Representation Theory

Arun Ram
Department of Mathematics and Statistics
University of Melbourne
Parkville, VIC 3010 Australia
aram@unimelb.edu.au

Last updated: 2 October 2014

Lecture 3

Let d=span{ei|1id} where ei=(00100) with Md()-action by left multiplication.

(1) If M is a simple Md()-module then Md.
(2) If t:Md() is a trace then t=κ·Tr, with κ.
(3) 𝒵(Md())=·Id.
(4) Md() has one nonzero ideal.

Proof.

(1) Let M be a simple Md()-module. Let mM be nonzero. Since m=1·m=i=1d Eiim, then Eiim0 for some i. Let mj=Ejim for j=1,2,,d. Then N=span{m1,,md} is a submodule of M such that Ersmj= δsjmr. Since M is simple, N=M and M d mj ej is an Md()-module isomorphism.
(2) Let t:Md() be a trace. then t(Eij)= t(Ei1E1j)= t(E1jEi1)= δijt(E11). So t=t(E11)·Tr.

The algebra A=λAˆMdλ()

A has basis {Eijλ|1i,jdλ,λAˆ} with Eijλ Ersμ= δλμ δjr Eisλ. Define Aλ=span{eiλ|1idλ} with Eijμerλ= δλμδjr eiλ. Define Trλ:A by Trλ(Eijμ)=δλμδij. Define zλ=i=1dEiiλ, so that zλ2=zλ and 1=λAˆzλ. Define Iλ=zλA=span{Eijλ|1i,jdλ}.

(1) Aλ, λAˆ, are the simple A-modules.
(2) If t:A is a trace then t=λAˆ tλTrλ, with tλ.
(3) 𝒵(A)=span{zλ|λAˆ}.
(4) The minimal ideals of A are Iλ, λAˆ. The ideals of A are sums of Iλ.

The regular representation of A

A=λAˆ (Aλ)dλ sinceA= { ( 0 0 ) } . Note that Trλ: A a Tr(pλ(a)) where ρλ: A Mdλ() a ρλ(a), where ρλ(a) is the matrix of the action of a on Aλ. If t:A is the trace of the regular representation then t=λAˆ dλTrλ. The dual basis to {Eijλ|λAˆ,1i,jdλ} is { 1dλ Ejiλ |λAˆ, 1i,jdλ } since t(Eijλ1dμEsrμ)=δirδλμδjs. Thus Eijλ= μAˆ1r,sdμ dμAλ (1dμEsrμ)ji Ersμ. If B={b} is a basis of A and {b*} is the dual basis with respect to , defined by t. Then Eijλ= bBdλ Aλ(b*)ji b and b=λAˆ1i,jdλ Aλ(b)ij Eijλ.

(Artin-Wedderburn) Let A be a finite dimensional algebra such that the trace of the regular representation is nondegenerate. Then AλAˆ Mdλ() as algebras where A=λAˆ(Aλ)dλ as A-modules.

Proof.

By Maschke's theorem AλAˆ (Aλ)dλ as A-modules. Since A End(A) a aA is injective, ρ: A End(λAˆAλ) a λAˆρλ(a) is also injective and ρ is an algebra homomorphism.

Temperley-Lieb TL3=span{,,,,} has module M=span{,,} which we found was M=NP where N = span { , } and P = span { (1+(q+q-1)) - - } . We have ρ: TL3 M1() 0 0 and ρ: TL3 M2() (q+q-1100) (001q+q-1) Then TL3 λTL3Mdλ() (q+q-110000000) (0001q+q-10000) Let G be a group. The group algebra of G is the algebra G with basis G and multiplication determined by the multiplication in G. The map t:Ggiven by t(a)=a|1, the coefficient of 1 in a, is a trace on G. If Tr is the trace of the regular representation of G then Tr(g)= hG gh|h= { |G|, ifg=1, 0, ifg1 =|G|·t(g).

The braid group Bn is the group of braids with n strands with product b1b2= b1 b2

(Artin) Bn is presented by generators g1,,gn-1, where gi= 1 2 i i+1 n with relations gigi+1gi=gi+1gigi+1.

The symmetric group Sn is the quotient of Bn by the relations gi2=1. ( writesi= 1 2 i i+1 n since gi = =gi-1 = inSn )

The Iwahori-Hecke algebra is the quotient of Bn by the relations (Ti-q) (Ti+q-1)=0. ( Ti2= (q-q-1) Ti+1 ) . Let ei=q-Tiin Hn. Then TiTi+1Ti=Ti+1TiTi+1 becomes eiei+1ei- ei+1eiei+1 =ei-ei+1. The map Bn Hn TLn ei ei Ti Ti are surjective algebra homomorphisms.

Remark Tw02= 𝒵(Bn) and Tw02= yε1 yεn, where yεi= i and yεi yεj= = =yεj yεi, for 1i,jn.

So the image of [yε1,,yεn] is a large commutative subalgebra of Hn (or TLn).

Note: Bn Bn+1 b b gives inclusions H1 H2H3, TL1 TL2TL3,and S1 S2S3.

Pullback functors

Suppose AφR is an algebra homomorphism. Then we get a functor {R-modules} {A-modules} M φ*(M) where φ*(M)=M, as vector spaces and the A-action is given by a·m=φ(a)m, for aA, mM.

The map HnπTLn gives a functor {TLn-modules} {Hn-modules} which takes simple modules to simple modules. The map π is surjective and the map π* is injective. The map TL3ιTL4 gives {TL4-modules} ι* {TL3-modules}. The functor ι* is Restriction, ResTL3TL4.

Adjoint functors

Let F:{A-modules}{B-modules} be a functor. The adjoint functor F:{B-modules}{A-modules} is determined by HomB-mod (FM,N) HomA-mod (M,FM). The adjoint functor to ResAB is induction IndAB. IndAB: {A-modules} {B-modules}. It is given explicitly by IndAB(M)= BAM, where BAM is generated by bm, bB, mM, with relations bam=bam and bilinearity.

Bratelli diagrams

Let A1A2A3 be a sequence of inclusions of semisimple algebras. The Bratelli diagram for A1A2 is the graph with vertices on levelk:Aˆk, where Aˆk is an index for the simple Ak-modules and mλμedges connecting λandμ (λAˆk,μAˆk-1) if ResAk-1Ak (Akλ)= μAˆk-1 (Ak-1μ)mλμ.

The Bratelli diagram for TL1TL2 has vertices on levelk: {partitions ofkwith2rows} =TLˆk and an edge λ-μ if μ is obtained from λ by removing a box.

A partition is a collection of boxes in a corner. λ= =(442211) Write λ=(λ1,,λ) with λ1λ and λi=# of boxes in row i.

(a) The Bratelli diagram for H1H2H3 has Hˆk={partitions withkboxes} andλ-μ if μ is obtained from λ by removing a box.
(b) The Bratelli diagram for S1S2S3 has Sˆk= {partitions withkboxes} andλ-μ if μ is obtained from λ by removing a box.

The algebra U𝔰𝔩2

A Lie algebra is a vector space 𝔤 with a bracket [,]:𝔤𝔤𝔤 such that

(1) [x,y]=-[y-x], for x,y𝔤,
(2) [[x,y],z]+[[z,x],y]+[[y,z],x]=0, for x,y,z𝔤.

A Lie algebra is not an algebra.

The enveloping algebra of 𝔤 is the algebra U𝔤 generated by the vector space 𝔤, with the relations xy=yx+[x,y], for x,y𝔤.

The Lie algebra 𝔰𝔩2 is the vector space 𝔰𝔩2= {aM2()|tra=0} with bracket [a,b]=ab-ba, for a,b𝔰𝔩2 (where the product on the RHS is matrix multiplication).

The enveloping algebra of 𝔰𝔩2 is the algebra U𝔰𝔩2 generated by x,y,k with relations xy=yx+k, kx=xk+2x, ky=yk-2y.

The Lie algebra 𝔰𝔩2 is presented by generators x=(0100), y=(0010), k=(100-1) and relations [x,y]=k, [k,x]=2x, [k,y]=-2y.

HW: Show that U𝔰𝔩2 has basis { ym1 km2 xm3 | m1,m2,m3 0 } . Hence dim(U𝔰𝔩2)=3.

Notes and References

These are a typed copy of Lecture 3 from a series of handwritten lecture notes for the class Representation Theory given on August 12, 2008.

page history