Group Actions

Prelim: Group Isomorphisms and Homomorphisms
Needed for Burnsides Lemma

Group Actions Introduction

Definition:

Let G be a group and X be a set. Then we can define a group action as:

G×XX(g,x)=gx

Additionally, we have it so ex=x and g1(g2x)=(g1g2)x for all gG and xX.
Essentially we have Identity and associativity property.

Example:

Take the Dihedral group Dn acting on a regular n-gon. This is a group action where the elements of Dn act on the vertices of the n-gon (which is a set of n elements).

For example, we can take a look at D4 acting on the vertices of a square; X={1,2,3,4}. Then we know r=(1234) and this acts on the vertex 1 by sending it to 2. i.e. (1234)1=2. Clearly:

D4×XXs.tX={1,2,3,4}

Which is a group action.

Acting by conjugation

A group can act on itself by conjugation through the following definition:
Let G be a group and let X=G. Then a group acts on itself via conjugation via:

G×XXs.t.(g,x)=gxg1

This works as we can prove identity and associativity through the following:

  1. Identity:
    1. e1=e so (e,x)=exe=x.
  2. Associativity
    1. Let g1,g2G.
    2. Then g1(g2x)=g1(g2xg21)=g1g2xg21g11=g1g2x(g1g2)1=(g1g2)x

Orbits and Stabilizers

Orbits

Two elements in a set are G-equivalent if an element of G can take the first element to the second. Rigorously, x is G-equivalent to y if gG such that x=gy. This creates an equivalence relation where all the classes are called orbits, denoted O and the specific orbit of x (all the places where G takes x) is denoted Ox.

To reiterate, you can think of Ox as all the places G takes x

Lets prove G-equvilance is a equivalence relation:
  1. Reflexive: x~x
    1. Clearly any group contains the identity, and we have defined group actions such that ex=x. Thus x~x
  2. Symmetry:
    1. Assume x~y. There there exists g s.t. x=gy. However, this means y=g1x and as G is a group and has inverses clearly the inverse of g is in G. Thus y~x
  3. Transitivity:
    1. Assume x ~ y and y ~ g. Then we have it so x=g1y and y=g2z.
    2. From here, we can show x=g1y=g1g2z.
    3. Then x=g1g2z and g1g2G as it is a group. thus x~z

Stabilizer

Stabilizers are the subgroup of elements gG that do not move an element xX. i.e. collection of g where gx=x. This is denoted as Gx for all elements in G which fix x. This creates a subgroup in G.

Proving that Gx is a subgroup of G:
  1. Identity:
    1. By definition e fixes every element, thus eGx
  2. Closure:
    1. Let g1,g2Gx. Then, (g1g2)x=g1(g2x)=g1x=x as both elements fix x. Thus g1g2Gx
  3. Inverses:
    1. Suppose gG. Then gx=x, so x=g1x and we have it that g1 is also in Gx.

Example

Lets consider D4 acting on the vertices of a square. What are the orbits and stabilizers?
First note: $$D_4 = {e,r^1,r^2,r^3,s,sr^1,sr^2,sr^3}$$
Which I drew out and represented as permutations in S4 here

e=(1),r1=(1234),r2=(13)(24),r3=(1432)s=(21)(34),sr=(13),sr2=(23)(14),sr3=(42)

and $$X = {1,2,3,4}$$
for the O1 we see that 1 can go anywhere through just the rotations. So we only have one orbit and all the vertices are G-equivalent. $$\mathcal{O}_1 = \mathcal{O}_2 = \mathcal{O}_3 = \mathcal{O}_4 = {1,2,3,4}$$ Then, for the stabilizer of 1, or G1={e,sr3}. Clearly these elements fix r, and (sr3)2=e so this is a subgroup isomorphic to Z2. I put the stabilizer groups for every element bellow:

G1={e,sr3}G2={e,sr}G3={e,sr3}G4={e,sr}

You can observe all these stabilizer groups are isomorphic to Z2 and are all subgroups of D4.

Connection Ox∣=[G:Gx]

The size of the orbit of x is the same as the number of left cosets of the stabilizer group in G. We can show this here:
#review proof if you have time.
For x=1

  1. we saw that O1={1,2,3,4} and so O1∣=4.
  2. The size of D4∣=2(4)=8
  3. Lastly, size of G1∣=2. So we know there are 4 cosets of G1 in D4
  4. Thus, O1∣=[G:G1]=4