424

From Mathsoc wiki

This page is far from complete! Can you help by adding anything to it (such as notes or links to useful sites)?

Ma424 Group Representations

Lecturer: Dr. Timothy Murphy

Website: Link

Finite groups

Examples of finite groups:

  • The cyclic group C_n = \mathbb{Z}/n = \lbrace e,a,a^2, \dots , a^{n-1} \rbrace such that a^n=e.
  • The symmetric group S_n consisting of permutations of n elements. This group is generated by the transpositions.
  • The alternating group A_n consisting of the even permutations in S_n (a permutation is said to be even if it can be written as an even number of transpositions).
  • The dihedral group D_n consisting of the symmetries of the regular n-gon. This consists of n rotations (forming a subgroup isomorphic to C_n) and n reflections. The group can be written as
D_n = \left \langle s,t : s^n = t^2 = 1, (st)^2 =1 \right \rangle

Definition (Group representation): A representation of a group G in a vector space V over a field k is defined by a homomorphism

\alpha: G \rightarrow GL(V)

from G to the general linear group on V - this is the group of all invertible linear maps on V. Given a basis for V we can express each operator in GL(V) as a matrix, and the representation can be thought of as a homomorphism from G to the group GL(n,k) of invertible n \times n matrices over the field k, in which we give matrices A(g) for all g \in G, such that

A(gh) = A(g) A(h) \qquad A(e) = \mathbb{I}

Obviously this works best in practice for small matrices, as everyone knows it's impossible to multiply 4 \times 4 matrices.

The one dimensional representation of G is a homomorphism from G to the field k without the zero element (here k = \mathbb{R} or \mathbb{C}). Note that is constant on conjugacy classes (one-dimensional multiplication commutes). The trivial representation is the representation defined by mapping every element of G to 1.

Definition (Equivalent representations): Two representations \alpha, \beta of a group G in vector spaces U,V are equivalent if we can find a linear map t from U to V preserving the action of G:

t(gu) = g(tu) \forall g \in G, u \in U

In matrix terms, suppose that

\alpha: g \mapsto A(g) \quad \beta: g \mapsto B(g)

Then the two representations are equivalent if there is an invertible matrix P such that

B(g) = PA(g)P^{-1} \quad \forall g \in G

Definition (Stable): A subspace U of a vector space V is said to be stable under the action of a group G if

g \in G, u \in U \Rightarrow gu \in U

Definition (Simple representation): A representation \alpha of a group G in a vector space V over k is said to be simple if no proper subspace of V is stable under G, i.e. the only stable subspaces are 0 and G itself.

All one dimensional representations are simple (as a one-dimensional space has no proper subspaces). A simple representation of a finite abelian group over \mathbb{C} is one dimensional.

Representations can be added, multiplied (using the tensor product), and have a conjugacy operation.

Definition (Semisimple representation): A representation \alpha of G is said to be semisimple if it is expressible as a sum of simple representations.

If a vector space V is the sum of simple subspaces then V is semisimple. We also have that V is semisimple if and only if each stable subspace U of V has at least one complementary stable subspace W, meaning that V = U \oplus W. (Proofs use sums of simple subspaces, relying on fact that S_i \cap S_j = 0 or S_j for S_j simple).

Maschke's Theorem: Every representation of a finite group G over \mathbb{R} or \mathbb{C} is semisimple. (Proof uses an averaged positive definite form on the vector space to define a complement for a stable subspace U).

Definition (Intertwining number): The intertwining number I(\alpha, \beta) of two representations \alpha,\beta of G over k in vector spaces U, V respectively is the dimension of space of G-maps from U to V,

I(\alpha, \beta) = \mbox{dim}\,\mbox{hom}^G\,(U,V)

If \alpha and \beta are simple representations, then their intertwining number is 0 if \alpha \neq \beta, and \geq 1 if \alpha = \beta. If k=\mathbb{C} then I(\alpha, \alpha)=1.

Definition (Character): The character \chi = \chi_{\alpha} of a representation \alpha of G over k is the function \chi: G \rightarrow k defined by

\chi(g) = \mbox{tr}\,\Big(\alpha(g)\Big)

We have the following formula for the intertwining number of two representations:

I(\alpha, \beta) = \frac{1}{||G||} \sum_{g\in G} \chi_{\alpha}(g^{-1}) \chi_{\beta} (g)

Note that \chi_{\alpha^*}(g) = \chi_{\alpha}(g^{-1}) = \overline{\chi_{\alpha}(g)} when k=\mathbb{C} or \mathbb{R} (obviously then the conjugation does nothing), and that \alpha = \beta \Leftrightarrow \chi_{\alpha}(g) = \chi_{\beta}(g) \, \forall g \in G. Characters are also constant on conjugacy classes.

Definition (Regular representation): The regular representation of a group is the permutation representation defined by the action of the group on itself, (g,x) \mapsto gx. The character of the regular representation is ||G|| if g=e and 0 otherwise. Every simple representation \sigma occurs in the regular representation; when k=\mathbb{C} they occur \mbox{dim}\,\sigma times in the regular representation. From this we can derive that (when k=\mathbb{C}) the dimensions of the simple representations \sigma_i satisfy

\mbox{dim}^2\,\sigma_1 + \dots + \mbox{dim}^2\,\sigma_r = ||G||

Compact groups

Much of the theory of finite groups can be passed over to compact groups once we have established the existence of a unique measure on the group.

Definition (Topological group): A topological space G with a group structure defined on it, such that the group operations (x,y) \mapsto xy and x \mapsto x^{-1} are continuous.

Definition (Hausdorff): A topological space X is said to be Hausdorff if given any two points x,y \in X there exist open sets U,V in X such that

x \in U, y \in V, U \cap V = \emptyset

Definition (Compact): A space X is said to be compact if it is Hausdorff and every open covering

X = \bigcup_{i\in I} U_i

has a finite subcovering, X=U_{i_1} \cup \dots \cup U_{i_r}

Note that a subspace X of Euclidean space E^n is compact iff X is closed and bounded.

Examples of compact groups:

  • The orthogonal group O(n) of real n\times n matrices T such that T^tT=\mathbb{I}
  • The special orthogonal group of SO(n) of real n\times n matrices T such that T^tT=\mathbb{I} and \det T = 1
  • The unitary and special unitary groups, whose definitions mirror the above, except for complex matrices such that T^{\dagger}T = \mathbb{I}.
  • The symplectic group of quaternionic matrices whose conjugate transposes are their inverses.

We can define integration on a compact group using a measure.

Definition (Measure): A measure \mu on X is a continuous linear functional \mu: C(X,k) \rightarrow k, where C(X,k) is the vector space of functions f:X\rightarrow k. We write

\mu(f) = \int_X f d \mu

Haar's theorem (for compact groups): Let G be a compact group. Then there exists a unique real measure \mu on G such that

  1. \mu is invariant under G, \int_G (gf) d \mu = \int_G f d \mu, for all g \in G, f \in C(G,\mathbb{R}).
  2. \mu is normalised such that G has volume 1, i.e. \int_G 1 d \mu = 1.

and moreover

  1. the measure is strictly positive: f(x) \geq 0 \, \forall x \Rightarrow \int f d \mu \geq 0 with equality iff f(g) = 0 \, \forall g
  2. \left| \int_G f d \mu \right| \leq | f |

(The reading week problem was to cover Haar's theorem for locally compact groups rather than compact groups)