:

## The Burnside legacy

Greetings to all.

It is a new year (for some time now) and I am writing to indicate that the mathematical adventures with cycle index computations and Maple continue!

Here are the previous installments:

My purpose this time is to alert readers who might be interested to a new cycle index computation that is neither an application of the classical form of the Polya Enumeration Theorem (PET) nor of Power Group Enumeration. The former counts objects being distributed into slots with a group acting on the slots and the latter objects going into slots with a second group which permutes the objects in addition to the slots being permuted. What I am about to present treats a third possible case: when the slot permutation group and the object permutation group are one and the same and act simultaneously (not exactly the same but induced by the action of a single group).

This requires quite radical proceedings in the etymological sense of the word, which is to go back to the roots of a problem. It seems that after working with the PET sooner or later one is confronted with enumeration problems that demand the original unmitigated power of Burnside's lemma, sometimes called the lemma that is not Burnside's. This is the case with the following problem. Suppose you have an N-by-N matrix whose entries are values from 1 to N, with all assignments allowed and the symmetric group on N elements acts on the row and column indices permuting rows and columns as well as the entries simultaneously. We ask how many such matrices there are taking these double symmetries into account. This also counts the number of closed binary operations on a set of N elemnents and there is a discussion as well as the Maple code (quite simple in my opinion and no more than a few lines) that solves this problem at the following Math Stackexchange link, which uses Lovasz Formula for the cycle index of the symmetric group which some readers may remember.

In continuing the saga of Polya and Burnside exploration I have often reflected on how best to encapsulate these techniques in a Maple package. With this latest installment it would appear that a command to do Burnside enumeration probably ought to be part of such a package.

Best regards,

Marko Riedel

﻿