos:= Position( OrdersClassRepresentatives( t ), 59 );
152
gap> Set( pi, x -> Maximum( ApproxP( [ x ], spos ) ) );
[ 1/3385007637938037777290625 ]
</pre></div>

<a id="X84E9D10F80A74A53" name="X84E9D10F80A74A53"></a>

<h5>11.4-3 Automorphism Groups of Sporadic Simple Groups</h5>

Next we consider the automorphism groups of the sporadic simple groups. There are exactly 12 cases where nontrivial outer automorphisms exist, and then the simple group S has index 2 in its automorphism group G.

<div class="example"><pre>
gap> sporautnames:= AllCharacterTableNames( IsSporadicSimple, true,
> IsDuplicateTable, false,
> OfThose, AutomorphismGroup );;
gap> sporautnames:= Difference( sporautnames, spornames );
[ "F3+.2", "Fi22.2", "HN.2", "HS.2", "He.2", "J2.2", "J3.2", "M12.2",
 "M22.2", "McL.2", "ON.2", "Suz.2" ]
</pre></div>

First we compute the values σ^'(G,s), for the same s ∈ S that were chosen for the simple group S in Section <a href="chap11.html#X86CE51E180A3D4ED">11.4-1</a>.

For six of the groups G in question, the character tables of all maximal subgroups are available in the GAP Character Table Library. In these cases, the values σ^'( G, s ) can be computed using <code class="code">ProbGenInfoAlmostSimple</code>.

(The above statement can meanwhile be replaced by the statement that the character tables of all maximal subgroups are available for all twelve groups. We show the table results for all these groups but keep the individual computations from the original computations.)

<div class="example"><pre>
gap> sporautinfo:= [];;
gap> fails:= [];;
gap> for name in sporautnames do
> tbl:= CharacterTable( name{ [ 1 .. Position( name, '.' ) - 1 ] } );
> tblG:= CharacterTable( name );
> info:= ProbGenInfoSimple( tbl );
> info:= ProbGenInfoAlmostSimple( tbl, tblG,
> List( info[4], x -> Position( AtlasClassNames( tbl ), x ) ) );
> if info = fail then
> Add( fails, name );
> else
> Add( sporautinfo, info );
> fi;
> od;
gap> PrintFormattedArray( sporautinfo );
 F3+.2 0 [ "29AB" ] [ 1 ]
 Fi22.2 251/3861 [ "16AB" ] [ 7 ]
 HN.2 1/6875 [ "19AB" ] [ 1 ]
 HS.2 36/275 [ "15A" ] [ 2 ]
 He.2 37/9520 [ "14CD" ] [ 3 ]
 J2.2 1/15 [ "10CD" ] [ 3 ]
 J3.2 1/1080 [ "19AB" ] [ 1 ]
 M12.2 4/99 [ "10A" ] [ 1 ]
 M22.2 1/21 [ "11AB" ] [ 1 ]
 McL.2 1/63 [ "15AB", "30AB" ] [ 3, 3 ]
 ON.2 1/84672 [ "31AB" ] [ 1 ]
 Suz.2 661/46332 [ "14A" ] [ 3 ]
</pre></div>

Note that for S = McL, the bound σ^'(G,s) for G = S.2 (in the second column) is worse than the bound for the simple group S.

The structures and multiplicities of the maximal subgroups containing s are as follows.

<div class="example"><pre>
gap> for entry in sporautinfo do
> DisplayProbGenMaxesInfo( CharacterTable( entry[1] ), entry[3] );
> od;
F3+.2, 29AB: F3+ (1)
 frob (1)
Fi22.2, 16AB: Fi22 (1)
 Fi22.2M4 (1)
 (2x2^(1+8)):(U4(2):2x2) (1)
 2F4(2)'.2 (4)
 2^(5+8):(S3xS6) (1)
HN.2, 19AB: HN (1)
 U3(8).6 (1)
HS.2, 15A: HS (1)
 S8x2 (1)
 5:4xS5 (1)
He.2, 14CD: He (1)
 2^(1+6)_+.L3(2).2 (1)
 7^2:2.L2(7).2 (1)
 7^(1+2):(S3x6) (1)
J2.2, 10CD: J2 (1)
 2^(1+4).S5 (1)
 (A5xD10).2 (1)
 5^2:(4xS3) (1)
J3.2, 19AB: J3 (1)
 19:18 (1)
M12.2, 10A: M12 (1)
 (2^2xA5):2 (1)
M22.2, 11AB: M22 (1)
 L2(11).2 (1)
McL.2, 15AB: McL (1)
 3^(1+4):4S5 (1)
 Isoclinic(2.A8.2) (1)
 5^(1+2):(24:2) (1)
McL.2, 30AB: McL (1)
 3^(1+4):4S5 (1)
 Isoclinic(2.A8.2) (1)
 5^(1+2):(24:2) (1)
ON.2, 31AB: ON (1)
 31:30 (1)
Suz.2, 14A: Suz (1)
 J2.2x2 (2)
 (A4xL3(4):2_3):2 (1)
</pre></div>

Note that the maximal subgroups L_2(19) of J_3 do not extend to J_3.2 and that a class of maximal subgroups of the type 19:18 appears in J_3.2 whose intersection with J_3 is not maximal in J_3. Similarly, the maximal subgroups A_6.2^2 of M_12 do not extend to M_12.2.

For the other six groups, we use individual computations.

In the case S = Fi_24^', the unique maximal subgroup 29:14 that contains an element s of order 29 extends to a group of the type 29:28 in Fi_24, which is a nonsplit extension of 29:14.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "Fi24'.2" ), "29AB",
> [ CharacterTable( "29:28" ) ], [ 1 ], "outer" );
0
</pre></div>

In the case S = Fi_22, there are four classes of maximal subgroups that contain s of order 16. They extend to G = Fi_22.2, and none of the novelties in G (i. e., subgroups of G that are maximal in G but whose intersections with S are not maximal in S) contains s, cf. <a href="chapBib.html#biBCCN85">[CCN+85, p. 163]</a>.

<div class="example"><pre>
gap> 16 in OrdersClassRepresentatives( CharacterTable( "U4(2).2" ) );
false
gap> 16 in OrdersClassRepresentatives( CharacterTable( "G2(3).2" ) );
false
</pre></div>

The character tables of three of the four extensions are available in the GAP Character Table Library. The permutation character on the cosets of the fourth extension can be obtained as the extension of the permutation character of S on the cosets of its maximal subgroup of the type 2^5+8:(S_3 × A_6).

<div class="example"><pre>
gap> t2:= CharacterTable( "Fi22.2" );;
gap> prim:= List( [ "Fi22.2M4", "(2x2^(1+8)):(U4(2):2x2)", "2F4(2)" ],
> n -> PossiblePermutationCharacters( CharacterTable( n ), t2 ) );;
gap> t:= CharacterTable( "Fi22" );;
gap> pi:= PossiblePermutationCharacters(
> CharacterTable( "2^(5+8):(S3xA6)" ), t );
[ Character( CharacterTable( "Fi22" ),
 [ 3648645, 56133, 10629, 2245, 567, 729, 405, 81, 549, 165, 133,
 37, 69, 20, 27, 81, 9, 39, 81, 19, 1, 13, 33, 13, 1, 0, 13, 13,
 5, 1, 0, 0, 0, 8, 4, 0, 0, 9, 3, 15, 3, 1, 1, 1, 1, 3, 3, 1, 0,
 0, 0, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 2 ] ) ]
gap> torso:= CompositionMaps( pi[1], InverseMap( GetFusionMap( t, t2 ) ) );
[ 3648645, 56133, 10629, 2245, 567, 729, 405, 81, 549, 165, 133, 37,
 69, 20, 27, 81, 9, 39, 81, 19, 1, 13, 33, 13, 1, 0, 13, 13, 5, 1,
 0, 0, 0, 8, 4, 0, 9, 3, 15, 3, 1, 1, 1, 3, 3, 1, 0, 0, 2, 1, 0, 0,
 0, 0, 0, 0, 1, 1, 2 ]
gap> ext:= PermChars( t2, rec( torso:= torso ) );;
gap> Add( prim, ext );
gap> prim:= Concatenation( prim );; Length( prim );
4
gap> spos:= Position( OrdersClassRepresentatives( t2 ), 16 );;
gap> List( prim, x -> x[ spos ] );
[ 1, 1, 4, 1 ]
gap> sigma:= ApproxP( prim, spos );;
gap> Maximum( sigma{ Difference( PositionsProperty(
> OrdersClassRepresentatives( t2 ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t2 ) ) } );
251/3861
</pre></div>

In the case S = HN, the unique maximal subgroup U_3(8).3 that contains the fixed element s of order 19 extends to a group of the type U_3(8).6 in HN.2.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "HN.2" ), "19AB",
> [ CharacterTable( "U3(8).6" ) ], [ 1 ], "outer" );
1/6875
</pre></div>

In the case S = HS, there are two classes of maximal subgroups that contain s of order 15. They extend to G = HS.2, and none of the novelties in G contains s (cf. <a href="chapBib.html#biBCCN85">[CCN+85, p. 80]</a>).

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "HS.2" ), "15A",
> [ CharacterTable( "S8x2" ),
> CharacterTable( "5:4" ) * CharacterTable( "A5.2" ) ], [ 1, 1 ],
> "outer" );
36/275
</pre></div>

In the case S = He, there are three classes of maximal subgroups that contain s in the class <code class="code">14C</code>. They extend to G = He.2, and none of the novelties in G contains s (cf. <a href="chapBib.html#biBCCN85">[CCN+85, p. 104]</a>). We compute the extensions of the corresponding primitive permutation characters of S.

<div class="example"><pre>
gap> t:= CharacterTable( "He" );;
gap> t2:= CharacterTable( "He.2" );;
gap> prim:= PrimitivePermutationCharacters( t );;
gap> spos:= Position( AtlasClassNames( t ), "14C" );;
gap> prim:= Filtered( prim, x -> x[ spos ] <> 0 );;
gap> map:= InverseMap( GetFusionMap( t, t2 ) );;
gap> torso:= List( prim, pi -> CompositionMaps( pi, map ) );
[ [ 187425, 945, 449, 0, 21, 21, 25, 25, 0, 0, 5, 0, 0, 7, 1, 0, 0,
 1, 0, 1, 0, 0, 0, 0, 0, 0 ],
 [ 244800, 0, 64, 0, 84, 0, 0, 16, 0, 0, 4, 24, 45, 3, 4, 0, 0, 0,
 0, 1, 0, 0, 0, 0, 0, 0 ],
 [ 652800, 0, 512, 120, 72, 0, 0, 0, 0, 0, 8, 8, 22, 1, 0, 0, 0, 0,
 0, 1, 0, 0, 1, 1, 2, 0 ] ]
gap> ext:= List( torso, x -> PermChars( t2, rec( torso:= x ) ) );
[ [ Character( CharacterTable( "He.2" ),
 [ 187425, 945, 449, 0, 21, 21, 25, 25, 0, 0, 5, 0, 0, 7, 1, 0,
 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 315, 15, 0, 0, 3, 7, 7, 3, 0,
 0, 0, 1, 1, 0, 1, 1, 0, 0, 0 ] ) ],
 [ Character( CharacterTable( "He.2" ),
 [ 244800, 0, 64, 0, 84, 0, 0, 16, 0, 0, 4, 24, 45, 3, 4, 0, 0,
 0, 0, 1, 0, 0, 0, 0, 0, 0, 360, 0, 0, 0, 6, 0, 0, 0, 0, 0,
 3, 2, 2, 0, 0, 0, 0, 0, 0 ] ) ],
 [ Character( CharacterTable( "He.2" ),
 [ 652800, 0, 512, 120, 72, 0, 0, 0, 0, 0, 8, 8, 22, 1, 0, 0, 0,
 0, 0, 1, 0, 0, 1, 1, 2, 0, 480, 0, 120, 0, 12, 0, 0, 0, 0,
 0, 4, 0, 0, 0, 0, 0, 0, 1, 1 ] ) ] ]
gap> spos:= Position( AtlasClassNames( t2 ), "14CD" );;
gap> sigma:= ApproxP( Concatenation( ext ), spos );;
gap> Maximum( sigma{ Difference( PositionsProperty(
> OrdersClassRepresentatives( t2 ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t2 ) ) } );
37/9520
</pre></div>

In the case S = O'N, the two classes of maximal subgroups of the type L_2(31) do not extend to G = O'N.2, and a class of novelties of the structure 31:30 appears (see <a href="chapBib.html#biBCCN85">[CCN+85, p. 132]</a>).

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "ON.2" ), "31AB",
> [ CharacterTable( "P:Q", [ 31, 30 ] ) ], [ 1 ], "outer" );
1/84672
</pre></div>

Now we consider also σ(G,hats), for suitable hats ∈ G ∖ S; this yields lower bounds for the spread of the nonsimple groups G. (These results are shown in the last two columns of <a href="chapBib.html#biBBGK">[BGK08, Table 9]</a>.)

As above, we use the known character tables of the maximal subgroups in order to compute the optimal choice for hats ∈ G ∖ S. (We may use the function <code class="code">ProbGenInfoSimple</code> although the groups are not simple; all we need is that the relevant maximal subgroups are self-normalizing.)

<div class="example"><pre>
gap> sporautinfo2:= [];;
gap> for name in List( sporautinfo, x -> x[1] ) do
> Add( sporautinfo2, ProbGenInfoSimple( CharacterTable( name ) ) );
> od;
gap> PrintFormattedArray( sporautinfo2 );
 F3+.2 19/5684 299 [ "42E" ] [ 10 ]
 Fi22.2 1165/20592 17 [ "24G" ] [ 3 ]
 HN.2 1/1425 1424 [ "24B" ] [ 4 ]
 HS.2 21/550 26 [ "20C" ] [ 4 ]
 He.2 33/4165 126 [ "24A" ] [ 2 ]
 J2.2 1/15 14 [ "14A" ] [ 1 ]
 J3.2 77/10260 133 [ "34A" ] [ 1 ]
 M12.2 113/495 4 [ "12B" ] [ 3 ]
 M22.2 8/33 4 [ "10A" ] [ 4 ]
 McL.2 1/135 134 [ "22A" ] [ 1 ]
 ON.2 61/109368 1792 [ "22A", "38A" ] [ 1, 1 ]
 Suz.2 1/351 350 [ "28A" ] [ 1 ]
gap> for entry in sporautinfo2 do
> DisplayProbGenMaxesInfo( CharacterTable( entry[1] ), entry[4] );
> od;
F3+.2, 42E: 2^12.M24 (2)
 2^2.U6(2):S3x2 (1)
 2^(3+12).(L3(2)xS6) (2)
 (S3xS3xG2(3)):2 (1)
 S6xL2(8):3 (1)
 7:6xS7 (1)
 7^(1+2)_+:(6xS3).2 (2)
Fi22.2, 24G: Fi22.2M4 (1)
 2^(5+8):(S3xS6) (1)
 3^5:(2xU4(2).2) (1)
HN.2, 24B: 2^(1+8)_+.(A5xA5).2^2 (1)
 5^2.5.5^2.4S5 (2)
 HN.2M13 (1)
HS.2, 20C: (2xA6.2^2).2 (1)
 HS.2N5 (2)
 5:4xS5 (1)
He.2, 24A: 2^(1+6)_+.L3(2).2 (1)
 S4xL3(2).2 (1)
J2.2, 14A: L3(2).2x2 (1)
J3.2, 34A: L2(17)x2 (1)
M12.2, 12B: L2(11).2 (1)
 2^3.(S4x2) (1)
 3^(1+2):D8 (1)
M22.2, 10A: M22.2M4 (1)
 A6.2^2 (1)
 L2(11).2 (2)
McL.2, 22A: 2xM11 (1)
ON.2, 22A: J1x2 (1)
ON.2, 38A: J1x2 (1)
Suz.2, 28A: (A4xL3(4):2_3):2 (1)
</pre></div>

In the other six cases, we do not have the complete lists of primitive permutation characters, so we choose a suitable element hats for each group. It is sufficient to prescribe |hats|, as follows.

<div class="example"><pre>
gap> sporautchoices:= [
> [ "Fi22", "Fi22.2", 42 ],
> [ "Fi24'", "Fi24'.2", 46 ],
> [ "He", "He.2", 42 ],
> [ "HN", "HN.2", 44 ],
> [ "HS", "HS.2", 30 ],
> [ "ON", "ON.2", 38 ], ];;
</pre></div>

First we list the maximal subgroups of the corresponding simple groups that contain the square of hats.

<div class="example"><pre>
gap> for triple in sporautchoices do
> tbl:= CharacterTable( triple[1] );
> tbl2:= CharacterTable( triple[2] );
> spos2:= PowerMap( tbl2, 2,
> Position( OrdersClassRepresentatives( tbl2 ), triple[3] ) );
> spos:= Position( GetFusionMap( tbl, tbl2 ), spos2 );
> DisplayProbGenMaxesInfo( tbl, AtlasClassNames( tbl ){ [ spos ] } );
> od;
Fi22, 21A: O8+(2).3.2 (1)
 S3xU4(3).2_2 (1)
 A10.2 (1)
 A10.2 (1)
F3+, 23A: Fi23 (1)
 F3+M7 (1)
He, 21B: 3.A7.2 (1)
 7^(1+2):(S3x3) (1)
 7:3xpsl(3,2) (2)
HN, 22A: 2.HS.2 (1)
HS, 15A: A8.2 (1)
 5:4xa5 (1)
ON, 19B: L3(7).2 (1)
 ONM2 (1)
 J1 (1)
</pre></div>

According to <a href="chapBib.html#biBCCN85">[CCN+85]</a>, exactly the following maximal subgroups of the simple group S in the above list do not extend to Aut(S): The two S_10 type subgroups of Fi_22 and the two L_3(7).2 type subgroups of O'N.

Furthermore, the following maximal subgroups of Aut(S) with the property that the intersection with S is not maximal in S have to be considered whether they contain s^': G_2(3).2 and 3^5:(2 × U_4(2).2) in Fi_22.2. (Note that the order of the 7^1+2_+:(3 × D_16) type subgroup in O'N.2 is obviously not divisible by 19.)

<div class="example"><pre>
gap> 42 in OrdersClassRepresentatives( CharacterTable( "G2(3).2" ) );
false
gap> Size( CharacterTable( "U4(2)" ) ) mod 7 = 0;
false
</pre></div>

So we take the extensions of the above maximal subgroups, as described in <a href="chapBib.html#biBCCN85">[CCN+85]</a>.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "Fi22.2" ), "42A",
> [ CharacterTable( "O8+(2).3.2" ) * CharacterTable( "Cyclic", 2 ),
> CharacterTable( "S3" ) * CharacterTable( "U4(3).(2^2)_{122}" ) ],
> [ 1, 1 ] );
163/1170
gap> SigmaFromMaxes( CharacterTable( "Fi24'.2" ), "46A",
> [ CharacterTable( "Fi23" ) * CharacterTable( "Cyclic", 2 ),
> CharacterTable( "2^12.M24" ) ],
> [ 1, 1 ] );
566/5481
gap> SigmaFromMaxes( CharacterTable( "He.2" ), "42A",
> [ CharacterTable( "3.A7.2" ) * CharacterTable( "Cyclic", 2 ),
> CharacterTable( "7^(1+2):(S3x6)" ),
> CharacterTable( "7:6" ) * CharacterTable( "L3(2)" ) ],
> [ 1, 1, 1 ] );
1/119
gap> SigmaFromMaxes( CharacterTable( "HN.2" ), "44A",
> [ CharacterTable( "4.HS.2" ) ],
> [ 1 ] );
997/192375
gap> SigmaFromMaxes( CharacterTable( "HS.2" ), "30A",
> [ CharacterTable( "S8" ) * CharacterTable( "C2" ),
> CharacterTable( "5:4" ) * CharacterTable( "S5" ) ],
> [ 1, 1 ] );
36/275
gap> SigmaFromMaxes( CharacterTable( "ON.2" ), "38A",
> [ CharacterTable( "J1" ) * CharacterTable( "C2" ) ],
> [ 1 ] );
61/109368
</pre></div>

<a id="X80DA58F187CDCF5F" name="X80DA58F187CDCF5F"></a>

<h5>11.4-4 Other Simple Groups – Easy Cases</h5>

We are interested in simple groups G for which <code class="code">ProbGenInfoSimple</code> does not guarantee ��/~(G) ≥ 3. So we examine the remaining tables of simple groups in the GAP Character Table Library, and distinguish the following three cases: Either <code class="code">ProbGenInfoSimple</code> yields the lower bound at least three, or a smaller bound, or the computation of a lower bound fails because not enough information is available to compute the primitive permutation characters.

<div class="example"><pre>
gap> names:= AllCharacterTableNames( IsSimple, true, IsAbelian, false,
> IsDuplicateTable, false );;
gap> names:= Difference( names, spornames );;
gap> fails:= [];;
gap> lessthan3:= [];;
gap> atleast3:= [];;
gap> for name in names do
> tbl:= CharacterTable( name );
> info:= ProbGenInfoSimple( tbl );
> if info = fail then
> Add( fails, name );
> elif info[3] < 3 then
> Add( lessthan3, info );
> else
> Add( atleast3, info );
> fi;
> od;
</pre></div>

For the following simple groups, (currently) not enough information is available in the GAP Character Table Library and in the GAP Library of Tables of Marks, for computing a lower bound for σ(G). Some of these groups will be dealt with in later sections, and for the other groups, the bounds derived with theoretical arguments in <a href="chapBib.html#biBBGK">[BGK08]</a> are sufficient, so we need no GAP computations for them.

<div class="example"><pre>
gap> fails;
[ "2E6(2)", "2F4(8)", "3D4(3)", "3D4(4)", "A14", "A15", "A16", "A17",
 "A18", "A19", "E6(2)", "F4(3)", "G2(7)", "L4(4)", "L4(5)", "L4(9)",
 "L5(3)", "L8(2)", "O10+(2)", "O10+(3)", "O10-(2)", "O10-(3)",
 "O12+(2)", "O12+(3)", "O12-(2)", "O12-(3)", "O7(5)", "O8+(7)",
 "O8-(3)", "O9(3)", "R(27)", "S10(2)", "S12(2)", "S4(7)", "S4(8)",
 "S4(9)", "S6(4)", "S6(5)", "S8(3)", "U4(4)", "U4(5)", "U5(3)",
 "U5(4)", "U6(4)", "U7(2)" ]
</pre></div>

The following simple groups appear in <a href="chapBib.html#biBBGK">[BGK08, Table 1–6]</a>. More detailed computations can be found in the sections <a href="chap11.html#X7B5321337B28100B">11.5-2</a>, <a href="chap11.html#X82C3B4287B0C7BEE">11.5-3</a>, <a href="chap11.html#X85B3C7217B105D4D">11.5-4</a>, <a href="chap11.html#X7F80F2527C424AA4">11.5-12</a>, <a href="chap11.html#X78F0815B86253A1F">11.5-13</a>, <a href="chap11.html#X82CFBAF07D3487A0">11.5-20</a>, <a href="chap11.html#X7A03F8EC839AF0B5">11.5-23</a>, <a href="chap11.html#X7D738BE5804CF22E">11.5-24</a>.

<div class="example"><pre>
gap> PrintFormattedArray( lessthan3 );
 A5 1/3 2 [ "5A" ] [ 1 ]
 A6 2/3 1 [ "5A" ] [ 2 ]
 A7 2/5 2 [ "7A" ] [ 2 ]
 O7(3) 199/351 1 [ "14A" ] [ 3 ]
 O8+(2) 334/315 0 [ "15A", "15B", "15C" ] [ 7, 7, 7 ]
 O8+(3) 863/1820 2 [ "20A", "20B", "20C" ] [ 8, 8, 8 ]
 S6(2) 4/7 1 [ "9A" ] [ 4 ]
 S8(2) 8/15 1 [ "17A" ] [ 3 ]
 U4(2) 21/40 1 [ "12A" ] [ 2 ]
 U4(3) 53/135 2 [ "7A" ] [ 7 ]
</pre></div>

For the following simple groups G, the inequality σ(G) < 1/3 follows from the loop above. The columns show the name of G, the values σ(G) and ��/~(G), the class names of s for which these values are attained, and |��(G,s)|.

(We increase the line length for this table. Even with this width, the entry for the group L_7(2) would not fit on one screen line, we show it separately below.)

<div class="example"><pre>
gap> oldsize:= SizeScreen();;
gap> SizeScreen( [ 80 ] );;
gap> PrintFormattedArray( Filtered( atleast3, l -> l[1] <> "L7(2)" ) );
 2F4(2)' 118/1755 14 [ "16A" ] [ 2 ]
 3D4(2) 1/5292 5291 [ "13A" ] [ 1 ]
 A10 3/10 3 [ "21A" ] [ 1 ]
 A11 2/105 52 [ "11A" ] [ 2 ]
 A12 2/9 4 [ "35A" ] [ 1 ]
 A13 4/1155 288 [ "13A" ] [ 5 ]
 A8 3/14 4 [ "15A" ] [ 1 ]
 A9 9/35 3 [ "9A", "9B" ] [ 4, 4 ]
 F4(2) 9/595 66 [ "13A" ] [ 5 ]
 G2(3) 1/7 6 [ "13A" ] [ 3 ]
 G2(4) 1/21 20 [ "13A" ] [ 2 ]
 G2(5) 1/31 30 [ "7A", "21A" ] [ 10, 1 ]
 L2(101) 1/101 100 [ "51A", "17A" ] [ 1, 1 ]
 L2(103) 53/5253 99 [ "52A", "26A", "13A" ] [ 1, 1, 1 ]
 L2(107) 55/5671 103 [ "54A", "27A", "18A", "9A", "6A" ] [ 1, 1, 1, 1, 1 ]
 L2(109) 1/109 108 [ "55A", "11A" ] [ 1, 1 ]
 L2(11) 7/55 7 [ "6A" ] [ 1 ]
 L2(113) 1/113 112 [ "57A", "19A" ] [ 1, 1 ]
 L2(121) 1/121 120 [ "61A" ] [ 1 ]
 L2(125) 1/125 124 [ "63A", "21A", "9A", "7A" ] [ 1, 1, 1, 1 ]
 L2(13) 1/13 12 [ "7A" ] [ 1 ]
 L2(16) 1/15 14 [ "17A" ] [ 1 ]
 L2(17) 1/17 16 [ "9A" ] [ 1 ]
 L2(19) 11/171 15 [ "10A" ] [ 1 ]
 L2(23) 13/253 19 [ "6A", "12A" ] [ 1, 1 ]
 L2(25) 1/25 24 [ "13A" ] [ 1 ]
 L2(27) 5/117 23 [ "7A", "14A" ] [ 1, 1 ]
 L2(29) 1/29 28 [ "15A" ] [ 1 ]
 L2(31) 17/465 27 [ "8A", "16A" ] [ 1, 1 ]
 L2(32) 1/31 30 [ "3A", "11A", "33A" ] [ 1, 1, 1 ]
 L2(37) 1/37 36 [ "19A" ] [ 1 ]
 L2(41) 1/41 40 [ "21A", "7A" ] [ 1, 1 ]
 L2(43) 23/903 39 [ "22A", "11A" ] [ 1, 1 ]
 L2(47) 25/1081 43 [ "24A", "12A", "8A", "6A" ] [ 1, 1, 1, 1 ]
 L2(49) 1/49 48 [ "25A" ] [ 1 ]
 L2(53) 1/53 52 [ "27A", "9A" ] [ 1, 1 ]
 L2(59) 31/1711 55 [ "30A", "15A", "10A", "6A" ] [ 1, 1, 1, 1 ]
 L2(61) 1/61 60 [ "31A" ] [ 1 ]
 L2(64) 1/63 62 [ "65A", "13A" ] [ 1, 1 ]
 L2(67) 35/2211 63 [ "34A", "17A" ] [ 1, 1 ]
 L2(71) 37/2485 67 [ "36A", "18A", "12A", "9A", "6A" ] [ 1, 1, 1, 1, 1 ]
 L2(73) 1/73 72 [ "37A" ] [ 1 ]
 L2(79) 41/3081 75 [ "40A", "20A", "10A", "8A" ] [ 1, 1, 1, 1 ]
 L2(8) 1/7 6 [ "3A", "9A" ] [ 1, 1 ]
 L2(81) 1/81 80 [ "41A" ] [ 1 ]
 L2(83) 43/3403 79 [ "42A", "21A", "14A", "7A", "6A" ] [ 1, 1, 1, 1, 1 ]
 L2(89) 1/89 88 [ "45A", "15A", "9A" ] [ 1, 1, 1 ]
 L2(97) 1/97 96 [ "49A", "7A" ] [ 1, 1 ]
 L3(11) 1/6655 6654 [ "19A", "133A" ] [ 1, 1 ]
 L3(2) 1/4 3 [ "7A" ] [ 1 ]
 L3(3) 1/24 23 [ "13A" ] [ 1 ]
 L3(4) 1/5 4 [ "7A" ] [ 3 ]
 L3(5) 1/250 249 [ "31A" ] [ 1 ]
 L3(7) 1/1372 1371 [ "19A" ] [ 1 ]
 L3(8) 1/1792 1791 [ "73A" ] [ 1 ]
 L3(9) 1/2880 2879 [ "91A" ] [ 1 ]
 L4(3) 53/1053 19 [ "20A" ] [ 1 ]
 L5(2) 1/5376 5375 [ "31A" ] [ 1 ]
 L6(2) 365/55552 152 [ "21A", "63A" ] [ 2, 2 ]
 O8-(2) 1/63 62 [ "17A" ] [ 1 ]
 S4(4) 4/15 3 [ "17A" ] [ 2 ]
 S4(5) 1/5 4 [ "13A" ] [ 1 ]
 S6(3) 1/117 116 [ "14A" ] [ 2 ]
 Sz(32) 1/1271 1270 [ "5A", "25A" ] [ 1, 1 ]
 Sz(8) 1/91 90 [ "5A" ] [ 1 ]
 U3(11) 1/6655 6654 [ "37A" ] [ 1 ]
 U3(3) 16/63 3 [ "6A", "12A" ] [ 2, 2 ]
 U3(4) 1/160 159 [ "13A" ] [ 1 ]
 U3(5) 46/525 11 [ "10A" ] [ 2 ]
 U3(7) 1/1372 1371 [ "43A" ] [ 1 ]
 U3(8) 1/1792 1791 [ "19A" ] [ 1 ]
 U3(9) 1/3600 3599 [ "73A" ] [ 1 ]
 U5(2) 1/54 53 [ "11A" ] [ 1 ]
 U6(2) 5/21 4 [ "11A" ] [ 4 ]
gap> SizeScreen( oldsize );;
gap> First( atleast3, l -> l[1] = "L7(2)" );
[ "L7(2)", 1/4388290560, 4388290559, [ "127A" ], [ 1 ] ]
</pre></div>

It should be mentioned that <a href="chapBib.html#biBBW1">[BW75]</a> states the following lower bounds for the uniform spread of the groups L_2(q).

<div class="pcenter"><table class="GAPDocTable">
<tr>
<td class="tdleft">q-2</td>
<td class="tdleft">if 4 ≤ q is even,</td>
</tr>
<tr>
<td class="tdleft">q-1</td>
<td class="tdleft">if 11 ≤ q ≡ 1 mod 4,</td>
</tr>
<tr>
<td class="tdleft">q-4</td>
<td class="tdleft">if 11 ≤ q ≡ -1 mod 4.</td>
</tr>
</table> 
</div>

These bounds appear in the third column of the above table. Furthermore, <a href="chapBib.html#biBBW1">[BW75]</a> states that the (uniform) spread of alternating groups of even degree at least 8 is exactly 4.

For the sake of completeness, Table II gives an overview of the sets ��(G,s) for those cases in the above list that are needed in <a href="chapBib.html#biBBGK">[BGK08]</a> but that do not require a further discussion here. The structure of the maximal subgroups and the order of s in the table refer to the matrix groups not to the simple groups. The number of the subgroups has been shown above, the structure follows from <a href="chapBib.html#biBCCN85">[CCN+85]</a>.

<div class="pcenter"><table class="GAPDocTable">
<caption class="GAPDocTable">Table: Table II: Maximal subgroups/></caption>
<tr>
<td class="tdleft">G</td>
<td class="tdleft">��(G,s)</td>
<td class="tdright">|s|</td>
<td class="tdright">see <a href="chapBib.html#biBCCN85">[CCN+85]</a></td>
</tr>
<tr>
<td class="tdleft">SL(3,4) = 3.L_3(4)</td>
<td class="tdleft">3 × L_3(2), 3 × L_3(2), 3 × L_3(2)</td>
<td class="tdright">21</td>
<td class="tdright">p. 23</td>
</tr>
<tr>
<td class="tdleft">Ω^-(8,2) = O^-_8(2)</td>
<td class="tdleft">Ω^-(4,4).2 = L_2(16).2</td>
<td class="tdright">17</td>
<td class="tdright">p. 89</td>
</tr>
<tr>
<td class="tdleft">Sp(4,4) = S_4(4)</td>
<td class="tdleft">Ω^-(4,4).2 = L_2(16).2, Sp(2,16).2 = L_2(16).2</td>
<td class="tdright">17</td>
<td class="tdright">p. 44</td>
</tr>
<tr>
<td class="tdleft">Sp(6,3) = 2.S_6(3)</td>
<td class="tdleft">(4 × U_3(3)).2, Sp(2,17).3 = 2.L_2(27).3</td>
<td class="tdright">28</td>
<td class="tdright">p. 113</td>
</tr>
<tr>
<td class="tdleft">SU(3,3) = U_3(3)</td>
<td class="tdleft">3^1+2_+:8, GU(2,3) = 4.S_4</td>
<td class="tdright">6</td>
<td class="tdright">p. 14</td>
</tr>
<tr>
<td class="tdleft">SU(3,5) = 3.U_3(5)</td>
<td class="tdleft">3 × 5^1+2_+:8, GU(2,5) = 3 × 2S_5</td>
<td class="tdright">30</td>
<td class="tdright">p. 34</td>
</tr>
<tr>
<td class="tdleft">SU(5,2) = U_5(2)</td>
<td class="tdleft">L_2(11)</td>
<td class="tdright">11</td>
<td class="tdright">p. 73</td>
</tr>
</table> 
</div>

<a id="X7B1E26D586337487" name="X7B1E26D586337487"></a>

<h5>11.4-5 Automorphism Groups of other Simple Groups – Easy Cases</h5>

We deal with automorphic extensions of those simple groups that are listed in Table I and that have been treated successfully in Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>.

For the following groups, <code class="code">ProbGenInfoAlmostSimple</code> can be used because GAP can compute their primitive permutation characters.

<div class="example"><pre>
gap> list:= [
> [ "A5", "A5.2" ],
> [ "A6", "A6.2_1" ],
> [ "A6", "A6.2_2" ],
> [ "A6", "A6.2_3" ],
> [ "A7", "A7.2" ],
> [ "A8", "A8.2" ],
> [ "A9", "A9.2" ],
> [ "A11", "A11.2" ],
> [ "L3(2)", "L3(2).2" ],
> [ "L3(3)", "L3(3).2" ],
> [ "L3(4)", "L3(4).2_1" ],
> [ "L3(4)", "L3(4).2_2" ],
> [ "L3(4)", "L3(4).2_3" ],
> [ "L3(4)", "L3(4).3" ],
> [ "S4(4)", "S4(4).2" ],
> [ "U3(3)", "U3(3).2" ],
> [ "U3(5)", "U3(5).2" ],
> [ "U3(5)", "U3(5).3" ],
> [ "U4(2)", "U4(2).2" ],
> [ "U4(3)", "U4(3).2_1" ],
> [ "U4(3)", "U4(3).2_3" ],
> ];;
gap> autinfo:= [];;
gap> fails:= [];;
gap> for pair in list do
> tbl:= CharacterTable( pair[1] );
> tblG:= CharacterTable( pair[2] );
> info:= ProbGenInfoSimple( tbl );
> spos:= List( info[4], x -> Position( AtlasClassNames( tbl ), x ) );
> Add( autinfo, ProbGenInfoAlmostSimple( tbl, tblG, spos ) );
> od;
gap> PrintFormattedArray( autinfo );
 A5.2 0 [ "5AB" ] [ 1 ]
 A6.2_1 2/3 [ "5AB" ] [ 2 ]
 A6.2_2 1/6 [ "5A" ] [ 1 ]
 A6.2_3 0 [ "5AB" ] [ 1 ]
 A7.2 1/15 [ "7AB" ] [ 1 ]
 A8.2 13/28 [ "15AB" ] [ 1 ]
 A9.2 1/4 [ "9AB" ] [ 1 ]
 A11.2 1/945 [ "11AB" ] [ 1 ]
 L3(2).2 1/4 [ "7AB" ] [ 1 ]
 L3(3).2 1/18 [ "13AB" ] [ 1 ]
 L3(4).2_1 3/10 [ "7AB" ] [ 3 ]
 L3(4).2_2 11/60 [ "7A" ] [ 1 ]
 L3(4).2_3 1/12 [ "7AB" ] [ 1 ]
 L3(4).3 1/64 [ "7A" ] [ 1 ]
 S4(4).2 0 [ "17AB" ] [ 2 ]
 U3(3).2 2/7 [ "6A", "12AB" ] [ 2, 2 ]
 U3(5).2 2/21 [ "10A" ] [ 2 ]
 U3(5).3 46/525 [ "10A" ] [ 2 ]
 U4(2).2 16/45 [ "12AB" ] [ 2 ]
 U4(3).2_1 76/135 [ "7A" ] [ 3 ]
 U4(3).2_3 31/162 [ "7AB" ] [ 3 ]
</pre></div>

We see that from this list, the two groups A_6.2_1 = S_6 and U_4(3).2_1 require further computations (see Sections <a href="chap11.html#X82C3B4287B0C7BEE">11.5-3</a> and <a href="chap11.html#X7D738BE5804CF22E">11.5-24</a>, respectively) because the bound in the second column is larger than 1/2.

Also U_4(2) is not done by the above, because in <a href="chapBib.html#biBBGK">[BGK08, Table 4]</a>, an element s of order 9 is chosen for the simple group, see Section <a href="chap11.html#X7A03F8EC839AF0B5">11.5-23</a>.

Finally, we deal with automorphic extensions of the groups L_4(3), O_8^-(2), S_6(3), and U_5(2).

For S = L_4(3) and s ∈ S of order 20, we have ��(S,s) = { (4 × A_6):2 }, the subgroup has index 2106, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 69]</a>.

<div class="example"><pre>
gap> t:= CharacterTable( "L4(3)" );;
gap> prim:= PrimitivePermutationCharacters( t );;
gap> spos:= Position( AtlasClassNames( t ), "20A" );;
gap> prim:= Filtered( prim, x -> x[ spos ] <> 0 );
[ Character( CharacterTable( "L4(3)" ),
 [ 2106, 106, 42, 0, 27, 27, 0, 46, 6, 6, 1, 7, 7, 0, 3, 3, 0, 0, 0,
 1, 1, 1, 0, 0, 0, 0, 0, 1, 1 ] ) ]
</pre></div>

For the three automorphic extensions of the structure G = S.2, we compute the extensions of the permutation character, and the bounds σ^'(G,s).

<div class="example"><pre>
gap> for name in [ "L4(3).2_1", "L4(3).2_2", "L4(3).2_3" ] do
> t2:= CharacterTable( name );
> map:= InverseMap( GetFusionMap( t, t2 ) );
> torso:= List( prim, pi -> CompositionMaps( pi, map ) );
> ext:= Concatenation( List( torso,
> x -> PermChars( t2, rec( torso:= x ) ) ) );
> sigma:= ApproxP( ext, Position( OrdersClassRepresentatives( t2 ), 20 ) );
> max:= Maximum( sigma{ Difference( PositionsProperty(
> OrdersClassRepresentatives( t2 ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t2 ) ) } );
> Print( name, ":\n", ext, "\n", max, "\n" );
> od;
L4(3).2_1:
[ Character( CharacterTable( "L4(3).2_1" ),
 [ 2106, 106, 42, 0, 27, 0, 46, 6, 6, 1, 7, 0, 3, 0, 0, 1, 1, 0,
 0, 0, 0, 0, 1, 1, 0, 4, 0, 0, 6, 6, 6, 6, 2, 0, 0, 0, 0, 0, 0,
 1, 1, 1, 1 ] ) ]
0
L4(3).2_2:
[ Character( CharacterTable( "L4(3).2_2" ),
 [ 2106, 106, 42, 0, 27, 27, 0, 46, 6, 6, 1, 7, 7, 0, 3, 3, 0, 0,
 0, 1, 1, 1, 0, 0, 0, 1, 306, 306, 42, 6, 10, 10, 0, 0, 15, 15,
 3, 3, 3, 3, 0, 0, 1, 1, 0, 1, 1, 0, 0 ] ) ]
17/117
L4(3).2_3:
[ Character( CharacterTable( "L4(3).2_3" ),
 [ 2106, 106, 42, 0, 27, 0, 46, 6, 6, 1, 7, 0, 3, 0, 0, 1, 1, 0,
 0, 0, 1, 36, 0, 0, 6, 6, 2, 2, 2, 1, 1, 0, 0, 0 ] ) ]
2/117
</pre></div>

For S = O_8^-(2) and s ∈ S of order 17, we have ��(S,s) = { L_2(16).2 }, the subgroup extends to L_2(16).4 in S.2, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 89]</a>. This is a non-split extension, so σ^'(S.2,s) = 0 holds.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "O8-(2).2" ), "17AB",
> [ CharacterTable( "L2(16).4" ) ], [ 1 ], "outer" );
0
</pre></div>

For S = S_6(3) and s ∈ S irreducible of order 14, we have ��(S,s) = { (2 × U_3(3)).2, L_2(27).3 }. In G = S.2, the subgroups extend to (4 × U_3(3)).2 and L_2(27).6, respectively, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 113]</a>. In order to show that σ^'(G,s) = 7/3240 holds, we compute the primitive permutation characters of S (cf. Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>) and the unique extensions to G of those which are nonzero on s.

<div class="example"><pre>
gap> t:= CharacterTable( "S6(3)" );;
gap> t2:= CharacterTable( "S6(3).2" );;
gap> prim:= PrimitivePermutationCharacters( t );;
gap> spos:= Position( AtlasClassNames( t ), "14A" );;
gap> prim:= Filtered( prim, x -> x[ spos ] <> 0 );;
gap> map:= InverseMap( GetFusionMap( t, t2 ) );;
gap> torso:= List( prim, pi -> CompositionMaps( pi, map ) );;
gap> ext:= List( torso, pi -> PermChars( t2, rec( torso:= pi ) ) );
[ [ Character( CharacterTable( "S6(3).2" ),
 [ 155520, 0, 288, 0, 0, 0, 216, 54, 0, 0, 0, 0, 0, 0, 0, 0, 0,
 0, 0, 0, 0, 6, 1, 0, 0, 0, 0, 0, 6, 0, 0, 0, 0, 0, 0, 0, 0,
 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 144, 288, 0, 0, 0,
 6, 0, 0, 0, 0, 0, 0, 0, 0, 6, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0,
 0 ] ) ],
 [ Character( CharacterTable( "S6(3).2" ),
 [ 189540, 1620, 568, 0, 486, 0, 0, 27, 540, 84, 24, 0, 0, 0, 0,
 0, 54, 0, 0, 10, 0, 7, 1, 6, 6, 0, 0, 0, 0, 0, 0, 18, 0, 0,
 0, 6, 12, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 234, 64,
 30, 8, 0, 3, 90, 6, 0, 4, 10, 6, 0, 2, 1, 0, 0, 0, 0, 0, 0,
 0, 1, 1, 0, 0 ] ) ] ]
gap> spos:= Position( AtlasClassNames( t2 ), "14A" );;
gap> sigma:= ApproxP( Concatenation( ext ), spos );;
gap> Maximum( sigma{ Difference(
> PositionsProperty( OrdersClassRepresentatives( t2 ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t2 ) ) } );
7/3240
</pre></div>

For S = U_5(2) and s ∈ S of order 11, we have ��(S,s) = { L_2(11) }, the subgroup extends to L_2(11).2 in S.2, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 73]</a>.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "U5(2).2" ), "11AB",
> [ CharacterTable( "L2(11).2" ) ], [ 1 ], "outer" );
1/288
</pre></div>

Here we clean the workspace for the first time. This may save more than 100 megabytes, due to the fact that the caches for tables of marks and character tables are flushed.

<div class="example"><pre>
gap> CleanWorkspace();
</pre></div>

<a id="X78B856907ED13545" name="X78B856907ED13545"></a>

<h5>11.4-6 O_8^-(3)</h5>

We show that S = O_8^-(3) = Ω^-(8,3) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S of order 41, ��(S,s) consists of one group of the type L_2(81).2_1 = Ω^-(4,9).2.

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 1/567.

</dd>
</dl>
The only maximal subgroups of S containing elements of order 41 have the type L_2(81).2_1, and there is one class of these subgroups, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 141]</a>.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "O8-(3)" ), "41A",
> [ CharacterTable( "L2(81).2_1" ) ], [ 1 ] );
1/567
</pre></div>

<a id="X84AB334886DCA746" name="X84AB334886DCA746"></a>

<h5>11.4-7 O_10^+(2)</h5>

We show that S = O_10^+(2) = Ω^+(10,2) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>Fords ∈ S of order 45, ��(S,s) consists of one group of the type (A_5 × U_4(2)).2 = (Ω^-(4,2) × Ω^-(6,2)).2.

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 43/4216.

</dd>
<dt>(c)</dt>
<dd>For s as in (a), the maximal subgroup in (a) extends to S_5 × U_4(2).2 in G = Aut(S) = S.2, and σ^'(G,s) = 23/248.

</dd>
</dl>
The only maximal subgroups of S containing elements of order 45 are one class of groups H = (A_5 × U_4(2)):2, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 146]</a>. (Note that none of the groups S_8(2), O_8^+(2), L_5(2), O_8^-(2), and A_8 contains elements of order 45.) H extends to subgroups of the type H.2 = S_5 × U_4(2):2 in G, so we can compute 1_H^S = (1_H.2^G)_S.

<div class="example"><pre>
gap> ForAny( [ "S8(2)", "O8+(2)", "L5(2)", "O8-(2)", "A8" ],
> x -> 45 in OrdersClassRepresentatives( CharacterTable( x ) ) );
false
gap> t:= CharacterTable( "O10+(2)" );;
gap> t2:= CharacterTable( "O10+(2).2" );;
gap> s2:= CharacterTable( "A5.2" ) * CharacterTable( "U4(2).2" );
CharacterTable( "A5.2xU4(2).2" )
gap> pi:= PossiblePermutationCharacters( s2, t2 );;
gap> spos:= Position( OrdersClassRepresentatives( t2 ), 45 );;
gap> approx:= ApproxP( pi, spos );;
gap> Maximum( approx{ ClassPositionsOfDerivedSubgroup( t2 ) } );
43/4216
</pre></div>

Statement (c) follows from considering the outer classes of prime element order.

<div class="example"><pre>
gap> Maximum( approx{ Difference(
> PositionsProperty( OrdersClassRepresentatives( t2 ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t2 ) ) } );
23/248
</pre></div>

Alternatively, we can use <code class="code">SigmaFromMaxes</code>.

<div class="example"><pre>
gap> SigmaFromMaxes( t2, "45AB", [ s2 ], [ 1 ], "outer" );
23/248
</pre></div>

<a id="X84E3E4837BB93977" name="X84E3E4837BB93977"></a>

<h5>11.4-8 O_10^-(2)</h5>

We show that S = O_10^-(2) = Ω^-(10,2) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S of order 33, ��(S,s) consists of one group of the type 3 × U_5(2) = GU(5,2).

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 1/119.

</dd>
<dt>(c)</dt>
<dd>For s as in (a), the maximal subgroup in (a) extends to (3 × U_5(2)).2 in G, and σ^'(G,s) = 1/595.

</dd>
</dl>
The only maximal subgroups of S containing elements of order 11 have the types A_12 and 3 × U_5(2), see <a href="chapBib.html#biBCCN85">[CCN+85, p. 147]</a>. So 3 × U_5(2) is the unique class of subgroups containing elements of order 33. This shows statement (a), and statement (b) follows using <code class="code">SigmaFromMaxes</code>.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "O10-(2)" ), "33A",
> [ CharacterTable( "Cyclic", 3 ) * CharacterTable( "U5(2)" ) ], [ 1 ] );
1/119
</pre></div>

The structure of the maximal subgroup of G follows from <a href="chapBib.html#biBCCN85">[CCN+85, p. 147]</a>. We create its character table with a generic construction that is based on the fact that the outer automorphism acts nontrivially on the two direct factors; this determines the character table uniquely. (See <a href="chapBib.html#biBAuto">[Brec]</a> for details.)

<div class="example"><pre>
gap> tblG:= CharacterTable( "U5(2)" );;
gap> tblMG:= CharacterTable( "Cyclic", 3 ) * tblG;;
gap> tblGA:= CharacterTable( "U5(2).2" );;
gap> acts:= PossibleActionsForTypeMGA( tblMG, tblG, tblGA );;
gap> poss:= Concatenation( List( acts, pi ->
> PossibleCharacterTablesOfTypeMGA( tblMG, tblG, tblGA, pi,
> "(3xU5(2)).2" ) ) );
[ rec(
 MGfusMGA := [ 1, 2, 3, 4, 4, 5, 5, 6, 7, 8, 9, 10, 11, 12, 12,
 13, 13, 14, 14, 15, 15, 16, 17, 17, 18, 18, 19, 20, 21, 21,
 22, 22, 23, 23, 24, 24, 25, 25, 26, 27, 27, 28, 28, 29, 29,
 30, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43,
 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58,
 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73,
 74, 75, 76, 77, 31, 32, 33, 35, 34, 37, 36, 38, 39, 40, 41,
 42, 43, 45, 44, 47, 46, 49, 48, 51, 50, 52, 54, 53, 56, 55,
 57, 58, 60, 59, 62, 61, 64, 63, 66, 65, 68, 67, 69, 71, 70,
 73, 72, 75, 74, 77, 76 ],
 table := CharacterTable( "(3xU5(2)).2" ) ) ]
</pre></div>

Now statement (c) follows using <code class="code">SigmaFromMaxes</code>.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "O10-(2).2" ), "33AB",
> [ poss[1].table ], [ 1 ], "outer" );
1/595
</pre></div>

<a id="X8307367E7C7C3BCE" name="X8307367E7C7C3BCE"></a>

<h5>11.4-9 O_12^+(2)</h5>

We show that S = O_12^+(2) = Ω^+(12,2) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S of the type 4^- perp 8^- (i. e., s decomposes the natural 12-dimensional module for GO^+_12(2) = S.2 into an orthogonal sum of two irreducible modules of the dimensions 4 and 8, respectively) and of order 85, ��(S,s) consists of one group of the type G_8 = (Ω^-(4,2) × Ω^-(8,2)).2 and two groups of the type L_4(4).2^2 = Ω^+(6,4).2^2 that are conjugate in G = Aut(S) = S.2 = SO^+(12,2) but not conjugate in S.

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 7675/1031184.

</dd>
<dt>(c)</dt>
<dd>σ^'(G,s) = 73/1008.

</dd>
</dl>
The element s is a ppd(12,2;8)-element in the sense of <a href="chapBib.html#biBGPPS">[GPPS99]</a>, so the maximal subgroups of S that contain s are among the nine cases (2.1)–(2.9) listed in this paper; in the notation of this paper, we have q = 2, d = 12, e = 8, and r = 17. Case (2.1) does not occur for orthogonal groups and q = 2, according to <a href="chapBib.html#biBKlL90">[KL90]</a>; case (2.2) contributes a unique maximal subgroup, the stabilizer G_8 of the orthogonal decomposition; the cases (2.3), (2.4) (a), (2.5), and (2.6) (a) do not occur because r ≠ e+1 in our situation; case (2.4) (b) describes extension field type subgroups that are contained in ΓL(6,4), which yields the candidates GU(6,2).2 ≅ 3.U_6(2).S_3 –but 3.U_6(2).3 does not contain elements of order 85– and Ω^+(6,4).2^2 ≅ L_4(4).2^2 (two classes by <a href="chapBib.html#biBKlL90">[KL90, Prop. 4.3.14]</a>); the cases (2.6) (b)–(c) and (2.8) do not occur because they require d ≤ 8; case (2.7) does not occur because <a href="chapBib.html#biBGPPS">[GPPS99, Table 5]</a> contains no entry for r = 2e+1 = 17; finally, case (2.9) does not occur because it requires e ∈ { d-1, d } in the case r = 2e+1.

So we need the permutation characters of the actions on the cosets of L_4(4).2^2 (two classes) and G_8. According to <a href="chapBib.html#biBKlL90">[KL90, Prop. 4.1.6]</a>, G_8 has the structure (Ω^-(4,2) × Ω^-(8,2)).2.

Newer versions of the GAP Character Table Library contain the character table of S, but it is still easier to work with the table of G, which was already available at the times when the first version of these examples was created.

The two classes of L_4(4).2^2 type subgroups in S are fused in G. This can be seen from the fact that inducing the trivial character of a subgroup H_1 = L_4(4).2^2 of S to G yields a character ψ whose values are not all even; note that if H_1 would extend in G to a subgroup of twice the size of H_1 then ψ would be induced from a degree two character of this subgroup whose values are all even, and induction preserves this property.

<div class="example"><pre>
gap> t:= CharacterTable( "O12+(2).2" );;
gap> h1:= CharacterTable( "L4(4).2^2" );;
gap> psi:= PossiblePermutationCharacters( h1, t );;
gap> Length( psi );
1
gap> ForAny( psi[1], IsOddInt );
true
</pre></div>

The fixed element s of order 85 is contained in a member of each of the two conjugacy classes of the type L_4(4).2^2 in S, since S contains only one class of subgroups of the order 85; note that the order of the Sylow 17 centralizer (in both S and G) is not divisible by 25.

<div class="example"><pre>
gap> SizesCentralizers( t ){ PositionsProperty(
> OrdersClassRepresentatives( t ), x -> x = 17 ) } / 25;
[ 408/5, 408/5 ]
</pre></div>

This implies that the restriction of ψ to S is the sum ψ_S = π_1 + π_2, say, of the first two interesting permutation characters of S.

The subgroup G_8 of S extends to a group of the structure H_2 = Ω^-(4,2).2 × Ω^-(8,2).2 in G, inducing the trivial characters of H_2 to G yields a permutation character φ of G whose restriction to S is the third permutation character φ_S = π_3, say.

<div class="example"><pre>
gap> h2:= CharacterTable( "S5" ) * CharacterTable( "O8-(2).2" );;
gap> phi:= PossiblePermutationCharacters( h2, t );;
gap> Length( phi );
1
</pre></div>

We have π_1(1) = π_2(1) and π_1(s) = π_2(s), the latter again because S contains only one class of subgroups of order 85.

Now statement (a) follows from the fact that π_i(s) = 1 holds for 1 ≤ i ≤ 3.

<div class="example"><pre>
gap> prim:= Concatenation( psi, phi );;
gap> spos:= Position( OrdersClassRepresentatives( t ), 85 );
213
gap> List( prim, x -> x[ spos ] );
[ 2, 1 ]
</pre></div>

For statement (b), we compute σ(S,s). Note that we have to consider only classes inside S = G^', and that

σ( g, s ) = ∑_i=1^3 fracπ_i(s) ⋅ π_i(g)π_i(1) = fracψ(s) ⋅ ψ(g)ψ(1) + fracφ(s) ⋅ φ(g)φ(1)

holds for g ∈ S^×, so the characters ψ and φ are sufficient.

<div class="example"><pre>
gap> approx:= ApproxP( prim, spos );;
gap> Maximum( approx{ ClassPositionsOfDerivedSubgroup( t ) } );
7675/1031184
</pre></div>

Statement (c) follows from considering the outer involution classes. Note that by <a href="chapBib.html#biBBGK">[BGK08, Remark after Proposition 5.14]</a>, only the subgroup H_2 need to be considered, no novelties appear.

<div class="example"><pre>
gap> Maximum( approx{ Difference(
> PositionsProperty( OrdersClassRepresentatives( t ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t ) ) } );
73/1008
</pre></div>

<a id="X834FE1B58119A5FF" name="X834FE1B58119A5FF"></a>

<h5>11.4-10 O_12^-(2)</h5>

We show that S = O_12^-(2) = Ω^-(12,2) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S irreducible of order 2^6+1 = 65, ��(S,s) consists of two groups of the types U_4(4).2 = Ω^-(6,4).2 and L_2(64).3 = Ω^-(4,8).3, respectively.

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 1/1023.

</dd>
<dt>(c)</dt>
<dd>σ^'(Aut(S),s) = 1/347820.

</dd>
</dl>
By <a href="chapBib.html#biBBe00">[Ber00]</a>, ��(S,s) consists of extension field subgroups, which have the structures U_4(4).2 and L_2(64).3, respectively, and by <a href="chapBib.html#biBKlL90">[KL90, Prop. 4.3.16]</a>, there is just one class of each of these types.

Newer versions of the GAP Character Table Library contain the character table of S, but using this table for the computations is not easier than using the table of G = Aut(S) = O_12^-(2).2, which was already available at the times when the first version of these examples was created. So we compute the permutation characters π_1, π_2 of the extensions of the groups in ��(S,s) to G –these maximal subgroups have the structures U_4(4).4 and L_2(64).6, respectively– and compute the fixed point ratios of the restrictions to S.

<div class="example"><pre>
gap> t:= CharacterTable( "O12-(2).2" );;
gap> s1:= CharacterTable( "U4(4).4" );;
gap> pi1:= PossiblePermutationCharacters( s1, t );;
gap> s2:= CharacterTable( "L2(64).6" );;
gap> pi2:= PossiblePermutationCharacters( s2, t );;
gap> prim:= Concatenation( pi1, pi2 );; Length( prim );
2
</pre></div>

Now statement (a) follows from the fact that π_1(s) = π_2(s) = 1 holds.

<div class="example"><pre>
gap> spos:= Position( OrdersClassRepresentatives( t ), 65 );;
gap> List( prim, x -> x[ spos ] );
[ 1, 1 ]
</pre></div>

For statement (b), we compute σ(S,s); note that we have to consider only classes inside S = G^'.

<div class="example"><pre>
gap> approx:= ApproxP( prim, spos );;
gap> Maximum( approx{ ClassPositionsOfDerivedSubgroup( t ) } );
1/1023
</pre></div>

Statement (c) follows from the values on the outer involution classes.

<div class="example"><pre>
gap> Maximum( approx{ Difference(
> PositionsProperty( OrdersClassRepresentatives( t ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t ) ) } );
1/347820
</pre></div>

<a id="X7C5980A385C088FA" name="X7C5980A385C088FA"></a>

<h5>11.4-11 S_6(4)</h5>

We show that S = S_6(4) = Sp(6,4) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S irreducible of order 65, ��(S,s) consists of two groups of the types U_4(4).2 = Ω^-(6,4).2 and L_2(64).3 = Sp(2,64).3, respectively.

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 16/63.

</dd>
<dt>(c)</dt>
<dd>σ^'(Aut(S),s) = 0.

</dd>
</dl>
By <a href="chapBib.html#biBBe00">[Ber00]</a>, the element s is contained in maximal subgroups of the given types, and by <a href="chapBib.html#biBKlL90">[KL90, Prop. 4.3.10, 4.8.6]</a>, there is exactly one class of these subgroups.

The character tables of these two subgroups are currently not contained in the GAP Character Table Library. We compute the permutation character induced from the first subgroup as the unique character of the right degree that is combinatorially possible (cf. <a href="chapBib.html#biBBP98copy">[BP98]</a>).

<div class="example"><pre>
gap> t:= CharacterTable( "S6(4)" );;
gap> degree:= Size( t ) / ( 2 * Size( CharacterTable( "U4(4)" ) ) );;
gap> pi1:= PermChars( t, rec( torso:= [ degree ] ) );;
gap> Length( pi1 );
1
</pre></div>

The index of the second subgroup is too large for this simpleminded approach; therefore, we first restrict the set of possible irreducible constituents of the permutation character to those of 1_H^G, where H is the derived subgroup of L_2(64).3, for which the character table is available.

<div class="example"><pre>
gap> CharacterTable( "L2(64).3" ); CharacterTable( "U4(4).2" );
fail
fail
gap> s:= CharacterTable( "L2(64)" );;
gap> subpi:= PossiblePermutationCharacters( s, t );;
gap> Length( subpi );
1
gap> scp:= MatScalarProducts( t, Irr( t ), subpi );;
gap> nonzero:= PositionsProperty( scp[1], x -> x <> 0 );
[ 1, 11, 13, 14, 17, 18, 32, 33, 56, 58, 59, 73, 74, 77, 78, 79, 80,
 93, 95, 96, 103, 116, 117, 119, 120 ]
gap> const:= RationalizedMat( Irr( t ){ nonzero } );;
gap> degree:= Size( t ) / ( 3 * Size( s ) );
5222400
gap> pi2:= PermChars( t, rec( torso:= [ degree ], chars:= const ) );;
gap> Length( pi2 );
1
gap> prim:= Concatenation( pi1, pi2 );;
</pre></div>

Now statement (a) follows from the fact that π_1(s) = π_2(s) = 1 holds.

<div class="example"><pre>
gap> spos:= Position( OrdersClassRepresentatives( t ), 65 );;
gap> List( prim, x -> x[ spos ] );
[ 1, 1 ]
</pre></div>

For statement (b), we compute σ(G,s).

<div class="example"><pre>
gap> Maximum( ApproxP( prim, spos ) );
16/63
</pre></div>

In order to prove statement (c), we have to consider only the extensions of the above permutation characters of S to Aut(S) ≅ S.2 (cf. <a href="chapBib.html#biBBGK">[BGK08, Section 2.2]</a>).

<div class="example"><pre>
gap> t2:= CharacterTable( "S6(4).2" );;
gap> tfust2:= GetFusionMap( t, t2 );;
gap> cand:= List( prim, x -> CompositionMaps( x, InverseMap( tfust2 ) ) );;
gap> ext:= List( cand, pi -> PermChars( t2, rec( torso:= pi ) ) );
[ [ Character( CharacterTable( "S6(4).2" ),
 [ 2016, 512, 96, 128, 32, 120, 0, 6, 16, 40, 24, 0, 8, 136, 1,
 6, 6, 1, 32, 0, 8, 6, 2, 0, 2, 0, 0, 4, 0, 16, 32, 1, 8, 2,
 6, 2, 1, 2, 4, 0, 0, 1, 6, 0, 1, 10, 0, 1, 1, 0, 10, 10, 4,
 0, 1, 0, 2, 0, 2, 1, 2, 2, 1, 1, 0, 0, 0, 1, 1, 1, 1, 0, 0,
 0, 0, 0, 32, 0, 0, 8, 0, 0, 0, 0, 8, 8, 0, 0, 0, 0, 8, 0,
 0, 0, 2, 2, 0, 2, 2, 0, 2, 2, 2, 0, 0 ] ) ],
 [ Character( CharacterTable( "S6(4).2" ),
 [ 5222400, 0, 0, 0, 1280, 0, 960, 120, 0, 0, 0, 0, 0, 0, 1600,
 0, 0, 0, 0, 0, 0, 0, 0, 0, 8, 1, 0, 0, 15, 0, 0, 0, 0, 0,
 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 10, 0, 0, 0, 0,
 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0,
 0, 0, 960, 0, 0, 0, 16, 0, 24, 12, 0, 0, 0, 0, 0, 0, 0, 0,
 0, 0, 0, 0, 4, 1, 0, 0, 3, 0, 0, 0, 0, 0 ] ) ] ]
gap> spos2:= Position( OrdersClassRepresentatives( t2 ), 65 );;
gap> sigma:= ApproxP( Concatenation( ext ), spos2 );;
gap> Maximum( approx{ Difference(
> PositionsProperty( OrdersClassRepresentatives( t2 ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t2 ) ) } );
0
</pre></div>

For the simple group, we can alternatively consider a reducible element s: 2 perp 4 of order 85, which is a multiple of the primitive prime divisor r = 17 of 4^4-1. So we have e = 4, d = 6, and q = 4, in the terminology of <a href="chapBib.html#biBGPPS">[GPPS99]</a>. Then ��(S,s) consists of two groups, of the types Ω^+(6,4).2 ≅ L_4(4).2_2 and Sp(2,4) × Sp(4,4). This can be shown by checking <a href="chapBib.html#biBGPPS">[GPPS99, Ex. 2.1–2.9]</a>. Ex. 2.1 yields the candidates Ω^±(6,4).2, but only Ω^+(6,4).2 contains elements of order 85. Ex. 2.2 yields the stabilizer of a two-dimensional subspace, which has the structure Sp(2,4) × Sp(4,4), by <a href="chapBib.html#biBKlL90">[KL90]</a>. All other cases except Ex. 2.4 (b) are excluded by the fact that r = 4e+1, and Ex. 2.4 (b) does not apply because d/gcd(d,e) is odd.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "S6(4)" ), "85A",
> [ CharacterTable( "L4(4).2_2" ),
> CharacterTable( "A5" ) * CharacterTable( "S4(4)" ) ], [ 1, 1 ] );
142/455
</pre></div>

This bound is not as good as the one obtained from the irreducible element of order 65 used above.

<div class="example"><pre>
gap> 16/63 < 142/455;
true
</pre></div>

<a id="X829EDF7F7C0BCB8E" name="X829EDF7F7C0BCB8E"></a>

<h5>11.4-12 ∗ S_6(5)</h5>

We show that S = S_6(5) = PSp(6,5) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S of the type 2 perp 4 (i. e., the preimage of s in Sp(6,5) = 2.G decomposes the natural 6-dimensional module for Sp(6,5) into an orthogonal sum of two irreducible modules of the dimensions 2 and 4, respectively) and of order 78, ��(S,s) consists of one group of the type G_2 = 2.(PSp(2,5) × PSp(4,5)).

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 9/217.

</dd>
</dl>
The order of s is a multiple of the primitive prime divisor r = 13 of 5^4-1, so we have e = 4, d = 6, and q = 5, in the terminology of <a href="chapBib.html#biBGPPS">[GPPS99]</a>. We check <a href="chapBib.html#biBGPPS">[GPPS99, Ex. 2.1–2.9]</a>. Ex. 2.1 does not apply because the classes C_5 and C_8 are empty by <a href="chapBib.html#biBKlL90">[KL90, Table 3.5.C]</a>, Ex. 2.2 yields exactly the stabilizer G_2 of a 2-dimensional subspace, Ex. 2.4 (b) does not apply because d/gcd(d,e) is odd, and all other cases are excluded by the fact that r = 3e+1.

The group G_2 has the structure 2.(PSp(2,5) × PSp(4,5)), which is a central product of Sp(2,5) ≅ 2.A_5 and Sp(4,5) = 2.S_4(5) (see <a href="chapBib.html#biBKlL90">[KL90, Prop. 4.1.3]</a>). The character table of G_2 can be derived from that of the direct product of 2.A_5 and 2.S_4(5), by factoring out the diagonal central subgroup of order two.

<div class="example"><pre>
gap> t:= CharacterTable( "S6(5)" );;
gap> s1:= CharacterTable( "2.A5" );;
gap> s2:= CharacterTable( "2.S4(5)" );;
gap> dp:= s1 * s2;
CharacterTable( "2.A5x2.S4(5)" )
gap> c:= Difference( ClassPositionsOfCentre( dp ), Union(
> GetFusionMap( s1, dp ), GetFusionMap( s2, dp ) ) );
[ 62 ]
gap> s:= dp / c;
CharacterTable( "2.A5x2.S4(5)/[ 1, 62 ]" )
</pre></div>

Now we compute σ(S,s).

<div class="example"><pre>
gap> SigmaFromMaxes( t, "78A", [ s ], [ 1 ] );
9/217
</pre></div>

<a id="X85162B297E4B67EB" name="X85162B297E4B67EB"></a>

<h5>11.4-13 S_8(3)</h5>

We show that S = S_8(3) = PSp(8,3) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S irreducible of order 41, ��(S,s) consists of one group M of the type S_4(9).2 = PSp(4,9).2.

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 1/546.

</dd>
<dt>(c)</dt>
<dd>The preimage of s in the matrix group 2.S_8(3) = Sp(8,3) can be chosen of order 82, and the preimage of M is 2.S_4(9).2 = Sp(4,9).2.

</dd>
</dl>
By <a href="chapBib.html#biBBe00">[Ber00]</a>, the only maximal subgroups of S that contain irreducible elements of order (3^4+1)/2 = 41 are of extension field type, and by <a href="chapBib.html#biBKlL90">[KL90, Prop. 4.3.10]</a>, these groups have the structure S_4(9).2 and there is exactly one class of these groups.

The group U = S_4(9) has three nontrivial outer automorphisms, the character table of the subgroup U.2 in question has the identifier <code class="code">"S4(9).2_1"</code>, which follows from the fact that the extensions of U by the other two outer automorphisms do not admit a class fusion into S.

<div class="example"><pre>
gap> t:= CharacterTable( "S8(3)" );;
gap> pi:= List( [ "S4(9).2_1", "S4(9).2_2", "S4(9).2_3" ],
> name -> PossiblePermutationCharacters(
> CharacterTable( name ), t ) );;
gap> List( pi, Length );
[ 1, 0, 0 ]
</pre></div>

Now statement (a) follows from the fact that (1_U.2)^S(s) = 1 holds.

<div class="example"><pre>
gap> spos:= Position( OrdersClassRepresentatives( t ), 41 );;
gap> pi[1][1][ spos ];
1
</pre></div>

Now we compute σ(S,s) in order to show statement (b).

<div class="example"><pre>
gap> Maximum( ApproxP( pi[1], spos ) );
1/546
</pre></div>

Statement (c) is clear from the description of extension field type subgroups in <a href="chapBib.html#biBKlL90">[KL90]</a>.

<a id="X8495C2BF7B6EFFEF" name="X8495C2BF7B6EFFEF"></a>

<h5>11.4-14 U_4(4)</h5>

We show that S = U_4(4) = SU(4,4) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S of the type 1 perp 3 (i. e., s decomposes the natural 4-dimensional module for SU(4,4) into an orthogonal sum of two irreducible modules of the dimensions 1 and 3, respectively) and of order 4^3+1 = 65, ��(S,s) consists of one group of the type G_1 = 5 × U_3(4) = GU(3,4).

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 209/3264.

</dd>
</dl>
By <a href="chapBib.html#biBMSW94">[MSW94]</a>, the only maximal subgroups of S that contain s are one class of stabilizers H ≅ 5 × U_3(4) of this decomposition, and clearly there is only one such group containing s.

Note that H has index 3264 in S, since S has two orbits on the 1-dimensional subspaces, of lengths 1105 and 3264, respectively, and elements of order 13 = 65/5 lie in the stabilizers of points in the latter orbit.

<div class="example"><pre>
gap> g:= SU(4,4);;
gap> orbs:= OrbitsDomain( g, NormedRowVectors( GF(16)^4 ), OnLines );;
gap> orblen:= List( orbs, Length );
[ 1105, 3264 ]
gap> List( orblen, x -> x mod 13 );
[ 0, 1 ]
</pre></div>

We compute the permutation character 1_G_1^S; there is exactly one combinatorially possible permutation character of degree 3264 (cf. <a href="chapBib.html#biBBP98copy">[BP98]</a>).

<div class="example"><pre>
gap> t:= CharacterTable( "U4(4)" );;
gap> pi:= PermChars( t, rec( torso:= [ orblen[2] ] ) );;
gap> Length( pi );
1
</pre></div>

Now we compute σ(S,s).

<div class="example"><pre>
gap> spos:= Position( OrdersClassRepresentatives( t ), 65 );;
gap> Maximum( ApproxP( pi, spos ) );
209/3264
</pre></div>

<a id="X7A3BB5AA83A2BDF3" name="X7A3BB5AA83A2BDF3"></a>

<h5>11.4-15 U_6(2)</h5>

We show that S = U_6(2) = PSU(6,2) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S of order 11, ��(S,s) consists of one group of the type U_5(2) = SU(5,2) and three groups of the type M_22.

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 5/21.

</dd>
<dt>(c)</dt>
<dd>The preimage of s in the matrix group SU(6,2) = 3.U_6(2) can be chosen of order 33, and the preimages of the groups in ��(S,s) have the structures 3 × U_5(2) ≅ GU(5,2) and 3.M_22, respectively.

</dd>
<dt>(d)</dt>
<dd>With s as in (a), the automorphic extensions S.2, S.3 of S satisfy σ^'(S.2,s) = 5/96 and σ^'(S.3,s) = 59/224.

</dd>
</dl>
According to the list of maximal subgroups of S in <a href="chapBib.html#biBCCN85">[CCN+85, p. 115]</a>, s is contained exactly in maximal subgroups of the types U_5(2) (one class) and M_22 (three classes).

The permutation character of the action on the cosets of U_5(2) type subgroups is uniquely determined by the character tables. We get three possibilities for the permutation character on the cosets of M_22 type subgroups; they correspond to the three classes of such subgroups, because each of these classes contains elements in exactly one of the conjugacy classes <code class="code">4C</code>, <code class="code">4D</code>, and <code class="code">4E</code> of elements in S, and these classes are fused under the outer automorphism of S of order three.

<div class="example"><pre>
gap> t:= CharacterTable( "U6(2)" );;
gap> s1:= CharacterTable( "U5(2)" );;
gap> pi1:= PossiblePermutationCharacters( s1, t );;
gap> Length( pi1 );
1
gap> s2:= CharacterTable( "M22" );;
gap> pi2:= PossiblePermutationCharacters( s2, t );
[ Character( CharacterTable( "U6(2)" ),
 [ 20736, 0, 384, 0, 0, 0, 54, 0, 0, 0, 0, 48, 0, 16, 6, 0, 0, 0, 0,
 0, 0, 6, 0, 2, 0, 0, 0, 4, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0,
 0, 0, 0, 0, 0, 0 ] ), Character( CharacterTable( "U6(2)" ),
 [ 20736, 0, 384, 0, 0, 0, 54, 0, 0, 0, 48, 0, 0, 16, 6, 0, 0, 0, 0,
 0, 0, 6, 0, 2, 0, 0, 4, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0,
 0, 0, 0, 0, 0, 0 ] ), Character( CharacterTable( "U6(2)" ),
 [ 20736, 0, 384, 0, 0, 0, 54, 0, 0, 48, 0, 0, 0, 16, 6, 0, 0, 0, 0,
 0, 0, 6, 0, 2, 0, 4, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0,
 0, 0, 0, 0, 0, 0 ] ) ]
gap> imgs:= Set( pi2, x -> Position( x, 48 ) );
[ 10, 11, 12 ]
gap> AtlasClassNames( t ){ imgs };
[ "4C", "4D", "4E" ]
gap> GetFusionMap( t, CharacterTable( "U6(2).3" ) ){ imgs };
[ 10, 10, 10 ]
gap> prim:= Concatenation( pi1, pi2 );;
</pre></div>

Now statement (a) follows from the fact that the permutation characters have the value 1 on s.

<div class="example"><pre>
gap> spos:= Position( OrdersClassRepresentatives( t ), 11 );;
gap> List( prim, x -> x[ spos ] );
[ 1, 1, 1, 1 ]
</pre></div>

For statement (b), we compute σ(S,s).

<div class="example"><pre>
gap> Maximum( ApproxP( prim, spos ) );
5/21
</pre></div>

Statement (c) follows from <a href="chapBib.html#biBCCN85">[CCN+85]</a>, plus the information that 3.U_6(2) does not contain groups of the structure 3 × M_22.

<div class="example"><pre>
gap> PossibleClassFusions(
> CharacterTable( "Cyclic", 3 ) * CharacterTable( "M22" ),
> CharacterTable( "3.U6(2)" ) );
[ ]
</pre></div>

For statement (d), we need that the relevant maximal subgroups of S.2 are U_5(2).2 and one subgroup M_22.2, and that the relevant maximal subgroup of S.3 is U_5(2) × 3, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 115]</a>.

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "U6(2).2" ), "11AB",
> [ CharacterTable( "U5(2).2" ), CharacterTable( "M22.2" ) ],
> [ 1, 1 ], "outer" );
5/96
gap> SigmaFromMaxes( CharacterTable( "U6(2).3" ), "11A",
> [ CharacterTable( "U5(2)" ) * CharacterTable( "Cyclic", 3 ) ],
> [ 1 ], "outer" );
59/224
</pre></div>

<a id="X8237B8617D6F6027" name="X8237B8617D6F6027"></a>

<h4>11.5 Computations using Groups</h4>

Before we start the computations using groups, we clean the workspace.

<div class="example"><pre>
gap> CleanWorkspace();
</pre></div>

<a id="X815320787B601000" name="X815320787B601000"></a>

<h5>11.5-1 A_2m+1, 2 ≤ m ≤ 11</h5>

For alternating groups of odd degree n = 2m+1, we choose s to be an n-cycle. The interesting cases in <a href="chapBib.html#biBBGK">[BGK08, Proposition 6.7]</a> are 5 ≤ n ≤ 23.

In each case, we compute representatives of the maximal subgroups of A_n, consider only those that contain an n-cycle, and then compute the permutation characters. Additionally, we show also the names that are used for the subgroups in the GAP Library of Transitive Groups, see <a href="chapBib.html#biBHulpkeTG">[Hul05]</a> and the documentation of this library in the GAP Reference Manual.

<div class="example"><pre>
gap> PrimitivesInfoForOddDegreeAlternatingGroup:= function( n )
> local G, max, cycle, spos, prim, nonz;
> 
> G:= AlternatingGroup( n );
> 
> # Compute representatives of the classes of maximal subgroups.
> max:= MaximalSubgroupClassReps( G );
> 
> # Omit subgroups that cannot contain an `n'-cycle.
> max:= Filtered( max, m -> IsTransitive( m, [ 1 .. n ] ) );
> 
> # Compute the permutation characters.
> cycle:= [];
> cycle[ n-1 ]:= 1;
> spos:= PositionProperty( ConjugacyClasses( CharacterTable( G ) ),
> c -> CycleStructurePerm( Representative( c ) ) = cycle );
> prim:= List( max, m -> TrivialCharacter( m )^G );
> nonz:= PositionsProperty( prim, x -> x[ spos ] <> 0 );
> 
> # Compute the subgroup names and the multiplicities.
> return rec( spos := spos,
> prim := prim{ nonz },
> grps := List( max{ nonz },
> m -> TransitiveGroup( n,
> TransitiveIdentification( m ) ) ),
> mult := List( prim{ nonz }, x -> x[ spos ] ) );
> end;;
</pre></div>

The sets ��/~(s) and the values σ(A_n,s) are as follows. For each degree in question, the first list shows names for representatives of the conjugacy classes of maximal subgroups containing a fixed n-cycle, and the second list shows the number of conjugates in each class.

<div class="example"><pre>
gap> for n in [ 5, 7 .. 23 ] do
> prim:= PrimitivesInfoForOddDegreeAlternatingGroup( n );
> bound:= Maximum( ApproxP( prim.prim, prim.spos ) );
> Print( n, ": ", prim.grps, ", ", prim.mult, ", ", bound, "\n" );
> od;
5: [ D(5) = 5:2 ], [ 1 ], 1/3
7: [ L(7) = L(3,2), L(7) = L(3,2) ], [ 1, 1 ], 2/5
9: [ 1/2[S(3)^3]S(3), L(9):3=P|L(2,8) ], [ 1, 3 ], 9/35
11: [ M(11), M(11) ], [ 1, 1 ], 2/105
13: [ F_78(13)=13:6, L(13)=PSL(3,3), L(13)=PSL(3,3) ], [ 1, 2, 2 ], 4/
1155
15: [ 1/2[S(3)^5]S(5), 1/2[S(5)^3]S(3), L(15)=A_8(15)=PSL(4,2),
 L(15)=A_8(15)=PSL(4,2) ], [ 1, 1, 1, 1 ], 29/273
17: [ L(17):4=PYL(2,16), L(17):4=PYL(2,16) ], [ 1, 1 ], 2/135135
19: [ F_171(19)=19:9 ], [ 1 ], 1/6098892800
21: [ t21n150, t21n161, t21n91 ], [ 1, 1, 2 ], 29/285
23: [ M(23), M(23) ], [ 1, 1 ], 2/130945815
</pre></div>

In the above output, a subgroup printed as <code class="code">1/2[S(</code>n_1<code class="code">)^</code>n_2<code class="code">]S(</code>n_2<code class="code">)</code>, <code class="code">1/2[S(</code>n_1<code class="code">)^</code>n_2<code class="code">]S(</code>n_2<code class="code">)</code>, where n = n_1 n_2 holds, denotes the intersection of A_n with the wreath product S_n_1 ≀ S_n_2 ≤ S_n. (Note that the Atlas denotes the subgroup <code class="code">1/2[S(3)^3]S(3)</code> of A_9 as 3^3:S_4.) The groups printed as <code class="code">P|L(2,8)</code> and <code class="code">PYL(2,16)</code> denote PΓL(2,8) and PΓL(2,16), respectively. And the three subgroups of A_21 have the structures (S_3 ≀ S_7) ∩ A_21, (S_7 ≀ S_3) ∩ A_21, and PGL(3,4), respectively.

Note that A_9 contains two conjugacy classes of maximal subgroups of the type PΓL(2,8) ≅ L_2(8):3, and that each 9-cycle in A_9 is contained in exactly three conjugate subgroups of this type. For n ∈ { 13, 15, 17 }, A_n contains two conjugacy classes of isomorphic maximal subgroups of linear type, and each n-cycle is contained in subgroups from each class. Finally, A_21 contains only one class of maximal subgroups of linear type.

For the two groups A_5 and A_7, the values computed above are not sufficient. See Section <a href="chap11.html#X7B5321337B28100B">11.5-2</a> and <a href="chap11.html#X85B3C7217B105D4D">11.5-4</a> for a further treatment.

The above computations look like a brute-force approach, but note that the computation of the maximal subgroups of alternating and symmetric groups in GAP uses the classification of these subgroups, and also the conjugacy classes of elements in alternating and symmetric groups can be computed cheaply.

Alternative (character-theoretic) computations for n ∈ { 5, 7, 9, 11, 13 } were shown in Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>. (A hand calculation for the case n = 19 can be found in <a href="chapBib.html#biBBW1">[BW75]</a>.)

<a id="X7B5321337B28100B" name="X7B5321337B28100B"></a>

<h5>11.5-2 A_5</h5>

We show that S = A_5 satisfies the following.

<dl>
<dt>(a)</dt>
<dd>σ(S) = 1/3, and this value is attained exactly for σ(S,s) with s of order 5.

</dd>
<dt>(b)</dt>
<dd>For s ∈ S of order 5, ��(S,s) consists of one group of the type D_10.

</dd>
<dt>(c)</dt>
<dd>P(S) = 1/3, and this value is attained exactly for P(S,s) with s of order 5.

</dd>
<dt>(d)</dt>
<dd>Each element in S together with one of (1,2)(3,4), (1,3)(2,4), (1,4)(2,3) generates a proper subgroup of S.

</dd>
<dt>(e)</dt>
<dd>Both the spread and the uniform spread of S is exactly two (see <a href="chapBib.html#biBBW1">[BW75]</a>), with s of order 5.

</dd>
</dl>
Statement (a) follows from inspection of the primitive permutation characters, cf. Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>.

<div class="example"><pre>
gap> t:= CharacterTable( "A5" );;
gap> ProbGenInfoSimple( t );
[ "A5", 1/3, 2, [ "5A" ], [ 1 ] ]
</pre></div>

Statement (b) can be read off from the primitive permutation characters, and the fact that the unique class of maximal subgroups that contain elements of order 5 consists of groups of the structure D_10, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 2]</a>.

<div class="example"><pre>
gap> OrdersClassRepresentatives( t );
[ 1, 2, 3, 5, 5 ]
gap> PrimitivePermutationCharacters( t );
[ Character( CharacterTable( "A5" ), [ 5, 1, 2, 0, 0 ] ),
 Character( CharacterTable( "A5" ), [ 6, 2, 0, 1, 1 ] ),
 Character( CharacterTable( "A5" ), [ 10, 2, 1, 0, 0 ] ) ]
</pre></div>

For statement (c), we compute that for all nonidentity elements s ∈ S and involutions g ∈ S, P(g,s) ≥ 1/3 holds, with equality if and only if s has order 5. We actually compute, for class representatives s, the proportion of involutions g such that ⟨ g, s ⟩ ≠ S holds.

<div class="example"><pre>
gap> g:= AlternatingGroup( 5 );;
gap> inv:= g.1^2 * g.2;
(1,4)(2,5)
gap> cclreps:= List( ConjugacyClasses( g ), Representative );;
gap> SortParallel( List( cclreps, Order ), cclreps );
gap> List( cclreps, Order );
[ 1, 2, 3, 5, 5 ]
gap> Size( ConjugacyClass( g, inv ) );
15
gap> prop:= List( cclreps,
> r -> RatioOfNongenerationTransPermGroup( g, inv, r ) );
[ 1, 1, 3/5, 1/3, 1/3 ]
gap> Minimum( prop );
1/3
</pre></div>

Statement (d) follows by explicit computations.

<div class="example"><pre>
gap> triple:= [ (1,2)(3,4), (1,3)(2,4), (1,4)(2,3) ];;
gap> CommonGeneratorWithGivenElements( g, cclreps, triple );
fail
</pre></div>

As for statement (e), we know from (a) that the uniform spread of S is at least two, and from (d) that the spread is less than three.

<a id="X82C3B4287B0C7BEE" name="X82C3B4287B0C7BEE"></a>

<h5>11.5-3 A_6</h5>

We show that S = A_6 satisfies the following.

<dl>
<dt>(a)</dt>
<dd>σ(S) = 2/3, and this value is attained exactly for σ(S,s) with s of order 5.

</dd>
<dt>(b)</dt>
<dd>For s of order 5, ��(S,s) consists of two nonconjugate groups of the type A_5.

</dd>
<dt>(c)</dt>
<dd>P(S) = 5/9, and this value is attained exactly for P(S,s) with s of order 5.

</dd>
<dt>(d)</dt>
<dd>Each element in S together with one of (1,2)(3,4), (1,3)(2,4), (1,4)(2,3) generates a proper subgroup of S.

</dd>
<dt>(e)</dt>
<dd>Both the spread and the uniform spread of S is exactly two (see <a href="chapBib.html#biBBW1">[BW75]</a>), with s of order 4.

</dd>
<dt>(f)</dt>
<dd>For x, y ∈ S_6^×, there is s ∈ S_6 such that S ⊆ ⟨ x, s ⟩ ∩ ⟨ y, s ⟩. It is not possible to find s ∈ S with this property, or s in a prescribed conjugacy class of S_6.

</dd>
<dt>(g)</dt>
<dd>σ( PGL(2,9) ) = 1/6 and σ( M_10 ) = 1/9, with s of order 10 and 8, respectively.

</dd>
</dl>
(Note that in this example, the optimal choice of s for P(S) cannot be used to obtain the result on the exact spread.)

Statement (a) follows from inspection of the primitive permutation characters, cf. Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>.

<div class="example"><pre>
gap> t:= CharacterTable( "A6" );;
gap> ProbGenInfoSimple( t );
[ "A6", 2/3, 1, [ "5A" ], [ 2 ] ]
</pre></div>

Statement (b) can be read off from the permutation characters, and the fact that the two classes of maximal subgroups that contain elements of order 5 consist of groups of the structure A_5, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 4]</a>.

<div class="example"><pre>
gap> OrdersClassRepresentatives( t );
[ 1, 2, 3, 3, 4, 5, 5 ]
gap> prim:= PrimitivePermutationCharacters( t );
[ Character( CharacterTable( "A6" ), [ 6, 2, 3, 0, 0, 1, 1 ] ),
 Character( CharacterTable( "A6" ), [ 6, 2, 0, 3, 0, 1, 1 ] ),
 Character( CharacterTable( "A6" ), [ 10, 2, 1, 1, 2, 0, 0 ] ),
 Character( CharacterTable( "A6" ), [ 15, 3, 3, 0, 1, 0, 0 ] ),
 Character( CharacterTable( "A6" ), [ 15, 3, 0, 3, 1, 0, 0 ] ) ]
</pre></div>

For statement (c), we first compute that for all nonidentity elements s ∈ S and involutions g ∈ S, P(g,s) ≥ 5/9 holds, with equality if and only if s has order 5. We actually compute, for class representatives s, the proportion of involutions g such that ⟨ g, s ⟩ ≠ S holds.

<div class="example"><pre>
gap> S:= AlternatingGroup( 6 );;
gap> inv:= (S.1*S.2)^2;
(1,3)(2,5)
gap> cclreps:= List( ConjugacyClasses( S ), Representative );;
gap> SortParallel( List( cclreps, Order ), cclreps );
gap> List( cclreps, Order );
[ 1, 2, 3, 3, 4, 5, 5 ]
gap> C:= ConjugacyClass( S, inv );;
gap> Size( C );
45
gap> prop:= List( cclreps,
> r -> RatioOfNongenerationTransPermGroup( S, inv, r ) );
[ 1, 1, 1, 1, 29/45, 5/9, 5/9 ]
gap> Minimum( prop );
5/9
</pre></div>

Now statement (c) follows from the fact that for g ∈ S of order larger than two, σ(S,g) ≤ 1/2 < 5/9 holds.

<div class="example"><pre>
gap> ApproxP( prim, 6 );
[ 0, 2/3, 1/2, 1/2, 0, 1/3, 1/3 ]
</pre></div>

Statement (d) follows by explicit computations.

<div class="example"><pre>
gap> triple:= [ (1,2)(3,4), (1,3)(2,4), (1,4)(2,3) ];;
gap> CommonGeneratorWithGivenElements( S, cclreps, triple );
fail
</pre></div>

An alternative triple to that in statement (d) is the one given in <a href="chapBib.html#biBBW1">[BW75]</a>.

<div class="example"><pre>
gap> triple:= [ (1,3)(2,4), (1,5)(2,6), (3,6)(4,5) ];;
gap> CommonGeneratorWithGivenElements( S, cclreps, triple );
fail
</pre></div>

Of course we can also construct such a triple, as follows.

<div class="example"><pre>
gap> TripleWithProperty( [ [ inv ], C, C ],
> l -> ForAll( S, elm ->
> ForAny( l, x -> not IsGeneratorsOfTransPermGroup( S, [ elm, x ] ) ) ) );
[ (1,3)(2,5), (1,3)(2,6), (1,3)(2,4) ]
</pre></div>

For statement (e), we use the random approach described in Section <a href="chap11.html#X83DACCF07EF62FAE">11.3-3</a>.

<div class="example"><pre>
gap> s:= (1,2,3,4)(5,6);;
gap> reps:= Filtered( cclreps, x -> Order( x ) > 1 );;
gap> ResetGlobalRandomNumberGenerators();
gap> for pair in UnorderedTuples( reps, 2 ) do
> if RandomCheckUniformSpread( S, pair, s, 40 ) <> true then
> Print( "#E nongeneration!\n" );
> fi;
> od;
</pre></div>

We get no output, so a suitable element of order 4 works in all cases. Note that we cannot use an element of order 5, because it fixes a point in the natural permutation representation, and we may take x_1 = (1,2,3) and x_2 = (4,5,6). With this argument, only elements of order 4 and double 3-cycles are possible choices for s, and the latter are excluded by the fact that an outer automorphism maps the class of double s-cycles in A_6 to the class of 3-cycles. So no element in A_6 of order different from 4 works.

Next we show statement (f). Already in A_6.2_1 = S_6, elements s of order 4 do in general not work because they do not generate with transpositions.

<div class="example"><pre>
gap> G:= SymmetricGroup( 6 );;
gap> RatioOfNongenerationTransPermGroup( G, s, (1,2) );
1
</pre></div>

Also, choosing s from a prescribed conjugacy class of S_6 (that is, also s outside A_6 is allowed) with the property that A_6 ⊆ ⟨ x, s ⟩ ∩ ⟨ y, s ⟩ is not possible. Note that only 6-cycles are possible for s if x and y are commuting transpositions, and –applying the outer automorphism– no 6-cycle works for two commuting fixed-point free involutions. (The group is small enough for a brute force test.)

<div class="example"><pre>
gap> goods:= Filtered( Elements( G ),
> s -> IsGeneratorsOfTransPermGroup( G, [ s, (1,2) ] ) and
> IsGeneratorsOfTransPermGroup( G, [ s, (3,4) ] ) );;
gap> Collected( List( goods, CycleStructurePerm ) );
[ [ [ ,,,, 1 ], 24 ] ]
gap> goods:= Filtered( Elements( G ),
> s -> IsGeneratorsOfTransPermGroup( G, [ s, (1,2)(3,4)(5,6) ] ) and
> IsGeneratorsOfTransPermGroup( G, [ s, (1,3)(2,4)(5,6) ] ) );;
gap> Collected( List( goods, CycleStructurePerm ) );
[ [ [ 1, 1 ], 24 ] ]
</pre></div>

However, for each pair of nonidentity element x, y ∈ S_6, there is s ∈ S_6 such that ⟨ x, s ⟩ and ⟨ y, s ⟩ both contain A_6. (If s works for the pair (x,y) then s^g works for (x^g,y^g), so it is sufficient to consider only orbit representatives (x,y) under the conjugation action of G on pairs. Thus we check conjugacy class representatives x and, for fixed x, representatives of orbits of C_G(x) on the classes y^G, i. e., representatives of C_G(y)-C_G(x)-double cosets in G. Moreover, clearly we can restrict the checks to elements x, y of prime order.)

<div class="example"><pre>
gap> Sgens:= GeneratorsOfGroup( S );;
gap> primord:= Filtered( List( ConjugacyClasses( G ), Representative ),
> x -> IsPrimeInt( Order( x ) ) );;
gap> for x in primord do
> for y in primord do
> for pair in DoubleCosetRepsAndSizes( G, Centralizer( G, y ),
> Centralizer( G, x ) ) do
> if not ForAny( G, s -> IsSubset( Group( x,s ), S ) and 
> IsSubset( Group( y^pair[1], s ), S ) ) then
> Error( [ x, y ] );
> fi;
> od;
> od;
> od;
</pre></div>

In other words, the spread of S_6 is 2 but the uniform spread of S_6 is not 2 but only 1.

We cannot always find s ∈ A_6 with the required property: If x is a transposition then any s with S ⊆⟨ x, s ⟩ must be a 5-cycle.

<div class="example"><pre>
gap> filt:= Filtered( S, s -> IsSubset( Group( (1,2), s ), S ) );;
gap> Collected( List( filt, Order ) );
[ [ 5, 48 ] ]
</pre></div>

Moreover, clearly such s fixes one of the moved points of x, so we may prescribe a transposition y ≠ x that commutes with x, it satisfies S ⊈⟨ y, s ⟩.

For the other two automorphic extensions A_6.2_2 = PGL(2,9) and A_6.2_3 = M_10, we compute the character-theoretic bounds σ(A_6.2_2) = 1/6 and σ(A_6.2_3) = 1/9, which shows statement (g).

<div class="example"><pre>
gap> ProbGenInfoSimple( CharacterTable( "A6.2_2" ) );
[ "A6.2_2", 1/6, 5, [ "10A" ], [ 1 ] ]
gap> ProbGenInfoSimple( CharacterTable( "A6.2_3" ) );
[ "A6.2_3", 1/9, 8, [ "8C" ], [ 1 ] ]
</pre></div>

Note that σ^'( PGL(2,9), s ) = 1/6, with s of order 5, and σ^'( M_10, s ) = 0 for any s ∈ A_6 since M_10 is a non-split extension of A_6.

<div class="example"><pre>
gap> t:= CharacterTable( "A6" );;
gap> t2:= CharacterTable( "A6.2_2" );;
gap> spos:= PositionsProperty( OrdersClassRepresentatives( t ), x -> x = 5 );;
gap> ProbGenInfoAlmostSimple( t, t2, spos );
[ "A6.2_2", 1/6, [ "5A", "5B" ], [ 1, 1 ] ]
</pre></div>

<a id="X85B3C7217B105D4D" name="X85B3C7217B105D4D"></a>

<h5>11.5-4 A_7</h5>

We show that S = A_7 satisfies the following.

<dl>
<dt>(a)</dt>
<dd>σ(S) = 2/5, and this value is attained exactly for σ(S,s) with s of order 7.

</dd>
<dt>(b)</dt>
<dd>For s of order 7, ��(S,s) consists of two nonconjugate subgroups of the type L_2(7).

</dd>
<dt>(c)</dt>
<dd>P(S) = 2/5, and this value is attained exactly for P(S,s) with s of order 7.

</dd>
<dt>(d)</dt>
<dd>The uniform spread of S is exactly three, with s of order 7.

</dd>
</dl>
Statement (a) follows from inspection of the primitive permutation characters, cf. Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>.

<div class="example"><pre>
gap> t:= CharacterTable( "A7" );;
gap> ProbGenInfoSimple( t );
[ "A7", 2/5, 2, [ "7A" ], [ 2 ] ]
</pre></div>

Statement (b) can be read off from the permutation characters, and the fact that the two classes of maximal subgroups that contain elements of order 7 consist of groups of the structure L_2(7), see <a href="chapBib.html#biBCCN85">[CCN+85, p. 10]</a>.

<div class="example"><pre>
gap> OrdersClassRepresentatives( t );
[ 1, 2, 3, 3, 4, 5, 6, 7, 7 ]
gap> prim:= PrimitivePermutationCharacters( t );
[ Character( CharacterTable( "A7" ), [ 7, 3, 4, 1, 1, 2, 0, 0, 0 ] ),
 Character( CharacterTable( "A7" ), [ 15, 3, 0, 3, 1, 0, 0, 1, 1 ] ),
 Character( CharacterTable( "A7" ), [ 15, 3, 0, 3, 1, 0, 0, 1, 1 ] ),
 Character( CharacterTable( "A7" ), [ 21, 5, 6, 0, 1, 1, 2, 0, 0 ] ),
 Character( CharacterTable( "A7" ), [ 35, 7, 5, 2, 1, 0, 1, 0, 0 ] )
]
</pre></div>

For statement (c), we compute that for all nonidentity elements s ∈ S and involutions g ∈ S, P(g,s) ≥ 2/5 holds, with equality if and only if s has order 7. We actually compute, for class representatives s, the proportion of involutions g such that ⟨ g, s ⟩ ≠ S holds.

<div class="example"><pre>
gap> g:= AlternatingGroup( 7 );;
gap> inv:= (g.1^3*g.2)^3;
(2,6)(3,7)
gap> ccl:= List( ConjugacyClasses( g ), Representative );;
gap> SortParallel( List( ccl, Order ), ccl );
gap> List( ccl, Order );
[ 1, 2, 3, 3, 4, 5, 6, 7, 7 ]
gap> Size( ConjugacyClass( g, inv ) );
105
gap> prop:= List( ccl, r -> RatioOfNongenerationTransPermGroup( g, inv, r ) );
[ 1, 1, 1, 1, 89/105, 17/21, 19/35, 2/5, 2/5 ]
gap> Minimum( prop );
2/5
</pre></div>

For statement (d), we use the random approach described in Section <a href="chap11.html#X83DACCF07EF62FAE">11.3-3</a>. By the character-theoretic bounds, it suffices to consider triples of elements in the classes <code class="code">2A</code> or <code class="code">3B</code>.

<div class="example"><pre>
gap> OrdersClassRepresentatives( t );
[ 1, 2, 3, 3, 4, 5, 6, 7, 7 ]
gap> spos:= Position( OrdersClassRepresentatives( t ), 7 );;
gap> SizesCentralizers( t );
[ 2520, 24, 36, 9, 4, 5, 12, 7, 7 ]
gap> ApproxP( prim, spos );
[ 0, 2/5, 0, 2/5, 2/15, 0, 0, 2/15, 2/15 ]
gap> s:= (1,2,3,4,5,6,7);;
gap> 3B:= (1,2,3)(4,5,6);;
gap> C3B:= ConjugacyClass( g, 3B );;
gap> Size( C3B );
280
gap> ResetGlobalRandomNumberGenerators();
gap> for triple in UnorderedTuples( [ inv, 3B ], 3 ) do
> if RandomCheckUniformSpread( g, triple, s, 80 ) <> true then
> Print( "#E nongeneration!\n" );
> fi;
> od;
</pre></div>

We get no output, so the uniform spread of S is at least three.

Alternatively, we can use the lemma from Section <a href="chap11.html#X79D7312484E78274">11.2-2</a>; this approach is technically more involved but faster. We work with the diagonal product of the two degree 15 representations of S, which is constructed from the information stored in the GAP Library of Tables of Marks.

<div class="example"><pre>
gap> tom:= TableOfMarks( "A7" );;
gap> a7:= UnderlyingGroup( tom );;
gap> tommaxes:= MaximalSubgroupsTom( tom );
[ [ 39, 38, 37, 36, 35 ], [ 7, 15, 15, 21, 35 ] ]
gap> index15:= List( tommaxes[1]{ [ 2, 3 ] },
> i -> RepresentativeTom( tom, i ) );
[ Group([ (1,3)(2,7), (1,5,7)(3,4,6) ]),
 Group([ (1,4)(2,3), (2,4,6)(3,5,7) ]) ]
gap> deg15:= List( index15, s -> RightTransversal( a7, s ) );;
gap> reps:= List( deg15, l -> Action( a7, l, OnRight ) );
[ Group([ (1,5,7)(2,9,10)(3,11,4)(6,12,8)(13,14,15), (1,8,15,5,12)
 (2,13,11,3,10)(4,14,9,7,6) ]),
 Group([ (1,2,3)(4,6,5)(7,8,9)(10,12,11)(13,15,14), (1,12,3,13,10)
 (2,9,15,4,11)(5,6,14,7,8) ]) ]
gap> g:= DiagonalProductOfPermGroups( reps );;
gap> ResetGlobalRandomNumberGenerators();
gap> repeat s:= Random( g );
> until Order( s ) = 7;
gap> NrMovedPoints( s );
28
gap> mpg:= MovedPoints( g );;
gap> fixs:= Difference( mpg, MovedPoints( s ) );;
gap> orb_s:= Orbit( g, fixs, OnSets );;
gap> Length( orb_s );
120
gap> SizesCentralizers( t );
[ 2520, 24, 36, 9, 4, 5, 12, 7, 7 ]
gap> repeat 2a:= Random( g ); until Order( 2a ) = 2;
gap> repeat 3b:= Random( g );
> until Order( 3b ) = 3 and Size( Centralizer( g, 3b ) ) = 9;
gap> orb2a:= Orbit( g, Difference( mpg, MovedPoints( 2a ) ), OnSets );;
gap> orb3b:= Orbit( g, Difference( mpg, MovedPoints( 3b ) ), OnSets );;
gap> orb2aor3b:= Union( orb2a, orb3b );;
gap> TripleWithProperty( [ [ orb2a[1], orb3b[1] ], orb2aor3b, orb2aor3b ],
> l -> ForAll( orb_s,
> f -> not IsEmpty( Intersection( Union( l ), f ) ) ) );
fail
</pre></div>

It remains to show that for any choice of s ∈ S, a quadruple of elements in S^× exists such that s generates a proper subgroup of S together with at least one of these elements.

First we observe (without using GAP) that there is a pair of 3-cycles whose fixed points cover the seven points of the natural permutation representation. This implies the statement for all elements s ∈ S that fix a point in this representation. So it remains to consider elements s of the orders six and seven.

For the order seven element, the above setup and the lemma from Section <a href="chap11.html#X79D7312484E78274">11.2-2</a> can be used.

<div class="example"><pre>
gap> QuadrupleWithProperty( [ [ orb2a[1] ], orb2a, orb2a, orb2a ],
> l -> ForAll( orb_s,
> f -> not IsEmpty( Intersection( Union( l ), f ) ) ) );
[ [ 2, 5, 12, 18, 19, 26 ], [ 7, 8, 9, 16, 21, 25 ],
 [ 1, 6, 10, 17, 20, 27 ], [ 13, 14, 15, 28, 29, 30 ] ]
</pre></div>

For the order six element, we use the diagonal product of the primitive permutation representations of the degrees 21 and 35.

<div class="example"><pre>
gap> has6A:= List( tommaxes[1]{ [ 4, 5 ] },
> i -> RepresentativeTom( tom, i ) );
[ Group([ (1,2)(3,7), (2,6,5,4)(3,7) ]),
 Group([ (2,3)(5,7), (1,2)(4,5,6,7), (2,3)(5,6) ]) ]
gap> trans:= List( has6A, s -> RightTransversal( a7, s ) );;
gap> reps:= List( trans, l -> Action( a7, l, OnRight ) );
[ Group([ (1,16,12)(2,17,13)(3,18,11)(4,19,14)(15,20,21), (1,4,7,9,10)
 (2,5,8,3,6)(11,12,15,14,13)(16,20,19,17,18) ]),
 Group([ (2,16,6)(3,17,7)(4,18,8)(5,19,9)(10,20,26)(11,21,27)
 (12,22,28)(13,23,29)(14,24,30)(15,25,31), (1,2,3,4,5)
 (6,10,13,15,9)(7,11,14,8,12)(16,20,23,25,19)(17,21,24,18,22)
 (26,32,35,31,28)(27,33,29,34,30) ]) ]
gap> g:= DiagonalProductOfPermGroups( reps );;
gap> repeat s:= Random( g );
> until Order( s ) = 6;
gap> NrMovedPoints( s );
53
gap> mpg:= MovedPoints( g );;
gap> fixs:= Difference( mpg, MovedPoints( s ) );;
gap> orb_s:= Orbit( g, fixs, OnSets );;
gap> Length( orb_s );
105
gap> repeat 3a:= Random( g );
> until Order( 3a ) = 3 and Size( Centralizer( g, 3a ) ) = 36;
gap> orb3a:= Orbit( g, Difference( mpg, MovedPoints( 3a ) ), OnSets );;
gap> Length( orb3a );
35
gap> TripleWithProperty( [ [ orb3a[1] ], orb3a, orb3a ],
> l -> ForAll( orb_s,
> f -> not IsEmpty( Intersection( Union( l ), f ) ) ) );
[ [ 1, 4, 6, 12, 14, 15, 34, 37, 40, 43, 49 ],
 [ 1, 4, 6, 16, 19, 20, 27, 30, 33, 44, 49 ],
 [ 2, 3, 4, 5, 7, 9, 26, 47, 48, 50, 53 ] ]
</pre></div>

So we have found not only a quadruple but even a triple of 3-cycles that excludes candidates s of order six.

<a id="X84EA645A82E2BAFB" name="X84EA645A82E2BAFB"></a>

<h5>11.5-5 L_d(q)</h5>

In the treatment of small dimensional linear groups S = SL(d,q), <a href="chapBib.html#biBBGK">[BGK08]</a> uses a Singer element s of order (q^d-1)/(q-1). (So the order of the corresponding element in PSL(d,q) = (q^d-1)/[(q-1) gcd(d,q-1)].) By <a href="chapBib.html#biBBe00">[Ber00]</a>, ��(S,s) consists of extension field type subgroups, except in the cases d = 2, q ∈ { 2, 5, 7, 9 }, and (d,q) = (3,4). These subgroups have the structure GL(d/p,q^p):α_q ∩ S, for prime divisors p of d, where α_q denotes the Frobenius automorphism that acts on matrices by raising each entry to the q-th power. (If q is a prime then we have GL(d/p,q^p):α_q = ΓL(d/p,q^p).) Since s acts irreducibly, it is contained in at most one conjugate of each class of extension field type subgroups (cf. <a href="chapBib.html#biBBGK">[BGK08, Lemma 2.12]</a>).

First we write a GAP function <code class="code">RelativeSigmaL</code> that takes a positive integer d and a basis B of the field extension of degree n over the field with q elements, and returns the group GL(d,q^n):α_q, as a subgroup of GL(dn,q).

<div class="example"><pre>
gap> RelativeSigmaL:= function( d, B )
> local n, F, q, glgens, diag, pi, frob, i;
> 
> n:= Length( B );
> F:= LeftActingDomain( UnderlyingLeftModule( B ) );
> q:= Size( F );
> 
> # Create the generating matrices inside the linear subgroup.
> glgens:= List( GeneratorsOfGroup( SL( d, q^n ) ),
> m -> BlownUpMat( B, m ) );
> 
> # Create the matrix of a diagonal part that maps to determinant 1.
> diag:= IdentityMat( d*n, F );
> diag{ [ 1 .. n ] }{ [ 1 .. n ] }:= BlownUpMat( B, [ [ Z(q^n)^(q-1) ] ] );
> Add( glgens, diag );
> 
> # Create the matrix that realizes the Frobenius action,
> # and adjust the determinant.
> pi:= List( B, b -> Coefficients( B, b^q ) );
> frob:= NullMat( d*n, d*n, F );
> for i in [ 0 .. d-1 ] do
> frob{ [ 1 .. n ] + i*n }{ [ 1 .. n ] + i*n }:= pi;
> od;
> diag:= IdentityMat( d*n, F );
> diag{ [ 1 .. n ] }{ [ 1 .. n ] }:= BlownUpMat( B, [ [ Z(q^n) ] ] );
> diag:= diag^LogFFE( Inverse( Determinant( frob ) ), Determinant( diag ) );
> 
> # Return the result.
> return Group( Concatenation( glgens, [ diag * frob ] ) );
> end;;
</pre></div>

The next function computes σ(SL(d,q),s), by computing the sum of μ(g,S/(GL(d/p,q^p):α_q ∩ S)), for prime divisors p of d, and taking the maximum over g ∈ S^×. The computations take place in a permutation representation of PSL(d,q).

<div class="example"><pre>
gap> ApproxPForSL:= function( d, q )
> local G, epi, PG, primes, maxes, names, ccl;
> 
> # Check whether this is an admissible case (see [Be00]).
> if ( d = 2 and q in [ 2, 5, 7, 9 ] ) or ( d = 3 and q = 4 ) then
> return fail;
> fi;
> 
> # Create the group SL(d,q), and the map to PSL(d,q).
> G:= SL( d, q );
> epi:= ActionHomomorphism( G, NormedRowVectors( GF(q)^d ), OnLines );
> PG:= ImagesSource( epi );
> 
> # Create the subgroups corresponding to the prime divisors of `d'.
> primes:= PrimeDivisors( d );
> maxes:= List( primes, p -> RelativeSigmaL( d/p,
> Basis( AsField( GF(q), GF(q^p) ) ) ) );
> names:= List( primes, p -> Concatenation( "GL(", String( d/p ), ",",
> String( q^p ), ").", String( p ) ) );
> if 2 < q then
> names:= List( names, name -> Concatenation( name, " cap G" ) );
> fi;
> 
> # Compute the conjugacy classes of prime order elements in the maxes.
> # (In order to avoid computing all conjugacy classes of these subgroups,
> # we work in Sylow subgroups.)
> ccl:= List( List( maxes, x -> ImagesSet( epi, x ) ),
> M -> ClassesOfPrimeOrder( M, PrimeDivisors( Size( M ) ),
> TrivialSubgroup( M ) ) );
> 
> return [ names, UpperBoundFixedPointRatios( PG, ccl, true )[1] ];
> end;;
</pre></div>

We apply this function to the cases that are interesting in <a href="chapBib.html#biBBGK">[BGK08, Section 5.12]</a>.

<div class="example"><pre>
gap> pairs:= [ [ 3, 2 ], [ 3, 3 ], [ 4, 2 ], [ 4, 3 ], [ 4, 4 ],
> [ 6, 2 ], [ 6, 3 ], [ 6, 4 ], [ 6, 5 ], [ 8, 2 ], [ 10, 2 ] ];;
gap> array:= [];;
gap> for pair in pairs do
> d:= pair[1]; q:= pair[2];
> approx:= ApproxPForSL( d, q );
> Add( array, [ Concatenation( "SL(", String(d), ",", String(q), ")" ),
> (q^d-1)/(q-1),
> approx[1], approx[2] ] );
> od;
gap> oldsize:= SizeScreen();;
gap> SizeScreen( [ 80 ] );;
gap> PrintFormattedArray( array );
 SL(3,2) 7 [ "GL(1,8).3" ] 1/4
 SL(3,3) 13 [ "GL(1,27).3 cap G" ] 1/24
 SL(4,2) 15 [ "GL(2,4).2" ] 3/14
 SL(4,3) 40 [ "GL(2,9).2 cap G" ] 53/1053
 SL(4,4) 85 [ "GL(2,16).2 cap G" ] 1/108
 SL(6,2) 63 [ "GL(3,4).2", "GL(2,8).3" ] 365/55552
 SL(6,3) 364 [ "GL(3,9).2 cap G", "GL(2,27).3 cap G" ] 22843/123845436
 SL(6,4) 1365 [ "GL(3,16).2 cap G", "GL(2,64).3 cap G" ] 1/85932
 SL(6,5) 3906 [ "GL(3,25).2 cap G", "GL(2,125).3 cap G" ] 1/484220
 SL(8,2) 255 [ "GL(4,4).2" ] 1/7874
 SL(10,2) 1023 [ "GL(5,4).2", "GL(2,32).5" ] 1/129794
gap> SizeScreen( oldsize );;
</pre></div>

The only missing case for <a href="chapBib.html#biBBGK">[BGK08]</a> is S = L_3(4), for which ��(S,s) consists of three groups of the type L_3(2) (see <a href="chapBib.html#biBCCN85">[CCN+85, p. 23]</a>). The group L_3(4) has been considered already in Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>, where σ(S,s) = 1/5 has been proved. Also the cases SL(3,3), SL(4,2) ≅ A_8, and SL(4,3) have been handled there.

An alternative character-theoretic proof for S = L_6(2) looks as follows. In this case, the subgroups in ��(S,s) have the types ΓL(3,4) ≅ GL(3,4).2 ≅ 3.L_3(4).3.2_2 and ΓL(2,8) ≅ GL(2,8).3 ≅ (7 × L_2(8)).3.

<div class="example"><pre>
gap> t:= CharacterTable( "L6(2)" );;
gap> s1:= CharacterTable( "3.L3(4).3.2_2" );;
gap> s2:= CharacterTable( "(7xL2(8)).3" );;
gap> SigmaFromMaxes( t, "63A", [ s1, s2 ], [ 1, 1 ] );
365/55552
</pre></div>

<a id="X855460BE787188B9" name="X855460BE787188B9"></a>

<h5>11.5-6 ∗ L_d(q) with prime d</h5>

For S = SL(d,q) with prime dimension d, and s ∈ S a Singer cycle, we have ��(S,s) = { M }, where M = N_S(⟨ s ⟩) ≅ ΓL(1,q^d) ∩ S. So

σ(g,s) = μ(g,S/M) = |g^S ∩ M|/|g^S| < |M|/|g^S| ≤ (q^d-1) ⋅ d/|g^S|

holds for any g ∈ S ∖ Z(S), which implies σ( S, s ) < max{ (q^d-1) ⋅ d/|g^S|; g ∈ S ∖ Z(S) }. The right hand side of this inequality is returned by the following function. In <a href="chapBib.html#biBBGK">[BGK08, Lemma 3.8]</a>, the global upper bound 1/q^d is derived for primes d ≥ 5.

<div class="example"><pre>
gap> UpperBoundForSL:= function( d, q )
> local G, Msize, ccl;
> 
> if not IsPrimeInt( d ) then
> Error( "<d> must be a prime" );
> fi;
> 
> G:= SL( d, q );
> Msize:= (q^d-1) * d;
> ccl:= Filtered( ConjugacyClasses( G ),
> c -> Msize mod Order( Representative( c ) ) = 0
> and Size( c ) <> 1 );
> 
> return Msize / Minimum( List( ccl, Size ) );
> end;;
</pre></div>

The interesting values are (d,q) with d ∈ { 5, 7, 11 } and q ∈ { 2, 3, 4 }, and perhaps also (d,q) ∈ { (3,2), (3,3) }. (Here we exclude SL(11,4) because writing down the conjugacy classes of this group would exceed the permitted memory.)

<div class="example"><pre>
gap> NrConjugacyClasses( SL(11,4) );
1397660
gap> pairs:= [ [ 3, 2 ], [ 3, 3 ], [ 5, 2 ], [ 5, 3 ], [ 5, 4 ],
> [ 7, 2 ], [ 7, 3 ], [ 7, 4 ],
> [ 11, 2 ], [ 11, 3 ] ];;
gap> array:= [];;
gap> for pair in pairs do
> d:= pair[1]; q:= pair[2];
> approx:= UpperBoundForSL( d, q );
> Add( array, [ Concatenation( "SL(", String(d), ",", String(q), ")" ),
> (q^d-1)/(q-1),
> approx ] );
> od;
gap> PrintFormattedArray( array );
 SL(3,2) 7 7/8
 SL(3,3) 13 3/4
 SL(5,2) 31 31/64512
 SL(5,3) 121 10/81
 SL(5,4) 341 15/256
 SL(7,2) 127 7/9142272
 SL(7,3) 1093 14/729
 SL(7,4) 5461 21/4096
 SL(11,2) 2047 2047/34112245508649716682268134604800
 SL(11,3) 88573 22/59049
</pre></div>

The exact values are clearly better than the above bounds. We compute them for L_5(2) and L_7(2). In the latter case, the class fusion of the 127:7 type subgroup M is not uniquely determined by the character tables; here we use the additional information that the elements of order 7 in M have centralizer order 49 in L_7(2). (See Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a> for the examples with d = 3.)

<div class="example"><pre>
gap> SigmaFromMaxes( CharacterTable( "L5(2)" ), "31A",
> [ CharacterTable( "31:5" ) ], [ 1 ] );
1/5376
gap> t:= CharacterTable( "L7(2)" );;
gap> s:= CharacterTable( "P:Q", [ 127, 7 ] );;
gap> pi:= PossiblePermutationCharacters( s, t );;
gap> Length( pi );
2
gap> ord7:= PositionsProperty( OrdersClassRepresentatives( t ), x -> x = 7 );
[ 38, 45, 76, 77, 83 ]
gap> sizes:= SizesCentralizers( t ){ ord7 };
[ 141120, 141120, 3528, 3528, 49 ]
gap> List( pi, x -> x[83] );
[ 42, 0 ]
gap> spos:= Position( OrdersClassRepresentatives( t ), 127 );;
gap> Maximum( ApproxP( pi{ [ 1 ] }, spos ) );
1/4388290560
</pre></div>

<a id="X7EA88CEF81962F3F" name="X7EA88CEF81962F3F"></a>

<h5>11.5-7 Automorphic Extensions of L_d(q)</h5>

For the following values of d and q, automorphic extensions G of L_d(q) had to be checked for <a href="chapBib.html#biBBGK">[BGK08, Section 5.12]</a>.

(d,q) ∈ { (3,4), (6,2), (6,3), (6,4), (6,5), (10,2) }

The first case has been treated in Section <a href="chap11.html#X7B1E26D586337487">11.4-5</a>. For the other cases, we compute σ^'(G,s) below.

In any case, the extension by a graph automorphism occurs, which can be described by mapping each matrix in SL(d,q) to its inverse transpose. If q > 2, also extensions by diagonal automorphisms occur, which are induced by conjugation with elements in GL(d,q). If q is nonprime then also extensions by field automorphisms occur, which can be described by powering the matrix entries by roots of q. Finally, products (of prime order) of these three kinds of automorphisms have to be considered.

We start with the extension G of S = SL(d,q) by a graph automorphism. G can be embedded into GL(2d,q) by representing the matrix A ∈ S as a block diagonal matrix with diagonal blocks equal to A and A^-tr, and representing the graph automorphism by a permutation matrix that interchanges the two blocks. In order to construct the field extension type subgroups of G, we have to choose the basis of the field extension in such a way that the subgroup is normalized by the permutation matrix; a sufficient condition is that the matrices of the ��_q-linear mappings induced by the basis elements are symmetric.

(We do not give a function that computes a basis with this property from the parameters d and q. Instead, we only write down the bases that we will need.)

<div class="example"><pre>
gap> SymmetricBasis:= function( q, n )
> local vectors, B, issymmetric;
> 
> if q = 2 and n = 2 then
> vectors:= [ Z(2)^0, Z(2^2) ];
> elif q = 2 and n = 3 then
> vectors:= [ Z(2)^0, Z(2^3), Z(2^3)^5 ];
> elif q = 2 and n = 5 then
> vectors:= [ Z(2)^0, Z(2^5), Z(2^5)^4, Z(2^5)^25, Z(2^5)^26 ];
> elif q = 3 and n = 2 then
> vectors:= [ Z(3)^0, Z(3^2) ];
> elif q = 3 and n = 3 then
> vectors:= [ Z(3)^0, Z(3^3)^2, Z(3^3)^7 ];
> elif q = 4 and n = 2 then
> vectors:= [ Z(2)^0, Z(2^4)^3 ];
> elif q = 4 and n = 3 then
> vectors:= [ Z(2)^0, Z(2^3), Z(2^3)^5 ];
> elif q = 5 and n = 2 then
> vectors:= [ Z(5)^0, Z(5^2)^2 ];
> elif q = 5 and n = 3 then
> vectors:= [ Z(5)^0, Z(5^3)^9, Z(5^3)^27 ];
> else
> Error( "sorry, no basis for <q> and <n> stored" );
> fi;
> 
> B:= Basis( AsField( GF(q), GF(q^n) ), vectors );
> 
> # Check that the basis really has the required property.
> issymmetric:= M -> M = TransposedMat( M );
> if not ForAll( B, b -> issymmetric( BlownUpMat( B, [ [ b ] ] ) ) ) then
> Error( "wrong basis!" );
> fi;
> 
> # Return the result.
> return B;
> end;;
</pre></div>

In later examples, we will need similar embeddings of matrices. Therefore, we provide a more general function <code class="code">EmbeddedMatrix</code> that takes a field <code class="code">F</code>, a matrix <code class="code">mat</code>, and a function <code class="code">func</code>, and returns a block diagonal matrix over <code class="code">F</code> whose diagonal blocks are <code class="code">mat</code> and <code class="code">func( mat )</code>.

<div class="example"><pre>
gap> BindGlobal( "EmbeddedMatrix", function( F, mat, func )
> local d, result;
> 
> d:= Length( mat );
> result:= NullMat( 2*d, 2*d, F );
> result{ [ 1 .. d ] }{ [ 1 .. d ] }:= mat;
> result{ [ d+1 .. 2*d ] }{ [ d+1 .. 2*d ] }:= func( mat );
> 
> return result;
> end );
</pre></div>

The following function is similar to <code class="code">ApproxPForSL</code>, the differences are that the group G in question is not SL(d,q) but the extension of this group by a graph automorphism, and that σ^'(G,s) is computed not σ(G,s).

<div class="example"><pre>
gap> ApproxPForOuterClassesInExtensionOfSLByGraphAut:= function( d, q )
> local embedG, swap, G, orb, epi, PG, Gprime, primes, maxes, ccl, names;
> 
> # Check whether this is an admissible case (see [Be00],
> # note that a graph automorphism exists only for `d > 2').
> if d = 2 or ( d = 3 and q = 4 ) then
> return fail;
> fi;
> 
> # Provide a function that constructs a block diagonal matrix.
> embedG:= mat -> EmbeddedMatrix( GF( q ), mat,
> M -> TransposedMat( M^-1 ) );
> 
> # Create the matrix that exchanges the two blocks.
> swap:= NullMat( 2*d, 2*d, GF(q) );
> swap{ [ 1 .. d ] }{ [ d+1 .. 2*d ] }:= IdentityMat( d, GF(q) );
> swap{ [ d+1 .. 2*d ] }{ [ 1 .. d ] }:= IdentityMat( d, GF(q) );
> 
> # Create the group SL(d,q).2, and the map to the projective group.
> G:= ClosureGroupDefault( Group( List( GeneratorsOfGroup( SL( d, q ) ),
> embedG ) ),
> swap );
> orb:= Orbit( G, One( G )[1], OnLines );
> epi:= ActionHomomorphism( G, orb, OnLines );
> PG:= ImagesSource( epi );
> Gprime:= DerivedSubgroup( PG );
> 
> # Create the subgroups corresponding to the prime divisors of `d'.
> primes:= PrimeDivisors( d );
> maxes:= List( primes,
> p -> ClosureGroupDefault( Group( List( GeneratorsOfGroup(
> RelativeSigmaL( d/p, SymmetricBasis( q, p ) ) ),
> embedG ) ),
> swap ) );
> 
> # Compute conjugacy classes of outer involutions in the maxes.
> # (In order to avoid computing all conjugacy classes of these subgroups,
> # we work in the Sylow $2$ subgroups.)
> maxes:= List( maxes, M -> ImagesSet( epi, M ) );
> ccl:= List( maxes, M -> ClassesOfPrimeOrder( M, [ 2 ], Gprime ) );
> names:= List( primes, p -> Concatenation( "GL(", String( d/p ), ",",
> String( q^p ), ").", String( p ) ) );
> 
> return [ names, UpperBoundFixedPointRatios( PG, ccl, true )[1] ];
> end;;
</pre></div>

And these are the results for the groups we are interested in (and others).

<div class="example"><pre>
gap> ApproxPForOuterClassesInExtensionOfSLByGraphAut( 4, 3 );
[ [ "GL(2,9).2" ], 17/117 ]
gap> ApproxPForOuterClassesInExtensionOfSLByGraphAut( 4, 4 );
[ [ "GL(2,16).2" ], 73/1008 ]
gap> ApproxPForOuterClassesInExtensionOfSLByGraphAut( 6, 2 );
[ [ "GL(3,4).2", "GL(2,8).3" ], 41/1984 ]
gap> ApproxPForOuterClassesInExtensionOfSLByGraphAut( 6, 3 );
[ [ "GL(3,9).2", "GL(2,27).3" ], 541/352836 ]
gap> ApproxPForOuterClassesInExtensionOfSLByGraphAut( 6, 4 );
[ [ "GL(3,16).2", "GL(2,64).3" ], 3265/12570624 ]
gap> ApproxPForOuterClassesInExtensionOfSLByGraphAut( 6, 5 );
[ [ "GL(3,25).2", "GL(2,125).3" ], 13001/195250000 ]
gap> ApproxPForOuterClassesInExtensionOfSLByGraphAut( 8, 2 );
[ [ "GL(4,4).2" ], 367/1007872 ]
gap> ApproxPForOuterClassesInExtensionOfSLByGraphAut( 10, 2 );
[ [ "GL(5,4).2", "GL(2,32).5" ], 609281/476346056704 ]
</pre></div>

Now we consider diagonal automorphisms. We modify the approach for SL(d,q) by constructing the field extension type subgroups of GL(d,q) ...

<div class="example"><pre>
gap> RelativeGammaL:= function( d, B )
> local n, F, q, diag;
> 
> n:= Length( B );
> F:= LeftActingDomain( UnderlyingLeftModule( B ) );
> q:= Size( F );
> diag:= IdentityMat( d * n, F );
> diag{[ 1 .. n ]}{[ 1 .. n ]}:= BlownUpMat( B, [ [ Z(q^n) ] ] );
> return ClosureGroup( RelativeSigmaL( d, B ), diag );
> end;;
</pre></div>

... and counting the elements of prime order outside the simple group.

<div class="example"><pre>
gap> ApproxPForOuterClassesInGL:= function( d, q )
> local G, epi, PG, Gprime, primes, maxes, names;
> 
> # Check whether this is an admissible case (see [Be00]).
> if ( d = 2 and q in [ 2, 5, 7, 9 ] ) or ( d = 3 and q = 4 ) then
> return fail;
> fi;
> 
> # Create the group GL(d,q), and the map to PGL(d,q).
> G:= GL( d, q );
> epi:= ActionHomomorphism( G, NormedRowVectors( GF(q)^d ), OnLines );
> PG:= ImagesSource( epi );
> Gprime:= ImagesSet( epi, SL( d, q ) );
> 
> # Create the subgroups corresponding to the prime divisors of `d'.
> primes:= PrimeDivisors( d );
> maxes:= List( primes, p -> RelativeGammaL( d/p,
> Basis( AsField( GF(q), GF(q^p) ) ) ) );
> maxes:= List( maxes, M -> ImagesSet( epi, M ) );
> names:= List( primes, p -> Concatenation( "M(", String( d/p ), ",",
> String( q^p ), ")" ) );
> 
> return [ names,
> UpperBoundFixedPointRatios( PG, List( maxes,
> M -> ClassesOfPrimeOrder( M,
> PrimeDivisors( Index( PG, Gprime ) ), Gprime ) ),
> true )[1] ];
> end;;
</pre></div>

Here are the required results.

<div class="example"><pre>
gap> ApproxPForOuterClassesInGL( 6, 3 );
[ [ "M(3,9)", "M(2,27)" ], 41/882090 ]
gap> ApproxPForOuterClassesInGL( 4, 3 );
[ [ "M(2,9)" ], 0 ]
gap> ApproxPForOuterClassesInGL( 6, 4 );
[ [ "M(3,16)", "M(2,64)" ], 1/87296 ]
gap> ApproxPForOuterClassesInGL( 6, 5 );
[ [ "M(3,25)", "M(2,125)" ], 821563/756593750000 ]
</pre></div>

(Note that the extension field type subgroup in PGL(4,3) = L_4(3).2_1 is a non-split extension of its intersection with L_4(3), hence the zero value.)

Concerning extensions by Frobenius automorphisms, only the case (d,q) = (6,4) is interesting in <a href="chapBib.html#biBBGK">[BGK08]</a>. In fact, we would not need to compute anything for the extension G of S = SL(6,4) by the Frobenius map that squares each matrix entry. This is because ��^'(G,s) consists of the normalizers of the two subgroups of the types SL(3,16) and SL(2,64), and the former maximal subgroup is a non-split extension of its intersection with S, so only one maximal subgroup can contribute to σ^'(G,s), which is thus smaller than 1/2, by <a href="chapBib.html#biBBGK">[BGK08, Prop. 2.6]</a>.

However, it is easy enough to compute the exact value of σ^'(G,s). We work with the projective action of S on its natural module, and compute the permutation induced by the Frobenius map as the Frobenius action on the normed row vectors.

<div class="example"><pre>
gap> matgrp:= SL(6,4);;
gap> dom:= NormedRowVectors( GF(4)^6 );;
gap> Gprime:= Action( matgrp, dom, OnLines );;
gap> pi:= PermList( List( dom, v -> Position( dom, List( v, x -> x^2 ) ) ) );;
gap> G:= ClosureGroup( Gprime, pi );;
</pre></div>

Then we compute the maximal subgroups, the classes of outer involutions, and the bound, similar to the situation with graph automorphisms.

<div class="example"><pre>
gap> maxes:= List( [ 2, 3 ], p -> Normalizer( G,
> Action( RelativeSigmaL( 6/p,
> Basis( AsField( GF(4), GF(4^p) ) ) ), dom, OnLines ) ) );;
gap> ccl:= List( maxes, M -> ClassesOfPrimeOrder( M, [ 2 ], Gprime ) );;
gap> List( ccl, Length );
[ 0, 1 ]
gap> UpperBoundFixedPointRatios( G, ccl, true );
[ 1/34467840, true ]
</pre></div>

For (d,q) = (6,4), we have to consider also the extension G of S = SL(6,4) by the product α of the Frobenius map and the graph automorphism. We use the same approach as for the graph automorphism, i. e., we embed SL(6,4) into a 12-dimensional group of 6 × 6 block matrices, where the second block is the image of the first block under α, and describe α by the transposition of the two blocks.

First we construct the projective actions of S and G on an orbit of 1-spaces.

<div class="example"><pre>
gap> embedFG:= function( F, mat )
> return EmbeddedMatrix( F, mat,
> M -> List( TransposedMat( M^-1 ),
> row -> List( row, x -> x^2 ) ) );
> end;;
gap> d:= 6;; q:= 4;;
gap> alpha:= NullMat( 2*d, 2*d, GF(q) );;
gap> alpha{ [ 1 .. d ] }{ [ d+1 .. 2*d ] }:= IdentityMat( d, GF(q) );;
gap> alpha{ [ d+1 .. 2*d ] }{ [ 1 .. d ] }:= IdentityMat( d, GF(q) );;
gap> Gprime:= Group( List( GeneratorsOfGroup( SL(d,q) ),
> mat -> embedFG( GF(q), mat ) ) );;
gap> G:= ClosureGroupDefault( Gprime, alpha );;
gap> orb:= Orbit( G, One( G )[1], OnLines );;
gap> G:= Action( G, orb, OnLines );;
gap> Gprime:= Action( Gprime, orb, OnLines );;
</pre></div>

Next we construct the maximal subgroups, the classes of outer involutions, and the bound.

<div class="example"><pre>
gap> maxes:= List( PrimeDivisors( d ), p -> Group( List( GeneratorsOfGroup(
> RelativeSigmaL( d/p, Basis( AsField( GF(q), GF(q^p) ) ) ) ),
> mat -> embedFG( GF(q), mat ) ) ) );;
gap> maxes:= List( maxes, x -> Action( x, orb, OnLines ) );;
gap> maxes:= List( maxes, x -> Normalizer( G, x ) );;
gap> ccl:= List( maxes, M -> ClassesOfPrimeOrder( M, [ 2 ], Gprime ) );;
gap> List( ccl, Length );
[ 0, 1 ]
gap> UpperBoundFixedPointRatios( G, ccl, true );
[ 1/10792960, true ]
</pre></div>

The only missing cases are the extensions of SL(6,3) and SL(6,5) by the involutory outer automorphism that acts as the product of a diagonal and a graph automorphism.

In the case S = SL(6,3), we can directly write down the extension G.

<div class="example"><pre>
gap> d:= 6;; q:= 3;;
gap> diag:= IdentityMat( d, GF(q) );;
gap> diag[1][1]:= Z(q);;
gap> embedDG:= mat -> EmbeddedMatrix( GF(q), mat,
> M -> TransposedMat( M^-1 )^diag );;
gap> Gprime:= Group( List( GeneratorsOfGroup( SL(d,q) ), embedDG ) );;
gap> alpha:= NullMat( 2*d, 2*d, GF(q) );;
gap> alpha{ [ 1 .. d ] }{ [ d+1 .. 2*d ] }:= IdentityMat( d, GF(q) );;
gap> alpha{ [ d+1 .. 2*d ] }{ [ 1 .. d ] }:= IdentityMat( d, GF(q) );;
gap> G:= ClosureGroupDefault( Gprime, alpha );;
</pre></div>

The maximal subgroups are constructed as the normalizers in G of the extension field type subgroups in S. We work with a permutation representation of G.

<div class="example"><pre>
gap> maxes:= List( PrimeDivisors( d ), p -> Group( List( GeneratorsOfGroup(
> RelativeSigmaL( d/p, Basis( AsField( GF(q), GF(q^p) ) ) ) ),
> embedDG ) ) );;
gap> orb:= Orbit( G, One( G )[1], OnLines );;
gap> G:= Action( G, orb, OnLines );;
gap> Gprime:= Action( Gprime, orb, OnLines );;
gap> maxes:= List( maxes, M -> Normalizer( G, Action( M, orb, OnLines ) ) );;
gap> ccl:= List( maxes, M -> ClassesOfPrimeOrder( M, [ 2 ], Gprime ) );;
gap> List( ccl, Length );
[ 1, 1 ]
gap> UpperBoundFixedPointRatios( G, ccl, true );
[ 25/352836, true ]
</pre></div>

For S = SL(6,5), this approach does not work because we cannot realize the diagonal involution by an involutory matrix. Instead, we consider the extension of GL(6,5) ≅ 2.(2 × L_6(5)).2 by the graph automorphism α, which can be embedded into GL(12,5).

<div class="example"><pre>
gap> d:= 6;; q:= 5;;
gap> embedG:= mat -> EmbeddedMatrix( GF(q),
> mat, M -> TransposedMat( M^-1 ) );;
gap> Gprime:= Group( List( GeneratorsOfGroup( SL(d,q) ), embedG ) );;
gap> maxes:= List( PrimeDivisors( d ), p -> Group( List( GeneratorsOfGroup(
> RelativeSigmaL( d/p, Basis( AsField( GF(q), GF(q^p) ) ) ) ),
> embedG ) ) );;
gap> diag:= IdentityMat( d, GF(q) );;
gap> diag[1][1]:= Z(q);;
gap> diag:= embedG( diag );;
gap> alpha:= NullMat( 2*d, 2*d, GF(q) );;
gap> alpha{ [ 1 .. d ] }{ [ d+1 .. 2*d ] }:= IdentityMat( d, GF(q) );;
gap> alpha{ [ d+1 .. 2*d ] }{ [ 1 .. d ] }:= IdentityMat( d, GF(q) );;
gap> G:= ClosureGroupDefault( Gprime, alpha * diag );;
</pre></div>

Now we switch to the permutation action of this group on the 1-dimensional subspaces, thus factoring out the cyclic normal subgroup of order four. In this action, the involutory diagonal automorphism is represented by an involution, and we can proceed as above.

<div class="example"><pre>
gap> orb:= Orbit( G, One( G )[1], OnLines );;
gap> Gprime:= Action( Gprime, orb, OnLines );;
gap> G:= Action( G, orb, OnLines );;
gap> maxes:= List( maxes, M -> Action( M, orb, OnLines ) );;
gap> extmaxes:= List( maxes, M -> Normalizer( G, M ) );;
gap> ccl:= List( extmaxes, M -> ClassesOfPrimeOrder( M, [ 2 ], Gprime ) );;
gap> List( ccl, Length );
[ 2, 1 ]
gap> UpperBoundFixedPointRatios( G, ccl, true );
[ 3863/6052750000, true ]
</pre></div>

In the same way, we can recheck the values for the extensions of SL(6,5) by the diagonal or by the graph automorphism.

<div class="example"><pre>
gap> diag:= Permutation( diag, orb, OnLines );;
gap> G:= ClosureGroupDefault( Gprime, diag );;
gap> extmaxes:= List( maxes, M -> Normalizer( G, M ) );;
gap> ccl:= List( extmaxes, M -> ClassesOfPrimeOrder( M, [ 2 ], Gprime ) );;
gap> List( ccl, Length );
[ 3, 1 ]
gap> UpperBoundFixedPointRatios( G, ccl, true );
[ 821563/756593750000, true ]
gap> alpha:= Permutation( alpha, orb, OnLines );;
gap> G:= ClosureGroupDefault( Gprime, alpha );;
gap> extmaxes:= List( maxes, M -> Normalizer( G, M ) );;
gap> ccl:= List( extmaxes, M -> ClassesOfPrimeOrder( M, [ 2 ], Gprime ) );;
gap> List( ccl, Length );
[ 2, 2 ]
gap> UpperBoundFixedPointRatios( G, ccl, true );
[ 13001/195250000, true ]
</pre></div>

gap> t2:= CharacterTable( "L6(2).2" );; gap> map:= InverseMap( GetFusionMap( t, t2 ) );; gap> torso:= List( Concatenation( prim ), pi -> CompositionMaps( pi, map ) );; gap> ext:= List( torso, x -> PermChars( t2, rec( torso:= x ) ) ); [ [ Character( CharacterTable( "L6(2).2" ), [ 55552, 0, 128, 256, 337, 112, 22, 0, 0, 16, 0, 16, 2, 17, 0, 0, 8, 2, 4, 28, 0, 0, 0, 4, 1, 0, 1, 0, 0, 4, 0, 2, 2, 2, 1, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1120, 192, 32, 0, 0, 40, 13, 0, 4, 6, 0, 4, 4, 4, 0, 2, 8, 5, 0, 2, 0, 0, 0, 0, 1, 0, 1, 0 ] ) ], [ Character( CharacterTable( "L6(2).2" ), [ 1904640, 0, 0, 512, 960, 0, 120, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 8, 73, 24, 3, 0, 0, 15, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 1, 960, 960, 0, 0, 0, 0, 24, 0, 12, 12, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 3, 0, 0, 0 ] ) ] ] gap> sigma:= ApproxP( Concatenation( ext ), > Position( OrdersClassRepresentatives( t2 ), 63 ) );; gap> Maximum( sigma{ Difference( PositionsProperty( > OrdersClassRepresentatives( t2 ), IsPrimeInt ), > ClassPositionsOfDerivedSubgroup( t2 ) ) } ); 41/1984 -->

<a id="X7C8806DB8588BB51" name="X7C8806DB8588BB51"></a>

<h5>11.5-8 L_3(2)</h5>

We show that S = L_3(2) = SL(3,2) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>σ(S) = 1/4, and this value is attained exactly for σ(S,s) with s of order 7.

</dd>
<dt>(b)</dt>
<dd>For s of order 7, ��(S,s) consists of one group of the type 7:3.

</dd>
<dt>(c)</dt>
<dd>P(S) = 1/4, and this value is attained exactly for P(S,s) with s of order 7.

</dd>
<dt>(d)</dt>
<dd>The uniform spread of S is at exactly three, with s of order 7, and the spread of S is exactly four. (This had been left open in <a href="chapBib.html#biBBW1">[BW75]</a>.)

</dd>
</dl>
(Note that in this example, the spread and the uniform spread differ.)

Statement (a) follows from inspection of the primitive permutation characters, cf. Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>.

<div class="example"><pre>
gap> t:= CharacterTable( "L3(2)" );;
gap> ProbGenInfoSimple( t );
[ "L3(2)", 1/4, 3, [ "7A" ], [ 1 ] ]
</pre></div>

Statement (b) can be read off from the permutation characters, and the fact that the unique class of maximal subgroups that contain elements of order 7 consists of groups of the structure 7:3, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 3]</a>.

<div class="example"><pre>
gap> OrdersClassRepresentatives( t );
[ 1, 2, 3, 4, 7, 7 ]
gap> PrimitivePermutationCharacters( t );
[ Character( CharacterTable( "L3(2)" ), [ 7, 3, 1, 1, 0, 0 ] ),
 Character( CharacterTable( "L3(2)" ), [ 7, 3, 1, 1, 0, 0 ] ),
 Character( CharacterTable( "L3(2)" ), [ 8, 0, 2, 0, 1, 1 ] ) ]
</pre></div>

For the other statements, we will use the primitive permutation representations on 7 and 8 points of S (computed from the GAP Library of Tables of Marks), and their diagonal products of the degrees 14 and 15.

<div class="example"><pre>
gap> tom:= TableOfMarks( "L3(2)" );;
gap> g:= UnderlyingGroup( tom );
Group([ (2,4)(5,7), (1,2,3)(4,5,6) ])
gap> mx:= MaximalSubgroupsTom( tom );
[ [ 14, 13, 12 ], [ 7, 7, 8 ] ]
gap> maxes:= List( mx[1], i -> RepresentativeTom( tom, i ) );;
gap> tr:= List( maxes, s -> RightTransversal( g, s ) );;
gap> acts:= List( tr, x -> Action( g, x, OnRight ) );;
gap> g7:= acts[1];
Group([ (3,4)(6,7), (1,3,2)(4,6,5) ])
gap> g8:= acts[3];
Group([ (1,6)(2,5)(3,8)(4,7), (1,7,3)(2,5,8) ])
gap> g14:= DiagonalProductOfPermGroups( acts{ [ 1, 2 ] } );
Group([ (3,4)(6,7)(11,13)(12,14), (1,3,2)(4,6,5)(8,11,9)(10,12,13) ])
gap> g15:= DiagonalProductOfPermGroups( acts{ [ 2, 3 ] } );
Group([ (4,6)(5,7)(8,13)(9,12)(10,15)(11,14), (1,4,2)(3,5,6)(8,14,10)
 (9,12,15) ])
</pre></div>

First we compute that for all nonidentity elements s ∈ S and order three elements g ∈ S, P(g,s) ≥ 1/4 holds, with equality if and only if s has order 7; this implies statement (c). We actually compute, for class representatives s, the proportion of order three elements g such that ⟨ g, s ⟩ ≠ S holds.

<div class="example"><pre>
gap> ccl:= List( ConjugacyClasses( g7 ), Representative );;
gap> SortParallel( List( ccl, Order ), ccl );
gap> List( ccl, Order );
[ 1, 2, 3, 4, 7, 7 ]
gap> Size( ConjugacyClass( g7, ccl[3] ) );
56
gap> prop:= List( ccl,
> r -> RatioOfNongenerationTransPermGroup( g7, ccl[3], r ) );
[ 1, 5/7, 19/28, 2/7, 1/4, 1/4 ]
gap> Minimum( prop );
1/4
</pre></div>

Now we show that the uniform spread of S is less than four. In any of the primitive permutation representations of degree seven, we find three involutions whose sets of fixed points cover the seven points. The elements s of order different from 7 in S fix a point in this representation, so each such s generates a proper subgroup of S together with one of the three involutions.

<div class="example"><pre>
gap> x:= g7.1;
(3,4)(6,7)
gap> fix:= Difference( MovedPoints( g7 ), MovedPoints( x ) );
[ 1, 2, 5 ]
gap> orb:= Orbit( g7, fix, OnSets );
[ [ 1, 2, 5 ], [ 1, 3, 4 ], [ 2, 3, 6 ], [ 2, 4, 7 ], [ 1, 6, 7 ],
 [ 3, 5, 7 ], [ 4, 5, 6 ] ]
gap> Union( orb{ [ 1, 2, 5 ] } ) = [ 1 .. 7 ];
true
</pre></div>

So we still have to exclude elements s of order 7. In the primitive permutation representation of S on eight points, we find four elements of order three whose sets of fixed points cover the set of all points that are moved by S, so with each element of order seven in S, one of them generates an intransitive group.

<div class="example"><pre>
gap> three:= g8.2;
(1,7,3)(2,5,8)
gap> fix:= Difference( MovedPoints( g8 ), MovedPoints( three ) );
[ 4, 6 ]
gap> orb:= Orbit( g8, fix, OnSets );;
gap> QuadrupleWithProperty( [ [ fix ], orb, orb, orb ],
> list -> Union( list ) = [ 1 .. 8 ] );
[ [ 4, 6 ], [ 1, 7 ], [ 3, 8 ], [ 2, 5 ] ]
</pre></div>

Together with statement (a), this proves that the uniform spread of S is exactly three, with s of order seven.

Each element of S fixes a point in the permutation representation on 15 points. So for proving that the spread of S is less than five, it is sufficient to find a quintuple of elements whose sets of fixed points cover all 15 points. (From the permutation characters it is clear that four of these elements must have order three, and the fifth must be an involution.)

<div class="example"><pre>
gap> x:= g15.1;
(4,6)(5,7)(8,13)(9,12)(10,15)(11,14)
gap> fixx:= Difference( MovedPoints( g15 ), MovedPoints( x ) );
[ 1, 2, 3 ]
gap> orbx:= Orbit( g15, fixx, OnSets );
[ [ 1, 2, 3 ], [ 1, 4, 5 ], [ 1, 6, 7 ], [ 2, 4, 6 ], [ 3, 4, 7 ],
 [ 3, 5, 6 ], [ 2, 5, 7 ] ]
gap> y:= g15.2;
(1,4,2)(3,5,6)(8,14,10)(9,12,15)
gap> fixy:= Difference( MovedPoints( g15 ), MovedPoints( y ) );
[ 7, 11, 13 ]
gap> orby:= Orbit( g15, fixy, OnSets );;
gap> QuadrupleWithProperty( [ [ fixy ], orby, orby, orby ],
> l -> Difference( [ 1 .. 15 ], Union( l ) ) in orbx );
[ [ 7, 11, 13 ], [ 5, 8, 14 ], [ 1, 10, 15 ], [ 3, 9, 12 ] ]
</pre></div>

It remains to show that the spread of S is (at least) four. By the consideration of permutation characters, we know that we can find a suitable order seven element for all quadruples in question except perhaps quadruples of order three elements. We show that for each such case, we can choose s of order four. Since ��(S,s) consists of two subgroups of the type S_4, we work with the representation on 14 points.)

First we compute s and the S-orbit of its fixed points, and the S-orbit of the fixed points of an element x of order three. Then we prove that for each quadruple of conjugates of x, the union of their fixed points intersects the fixed points of at least one conjugate of s trivially.

<div class="example"><pre>
gap> ResetGlobalRandomNumberGenerators();
gap> repeat s:= Random( g14 );
> until Order( s ) = 4;
gap> s;
(1,3)(2,6,7,5)(9,11,10,12)(13,14)
gap> fixs:= Difference( MovedPoints( g14 ), MovedPoints( s ) );
[ 4, 8 ]
gap> orbs:= Orbit( g14, fixs, OnSets );;
gap> Length( orbs );
21
gap> three:= g14.2;
(1,3,2)(4,6,5)(8,11,9)(10,12,13)
gap> fix:= Difference( MovedPoints( g14 ), MovedPoints( three ) );
[ 7, 14 ]
gap> orb:= Orbit( g14, fix, OnSets );;
gap> Length( orb );
28
gap> QuadrupleWithProperty( [ [ fix ], orb, orb, orb ],
> l -> ForAll( orbs, o -> not IsEmpty( Intersection( o,
> Union( l ) ) ) ) );
fail
</pre></div>

By the lemma from Section <a href="chap11.html#X79D7312484E78274">11.2-2</a>, we are done.

<a id="X7B7061917ED3714D" name="X7B7061917ED3714D"></a>

<h5>11.5-9 M_11</h5>

We show that S = M_11 satisfies the following.

<dl>
<dt>(a)</dt>
<dd>σ(S) = 1/3, and this value is attained exactly for σ(S,s) with s of order 11.

</dd>
<dt>(b)</dt>
<dd>For s of order 11, ��(S,s) consists of one group of the type L_2(11).

</dd>
<dt>(c)</dt>
<dd>P(S) = 1/3, and this value is attained exactly for P(S,s) with s of order 11.

</dd>
<dt>(d)</dt>
<dd>Both the uniform spread and the spread of S is exactly three, with s of order 11.

</dd>
</dl>
Statement (a) follows from inspection of the primitive permutation characters, cf. Section <a href="chap11.html#X86CE51E180A3D4ED">11.4-1</a>.

<div class="example"><pre>
gap> t:= CharacterTable( "M11" );;
gap> ProbGenInfoSimple( t );
[ "M11", 1/3, 2, [ "11A" ], [ 1 ] ]
</pre></div>

Statement (b) can be read off from the permutation characters, and the fact that the unique class of maximal subgroups that contain elements of order 11 consists of groups of the structure L_2(11), see <a href="chapBib.html#biBCCN85">[CCN+85, p. 18]</a>.

<div class="example"><pre>
gap> OrdersClassRepresentatives( t );
[ 1, 2, 3, 4, 5, 6, 8, 8, 11, 11 ]
gap> PrimitivePermutationCharacters( t );
[ Character( CharacterTable( "M11" ),
 [ 11, 3, 2, 3, 1, 0, 1, 1, 0, 0 ] ),
 Character( CharacterTable( "M11" ),
 [ 12, 4, 3, 0, 2, 1, 0, 0, 1, 1 ] ),
 Character( CharacterTable( "M11" ),
 [ 55, 7, 1, 3, 0, 1, 1, 1, 0, 0 ] ),
 Character( CharacterTable( "M11" ),
 [ 66, 10, 3, 2, 1, 1, 0, 0, 0, 0 ] ),
 Character( CharacterTable( "M11" ),
 [ 165, 13, 3, 1, 0, 1, 1, 1, 0, 0 ] ) ]
gap> Maxes( t );
[ "A6.2_3", "L2(11)", "3^2:Q8.2", "A5.2", "2.S4" ]
</pre></div>

For the other statements, we will use the primitive permutation representations of S on 11 and 12 points (which are fetched from the Atlas of Group Representations <a href="chapBib.html#biBAGRv3">[WWT+]</a>), and their diagonal product.

<div class="example"><pre>
gap> gens11:= OneAtlasGeneratingSet( "M11", NrMovedPoints, 11 );
rec( charactername := "1a+10a", constituents := [ 1, 2 ],
 contents := "core",
 generators := [ (2,10)(4,11)(5,7)(8,9), (1,4,3,8)(2,5,6,9) ],
 groupname := "M11", id := "",
 identifier := [ "M11", [ "M11G1-p11B0.m1", "M11G1-p11B0.m2" ], 1,
 11 ], isPrimitive := true, maxnr := 1, p := 11, rankAction := 2,
 repname := "M11G1-p11B0", repnr := 1, size := 7920,
 stabilizer := "A6.2_3", standardization := 1, transitivity := 4,
 type := "perm" )
gap> g11:= GroupWithGenerators( gens11.generators );;
gap> gens12:= OneAtlasGeneratingSet( "M11", NrMovedPoints, 12 );;
gap> g12:= GroupWithGenerators( gens12.generators );;
gap> g23:= DiagonalProductOfPermGroups( [ g11, g12 ] );
Group([ (2,10)(4,11)(5,7)(8,9)(12,17)(13,20)(16,18)(19,21), (1,4,3,8)
 (2,5,6,9)(12,17,18,15)(13,19)(14,20)(16,22,23,21) ])
</pre></div>

First we compute that for all nonidentity elements s ∈ S and involutions g ∈ S, P(g,s) ≥ 1/3 holds, with equality if and only if s has order 11; this implies statement (c). We actually compute, for class representatives s, the proportion of involutions g such that ⟨ g, s ⟩ ≠ S holds.

<div class="example"><pre>
gap> inv:= g11.1;
(2,10)(4,11)(5,7)(8,9)
gap> ccl:= List( ConjugacyClasses( g11 ), Representative );;
gap> SortParallel( List( ccl, Order ), ccl );
gap> List( ccl, Order );
[ 1, 2, 3, 4, 5, 6, 8, 8, 11, 11 ]
gap> Size( ConjugacyClass( g11, inv ) );
165
gap> prop:= List( ccl,
> r -> RatioOfNongenerationTransPermGroup( g11, inv, r ) );
[ 1, 1, 1, 149/165, 25/33, 31/55, 23/55, 23/55, 1/3, 1/3 ]
gap> Minimum( prop );
1/3
</pre></div>

For the first part of statement (d), we have to deal only with the case of triples of involutions.

The 11-cycle s is contained in exactly one maximal subgroup of S, of index 12. By Corollary 1 in Section <a href="chap11.html#X79D7312484E78274">11.2-2</a>, it is enough to show that in the primitive degree 12 representation of S, the fixed points of no triple (x_1, x_2, x_3) of involutions in S can cover all twelve points; equivalenly (considering complements), we show that there is no triple such that the intersection of the sets of moved points is empty.

<div class="example"><pre>
gap> inv:= g12.1;
(1,6)(2,9)(5,7)(8,10)
gap> moved:= MovedPoints( inv );
[ 1, 2, 5, 6, 7, 8, 9, 10 ]
gap> orb12:= Orbit( g12, moved, OnSets );;
gap> Length( orb12 );
165
gap> TripleWithProperty( [ orb12{[1]}, orb12, orb12 ],
> list -> IsEmpty( Intersection( list ) ) );
fail
</pre></div>

This implies that the uniform spread of S is at least three.

Now we show that there is a quadruple consisting of one element of order three and three involutions whose fixed points cover all points in the degree 23 representation constructed above; since the permutation character of this representation is strictly positive, this implies that S does not have spread four, by Corollary 2 in Section <a href="chap11.html#X79D7312484E78274">11.2-2</a>, and we have proved statement (d).

<div class="example"><pre>
gap> inv:= g23.1;
(2,10)(4,11)(5,7)(8,9)(12,17)(13,20)(16,18)(19,21)
gap> moved:= MovedPoints( inv );
[ 2, 4, 5, 7, 8, 9, 10, 11, 12, 13, 16, 17, 18, 19, 20, 21 ]
gap> orb23:= Orbit( g23, moved, OnSets );;
gap> three:= ( g23.1*g23.2^2 )^2;
(2,6,10)(4,8,7)(5,9,11)(12,17,23)(15,18,16)(19,21,22)
gap> movedthree:= MovedPoints( three );;
gap> QuadrupleWithProperty( [ [ movedthree ], orb23, orb23, orb23 ],
> list -> IsEmpty( Intersection( list ) ) );
[ [ 2, 4, 5, 6, 7, 8, 9, 10, 11, 12, 15, 16, 17, 18, 19, 21, 22, 23 ],
 [ 1, 3, 4, 5, 6, 8, 9, 10, 12, 13, 14, 16, 17, 18, 20, 21 ],
 [ 1, 2, 3, 4, 5, 6, 7, 11, 12, 13, 14, 15, 18, 19, 20, 23 ],
 [ 1, 2, 3, 7, 8, 9, 10, 11, 13, 14, 15, 16, 17, 20, 22, 23 ] ]
</pre></div>

<a id="X82E0F48A7FF82BB3" name="X82E0F48A7FF82BB3"></a>

<h5>11.5-10 M_12</h5>

We show that S = M_12 satisfies the following.

<dl>
<dt>(a)</dt>
<dd>σ(S) = 1/3, and this value is attained exactly for σ(S,s) with s of order 10.

</dd>
<dt>(b)</dt>
<dd>For s ∈ S of order 10, ��(S,s) consists of two nonconjugate subgroups of the type A_6.2^2, and one group of the type 2 × S_5.

</dd>
<dt>(c)</dt>
<dd>P(S) = 31/99, and this value is attained exactly for P(S,s) with s of order 10.

</dd>
<dt>(d)</dt>
<dd>The uniform spread of S is at least three, with s of order 10.

</dd>
<dt>(e)</dt>
<dd>σ^'(Aut(S), s) = 4/99.

</dd>
</dl>
Statement (a) follows from inspection of the primitive permutation characters, cf. Section <a href="chap11.html#X86CE51E180A3D4ED">11.4-1</a>.

<div class="example"><pre>
gap> t:= CharacterTable( "M12" );;
gap> ProbGenInfoSimple( t );
[ "M12", 1/3, 2, [ "10A" ], [ 3 ] ]
</pre></div>

Statement (b) can be read off from the permutation characters, and the fact that the only classes of maximal subgroups that contain elements of order 10 consist of groups of the structures A_6.2^2 (two classes) and 2 × S_5 (one class), see <a href="chapBib.html#biBCCN85">[CCN+85, p. 33]</a>.

<div class="example"><pre>
gap> spos:= Position( OrdersClassRepresentatives( t ), 10 );
13
gap> prim:= PrimitivePermutationCharacters( t );;
gap> List( prim, x -> x{ [ 1, spos ] } );
[ [ 12, 0 ], [ 12, 0 ], [ 66, 1 ], [ 66, 1 ], [ 144, 0 ], [ 220, 0 ],
 [ 220, 0 ], [ 396, 1 ], [ 495, 0 ], [ 495, 0 ], [ 1320, 0 ] ]
gap> Maxes( t );
[ "M11", "M12M2", "A6.2^2", "M12M4", "L2(11)", "3^2.2.S4", "M12M7",
 "2xS5", "M8.S4", "4^2:D12", "A4xS3" ]
</pre></div>

For statement (c) (which implies statement (d)), we use the primitive permutation representation on 12 points.

<div class="example"><pre>
gap> g:= MathieuGroup( 12 );
Group([ (1,2,3,4,5,6,7,8,9,10,11), (3,7,11,8)(4,10,5,6), (1,12)(2,11)
 (3,6)(4,8)(5,9)(7,10) ])
</pre></div>

First we show that for s of order 10, P(S,s) = 31/99 holds.

<div class="example"><pre>
gap> approx:= ApproxP( prim, spos );
[ 0, 3/11, 1/3, 1/11, 1/132, 13/99, 13/99, 13/396, 1/132, 1/33, 1/33,
 1/33, 13/396, 0, 0 ]
gap> 2B:= g.2^2;
(3,11)(4,5)(6,10)(7,8)
gap> Size( ConjugacyClass( g, 2B ) );
495
gap> ResetGlobalRandomNumberGenerators();
gap> repeat s:= Random( g );
> until Order( s ) = 10;
gap> prop:= RatioOfNongenerationTransPermGroup( g, 2B, s );
31/99
gap> Filtered( approx, x -> x >= prop );
[ 1/3 ]
</pre></div>

Next we show that for s of order different from 10, P(g,s) is larger than 31/99 for suitable g ∈ S^×. Except for s in the class <code class="code">6A</code> (which fixes no point in the degree 12 representation), it suffices to consider g in the class <code class="code">2B</code> (with four fixed points).

<div class="example"><pre>
gap> x:= g.2^2;
(3,11)(4,5)(6,10)(7,8)
gap> ccl:= List( ConjugacyClasses( g ), Representative );;
gap> SortParallel( List( ccl, Order ), ccl );
gap> prop:= List( ccl, r -> RatioOfNongenerationTransPermGroup( g, x, r ) );;
gap> SortedList( prop );
[ 7/55, 31/99, 5/9, 5/9, 39/55, 383/495, 383/495, 43/55, 29/33, 1, 1,
 1, 1, 1, 1 ]
gap> bad:= Filtered( prop, x -> x < 31/99 );
[ 7/55 ]
gap> pos:= Position( prop, bad[1] );;
gap> [ Order( ccl[ pos ] ), NrMovedPoints( ccl[ pos ] ) ];
[ 6, 12 ]
</pre></div>

In the remaining case, we choose g in the class <code class="code">2A</code> (which is fixed point free).

<div class="example"><pre>
gap> x:= g.3;
(1,12)(2,11)(3,6)(4,8)(5,9)(7,10)
gap> s:= ccl[ pos ];;
gap> prop:= RatioOfNongenerationTransPermGroup( g, x, s );
17/33
gap> prop > 31/99;
true
</pre></div>

Statement (e) has been shown already in Section <a href="chap11.html#X84E9D10F80A74A53">11.4-3</a>.

<a id="X7FF2E8F27FBEB65C" name="X7FF2E8F27FBEB65C"></a>

<h5>11.5-11 O_7(3)</h5>

We show that S = O_7(3) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>σ(S) = 199/351, and this value is attained exactly for σ(S,s) with s of order 14.

</dd>
<dt>(b)</dt>
<dd>For s ∈ S of order 14, ��(S,s) consists of one group of the type 2.U_4(3).2_2 = Ω^-(6,3).2 and two nonconjugate groups of the type S_9.

</dd>
<dt>(c)</dt>
<dd>P(S) = 155/351, and this value is attained exactly for P(S,s) with s of order 14.

</dd>
<dt>(d)</dt>
<dd>The uniform spread of S is at least three, with s of order 14.

</dd>
<dt>(e)</dt>
<dd>σ^'(Aut(S), s) = 1/3.

</dd>
</dl>
Currently GAP provides neither the table of marks of S nor all character tables of its maximal subgroups. First we compute those primitive permutation characters of S that have the degrees 351 (point stabilizer 2.U_4(3).2_2), 364 (point stabilizer 3^5:U_4(2).2), 378 (point stabilizer L_4(3).2_2), 1080 (point stabilizer G_2(3), two classes), 1120 (point stabilizer 3^3+3:L_3(3)), 3159 (point stabilizer S_6(2), two classes), 12636 (point stabilizer S_9, two classes), 22113 (point stabilizer (2^2 × U_4(2)).2, which extends to D_8 × U_4(2).2 in O_7(3).2), and 28431 (point stabilizer 2^6:A_7).

(So we ignore the primitive permutation characters of the degrees 3640, 265356, and 331695. Note that the orders of the corresponding subgroups are not divisible by 7.)

<div class="example"><pre>
gap> t:= CharacterTable( "O7(3)" );;
gap> someprim:= [];;
gap> pi:= PossiblePermutationCharacters(
> CharacterTable( "2.U4(3).2_2" ), t );; Length( pi );
1
gap> Append( someprim, pi );
gap> pi:= PermChars( t, rec( torso:= [ 364 ] ) );; Length( pi );
1
gap> Append( someprim, pi );
gap> pi:= PossiblePermutationCharacters(
> CharacterTable( "L4(3).2_2" ), t );; Length( pi );
1
gap> Append( someprim, pi );
gap> pi:= PossiblePermutationCharacters( CharacterTable( "G2(3)" ), t );
[ Character( CharacterTable( "O7(3)" ),
 [ 1080, 0, 0, 24, 108, 0, 0, 0, 27, 18, 9, 0, 12, 4, 0, 0, 0, 0, 0,
 0, 0, 0, 12, 0, 0, 0, 0, 0, 3, 6, 0, 3, 2, 2, 2, 0, 0, 0, 3, 0,
 0, 0, 0, 0, 0, 4, 0, 3, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0 ] ),
 Character( CharacterTable( "O7(3)" ),
 [ 1080, 0, 0, 24, 108, 0, 0, 27, 0, 18, 9, 0, 12, 4, 0, 0, 0, 0, 0,
 0, 0, 0, 12, 0, 0, 0, 0, 3, 0, 0, 6, 3, 2, 2, 2, 0, 0, 3, 0, 0,
 0, 0, 0, 0, 0, 4, 3, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0 ] ) ]
gap> Append( someprim, pi );
gap> pi:= PermChars( t, rec( torso:= [ 1120 ] ) );; Length( pi );
1
gap> Append( someprim, pi );
gap> pi:= PossiblePermutationCharacters( CharacterTable( "S6(2)" ), t );
[ Character( CharacterTable( "O7(3)" ),
 [ 3159, 567, 135, 39, 0, 81, 0, 0, 27, 27, 0, 15, 3, 3, 7, 4, 0,
 27, 0, 0, 0, 0, 0, 9, 3, 0, 9, 0, 3, 9, 3, 0, 2, 1, 1, 0, 0, 0,
 3, 0, 2, 0, 0, 0, 3, 0, 0, 3, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0 ] ),
 Character( CharacterTable( "O7(3)" ),
 [ 3159, 567, 135, 39, 0, 81, 0, 27, 0, 27, 0, 15, 3, 3, 7, 4, 0,
 27, 0, 0, 0, 0, 0, 9, 3, 0, 9, 3, 0, 3, 9, 0, 2, 1, 1, 0, 0, 3,
 0, 0, 2, 0, 0, 0, 3, 0, 3, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0 ] ) ]
gap> Append( someprim, pi );
gap> pi:= PossiblePermutationCharacters( CharacterTable( "S9" ), t );
[ Character( CharacterTable( "O7(3)" ),
 [ 12636, 1296, 216, 84, 0, 81, 0, 0, 108, 27, 0, 6, 0, 12, 10, 1,
 0, 27, 0, 0, 0, 0, 0, 9, 3, 0, 9, 0, 12, 9, 3, 0, 1, 0, 2, 0,
 0, 0, 3, 1, 1, 0, 0, 0, 3, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0,
 1 ] ), Character( CharacterTable( "O7(3)" ),
 [ 12636, 1296, 216, 84, 0, 81, 0, 108, 0, 27, 0, 6, 0, 12, 10, 1,
 0, 27, 0, 0, 0, 0, 0, 9, 3, 0, 9, 12, 0, 3, 9, 0, 1, 0, 2, 0,
 0, 3, 0, 1, 1, 0, 0, 0, 3, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 0,
 1 ] ) ]
gap> Append( someprim, pi );
gap> t2:= CharacterTable( "O7(3).2" );;
gap> s2:= CharacterTable( "Dihedral", 8 ) * CharacterTable( "U4(2).2" );
CharacterTable( "Dihedral(8)xU4(2).2" )
gap> pi:= PossiblePermutationCharacters( s2, t2 );; Length( pi );
1
gap> pi:= RestrictedClassFunctions( pi, t );;
gap> Append( someprim, pi );
gap> pi:= PossiblePermutationCharacters(
> CharacterTable( "2^6:A7" ), t );; Length( pi );
1
gap> Append( someprim, pi );
gap> List( someprim, x -> x[1] );
[ 351, 364, 378, 1080, 1080, 1120, 3159, 3159, 12636, 12636, 22113,
 28431 ]
</pre></div>

Note that in the three cases where two possible permutation characters were found, there are in fact two classes of subgroups that induce different permutation characters. For the subgroups of the types G_2(3) and S_6(2), this is stated in <a href="chapBib.html#biBCCN85">[CCN+85, p. 109]</a>, and for the subgroups of the type S_9, this follows from the fact that each S_9 type subgroup in S contains elements in exactly one of the classes <code class="code">3D</code> or <code class="code">3E</code>, and these two classes are fused by the outer automorphism of S.

<div class="example"><pre>
gap> cl:= PositionsProperty( AtlasClassNames( t ),
> x -> x in [ "3D", "3E" ] );
[ 8, 9 ]
gap> List( Filtered( someprim, x -> x[1] = 12636 ), pi -> pi{ cl } );
[ [ 0, 108 ], [ 108, 0 ] ]
gap> GetFusionMap( t, t2 ){ cl };
[ 8, 8 ]
</pre></div>

Now we compute the lower bounds for σ( S, s^' ) that are given by the sublist <code class="code">someprim</code> of the primitive permutation characters.

<div class="example"><pre>
gap> spos:= Position( OrdersClassRepresentatives( t ), 14 );
52
gap> Maximum( ApproxP( someprim, spos ) );
199/351
</pre></div>

This shows that σ( S, s ) = 199/351 holds. For statement (a), we have to show that choosing s^' from another class than <code class="code">14A</code> yields a larger value for σ( S, s^' ).

<div class="example"><pre>
gap> approx:= List( [ 1 .. NrConjugacyClasses( t ) ],
> i -> Maximum( ApproxP( someprim, i ) ) );;
gap> PositionsProperty( approx, x -> x <= 199/351 );
[ 52 ]
</pre></div>

Statement (b) can be read off from the permutation characters.

<div class="example"><pre>
gap> pos:= PositionsProperty( someprim, x -> x[ spos ] <> 0 );
[ 1, 9, 10 ]
gap> List( someprim{ pos }, x -> x{ [ 1, spos ] } );
[ [ 351, 1 ], [ 12636, 1 ], [ 12636, 1 ] ]
</pre></div>

For statement (c), we first compute P(g, s) for g in the class <code class="code">2A</code>, via explicit computations with the group. For dealing with this case, we first construct a faithful permutation representation of O_7(3) from the natural matrix representation of SO(7,3).

<div class="example"><pre>
gap> so73:= SpecialOrthogonalGroup( 7, 3 );;
gap> o73:= DerivedSubgroup( so73 );;
gap> orbs:= OrbitsDomain( o73, Elements( GF(3)^7 ) );;
gap> Set( orbs, Length );
[ 1, 702, 728, 756 ]
gap> g:= Action( o73, First( orbs, x -> Length( x ) = 702 ) );;
gap> Size( g ) = Size( t );
true
</pre></div>

A <code class="code">2A</code> element g can be found as the 7-th power of any element of order 14 in S.

<div class="example"><pre>
gap> ResetGlobalRandomNumberGenerators();
gap> repeat s:= Random( g );
> until Order( s ) = 14;
gap> 2A:= s^7;;
gap> bad:= RatioOfNongenerationTransPermGroup( g, 2A, s );
155/351
gap> bad > 1/3;
true
gap> approx:= ApproxP( someprim, spos );;
gap> PositionsProperty( approx, x -> x >= 1/3 );
[ 2 ]
</pre></div>

This shows that P(g,s) = 155/351 > 1/3. Since σ( g, s ) < 1/3 for all nonidentity g not in the class <code class="code">2A</code>, we have P( S, s ) = 155/351. For statement (c), it remains to show that P( S, s^' ) is larger than 155/351 whenever s^' is not of order 14. First we compute P( g, s^' ), for g in the class <code class="code">2A</code>.

<div class="example"><pre>
gap> consider:= RepresentativesMaximallyCyclicSubgroups( t );
[ 18, 19, 25, 26, 27, 30, 31, 32, 34, 35, 38, 39, 41, 42, 43, 44, 45,
 46, 47, 48, 49, 50, 52, 53, 54, 56, 57, 58 ]
gap> Length( consider );
28
gap> consider:= ClassesPerhapsCorrespondingToTableColumns( g, t, consider );;
gap> Length( consider );
31
gap> consider:= List( consider, Representative );;
gap> SortParallel( List( consider, Order ), consider );
gap> app2A:= List( consider, c ->
> RatioOfNongenerationTransPermGroup( g, 2A, c ) );;
gap> SortedList( app2A );
[ 1/3, 1/3, 155/351, 191/351, 67/117, 23/39, 23/39, 85/117, 10/13,
 10/13, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
 1 ]
gap> test:= PositionsProperty( app2A, x -> x <= 155/351 );;
gap> List( test, i -> Order( consider[i] ) );
[ 13, 13, 14 ]
</pre></div>

We see that only for s^' in one of the two (algebraically conjugate) classes of element order 13, P( S, s^' ) has a chance to be smaller than 155/351. This possibility is now excluded by counting elements in the class <code class="code">3A</code> that do not generate S together with s^' of order 13.

<div class="example"><pre>
gap> C3A:= First( ConjugacyClasses( g ),
> c -> Order( Representative( c ) ) = 3 and Size( c ) = 7280 );;
gap> repeat ss:= Random( g );
> until Order( ss ) = 13;
gap> bad:= RatioOfNongenerationTransPermGroup( g, Representative( C3A ), ss );
17/35
gap> bad > 155/351;
true
</pre></div>

Now we show statement (d): For each triple (x_1, x_2, x_3) of nonidentity elements in S, there is an element s in the class <code class="code">14A</code> such that ⟨ x_i, s ⟩ = S holds for 1 ≤ i ≤ 3. We can read off from the character-theoretic data that only those triples have to be checked for which at least two elements are contained in the class <code class="code">2A</code>, and the third element lies in one of the classes <code class="code">2A</code>, <code class="code">2B</code>, <code class="code">3B</code>.

<div class="example"><pre>
gap> approx:= ApproxP( someprim, spos );;
gap> max:= Maximum( approx{ [ 3 .. Length( approx ) ] } );
59/351
gap> 155 + 2*59 < 351;
true
gap> third:= PositionsProperty( approx, x -> 2 * 155/351 + x >= 1 );
[ 2, 3, 6 ]
gap> ClassNames( t ){ third };
[ "2a", "2b", "3b" ]
</pre></div>

We can find elements in the classes <code class="code">2B</code> and <code class="code">3B</code> as powers of arbitrary elements of the orders 20 and 15, respectively.

<div class="example"><pre>
gap> ord20:= PositionsProperty( OrdersClassRepresentatives( t ),
> x -> x = 20 );
[ 58 ]
gap> PowerMap( t, 10 ){ ord20 };
[ 3 ]
gap> repeat x:= Random( g );
> until Order( x ) = 20;
gap> 2B:= x^10;;
gap> C2B:= ConjugacyClass( g, 2B );;
gap> ord15:= PositionsProperty( OrdersClassRepresentatives( t ),
> x -> x = 15 );
[ 53 ]
gap> PowerMap( t, 10 ){ ord15 };
[ 6 ]
gap> repeat x:= Random( g );
> until Order( x ) = 15;
gap> 3B:= x^5;;
gap> C3B:= ConjugacyClass( g, 3B );;
</pre></div>

The existence of s can be shown with the random approach described in Section <a href="chap11.html#X83DACCF07EF62FAE">11.3-3</a>.

<div class="example"><pre>
gap> repeat s:= Random( g );
> until Order( s ) = 14;
gap> RandomCheckUniformSpread( g, [ 2A, 2A, 2A ], s, 50 );
true
gap> RandomCheckUniformSpread( g, [ 2B, 2A, 2A ], s, 50 );
true
gap> RandomCheckUniformSpread( g, [ 3B, 2A, 2A ], s, 50 );
true
</pre></div>

Finally, we show statement (e). Let G = Aut(S) = S.2. By <a href="chapBib.html#biBCCN85">[CCN+85, p. 109]</a>, ��^'(G,s) consists of the extension of the 2.U_4(3).2_1 type subgroup. We compute the extension of the permutation character.

<div class="example"><pre>
gap> prim:= someprim{ [ 1 ] };
[ Character( CharacterTable( "O7(3)" ),
 [ 351, 127, 47, 15, 27, 45, 36, 0, 0, 9, 0, 15, 3, 3, 7, 6, 19, 19,
 10, 11, 12, 8, 3, 5, 3, 6, 1, 0, 0, 3, 3, 0, 1, 1, 1, 6, 3, 0,
 0, 2, 2, 0, 3, 0, 3, 3, 0, 0, 1, 0, 0, 1, 0, 4, 4, 1, 2, 0 ] ) ]
gap> spos:= Position( AtlasClassNames( t ), "14A" );;
gap> t2:= CharacterTable( "O7(3).2" );;
gap> map:= InverseMap( GetFusionMap( t, t2 ) );;
gap> torso:= List( prim, pi -> CompositionMaps( pi, map ) );;
gap> ext:= List( torso, x -> PermChars( t2, rec( torso:= x ) ) );
[ [ Character( CharacterTable( "O7(3).2" ),
 [ 351, 127, 47, 15, 27, 45, 36, 0, 9, 0, 15, 3, 3, 7, 6, 19,
 19, 10, 11, 12, 8, 3, 5, 3, 6, 1, 0, 3, 0, 1, 1, 1, 6, 3,
 0, 2, 2, 0, 3, 0, 3, 3, 0, 1, 0, 0, 1, 0, 4, 1, 2, 0, 117,
 37, 21, 45, 1, 13, 5, 1, 9, 9, 18, 15, 1, 7, 9, 6, 4, 0, 3,
 0, 3, 3, 6, 2, 2, 9, 6, 1, 3, 1, 4, 1, 2, 1, 1, 0, 3, 1, 0,
 0, 0, 0, 1, 1, 0, 0 ] ) ] ]
gap> approx:= ApproxP( Concatenation( ext ),
> Position( AtlasClassNames( t2 ), "14A" ) );;
gap> Maximum( approx{ Difference(
> PositionsProperty( OrdersClassRepresentatives( t2 ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t2 ) ) } );
1/3
</pre></div>

<a id="X7F80F2527C424AA4" name="X7F80F2527C424AA4"></a>

<h5>11.5-12 O_8^+(2)</h5>

We show that S = O_8^+(2) = Ω^+(8,2) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>σ(S) = 334/315, and this value is attained exactly for σ(S,s) with s of order 15.

</dd>
<dt>(b)</dt>
<dd>For s ∈ S of order 15, ��(S,s) consists of one group of the type S_6(2), two conjugate groups of the type 2^6:A_8, two conjugate groups of the type A_9, and one group of each of the types (3 × U_4(2)):2 = (3 × Ω^-(6,2)):2 and (A_5 × A_5):2^2 = (Ω^-(4,2) × Ω^-(4,2)):2^2.

</dd>
<dt>(c)</dt>
<dd>P(S) = 29/42, and this value is attained exactly for P(S,s) with s of order 15.

</dd>
<dt>(d)</dt>
<dd>Let x, y ∈ S such that x, y, x y lie in the unique involution class of length 1575 of S. (This is the class <code class="code">2A</code>.) Then each element in S together with one of x, y, x y generates a proper subgroup of S.

</dd>
<dt>(e)</dt>
<dd>Both the spread and the uniform spread of S is exactly two, with s of order 15.

</dd>
<dt>(f)</dt>
<dd>For each choice of s ∈ S, there is an extension S.2 such that for any element g in the (outer) class <code class="code">2F</code>, ⟨ s, g ⟩ does not contain S.

</dd>
<dt>(g)</dt>
<dd>For an element s of order 15 in S, either S is the only maximal subgroup of S.2 that contains s, or the maximal subgroups of S.2 that contain s are S and the extensions of the subgroups listed in statement (b); these groups have the structures S_6(2) × 2, 2^6:S_8 (twice), S_9 (twice), S_3 × U_4(2).2, and S_5 ≀ 2.

</dd>
<dt>(h)</dt>
<dd>For s ∈ S of order 15 and arbitrary g ∈ S.3 ∖ S, we have ⟨ s, g ⟩ = S.3.

</dd>
<dt>(i)</dt>
<dd>If x, y are nonidentity elements in Aut(S) then there is an element s of order 15 in S such that S ⊆ ⟨ x, s ⟩ ∩ ⟨ y, s ⟩.

</dd>
</dl>
Statement (a) follows from inspection of the primitive permutation characters, cf. Section <a href="chap11.html#X80DA58F187CDCF5F">11.4-4</a>.

<div class="example"><pre>
gap> t:= CharacterTable( "O8+(2)" );;
gap> ProbGenInfoSimple( t );
[ "O8+(2)", 334/315, 0, [ "15A", "15B", "15C" ], [ 7, 7, 7 ] ]
</pre></div>

Statement (b) can be read off from the permutation characters, and the fact that the only classes of maximal subgroups that contain elements of order 15 consist of groups of the structures as claimed, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 85]</a>.

<div class="example"><pre>
gap> prim:= PrimitivePermutationCharacters( t );;
gap> spos:= Position( OrdersClassRepresentatives( t ), 15 );;
gap> List( Filtered( prim, x -> x[ spos ] <> 0 ), l -> l{ [ 1, spos ] } );
[ [ 120, 1 ], [ 135, 2 ], [ 960, 2 ], [ 1120, 1 ], [ 12096, 1 ] ]
</pre></div>

For the remaining statements, we take a primitive permutation representation on 120 points, and assume that the permutation character is <code class="code">1a+35a+84a</code>. (See <a href="chapBib.html#biBCCN85">[CCN+85, p. 85]</a>, note that the three classes of maximal subgroups of index 120 in S are conjugate under triality.)

<div class="example"><pre>
gap> matgroup:= DerivedSubgroup( GeneralOrthogonalGroup( 1, 8, 2 ) );;
gap> points:= NormedRowVectors( GF(2)^8 );;
gap> orbs:= OrbitsDomain( matgroup, points );;
gap> List( orbs, Length );
[ 135, 120 ]
gap> g:= Action( matgroup, orbs[2] );;
gap> Size( g );
174182400
gap> pi:= Sum( Irr( t ){ [ 1, 3, 7 ] } );
Character( CharacterTable( "O8+(2)" ),
[ 120, 24, 32, 0, 0, 8, 36, 0, 0, 3, 6, 12, 4, 8, 0, 0, 0, 10, 0, 0,
 12, 0, 0, 8, 0, 0, 3, 6, 0, 0, 2, 0, 0, 2, 1, 2, 2, 3, 0, 0, 2, 0,
 0, 0, 0, 0, 3, 2, 0, 0, 1, 0, 0 ] )
</pre></div>

In order to show statement (c), we first observe that for s in the class <code class="code">15A</code> and g not in one of the classes <code class="code">2A</code>, <code class="code">2B</code>, <code class="code">3A</code>, σ(g,s) < 1/3 holds, and for the exceptional three classes, we have σ(g,s) > 1/2.

<div class="example"><pre>
gap> approx:= ApproxP( prim, spos );;
gap> testpos:= PositionsProperty( approx, x -> x >= 1/3 );
[ 2, 3, 7 ]
gap> AtlasClassNames( t ){ testpos };
[ "2A", "2B", "3A" ]
gap> approx{ testpos };
[ 254/315, 334/315, 1093/1120 ]
gap> ForAll( approx{ testpos }, x -> x > 1/2 );
true
</pre></div>

Now we compute the values P(g,s), for s in the class <code class="code">15A</code> and g in one of the classes <code class="code">2A</code>, <code class="code">2B</code>, <code class="code">3A</code>.

By our choice of the character of the permutation representation we use, the class <code class="code">15A</code> is determined as the unique class of element order 15 with one fixed point. (Note that the three classes of element order 15 in S are conjugate under triality.) A <code class="code">2A</code> element can be found as the fourth power of any element of order 8 in S, a <code class="code">3A</code> element can be found as the fifth power of a <code class="code">15A</code> element, and a <code class="code">2B</code> element can be found as the sixth power of an element of order 12, with 32 fixed points.

<div class="example"><pre>
gap> ResetGlobalRandomNumberGenerators();
gap> repeat s:= Random( g );
> until Order( s ) = 15 and NrMovedPoints( g ) = 1 + NrMovedPoints( s );
gap> 3A:= s^5;;
gap> repeat x:= Random( g ); until Order( x ) = 8;
gap> 2A:= x^4;;
gap> repeat x:= Random( g ); until Order( x ) = 12 and
> NrMovedPoints( g ) = 32 + NrMovedPoints( x^6 );
gap> 2B:= x^6;;
gap> prop15A:= List( [ 2A, 2B, 3A ],
> x -> RatioOfNongenerationTransPermGroup( g, x, s ) );
[ 23/35, 29/42, 149/224 ]
gap> Maximum( prop15A );
29/42
</pre></div>

This means that for s in the class <code class="code">15A</code>, we have P( S, s ) = 29/42, and the same holds for all s of order 15 since the three classes of element order 15 are conjugate under triality. Now we show that for s of order different from 15, the value P(g,s) is larger than 29/42, for g in one of the classes <code class="code">2A</code>, <code class="code">2B</code>, <code class="code">3A</code>, or their images under triality. This implies statement (c).

<div class="example"><pre>
gap> test:= List( [ 2A, 2B, 3A ], x -> ConjugacyClass( g, x ) );;
gap> ccl:= ConjugacyClasses( g );;
gap> consider:= Filtered( ccl, c -> Size( c ) in List( test, Size ) );;
gap> Length( consider );
7
gap> filt:= Filtered( ccl, c -> ForAll( consider, cc ->
> RatioOfNongenerationTransPermGroup( g, Representative( cc ),
> Representative( c ) ) <= 29/42 ) );;
gap> Length( filt );
3
gap> List( filt, c -> Order( Representative( c ) ) );
[ 15, 15, 15 ]
</pre></div>

Now we show statement (d). First we observe that all those Klein four groups in S whose involutions lie in the class <code class="code">2A</code> are conjugate in S. Note that this is the unique class of length 1575 in S, and also the unique class whose elements have 24 fixed points in the degree 120 permutation representation.

For that, we use the character table of S to read off that S contains exactly 14175 such subgroups, and we use the group to compute one such subgroup and its normalizer of index 14175.

<div class="example"><pre>
gap> SizesConjugacyClasses( t );
[ 1, 1575, 3780, 3780, 3780, 56700, 2240, 2240, 2240, 89600, 268800,
 37800, 340200, 907200, 907200, 907200, 2721600, 580608, 580608,
 580608, 100800, 100800, 100800, 604800, 604800, 604800, 806400,
 806400, 806400, 806400, 2419200, 2419200, 2419200, 7257600,
 24883200, 5443200, 5443200, 6451200, 6451200, 6451200, 8709120,
 8709120, 8709120, 1209600, 1209600, 1209600, 4838400, 7257600,
 7257600, 7257600, 11612160, 11612160, 11612160 ]
gap> NrPolyhedralSubgroups( t, 2, 2, 2 );
rec( number := 14175, type := "V4" )
gap> repeat x:= Random( g );
> until Order( x ) mod 2 = 0
> and NrMovedPoints( x^( Order(x)/2 ) ) = 120 - 24;
gap> x:= x^( Order(x)/2 );;
gap> repeat y:= x^Random( g );
> until NrMovedPoints( x*y ) = 120 - 24;
gap> v4:= SubgroupNC( g, [ x, y ] );;
gap> n:= Normalizer( g, v4 );;
gap> Index( g, n );
14175
</pre></div>

We verify that the triple has the required property.

<div class="example"><pre>
gap> maxorder:= RepresentativesMaximallyCyclicSubgroups( t );;
gap> maxorderreps:= List( ClassesPerhapsCorrespondingToTableColumns( g, t,
> maxorder ), Representative );;
gap> Length( maxorderreps );
28
gap> CommonGeneratorWithGivenElements( g, maxorderreps, [ x, y, x*y ] );
fail
</pre></div>

For the simple group S, it remains to show statement (e). We want to show that for any choice of two nonidentity elements x, y in S, there is an element s in the class <code class="code">15A</code> such that ⟨ s, x ⟩ = ⟨ s, y ⟩ = S holds. Only x, y in the classes given by the list <code class="code">testpos</code> must be considered, by the estimates σ(g,s).

We replace the values σ(g,s) by the exact values P(g,s), for g in one of these three classes. Each of the three classes is determined by its element order and its number of fixed points.

<div class="example"><pre>
gap> reps:= List( ccl, Representative );;
gap> bading:= List( testpos, i -> Filtered( reps,
> r -> Order( r ) = OrdersClassRepresentatives( t )[i] and
> NrMovedPoints( r ) = 120 - pi[i] ) );;
gap> List( bading, Length );
[ 1, 1, 1 ]
gap> bading:= List( bading, x -> x[1] );;
</pre></div>

For each pair (C_1, C_2) of classes represented by this list, we have to show that for any choice of elements x ∈ C_1, y ∈ C_2 there is s in the class <code class="code">15A</code> such that ⟨ s, x ⟩ = ⟨ s, y ⟩ = S holds. This is done with the random approach that is described in Section <a href="chap11.html#X83DACCF07EF62FAE">11.3-3</a>.

<div class="example"><pre>
gap> for pair in UnorderedTuples( bading, 2 ) do
> test:= RandomCheckUniformSpread( g, pair, s, 80 );
> if test <> true then
> Error( test );
> fi;
> od;
</pre></div>

We get no error message, so statement (e) holds.

Now we turn to the automorphic extensions of S. First we compute a permutation representation of SO^+(8,2) ≅ S.2 and an element g in the class <code class="code">2F</code>, which is the unique conjugacy class of size 120 in S.2.

<div class="example"><pre>
gap> matgrp:= SO(1,8,2);;
gap> g2:= Image( IsomorphismPermGroup( matgrp ) );;
gap> IsTransitive( g2, MovedPoints( g2 ) );
true
gap> repeat x:= Random( g2 ); until Order( x ) = 14;
gap> 2F:= x^7;;
gap> Size( ConjugacyClass( g2, 2F ) );
120
</pre></div>

Only for s in six conjugacy classes of S, there is a nonzero probability to have S.2 = ⟨ g, s ⟩.

<div class="example"><pre>
gap> der:= DerivedSubgroup( g2 );;
gap> cclreps:= List( ConjugacyClasses( der ), Representative );;
gap> nongen:= List( cclreps,
> x -> RatioOfNongenerationTransPermGroup( g2, 2F, x ) );;
gap> goodpos:= PositionsProperty( nongen, x -> x < 1 );;
gap> invariants:= List( goodpos, i -> [ Order( cclreps[i] ),
> Size( Centralizer( g2, cclreps[i] ) ), nongen[i] ] );;
gap> SortedList( invariants );
[ [ 10, 20, 1/3 ], [ 10, 20, 1/3 ], [ 12, 24, 2/5 ], [ 12, 24, 2/5 ],
 [ 15, 15, 0 ], [ 15, 15, 0 ] ]
</pre></div>

S contains three classes of element order 10, which are conjugate in S.3. For a fixed extension of the type S.2, the element s can be chosen only in two of these three classes, which means that there is another group of the type S.2 (more precisely, another subgroup of index three in S.S_3) in which this choice of s is not suitable –note that the general aim is to find s ∈ S uniformly for all automorphic extensions of S. Analogous statements hold for the other possibilities for s, so statement (f) follows.

Statement (g) follows from the list of maximal subgroups in <a href="chapBib.html#biBCCN85">[CCN+85, p. 85]</a>.

Statement (h) follows from the fact that S is the only maximal subgroup of S.3 that contains elements of order 15, according to the list of maximal subgroups in <a href="chapBib.html#biBCCN85">[CCN+85, p. 85]</a>. Alternatively, if we do not want to assume this information, we can use explicit computations, as follows. All we have to check is that any element in the classes <code class="code">3F</code> and <code class="code">3G</code> generates S.3 together with a fixed element of order 15 in S.

We compute a permutation representation of S.3 as the derived subgroup of a subgroup of the type S.S_3 inside the sporadic simple Fischer group Fi_22; these subgroups lie in the fourth class of maximal subgroups of Fi_22, see <a href="chapBib.html#biBCCN85">[CCN+85, p. 163]</a>. An element in the class <code class="code">3F</code> of S.3 can be found as a power of an order 21 element, and an element in the class <code class="code">3G</code> can be found as the fourth power of a <code class="code">12P</code> element.

<div class="example"><pre>
gap> aut:= Group( AtlasGenerators( "Fi22", 1, 4 ).generators );;
gap> Size( aut ) = 6 * Size( t );
true
gap> g3:= DerivedSubgroup( aut );;
gap> orbs:= OrbitsDomain( g3, MovedPoints( g3 ) );;
gap> List( orbs, Length );
[ 3150, 360 ]
gap> g3:= Action( g3, orbs[2] );;
gap> repeat s:= Random( g3 ); until Order( s ) = 15;
gap> repeat x:= Random( g3 ); until Order( x ) = 21;
gap> 3F:= x^7;;
gap> RatioOfNongenerationTransPermGroup( g3, 3F, s );
0
gap> repeat x:= Random( g3 );
> until Order( x ) = 12 and Size( Centralizer( g3, x^4 ) ) = 648;
gap> 3G:= x^4;;
gap> RatioOfNongenerationTransPermGroup( g3, 3G, s );
0
</pre></div>

Finally, consider statement (i). It implies that <a href="chapBib.html#biBBGK">[BGK08, Corollary 1.5]</a> holds for Ω^+(8,2), with s of order 15. Note that by part (f), s cannot be chosen in a prescribed conjugacy class of S that is independent of the elements x, y.

If x and y lie in S then statement (i) follows from part (e), and by part (g), the case that x or y lie in S.3 ∖ S is also not a problem. We now show that also x or y in S.2 ∖ S is not a problem. Here we have to deal with the cases that x and y lie in the same subgroup of index 3 in Aut(S) or in different such subgroups. Actually we show that for each index 3 subgroup H = S.2 < Aut(S), we can choose s from two of the three classes of element order 15 in S such that S is the only maximal subgroup of H that contains s, and thus ⟨ x, s ⟩ contains H, for any choice of x ∈ H ∖ S.

For that, we note that no novelty in S.2 contains elements of order 15, so all maximal subgroups of S.2 that contain such elements –besides S– have one of the indices 120, 135, 960, 1120, or 12096, and point stabilizers of the types S_6(2) × 2, 2^6:S_8, S_9, S_3 × U_4(2):2, or S_5 ≀ 2. We compute the corresponding permutation characters.

<div class="example"><pre>
gap> t2:= CharacterTable( "O8+(2).2" );;
gap> s:= CharacterTable( "S6(2)" ) * CharacterTable( "Cyclic", 2 );;
gap> pi:= PossiblePermutationCharacters( s, t2 );;
gap> prim:= pi;;
gap> pi:= PermChars( t2, rec( torso:= [ 135 ] ) );;
gap> Append( prim, pi );
gap> pi:= PossiblePermutationCharacters( CharacterTable( "A9.2" ), t2 );;
gap> Append( prim, pi );
gap> s:= CharacterTable( "Dihedral(6)" ) * CharacterTable( "U4(2).2" );;
gap> pi:= PossiblePermutationCharacters( s, t2 );;
gap> Append( prim, pi );
gap> s:= CharacterTableWreathSymmetric( CharacterTable( "S5" ), 2 );;
gap> pi:= PossiblePermutationCharacters( s, t2 );;
gap> Append( prim, pi );
gap> Length( prim );
5
gap> ord15:= PositionsProperty( OrdersClassRepresentatives( t2 ),
> x -> x = 15 );
[ 39, 40 ]
gap> List( prim, pi -> pi{ ord15 } );
[ [ 1, 0 ], [ 2, 0 ], [ 2, 0 ], [ 1, 0 ], [ 1, 0 ] ]
gap> List( ord15, i -> Maximum( ApproxP( prim, i ) ) );
[ 307/120, 0 ]
</pre></div>

Here it is appropriate to clean the workspace again.

<div class="example"><pre>
gap> CleanWorkspace();
</pre></div>

<a id="X78F0815B86253A1F" name="X78F0815B86253A1F"></a>

<h5>11.5-13 O_8^+(3)</h5>

We show that S = O_8^+(3) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>σ(S) = 863/1820, and this value is attained exactly for σ(S,s) with s of order 20.

</dd>
<dt>(b)</dt>
<dd>For s ∈ S of order 20, ��(S,s) consists of two nonconjugate groups of the type O_7(3) = Ω(7,3), two conjugate subgroups of the type 3^6:L_4(3), two nonconjugate subgroups of the type (A_4 × U_4(2)):2, and one subgroup of each of the types 2.U_4(3).(2^2)_122 and (A_6 × A_6):2^2.

</dd>
<dt>(c)</dt>
<dd>P(S) = 194/455, and this value is attained exactly for P(S,s) with s of order 20.

</dd>
<dt>(d)</dt>
<dd>The uniform spread of S is at least three, with s of order 20.

</dd>
<dt>(e)</dt>
<dd>The preimage of s in the matrix group 2.S = Ω^+(8,3) can be chosen of order 40, and then the maximal subgroups of 2.S containing s have the structures 2.O_7(3), 3^6:2.L_4(3), 4.U_4(3).2^2 = SU(4,3).2^2, 2.(A_4 × U_4(2)).2 = 2.(PSp(2,3) ⊗ PSp(4,3)).2, and 2.(A_6 × A_6):2^2 = 2.(Ω^-(4,3) × Ω^-(4,3)):2^2, respectively.

</dd>
<dt>(f)</dt>
<dd>For s ∈ S of order 20, we have P^'(S.2_1, s) ∈ { 83/567, 574/1215 }, P^'(S.2_2, s) ∈ { 0, 1 } (depending on the choice of s), and σ^'(S.3, s) = 0.

Furthermore, for any choice of s^' ∈ S, we have σ^'(S.2_2, s^') = 1 for some group S.2_2. However, if it is allowed to choose s from an Aut(S)-class of elements of order 20 (and not from a fixed S-class) then we can achieve σ(g,s) = 0 for any given g ∈ S.2_2 ∖ S.

</dd>
<dt>(g)</dt>
<dd>The maximal subgroups of S.2_1 that contain an element of order 20 are either S and the extensions of the subgroups listed in statement (b) or they are S and L_4(3).2^2, 3^6:L_4(3).2 (twice), 2.U_4(3).(2^2)_122.2, and (A_6 × A_6):2^2.2.

In the former case, the groups have the structures O_7(3):2 (twice), 3^6:(L_4(3) × 2) (twice), S_4 × U_4(2).2 (twice), 2.U_4(3).(2^2)_122.2, and (A_6 × A_6):2^2 × 2.

</dd>
</dl>
Statement (a) follows from inspection of the primitive permutation characters.

<div class="example"><pre>
gap> t:= CharacterTable( "O8+(3)" );;
gap> ProbGenInfoSimple( t );
[ "O8+(3)", 863/1820, 2, [ "20A", "20B", "20C" ], [ 8, 8, 8 ] ]
</pre></div>

Also statement (b) follows from the information provided by the character table of S (cf. <a href="chapBib.html#biBCCN85">[CCN+85, p. 140]</a>).

<div class="example"><pre>
gap> prim:= PrimitivePermutationCharacters( t );;
gap> ord:= OrdersClassRepresentatives( t );;
gap> spos:= Position( ord, 20 );;
gap> filt:= PositionsProperty( prim, x -> x[ spos ] <> 0 );
[ 1, 2, 7, 15, 18, 19, 24 ]
gap> Maxes( t ){ filt };
[ "O7(3)", "O8+(3)M2", "3^6:L4(3)", "2.U4(3).(2^2)_{122}",
 "(A4xU4(2)):2", "O8+(3)M19", "(A6xA6):2^2" ]
gap> prim{ filt }{ [ 1, spos ] };
[ [ 1080, 1 ], [ 1080, 1 ], [ 1120, 2 ], [ 189540, 1 ],
 [ 7960680, 1 ], [ 7960680, 1 ], [ 9552816, 1 ] ]
</pre></div>

For statement (c), we first show that P(S,s) = 194/455 holds. Since this value is larger than 1/3, we have to inspect only those classes g^S for which σ(g,s) ≥ 1/3 holds,

<div class="example"><pre>
gap> ord:= OrdersClassRepresentatives( t );;
gap> ord20:= PositionsProperty( ord, x -> x = 20 );;
gap> cand:= [];;
gap> for i in ord20 do
> approx:= ApproxP( prim, i );
> Add( cand, PositionsProperty( approx, x -> x >= 1/3 ) );
> od;
gap> cand;
[ [ 2, 6, 7, 10 ], [ 3, 6, 8, 11 ], [ 4, 6, 9, 12 ] ]
gap> AtlasClassNames( t ){ cand[1] };
[ "2A", "3A", "3B", "3E" ]
</pre></div>

The three possibilities form one orbit under the outer automorphism group of S.

<div class="example"><pre>
gap> t3:= CharacterTable( "O8+(3).3" );;
gap> tfust3:= GetFusionMap( t, t3 );;
gap> List( cand, x -> tfust3{ x } );
[ [ 2, 4, 5, 6 ], [ 2, 4, 5, 6 ], [ 2, 4, 5, 6 ] ]
</pre></div>

By symmetry, we may consider only the first possibility, and assume that s is in the class <code class="code">20A</code>.

We work with a permutation representation of degree 1080, and assume that the permutation character is <code class="code">1a+260a+819a</code>. (Note that all permutation characters of S of degree 1080 are conjugate under Aut(S).)

<div class="example"><pre>
gap> g:= Action( SO(1,8,3), NormedRowVectors( GF(3)^8 ), OnLines );;
gap> Size( g );
9904359628800
gap> g:= DerivedSubgroup( g );; Size( g );
4952179814400
gap> orbs:= OrbitsDomain( g, MovedPoints( g ) );;
gap> List( orbs, Length );
[ 1080, 1080, 1120 ]
gap> g:= Action( g, orbs[1] );;
gap> PositionProperty( Irr( t ), chi -> chi[1] = 819 );
9
gap> permchar:= Sum( Irr( t ){ [ 1, 2, 9 ] } );
Character( CharacterTable( "O8+(3)" ),
[ 1080, 128, 0, 0, 24, 108, 135, 0, 0, 108, 0, 0, 27, 27, 0, 0, 18,
 9, 12, 16, 0, 0, 4, 15, 0, 0, 20, 0, 0, 12, 11, 0, 0, 20, 0, 0, 15,
 0, 0, 12, 0, 0, 2, 0, 0, 3, 3, 0, 0, 6, 6, 0, 0, 3, 2, 2, 2, 18, 0,
 0, 9, 0, 0, 0, 0, 0, 0, 3, 3, 0, 0, 3, 0, 0, 12, 0, 0, 3, 0, 0, 0,
 0, 0, 4, 3, 3, 0, 0, 1, 0, 0, 4, 0, 0, 1, 1, 2, 0, 0, 0, 0, 0, 3,
 0, 0, 2, 0, 0, 5, 0, 0, 1, 0, 0 ] )
</pre></div>

Now we show that for s in the class <code class="code">20A</code> (which fixes one point), the proportion of nongenerating elements g in one of the classes <code class="code">2A</code>, <code class="code">3A</code>, <code class="code">3B</code>, <code class="code">3E</code> has the maximum 194/455, which is attained exactly for <code class="code">3A</code>. (We find a <code class="code">2A</code> element as a power of s, a <code class="code">3A</code> element as a power of any element of order 18, a <code class="code">3B</code> and a <code class="code">3E</code> element as elements with 135 and 108 fixed points, respectively, which occur as powers of suitable elements of order 15.)

<div class="example"><pre>
gap> permchar{ ord20 };
[ 1, 0, 0 ]
gap> AtlasClassNames( t )[ PowerMap( t, 10 )[ ord20[1] ] ];
"2A"
gap> ord18:= PositionsProperty( ord, x -> x = 18 );;
gap> Set( AtlasClassNames( t ){ PowerMap( t, 6 ){ ord18 } } );
[ "3A" ]
gap> ord15:= PositionsProperty( ord, x -> x = 15 );;
gap> PowerMap( t, 5 ){ ord15 };
[ 7, 8, 9, 10, 11, 12 ]
gap> AtlasClassNames( t ){ [ 7 .. 12 ] };
[ "3B", "3C", "3D", "3E", "3F", "3G" ]
gap> permchar{ [ 7 .. 12 ] };
[ 135, 0, 0, 108, 0, 0 ]
gap> mp:= NrMovedPoints( g );;
gap> ResetGlobalRandomNumberGenerators();
gap> repeat 20A:= Random( g );
> until Order( 20A ) = 20 and mp - NrMovedPoints( 20A ) = 1;
gap> 2A:= 20A^10;;
gap> repeat x:= Random( g ); until Order( x ) = 18;
gap> 3A:= x^6;;
gap> repeat x:= Random( g );
> until Order( x ) = 15 and mp - NrMovedPoints( x^5 ) = 135;
gap> 3B:= x^5;;
gap> repeat x:= Random( g );
> until Order( x ) = 15 and mp - NrMovedPoints( x^5 ) = 108;
gap> 3E:= x^5;;
gap> nongen:= List( [ 2A, 3A, 3B, 3E ],
> c -> RatioOfNongenerationTransPermGroup( g, c, 20A ) );
[ 3901/9477, 194/455, 451/1092, 451/1092 ]
gap> Maximum( nongen );
194/455
</pre></div>

Next we compute the values P(g,s), for g is in the class <code class="code">3A</code> and certain elements s. It is enough to consider representatives s of maximally cyclic subgroups in S, but here we can do better, as follows. Since <code class="code">3A</code> is the unique class of length 72800, it is fixed under Aut(S), so it is enough to consider one element s from each Aut(S)-orbit on the classes of S. We use the class fusion between the character tables of S and Aut(S) for computing orbit representatives.

<div class="example"><pre>
gap> maxorder:= RepresentativesMaximallyCyclicSubgroups( t );;
gap> Length( maxorder );
57
gap> autt:= CharacterTable( "O8+(3).S4" );;
gap> fus:= PossibleClassFusions( t, autt );;
gap> orbreps:= Set( fus, map -> Set( ProjectionMap( map ) ) );
[ [ 1, 2, 5, 6, 7, 13, 17, 18, 19, 20, 23, 24, 27, 30, 31, 37, 43,
 46, 50, 54, 55, 56, 57, 58, 64, 68, 72, 75, 78, 84, 85, 89, 95,
 96, 97, 100, 106, 112 ] ]
gap> totest:= Intersection( maxorder, orbreps[1] );
[ 43, 50, 54, 56, 57, 64, 68, 75, 78, 84, 85, 89, 95, 97, 100, 106,
 112 ]
gap> Length( totest );
17
gap> AtlasClassNames( t ){ totest };
[ "6Q", "6X", "6B1", "8A", "8B", "9G", "9K", "12A", "12D", "12J",
 "12K", "12O", "13A", "14A", "15A", "18A", "20A" ]
</pre></div>

This means that we have to test one element of each of the element orders 13, 14, 15, and 18 (note that we know already a bound for elements of order 20), plus certain elements of the orders 6, 8, 9, and 12 which can be identified by their centralizer orders and (for elements of order 6 and 8) perhaps the centralizer orders of some powers.

<div class="example"><pre>
gap> elementstotest:= [];;
gap> for elord in [ 13, 14, 15, 18 ] do
> repeat s:= Random( g );
> until Order( s ) = elord;
> Add( elementstotest, s );
> od;
</pre></div>

The next elements to be tested are in the classes <code class="code">6B1</code> (centralizer order 162), in one of <code class="code">9G</code>–<code class="code">9J</code> (centralizer order 729), in one of <code class="code">9K</code>–<code class="code">9N</code> (centralizer order 81), in one of <code class="code">12A</code>–<code class="code">12C</code> (centralizer order 1728), in one of <code class="code">12D</code>–<code class="code">12I</code> (centralizer order 432), in <code class="code">12J</code> (centralizer order 192), in one of <code class="code">12K</code>–<code class="code">12N</code> (centralizer order 108), and in one of <code class="code">12O</code>–<code class="code">12T</code> (centralizer order 72).

<div class="example"><pre>
gap> ordcent:= [ [ 6, 162 ], [ 9, 729 ], [ 9, 81 ], [ 12, 1728 ],
> [ 12, 432 ], [ 12, 192 ], [ 12, 108 ], [ 12, 72 ] ];;
gap> cents:= SizesCentralizers( t );;
gap> for pair in ordcent do
> Print( pair, ": ", AtlasClassNames( t ){
> Filtered( [ 1 .. Length( ord ) ],
> i -> ord[i] = pair[1] and cents[i] = pair[2] ) }, "\n" );
> repeat s:= Random( g );
> until Order( s ) = pair[1] and Size( Centralizer( g, s ) ) = pair[2];
> Add( elementstotest, s );
> od;
[ 6, 162 ]: [ "6B1" ]
[ 9, 729 ]: [ "9G", "9H", "9I", "9J" ]
[ 9, 81 ]: [ "9K", "9L", "9M", "9N" ]
[ 12, 1728 ]: [ "12A", "12B", "12C" ]
[ 12, 432 ]: [ "12D", "12E", "12F", "12G", "12H", "12I" ]
[ 12, 192 ]: [ "12J" ]
[ 12, 108 ]: [ "12K", "12L", "12M", "12N" ]
[ 12, 72 ]: [ "12O", "12P", "12Q", "12R", "12S", "12T" ]
</pre></div>

The next elements to be tested are in one of the classes <code class="code">6Q</code>–<code class="code">6S</code> (centralizer order 648).

<div class="example"><pre>
gap> AtlasClassNames( t ){ Filtered( [ 1 .. Length( ord ) ],
> i -> cents[i] = 648 and cents[ PowerMap( t, 2 )[i] ] = 52488
> and cents[ PowerMap( t, 3 )[i] ] = 26127360 ) };
[ "6Q", "6R", "6S" ]
gap> repeat s:= Random( g );
> until Order( s ) = 6 and Size( Centralizer( g, s ) ) = 648
> and Size( Centralizer( g, s^2 ) ) = 52488
> and Size( Centralizer( g, s^3 ) ) = 26127360;
gap> Add( elementstotest, s );
</pre></div>

The next elements to be tested are in the class <code class="code">6X</code>–<code class="code">6A1</code> (centralizer order 648).

<div class="example"><pre>
gap> AtlasClassNames( t ){ Filtered( [ 1 .. Length( ord ) ],
> i -> cents[i] = 648 and cents[ PowerMap( t, 2 )[i] ] = 52488
> and cents[ PowerMap( t, 3 )[i] ] = 331776 ) };
[ "6X", "6Y", "6Z", "6A1" ]
gap> repeat s:= Random( g );
> until Order( s ) = 6 and Size( Centralizer( g, s ) ) = 648
> and Size( Centralizer( g, s^2 ) ) = 52488
> and Size( Centralizer( g, s^3 ) ) = 331776;
gap> Add( elementstotest, s );
</pre></div>

Finally, we add elements from the classes <code class="code">8A</code> and <code class="code">8B</code>.

<div class="example"><pre>
gap> AtlasClassNames( t ){ Filtered( [ 1 .. Length( ord ) ],
> i -> ord[i] = 8 and cents[ PowerMap( t, 2 )[i] ] = 13824 ) };
[ "8A" ]
gap> repeat s:= Random( g );
> until Order( s ) = 8 and Size( Centralizer( g, s^2 ) ) = 13824;
gap> Add( elementstotest, s );
gap> AtlasClassNames( t ){ Filtered( [ 1 .. Length( ord ) ],
> i -> ord[i] = 8 and cents[ PowerMap( t, 2 )[i] ] = 1536 ) };
[ "8B" ]
gap> repeat s:= Random( g );
> until Order( s ) = 8 and Size( Centralizer( g, s^2 ) ) = 1536;
gap> Add( elementstotest, s );
</pre></div>

Now we compute the ratios. It turns out that from these candidates, only elements s of the orders 14 and 15 satisfy P(g,s) < 194/455.

<div class="example"><pre>
gap> nongen:= List( elementstotest,
> s -> RatioOfNongenerationTransPermGroup( g, 3A, s ) );;
gap> smaller:= PositionsProperty( nongen, x -> x < 194/455 );
[ 2, 3 ]
gap> nongen{ smaller };
[ 127/325, 1453/3640 ]
</pre></div>

So the only candidates for s that may be better than order 20 elements are elements of order 14 or 15. In order to exclude these two possibilities, we compute P(g,s) for s in the class <code class="code">14A</code> and g = s^7 in the class <code class="code">2A</code>, and for s in the class <code class="code">15A</code> and g in the class <code class="code">2A</code>, which yields values that are larger than 194/455.

<div class="example"><pre>
gap> repeat s:= Random( g );
> until Order( s ) = 14 and NrMovedPoints( s ) = 1078;
gap> 2A:= s^7;;
gap> nongen:= RatioOfNongenerationTransPermGroup( g, 2A, s );
1573/3645
gap> nongen > 194/455;
true
gap> repeat s:= Random( g );
> until Order( s ) = 15 and NrMovedPoints( s ) = 1080 - 3;
gap> nongen:= RatioOfNongenerationTransPermGroup( g, 2A, s );
490/1053
gap> nongen > 194/455;
true
</pre></div>

For statement (d), we show that for each triple of elements in the union of the classes <code class="code">2A</code>, <code class="code">3A</code>, <code class="code">3B</code>, <code class="code">3E</code> there is an element in the class <code class="code">20A</code> that generates S together with each element of the triple.

<div class="example"><pre>
gap> for tup in UnorderedTuples( [ 2A, 3A, 3B, 3E ], 3 ) do
> cl:= ShallowCopy( tup );
> test:= RandomCheckUniformSpread( g, cl, 20A, 100 );
> if test <> true then
> Error( test );
> fi;
> od;
</pre></div>

We get no error message, so statement (d) is true.

For statement (e), first we show that 2.S = Ω^+(8,3) contains elements of order 40 but S does not.

<div class="example"><pre>
gap> der:= DerivedSubgroup( SO(1,8,3) );;
gap> repeat x:= PseudoRandom( der ); until Order( x ) = 40;
gap> 40 in ord;
false
</pre></div>

Thus elements of order 40 must arise as preimages of order 20 elements under the natural epimorphism from 2.S to S, which means that we may choose an order 40 preimage hats of s. Then ��(2.S, hats) consists of central extensions of the subgroups listed in statement (b). The perfect subgroups O_7(3), L_4(3), 2.U_4(3), and U_4(2) of these groups must lift to their Schur double covers in 2.S because otherwise the preimages would not contain elements of order 40.

Next we consider the preimage of the subgroup U = (A_4 × U_4(2)).2 of S. We show that the preimages of the two direct factors A_4 and U_4(2) in U^' = A_4 × U_4(2) are Schur covers. For A_4, this follows from the fact that the preimage of U^' must contain elements of order 20, and that U_4(2) does not contain elements of order 10.

<div class="example"><pre>
gap> u42:= CharacterTable( "U4(2)" );;
gap> Filtered( OrdersClassRepresentatives( u42 ), x -> x mod 5 = 0 );
[ 5 ]
</pre></div>

In order to show that the U_4(2) type subgroup of U^' lifts to its double cover in 2.S, we note that the class <code class="code">2B</code> of U_4(2) lifts to a class of elements of order four in the double cover 2.U_4(2), and that the corresponding class of elements in U is S-conjugate to the class of involutions in the direct factor A_4 (which is the unique class of length three in U).

<div class="example"><pre>
gap> u:= CharacterTable( Maxes( t )[18] );
CharacterTable( "(A4xU4(2)):2" )
gap> 2u42:= CharacterTable( "2.U4(2)" );;
gap> OrdersClassRepresentatives( 2u42 )[4];
4
gap> GetFusionMap( 2u42, u42 )[4];
3
gap> OrdersClassRepresentatives( u42 )[3];
2
gap> List( PossibleClassFusions( u42, u ), x -> x[3] );
[ 8 ]
gap> PositionsProperty( SizesConjugacyClasses( u ), x -> x = 3 );
[ 2 ]
gap> ForAll( PossibleClassFusions( u, t ), x -> x[2] = x[8] );
true
</pre></div>

The last subgroup for which the structure of the preimage has to be shown is U = (A_6 × A_6):2^2. We claim that each of the A_6 type subgroups in the derived subgroup U^' = A_6 × A_6 lifts to its double cover in 2.S. Since all elements of order 20 in U lie in U^', at least one of the two direct factors must lift to its double cover, in order to give rise to an order 40 element in U. In fact both factors lift to the double cover since the two direct factors are interchanged by conjugation in U; the latter follows form tha fact that U has no normal subgroup of type A_6.

<div class="example"><pre>
gap> u:= CharacterTable( Maxes( t )[24] );
CharacterTable( "(A6xA6):2^2" )
gap> ClassPositionsOfDerivedSubgroup( u );
[ 1 .. 22 ]
gap> PositionsProperty( OrdersClassRepresentatives( u ), x -> x = 20 );
[ 8 ]
gap> List( ClassPositionsOfNormalSubgroups( u ),
> x -> Sum( SizesConjugacyClasses( u ){ x } ) );
[ 1, 129600, 259200, 259200, 259200, 518400 ]
</pre></div>

So statement (e) holds.

For statement (f), we have to consider the upward extensions S.2_1, S.2_2, and S.3.

First we look at S.2_1, an extension by an outer automorphism that acts as a double transposition in the outer automorphism group S_4. Note that the symmetry between the three classes of element oder 20 in S is broken in S.2_1, two of these classes have square roots in S.2_1, the third has not.

<div class="example"><pre>
gap> t2:= CharacterTable( "O8+(3).2_1" );;
gap> ord20:= PositionsProperty( OrdersClassRepresentatives( t2 ),
> x -> x = 20 );;
gap> ord20:= Intersection( ord20, ClassPositionsOfDerivedSubgroup( t2 ) );
[ 84, 85, 86 ]
gap> List( ord20, x -> x in PowerMap( t2, 2 ) );
[ false, true, true ]
</pre></div>

Changing the viewpoint, we see that for each class of element order 20 in S, there is a group of the type S.2_1 in which the elements in this class do not have square roots, and there are groups of this type in which these elements have square roots. So we have to deal with two different cases, and we do this by first collecting the permutation characters induced from all maximal subgroups of S.2_1 (other than S) that contain elements of order 20 in S, and then considering s in each of these classes of S.

We fix an embedding of S into S.2_1 in which the elements in the class <code class="code">20A</code> do not have square roots. This situation is given for the stored class fusion between the tables in the GAP Character Table Library.

<div class="example"><pre>
gap> tfust2:= GetFusionMap( t, t2 );;
gap> tfust2{ PositionsProperty( OrdersClassRepresentatives( t ),
> x -> x = 20 ) };
[ 84, 85, 86 ]
</pre></div>

The six different actions of S on the cosets of O_7(3) type subgroups induce pairwise different permutation characters that form an orbit under the action of Aut(S). Four of these characters cannot extend to S.2_1, the other two extend to permutation characters of S.2_1 on the cosets of O_7(3).2 type subgroups; these subgroups contain <code class="code">20A</code> elements.

<div class="example"><pre>
gap> primt2:= [];;
gap> poss:= PossiblePermutationCharacters( CharacterTable( "O7(3)" ), t );;
gap> invfus:= InverseMap( tfust2 );;
gap> List( poss, pi -> ForAll( CompositionMaps( pi, invfus ), IsInt ) );
[ false, false, false, false, true, true ]
gap> PossiblePermutationCharacters(
> CharacterTable( "O7(3)" ) * CharacterTable( "Cyclic", 2 ), t2 );
[ ]
gap> ext:= PossiblePermutationCharacters( CharacterTable( "O7(3).2" ), t2 );;
gap> List( ext, pi -> pi{ ord20 } );
[ [ 1, 0, 0 ], [ 1, 0, 0 ] ]
gap> Append( primt2, ext );
</pre></div>

The novelties in S.2_1 that arise from O_7(3) type subgroups of S have the structure L_4(3).2^2. These subgroups contain elements in the classes <code class="code">20B</code> and <code class="code">20C</code> of S.

<div class="example"><pre>
gap> ext:= PossiblePermutationCharacters( CharacterTable( "L4(3).2^2" ), t2 );;
gap> List( ext, pi -> pi{ ord20 } );
[ [ 0, 0, 1 ], [ 0, 1, 0 ] ]
gap> Append( primt2, ext );
</pre></div>

Note that from the possible permutation characters of S.2_1 on the cosets of L_4(3):2 × 2 type subgroups, we see that such subgroups must contain <code class="code">20A</code> elements, i. e., all such subgroups of S.2_1 lie inside O_7(3).2 type subgroups. This means that the structure description of these novelties in <a href="chapBib.html#biBCCN85">[CCN+85, p. 140]</a> is not correct. The correct structure is L_4(3).2^2.)

<div class="example"><pre>
gap> List( PossiblePermutationCharacters( CharacterTable( "L4(3).2_2" ) *
> CharacterTable( "Cyclic", 2 ), t2 ), pi -> pi{ ord20 } );
[ [ 1, 0, 0 ] ]
</pre></div>

All 3^6:L_4(3) type subgroups of S extend to S.2_1. We compute these permutation characters as the possible permutation characters of the right degree.

<div class="example"><pre>
gap> ext:= PermChars( t2, rec( torso:= [ 1120 ] ) );;
gap> List( ext, pi -> pi{ ord20 } );
[ [ 2, 0, 0 ], [ 0, 0, 2 ], [ 0, 2, 0 ] ]
gap> Append( primt2, ext );
</pre></div>

Also all 2.U_4(3).2^2 type subgroups of S extend to S.2_1. We compute the permutation characters as the extensions of the corresponding permutation characters of S.

<div class="example"><pre>
gap> filt:= Filtered( prim, x -> x[1] = 189540 );;
gap> cand:= List( filt, x -> CompositionMaps( x, invfus ) );;
gap> ext:= Concatenation( List( cand,
> pi -> PermChars( t2, rec( torso:= pi ) ) ) );;
gap> List( ext, x -> x{ ord20 } );
[ [ 1, 0, 0 ], [ 0, 1, 0 ], [ 0, 0, 1 ] ]
gap> Append( primt2, ext );
</pre></div>

The extensions of (A_4 × U_4(2)):2 type subgroups of S to S.2_1 have the type S_4 × U_4(2):2, they contain <code class="code">20A</code> elements.

<div class="example"><pre>
gap> ext:= PossiblePermutationCharacters( CharacterTable( "Symmetric", 4 ) *
> CharacterTable( "U4(2).2" ), t2 );;
gap> List( ext, x -> x{ ord20 } );
[ [ 1, 0, 0 ], [ 1, 0, 0 ] ]
gap> Append( primt2, ext );
</pre></div>

All (A_6 × A_6):2^2 type subgroups of S extend to S.2_1. We compute the permutation characters as the extensions of the corresponding permutation characters of S.

<div class="example"><pre>
gap> filt:= Filtered( prim, x -> x[1] = 9552816 );;
gap> cand:= List( filt, x -> CompositionMaps( x, InverseMap( tfust2 ) ));;
gap> ext:= Concatenation( List( cand,
> pi -> PermChars( t2, rec( torso:= pi ) ) ) );;
gap> List( ext, x -> x{ ord20 } );
[ [ 1, 0, 0 ], [ 0, 1, 0 ], [ 0, 0, 1 ] ]
gap> Append( primt2, ext );
</pre></div>

We have found all relevant permutation characters of S.2_1. This together with the list in <a href="chapBib.html#biBCCN85">[CCN+85, p. 140]</a> implies statement (g).

Now we compute the bounds σ^'(S.2_1, s).

<div class="example"><pre>
gap> Length( primt2 );
15
gap> approx:= List( ord20, x -> ApproxP( primt2, x ) );;
gap> outer:= Difference(
> PositionsProperty( OrdersClassRepresentatives( t2 ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t2 ) );;
gap> List( approx, l -> Maximum( l{ outer } ) );
[ 574/1215, 83/567, 83/567 ]
</pre></div>

Next we look at S.2_2, an extension by an outer automorphism that acts as a transposition in the outer automorphism group S_4. Similar to the above situation, the symmetry between the three classes of element oder 20 in S is broken also in S.2_2: The first is a conjugacy class of S.2_2, the other two classes are fused in S.2_2,

<div class="example"><pre>
gap> t2:= CharacterTable( "O8+(3).2_2" );;
gap> ord20:= PositionsProperty( OrdersClassRepresentatives( t2 ),
> x -> x = 20 );;
gap> ord20:= Intersection( ord20, ClassPositionsOfDerivedSubgroup( t2 ) );
[ 82, 83 ]
gap> tfust2:= GetFusionMap( t, t2 );;
gap> tfust2{ PositionsProperty( OrdersClassRepresentatives( t ),
> x -> x = 20 ) };
[ 82, 83, 83 ]
</pre></div>

Like in the case S.2_1, we compute the permutation characters induced from all maximal subgroups of S.2_2 (other than S) that contain elements of order 20 in S.

We fix the embedding of S into S.2_2 in which the class <code class="code">20A</code> of S is a class of S.2_2. This situation is given for the stored class fusion between the tables in the GAP Character Table Library.

Exactly two classes of O_7(3) type subgroups in S extend to S.2_2, these groups contain <code class="code">20A</code> elements.

<div class="example"><pre>
gap> primt2:= [];;
gap> ext:= PermChars( t2, rec( torso:= [ 1080 ] ) );;
gap> List( ext, pi -> pi{ ord20 } );
[ [ 1, 0 ], [ 1, 0 ] ]
gap> Append( primt2, ext );
</pre></div>

Only one class of 3^6:L_4(3) type subgroups extends to S.2_2. (Note that we need not consider the novelties of the type 3^3+6:(L_3(3) × 2), because the order of these groups is not divisible by 5.)

<div class="example"><pre>
gap> ext:= PermChars( t2, rec( torso:= [ 1120 ] ) );;
gap> List( ext, pi -> pi{ ord20 } );
[ [ 2, 0 ] ]
gap> Append( primt2, ext );
</pre></div>

Only one class of 2.U_4(3).2^2 type subgroups of S extends to S.2_2. We compute the permutation character as the extension of the corresponding permutation characters of S.

<div class="example"><pre>
gap> filt:= Filtered( prim, x -> x[1] = 189540 );;
gap> cand:= List( filt, x -> CompositionMaps( x, InverseMap( tfust2 ) ));;
gap> ext:= Concatenation( List( cand,
> pi -> PermChars( t2, rec( torso:= pi ) ) ) );;
gap> List( ext, x -> x{ ord20 } );
[ [ 1, 0 ] ]
gap> Append( primt2, ext );
</pre></div>

Two classes of (A_4 × U_4(2)):2 type subgroups of S extend to S.2_2.

<div class="example"><pre>
gap> filt:= Filtered( prim, x -> x[1] = 7960680 );;
gap> cand:= List( filt, x -> CompositionMaps( x, InverseMap( tfust2 ) ));;
gap> ext:= Concatenation( List( cand,
> pi -> PermChars( t2, rec( torso:= pi ) ) ) );;
gap> List( ext, x -> x{ ord20 } );
[ [ 1, 0 ], [ 1, 0 ] ]
gap> Append( primt2, ext );
</pre></div>

Exactly one class of (A_6 × A_6):2^2 type subgroups in S extends to S.2_2, and the extensions have the structure S_6 ≀ 2.

<div class="example"><pre>
gap> ext:= PossiblePermutationCharacters( CharacterTableWreathSymmetric(
> CharacterTable( "S6" ), 2 ), t2 );;
gap> List( ext, x -> x{ ord20 } );
[ [ 1, 0 ] ]
gap> Append( primt2, ext );
</pre></div>

We have found all relevant permutation characters of S.2_2, and compute the bounds σ^'(S.2_2, s).

<div class="example"><pre>
gap> Length( primt2 );
7
gap> approx:= List( ord20, x -> ApproxP( primt2, x ) );;
gap> outer:= Difference(
> PositionsProperty( OrdersClassRepresentatives( t2 ), IsPrimeInt ),
> ClassPositionsOfDerivedSubgroup( t2 ) );;
gap> List( approx, l -> Maximum( l{ outer } ) );
[ 14/9, 0 ]
</pre></div>

This means that there is an extension of the type S.2_2 in which s cannot be chosen such that the bound is less than 1/2. More precisely, we have σ(g,s) ≥ 1/2 exactly for g in the unique outer involution class of size 1080.

<div class="example"><pre>
gap> approx:= ApproxP( primt2, ord20[1] );;
gap> bad:= Filtered( outer, i -> approx[i] >= 1/2 );
[ 84 ]
gap> OrdersClassRepresentatives( t2 ){ bad };
[ 2 ]
gap> SizesConjugacyClasses( t2 ){ bad };
[ 1080 ]
gap> Number( SizesConjugacyClasses( t2 ), x -> x = 1080 );
1
</pre></div>

So we compute the proportion of elements in this class that generate S.2_2 together with an element s of order 20 in S. (As above, we have to consider two conjugacy classes.) For that, we first compute a permutation representation of S.2_2, using that S.2_2 is isomporphic to the two subgroups of index 2 in PGO^+(8,3) = O_8^+(3).2^2_122 that are different from PSO^+(8,3) = O_8^+(3).2_1, cf. <a href="chapBib.html#biBCCN85">[CCN+85, p. 140]</a>.

<div class="example"><pre>
gap> go:= GO(1,8,3);;
gap> so:= SO(1,8,3);;
gap> outerelm:= First( GeneratorsOfGroup( go ), x -> not x in so );;
gap> g2:= ClosureGroup( DerivedSubgroup( so ), outerelm );;
gap> Size( g2 );
19808719257600
gap> dom:= NormedRowVectors( GF(3)^8 );;
gap> orbs:= OrbitsDomain( g2, dom, OnLines );;
gap> List( orbs, Length );
[ 1080, 1080, 1120 ]
gap> act:= Action( g2, orbs[1], OnLines );;
</pre></div>

An involution g can be found as a power of one of the given generators.

<div class="example"><pre>
gap> Order( outerelm );
26
gap> g:= Permutation( outerelm^13, orbs[1], OnLines );;
gap> Size( ConjugacyClass( act, g ) );
1080
</pre></div>

Now we find the candidates for the elements s, and compute their ratios of nongeneration.

<div class="example"><pre>
gap> ord20;
[ 82, 83 ]
gap> SizesCentralizers( t2 ){ ord20 };
[ 40, 20 ]
gap> der:= DerivedSubgroup( act );;
gap> repeat 20A:= Random( der );
> until Order( 20A ) = 20 and Size( Centralizer( act, 20A ) ) = 40;
gap> RatioOfNongenerationTransPermGroup( act, g, 20A );
1
gap> repeat 20BC:= Random( der );
> until Order( 20BC ) = 20 and Size( Centralizer( act, 20BC ) ) = 20;
gap> RatioOfNongenerationTransPermGroup( act, g, 20BC );
0
</pre></div>

This means that for s in one S-class of elements of order 20, we have P^'(g, s) = 1, and s in the other two S-classes of elements of order 20 generates with any conjugate of g.

Concerning S.2_2, it remains to show that we cannot find a better element than s. For that, we first compute class representatives s^' in S, w.r.t. conjugacy in S.2_2, and then compute P^'( s^', g ). (It would be enough to check representatives of classes of maximal element order, but computing all classes is easy enough.)

<div class="example"><pre>
gap> ccl:= ConjugacyClasses( act );;
gap> der:= DerivedSubgroup( act );;
gap> reps:= Filtered( List( ccl, Representative ), x -> x in der );;
gap> Length( reps );
83
gap> ratios:= List( reps,
> s -> RatioOfNongenerationTransPermGroup( act, g, s ) );;
gap> cand:= PositionsProperty( ratios, x -> x < 1 );;
gap> ratios:= ratios{ cand };;
gap> SortParallel( ratios, cand );
gap> ratios;
[ 0, 1/10, 1/10, 16/135, 1/3, 1/3, 11/27, 7/15, 7/15 ]
</pre></div>

For S.2_2, it remains to show that there is no element s^' ∈ S such that P^'( s^'}^x, g ) < 1 holds for any x ∈ Aut(S) and g ∈ S.2_2. So we are done when we can show that each class given by <code class="code">cand</code> is conjugate in S.3 to a class outside <code class="code">cand</code>. The classes can be identified by element orders and centralizer orders.

<div class="example"><pre>
gap> invs:= List( cand,
> x -> [ Order( reps[x] ), Size( Centralizer( der, reps[x] ) ) ] );
[ [ 20, 20 ], [ 18, 108 ], [ 18, 108 ], [ 14, 28 ], [ 15, 45 ],
 [ 15, 45 ], [ 10, 40 ], [ 12, 72 ], [ 12, 72 ] ]
</pre></div>

Namely, <code class="code">cand</code> contains no full S.3-orbit of classes of the element orders 20, 18, 14, 15, and 10; also, <code class="code">cand</code> does not contain full S.3-orbits on the classes <code class="code">12O</code>–<code class="code">12T</code>.

Finally, we deal with S.3. The fact that no maximal subgroup of S containing an element of order 20 extends to S.3 follows either from the list of maximal subgroups of S in <a href="chapBib.html#biBCCN85">[CCN+85, p. 140]</a> or directly from the permutation characters.

<div class="example"><pre>
gap> t3:= CharacterTable( "O8+(3).3" );;
gap> tfust3:= GetFusionMap( t, t3 );;
gap> inv:= InverseMap( tfust3 );;
gap> filt:= PositionsProperty( prim, x -> x[ spos ] <> 0 );;
gap> ForAll( prim{ filt },
> pi -> ForAny( CompositionMaps( pi, inv ), IsList ) );
true
</pre></div>

So we have to consider only the classes of novelties in S.3, but the order of none of these groups is divisible by 20 –again see <a href="chapBib.html#biBCCN85">[CCN+85, p. 140]</a>). This means that any element in S.3 ∖ S together with an element of order 20 in S generates S.3. This is in fact stronger than statement (f), which claims this property only for elements of prime order in S.3 ∖ S (and their roots); note that S.3 ∖ S contains elements of the orders 9 and 27.

<div class="example"><pre>
gap> outer:= Difference( [ 1 .. NrConjugacyClasses( t3 ) ],
> ClassPositionsOfDerivedSubgroup( t3 ) );
[ 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69,
 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86,
 87, 88, 89, 90, 91, 92, 93, 94 ]
gap> Set( OrdersClassRepresentatives( t3 ){ outer } );
[ 3, 6, 9, 12, 18, 21, 24, 27, 36, 39 ]
</pre></div>

Before we turn to the next computations, we clean the workspace.

<div class="example"><pre>
gap> CleanWorkspace();
</pre></div>

<a id="X85BACC4A83F73392" name="X85BACC4A83F73392"></a>

<h5>11.5-14 O^+_8(4)</h5>

We show that S = O^+_8(4) = Ω^+(8,4) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For suitable s ∈ S of the type 2^- perp 6^- (i. e., s decomposes the natural 8-dimensional module for S into an orthogonal sum of two irreducible modules of the dimensions 2 and 6, respectively) and of order 65, ��(S,s) consists of exactly three pairwise nonconjugate subgroups of the type (5 × O^-_6(4)).2 = (5 × Ω^-(6,4)).2.

</dd>
<dt>(b)</dt>
<dd>σ( S, s ) ≤ 34817 / 1645056.

</dd>
<dt>(c)</dt>
<dd>In the extensions S.2_1 and S.3 of S by graph automorphisms, there is at most one maximal subgroup besides S that contains s. For the extension S.2_2 of S by a field automorphism, we have σ^'(S.2_2, s) = 0. In the extension S.2_3 of S by the product of an involutory graph automorphism and a field automorphism, there is a unique maximal subgroup besides S that contains s.

</dd>
</dl>
A safe source for determining ��(S,s) is <a href="chapBib.html#biBKle87">[Kle87]</a>. By inspection of the result matrix in this paper, we get that the only maximal subgroups of S that contain elements of order 65 occur in the rows 9–14 and 23–25; they have the isomorphism types S_6(4) = Sp(6,4) ≅ O_7(4) = Ω(7,4) and (5 × O_6^-(4)).2 = (5 × Ω^-(6,4)).2, respectively, and for each of these, there are three conjugacy classes of subgroups in S, which are conjugate under the triality graph automorphism of S.

We start with the natural matrix representation of S. For convenience, we compute an isomorphic permutation group on 5525 points.

<div class="example"><pre>
gap> q:= 4;; n:= 8;;
gap> G:= DerivedSubgroup( SO( 1, n, q ) );;
gap> points:= NormedRowVectors( GF(q)^n );;
gap> orbs:= OrbitsDomain( G, points, OnLines );;
gap> List( orbs, Length );
[ 5525, 16320 ]
gap> hom:= ActionHomomorphism( G, orbs[1], OnLines );;
gap> G:= Image( hom );;
</pre></div>

The group S contains exactly six conjugacy classes of (cyclic) subgroups of order 65; this follows from the fact that the centralizer of any Sylow 13 subgroup in S has the structure 5 × 5 × 13.

<div class="example"><pre>
gap> Collected( Factors( Size( G ) ) );
[ [ 2, 24 ], [ 3, 5 ], [ 5, 4 ], [ 7, 1 ], [ 13, 1 ], [ 17, 2 ] ]
gap> ResetGlobalRandomNumberGenerators();
gap> repeat x:= Random( G );
> until Order( x ) mod 13 = 0;
gap> x:= x^( Order( x ) / 13 );;
gap> c:= Centralizer( G, x );;
gap> IsAbelian( c ); AbelianInvariants( c );
true
[ 5, 5, 13 ]
</pre></div>

The group S_6(4) contains exactly one class of subgroups of order 65, since the conjugacy classes of elements of order 65 in S_6(4) are algebraically conjugate.

<div class="example"><pre>
gap> t:= CharacterTable( "S6(4)" );;
gap> ord65:= PositionsProperty( OrdersClassRepresentatives( t ),
> x -> x = 65 );
[ 105, 106, 107, 108, 109, 110, 111, 112 ]
gap> ord65 = ClassOrbit( t, ord65[1] );
true
</pre></div>

Thus there are at least three classes of order 65 elements in S that are not contained in S_6(4) type subgroups of S. So we choose such an element s, and have to consider only overgroups of the type (5 × Ω^-(6,4)).2.

The group Ω^-(6,4) ≅ U_4(4) contains exactly one class of subgroups of order 65.

<div class="example"><pre>
gap> t:= CharacterTable( "U4(4)" );;
gap> ords:= OrdersClassRepresentatives( t );;
gap> ord65:= PositionsProperty( ords, x -> x = 65 );;
gap> ord65 = ClassOrbit( t, ord65[1] );
true
</pre></div>

So 5 × Ω^-(6,4) contains exactly six such classes. Furthermore, subgroups in different classes are not S-conjugate.

<div class="example"><pre>
gap> syl5:= SylowSubgroup( c, 5 );;
gap> elms:= Filtered( Elements( syl5 ), y -> Order( y ) = 5 );;
gap> reps:= Set( elms, SmallestGeneratorPerm );; Length( reps );
6
gap> reps65:= List( reps, y -> SubgroupNC( G, [ y * x ] ) );;
gap> pairs:= Combinations( reps65, 2 );;
gap> ForAny( pairs, p -> IsConjugate( G, p[1], p[2] ) );
false
</pre></div>

We consider only subgroups M ≤ S in the three S-classes of the type (5 × Ω^-(6,4)).2.

<div class="example"><pre>
gap> cand:= List( reps, y -> Normalizer( G, SubgroupNC( G, [ y ] ) ) );;
gap> cand:= Filtered( cand, y -> Size( y ) = 10 * Size( t ) );;
gap> Length( cand );
3
</pre></div>

(Note that one of the members in ��(S,s) is the stabilizer in S of the orthogonal decomposition 2^- perp 6^-, the other two members are not reducible.)

By the above, the classes of subgroups of order 65 in each such M are in bijection with the corresponding classes in S. Since N_S(⟨ g ⟩) ⊆ M holds for any g ∈ M of order 65, also the conjugacy classes of elements of order 65 in M are in bijection with those in S.

<div class="example"><pre>
gap> norms:= List( reps65, y -> Normalizer( G, y ) );;
gap> ForAll( norms, y -> ForAll( cand, M -> IsSubset( M, y ) ) );
true
</pre></div>

As a consequence, we have g^S ∩ M = g^M and thus 1_M^S(g) = 1. This implies statement (a).

In order to show statement (b), we want to use the function <code class="code">UpperBoundFixedPointRatios</code> introduced in Section <a href="chap11.html#X83DACCF07EF62FAE">11.3-3</a>. For that, we first compute the conjugacy classes of the three class representatives M. (Since the groups have elementary abelian Sylow 5 subgroups of the order 5^4, computing all conjugacy classes appears to be faster than using <code class="code">ClassesOfPrimeOrder</code>.) Then we compute an upper bound for σ(S,s).

<div class="example"><pre>
gap> syl5:= SylowSubgroup( cand[1], 5 );;
gap> Size( syl5 ); IsElementaryAbelian( syl5 );
625
true
gap> UpperBoundFixedPointRatios( G, List( cand, ConjugacyClasses ), false );
[ 34817/1645056, false ]
</pre></div>

Remark:

Computing the exact value σ(S,s) in the above setup would require to test the S-conjugacy of certain order 5 elements in M. With the current GAP implementation, some of the relevant tests need several hours of CPU time.

An alternative approach would be to compute the permutation action of S on the cosets of M, of degree 6580224, and to count the fixed points of conjugacy class representatives of prime order. The currently available GAP library methods are not sufficient for computing this in reasonable time. <q>Ad-hoc code</q> for this special case works, but it seemed to be not appropriate to include it here.

In the proof of statement (c), again we consult the result matrix in <a href="chapBib.html#biBKle87">[Kle87]</a>. For S.3, the maximal subgroups are in the rows 4, 15, 22, 26, and 61. Only row 26 yields subgroups that contain elements s of order 65, they have the isomorphism type (5 × GU(3,4)).2 ≅ (5^2 × U_3(4)).2. Note that the conjugacy classes of the members in ��(S,s) are permuted by the outer automorphism of order 3, so none of the subgroups in ��(S,s) extends to S.3. By <a href="chapBib.html#biBBGK">[BGK08, Lemma 2.4 (2)]</a>, if there is a maximal subgroup of S.3 besides S that contains s then this subgroup is the normalizer in S.3 of the intersection of the three members of ��(S,s), i. e., s is contained in at most one such subgroup.

For S.2_1, only the rows 9 and 23 yield maximal subgroups containing elements of order 65, and since we had chosen s in such a way that row 9 was excluded already for the simple group, only extensions of the elements in ��(S,s) can appear. Exactly one of these three subgroups of S extends to S.2_1, so again we get just one maximal subgroup of S.2_1, besides S, that contains s.

All subgroups in ��(S,s) extend to S.2_2, see <a href="chapBib.html#biBKle87">[Kle87]</a>. We compute the extensions of the above subgroups M of S to S.2_2, by constructing the action of the field automorphism in the permutation representation we used for S. In other words, we compute the projective action of the Frobenius map.

<div class="example"><pre>
gap> frob:= PermList( List( orbs[1], v -> Position( orbs[1],
> List( v, x -> x^2 ) ) ) );;
gap> G2:= ClosureGroupDefault( G, frob );;
gap> cand2:= List( cand, M -> Normalizer( G2, M ) );;
gap> ccl:= List( cand2,
> M2 -> PcConjugacyClassReps( SylowSubgroup( M2, 2 ) ) );;
gap> List( ccl, l -> Number( l, x -> Order( x ) = 2 and not x in G ) );
[ 0, 0, 0 ]
</pre></div>

So in each case, the extension of M to its normalizer in S.2_2 is non-split. This implies σ^'(S.2_2,s) = 0.

Finally, in the extension of S by the product of a graph automorphism and the field automorphism, exactly that member of ��(S,s) is invariant that is invariant under the graph automorphism, hence statement (c) holds.

It is again time to clean the workspace.

<div class="example"><pre>
gap> CleanWorkspace();
</pre></div>

<a id="X86EC26F78609618E" name="X86EC26F78609618E"></a>

<h5>11.5-15 ∗ O_9(3)</h5>

The group S = O_9(3) = Ω_9(3) is the first member in the series dealt with in <a href="chapBib.html#biBBGK">[BGK08, Proposition 5.7]</a>, and serves as an example to illustrate this statement.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S of the type 1 perp 8^- (i. e., s decomposes the natural 9-dimensional module for S into an orthogonal sum of two irreducible modules of the dimensions 1 and 8, respectively) and of order (3^4 + 1)/2 = 41, ��(S,s) consists of one group of the type O_8^-(3).2_1 = PGO^-(8,3).

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 1/3.

</dd>
<dt>(c)</dt>
<dd>The uniform spread of S is at least three, with s of order 41.

</dd>
</dl>
By <a href="chapBib.html#biBMSW94">[MSW94]</a>, the only maximal subgroup of S that contains s is the stabilizer M of the orthogonal decomposition. The group 2 × O_8^-(3).2_1 = GO^-(8,3) embeds naturally into SO(9,3), its intersection with S is PGO^-(8,3). This proves statement (a).

The group M is the stabilizer of a 1-space, it has index 3240 in S.

<div class="example"><pre>
gap> g:= SO( 9, 3 );;
gap> g:= DerivedSubgroup( g );;
gap> Size( g );
65784756654489600
gap> orbs:= OrbitsDomain( g, NormedRowVectors( GF(3)^9 ), OnLines );;
gap> List( orbs, Length ) / 41;
[ 3240/41, 81, 80 ]
gap> Size( SO( 9, 3 ) ) / Size( GO( -1, 8, 3 ) );
3240
</pre></div>

So we compute the unique transitive permutation character of S that has degree 3240.

<div class="example"><pre>
gap> t:= CharacterTable( "O9(3)" );;
gap> pi:= PermChars( t, rec( torso:= [ 3240 ] ) );
[ Character( CharacterTable( "O9(3)" ),
 [ 3240, 1080, 380, 132, 48, 324, 378, 351, 0, 0, 54, 27, 54, 27, 0,
 118, 0, 36, 46, 18, 12, 2, 8, 45, 0, 108, 108, 135, 126, 0, 0,
 56, 0, 0, 36, 47, 38, 27, 39, 36, 24, 12, 18, 18, 15, 24, 2,
 18, 15, 9, 0, 0, 0, 2, 0, 18, 11, 3, 9, 6, 6, 9, 6, 3, 6, 3, 0,
 6, 16, 0, 4, 6, 2, 45, 36, 0, 0, 0, 0, 0, 0, 0, 9, 9, 6, 3, 0,
 0, 15, 13, 0, 5, 7, 36, 0, 10, 0, 10, 19, 6, 15, 0, 0, 0, 0,
 12, 3, 10, 0, 3, 3, 7, 0, 6, 6, 2, 8, 0, 4, 0, 2, 0, 1, 3, 0,
 0, 3, 0, 3, 2, 2, 3, 3, 6, 2, 2, 9, 6, 3, 0, 0, 18, 9, 0, 0,
 12, 0, 0, 8, 0, 6, 9, 5, 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 2, 1,
 3, 3, 1, 0, 0, 4, 1, 0, 0, 1, 0, 3, 3, 1, 1, 2, 2, 0, 0, 1, 3,
 4, 0, 1, 2, 0, 0, 1, 0, 4, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1,
 1, 1, 1, 1, 0, 0, 1, 1, 1, 0 ] ) ]
gap> spos:= Position( OrdersClassRepresentatives( t ), 41 );
208
gap> approx:= ApproxP( pi, spos );;
gap> Maximum( approx );
1/3
gap> PositionsProperty( approx, x -> x = 1/3 );
[ 2 ]
gap> SizesConjugacyClasses( t )[2];
3321
gap> OrdersClassRepresentatives( t )[2];
2
</pre></div>

We see that P( S, s ) = σ( S, s ) = 1/3 holds, and that σ( g, s ) attains this maximum only for g in one class of involutions in S; let us call this class <code class="code">2A</code>. (This class consists of the negatives of a class of reflections in GO(9,3).) This shows statement (b).

In order to show that the uniform spread of S is at least three, it suffices to show that for each triple of <code class="code">2A</code> elements, there is an element s of order 41 in S that generates S with each element of the triple.

We work with the primitive permutation representation of S on 3240 points. In this representation, s fixes exactly one point, and by statement (a), s generates S with x ∈ S if and only if x moves this point. Since the number of fixed points of each <code class="code">2A</code> involution in S is exactly one third of the moved points of S, it suffices to show that we cannot choose three such involutions with mutually disjoint fixed point sets. And this is shown particularly easily because it will turn out that already for any two different <code class="code">2A</code> involutions, the sets of fixed points of are never disjoint.

First we compute a <code class="code">2A</code> element, which is determined as an involution with exactly 1080 fixed points.

<div class="example"><pre>
gap> g:= Action( g, orbs[1], OnLines );;
gap> repeat
> repeat x:= Random( g ); ord:= Order( x ); until ord mod 2 = 0;
> y:= x^(ord/2);
> until NrMovedPoints( y ) = 3240 - 1080;
</pre></div>

Next we compute the sets of fixed points of the elements in the class <code class="code">2A</code>, by forming the S-orbit of the set of fixed points of the chosen <code class="code">2A</code> element.

<div class="example"><pre>
gap> fp:= Difference( MovedPoints( g ), MovedPoints( y ) );;
gap> orb:= Orbit( g, fp, OnSets );;
</pre></div>

Finally, we show that for any pair of <code class="code">2A</code> elements, their sets of fixed points intersect nontrivially. (Of course we can fix one of the two elements.) This proves statement (c).

<div class="example"><pre>
gap> ForAny( orb, l -> IsEmpty( Intersection( l, fp ) ) );
false
</pre></div>

<a id="X8393978A8773997E" name="X8393978A8773997E"></a>

<h5>11.5-16 O_10^-(3)</h5>

We show that the group S = O_10^-(3) = PΩ^-(10,3) satisfies the following.

<dl>
<dt>(a)</dt>
<dd>For s ∈ S irreducible of order (3^5 + 1)/4 = 61, ��(S,s) consists of one subgroup of the type SU(5,3) ≅ U_5(3).

</dd>
<dt>(b)</dt>
<dd>σ(S,s) = 1/1066.

</dd>
</dl>
By <a href="chapBib.html#biBBe00">[Ber00]</a>, the maximal subgroups of S containing s are of extension field type, and by <a href="chapBib.html#biBKlL90">[KL90, Prop. 4.3.18 and 4.3.20]</a>, these groups have the structure SU(5,3) = U_5(3) (which lift to 2 × U_5(3) < GU(5,3) in Ω^-(10,3) = 2.S) or Ω(5,9).2, but the order of the latter group is not divisible by |s|. Furthermore, by <a href="chapBib.html#biBBGK">[BGK08, Lemma 2.12 (b)]</a>, s is contained in only one member of the former class.

<div class="example"><pre>
gap> Size( GO(5,9) ) / 61;
6886425600/61
</pre></div>

When the first version of these computations was written, the character tables of both S and U_5(3) were not contained in the GAP Character Table Library, so we worked with the groups. Meanwhile the character tables are available, thus we can show also a character theoretic solution.)

<div class="example"><pre>
gap> t:= CharacterTable( "O10-(3)" ); s:= CharacterTable( "U5(3)" );
CharacterTable( "O10-(3)" )
CharacterTable( "U5(3)" )
gap> SigmaFromMaxes( t, "61A", [ s ], [ 1 ] );
1/1066
</pre></div>

(Now follow the computations with groups.)

The first step is the construction of the embedding of M = SU(5,3) into the matrix group 2.S, that is, we write the matrix generators of M as linear mappings on the natural module for 2.S, and then conjugate them such that the result matrices respect the bilinear form of 2.S. For convenience, we choose a basis for the field extension ��_9/��_3 such that the ��_3-linear mapping given by the invariant form of M is invariant under the ��_3-linear mappings given by the generators of M.

<div class="example"><pre>
gap> m:= SU(5,3);;
gap> so:= SO(-1,10,3);;
gap> omega:= DerivedSubgroup( so );;
gap> om:= InvariantBilinearForm( so ).matrix;;
gap> Display( om );
. 1 . . . . . . . .
1 . . . . . . . . .
. . 1 . . . . . . .
. . . 2 . . . . . .
. . . . 2 . . . . .
. . . . . 2 . . . .
. . . . . . 2 . . .
. . . . . . . 2 . .
. . . . . . . . 2 .
. . . . . . . . . 2
gap> b:= Basis( GF(9), [ Z(3)^0, Z(3^2)^2 ] );
Basis( GF(3^2), [ Z(3)^0, Z(3^2)^2 ] )
gap> blow:= List( GeneratorsOfGroup( m ), x -> BlownUpMat( b, x ) );;
--> --------------------

--> maximum size reached

--> --------------------

Messung V0.5 in Prozent

¤ Diese beiden folgenden Angebotsgruppen bietet das Unternehmen0.422Angebot (Wie Sie bei der Firma Beratungs- und Dienstleistungen beauftragen können 2026-04-28) ¤

*Eine klare Vorstellung vom Zielzustand

Wurzel

Suchen

Beweissystem der NASA

Beweissystem Isabelle

NIST Cobol Testsuite

Cephes Mathematical Library

Wiener Entwicklungsmethode

Haftungshinweis

Die Informationen auf dieser Webseite wurden nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit, noch Qualität der bereit gestellten Informationen zugesichert.