Two boundary Hecke Algebras and the combinatorics of type C

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

Last updated: 27 January 2015

The two-boundary Hecke algebra

In this section we define the two-boundary braid group and Hecke algebras and establish multiple presentations of each. The conversion between presentations is important for matching the algebraic approach to the representation theory with the Schur-Weyl duality approach which we give in Section 5.

For generators gi,gj, encode relations graphically by gi gj means gigj = gjgi, gi gj means gigjgi = gjgigj, and gi gj means gigjgigj = gjgigjgi. (2.1) For example, the group of signed permutations, 𝒲0= { bijectionsw: {-k,,-1,1,,k} {-k,,-1,1,,k} such thatw(-i)=-w(i) fori=1,,k } , (2.2) has a presentation by generators s0,s1,,sk-1, with relations s0 s1 s2 sk-2 sk-1 and si2=1fori=0,1,2,,k-1. (2.3)

The two-boundary braid group

The two-boundary braid group is the group k generated by T0,T1,,Tk, with relations T0 T1 T2 Tk-2 Tk-1 Tk , (2.4) Pictorially, the generators of k are identified with the braid diagrams Tk= ,T0= ,and Ti= i i+1 i i+1 fori=1,,k-1, (2.5) and the multiplication of braid diagrams is given by placing one diagram on top of another.

To make explicit the Schur-Weyl duality approach to representations of k appearing in Section 5, it is useful to move the rightmost pole to the left by conjugating by the diagram σ= . (2.6) Define Ti=σTi σ-1= i i+1 i i+1 ,Y1=σ T0σ-1= , (2.7) and X1= T1-1 T2-1 Tk-1-1 σTk σ-1 Tk-1T1= . (2.8) Define Z1=X1Y1and Zi=Ti-1 Ti-2T1X1 Y1T1Ti-1 = i i , (2.9) for i=2,,k.

The two-boundary braid group k is presented in the following three ways, using the notation defined in (2.1).

(a) k is presented by generators X1,Y1,Z1,T1,,Tk-1 and relations X1 T1 T2 Tk-2 Tk-1 (a1) Y1 T1 T2 Tk-2 Tk-1 (a2) Z1 T1 T2 Tk-2 Tk-1 (a3) and Z1=X1Y1. (a4)
(b) k is presented by generators X1,Y1,T1,,Tk-1 and relations (a1), (a2), and (T1X1T1-1) Y1=Y1(T1X1T1-1). (b3)
(c) k is presented by generators Z1, , Zk, Y1, T1, , Tk-1, and relations (a2), ZiZj=ZjZi fori,j=1, ,k, (c1) Y1Zi=ZiY1 fori=2,,k, and (c2) TiZj=ZjTi forji,i+1, withi=1,,k-1 ,andj=1,,k, (c3) and Zi+1=TiZi Tifori= 1,,k-1. (c4)

Proof.

If P1/2= (2.12) then P1/2Y1 P-1/2= = =Y1-1X1Y1, (2.13) and P1/2X1P-1/2= = =Y1. (2.14) Following these pictorial computations, the extended affine braid group is the group kext generated by k and P with the additional relations PX1P-1= Z1-1X1 Z1,PY1 P-1=Z1-1 Y1Z1, (2.15) PZ1P-1=Z1, PTiP-1=Ti, for i=1,,k-1. (2.16) The element Z0=PZ1Zk is central inkext (c0) since the group kext is a subgroup of the braid group on k+2 strands, and Z0 is the generator of the center of the braid group on k+2 strands (see [Gon2011, Theorem 4.2]). If𝒟={Z0j|j} thenkext=𝒟× k,with𝒟.

The two-boundary Hecke algebra Hkext

In this subsection we define the two-boundary Hecke algebra and relate it to the presentation of the affine Hecke algebra of type C that is found, for example, in [Lus1989, Proposition 3.6] and [Mac2003, (4.2.4)].

Fix a1,a2,b1,b2,t12×. The extended two-boundary Hecke algebra Hkext is the quotient of kext by the relations (X1-a1) (X1-a2)=0, (Y1-b1) (Y1-b2)=0, and (Ti-t12) (Ti+t-12) =0, (h) for i=1,,k-1. Let tk12=a112 (-a2)-12 andt012=b112 (-b2)-12. (2.17) With ZiHkext as in (2.9), define T0=b1-12 (-b2)-12Y1 ,andWi=- (a1a2b1b2)-12 Zi,fori=1,, k, and (2.18) W0=PW1Wk= (-1)k (a1a2b1b2)-k2 PZ1Zk=(-1)k (a1a2b1b2)-k2 Z0. (2.19) Then X1=Z1Y1-1 =a112 (-a2)12 W1T0-1. (2.20)

Fix t0,tk,t× and use notations for relations as defined in (2.1). The extended affine Hecke algebra Hkext defined in (h) is presented by generators, T0, T1, , Tk-1, W0, W1, , Wk and relations W0Z(Hkext), T1 T1 T2 Tk-2 Tk-1 (B1) WiWj=WjWi, fori,j=0, 1,,k; (B2) T0Wj=WjT0, forj1; (B3) TiWj=WjTi fori=1,,k-1 andj=1,,k withji,i+1; (B4) (T0-t012) (T0+t0-12) =0,and (Ti-t12) (Ti+t-12)=0 fori=1,,k-1. (H) For i=1,,k-1, TiWi=Wi+1 Ti+(t12-t-12) Wi-Wi+1 1-WiWi+1-1 andTiWi+1 =WiTi+ (t12-t-12) Wi+1-Wi 1-WiWi+1-1 , (C1) T0W1=W1-1 T0+ ( (t012-t0-12)+ (tk12-tk-12) W1-1 ) W1-W1-1 1-W1-2 ,and (C2)

Proof.

As vector spaces, Hkext= [W0±1,W1±1,,Wk±1] Hkfin, (2.21) where Hkfin is the subalgebra of Hkext generated by T0,T1,,Tk-1. The algebra Hkfin is the Iwahori-Hecke algebra of finite type Ck. If s0,s1,,sk-1 are the generators of 𝒲0 as given in (2.3) and Tw=Tsi1Tsi for a reduced expression w=si1si, then {Tw|w𝒲0} is a-basis ofHkfin. Thus (2.21) means that any element hHkext can be written uniquely as h=w𝒲0hw Tw,withhw [W0±1,W1±1,,Wk±1].

Let Wλ=W0λ0 W1λ1 W2λ2 Wkλkfor λ=(λ0,λ1,,λk) k+1. (2.22) Relations (C1) and (C2) produce an action of 𝒲0 on [W0±1,W1±1,,Wk±1]= span{Wλ|λ=(λ0,λ1,,λk)k+1}. Namely, for w𝒲0 and λk+1, wWλ=Wwλ,where s0λ=s0(λ0,λ1,,λk) =(λ0,-λ1,,λk) ,and siλ=si (λ0,λ1,,λk)= (λ0,λ1,,λi-1,λi+1,λi,λi+2,,λk), (2.23) for i=1,2,,k-1 (see [Ram2003, (1.12)]). With this notation, for λk+1, the relations (C1), and (C2) give TiWλ = WsiλTi+ (t12-t-12) Wλ-Wsiλ 1-WiWi+1-1 and (2.24) T0Wλ = Ws0λT0+ ( (t012-t0-12)+ (tk12-tk-12)W1-1 ) Wλ-Ws0λ 1-W1-2 , (2.25) and, replacing siλ by μ, WμTi = TiWsiμ+ (t12-t-12) Wμ-Wsiμ 1-WiWi+1-1 and (2.26) WμT0 = T0Ws0μ+ ( (t012-t0-12)+ (tk12-tk-12) W1-1 ) Wμ-Ws0μ 1-W1-2 ,forμk+1. (2.27)

The subalgebra HkHkext generated by W1,,Wk and T0,,Tk-1 is the affine Hecke algebra of type C considered, for example, in [Lus1989]. The following theorem determines the center of Hkext and shows that, as algebras, Hkext is a tensor product of Hk by the algebra of Laurent polynomials in one variable. It follows that the irreducible representations of Hkext are indexed by C××Hˆk, where Hˆk is an indexing set for the irreducible representations of Hk.

Let Hk be the subalgebra of Hkext generated by W1,,Wk and T0,,Tk-1. As algebras, Hkext [W0±1]Hk, (2.28) The center of Hkext is Z(Hkext)= [W0±1] [W1±1,,Wk±1]𝒲0, and Hkext is a free module of rank Card(𝒲0)2=22k(k!)2 over Z(Hkext).

Proof.

Weights of representations and intertwiners

Let t12× be such that (t12)1 for . All irreducible complex representations γ of the algebra [W0±1,W1±1,,Wk±1] are one-dimensional. Identify the sets 𝒞 = { irreducible representationsγof [W0±1,W1±1,,Wk±1] } { sequences (z,γ1,,γk) (×)k+1 } { sequences (ζ,c1,,ck) k+1 } (2.31) via γ(W0)=z= (-1)ktζ andγ(Wi) =γi=-tci fori=1,,k (2.32) (the strange choice of sign is an artifact of equations (5.31) and (5.32) and an effort to make the combinatorics of contents of boxes Section 5 optimally helpful). The action of 𝒲0 from (2.23) induces an action of 𝒲0 on 𝒞 by (wγ)(Wλ)= γ(Ww-1λ), forw𝒲0and λk+1. (2.33) Equivalently, on sequences (ζ,c1,,ck), this action is given by w(ζ,c1,,ck)= (ζ,cw-1(1),,cw-1(k)), forw𝒲0. (2.34)

Extend the algebra Hkext to include rational functions in W1,,Wk, defining Hˆkext= [W0±1] (W1,,Wk) Hkfin, where Hkfin is the subalgebra of Hkext generated by T0,T1,,Tk-1. The intertwining operators for Hˆkext are τ0=T0- (t012-t0-12)+ (tk12-tk-12) W1-1 1-W1-2 andτi=Ti- t12-t-12 1-WiWi+1-1 , (2.35) for i=1,2,,k-1. Proposition 2.4 shows that these elements satisfy τ0Wλ=Ws0λτ0 and τiWλ=Wsiλτi so that, for w𝒲0 and λ=(λ0,,λk)k+1, τwWλ= Wwλτw, whereτw= τi1 τi (2.36) for a reduced expression w=si1si.

Each Hkext-module M can be written as M=γ𝒞Mγgen, where for each γ=(z,γ1,,γk)𝒞, Mγgen= { mM| there existsN>0such that(W0-z)Nm=0 and(Wi-γi)Nm=0fori=1,,k } (2.37) is the generalized weight space associated to γ. The intertwiners (2.35) define vector space homomorphisms τ0:Mγgen Ms0γgen andτi:Mγgen Msiγgen, fori=1,,k-1, (2.38) where τ0is defined only whenγ11 , so that(1-W1-1)-1 well-defined onMγgenand τiis defined only whenγi γi+1so that (1-WiWi+1-1)-1 is well-defined onMγgen, for i=1,,k-1.

(Intertwiner presentation) The algebra Hˆkext is generated by τ0,,τk, W0, and (W1,,Wk) with relations τ0 τ1 τ2 τk-2 τk-1 (2.39) in the notation of (2.1); τ0W1=W1-1 τ0andτ0 Wj=Wjτ0for j1; (2.40) for i=1,,k-1, τiWi=Wi+1τi andτiWi+1 =Wiτifori >0,andτi Wj=Wjτifor ji,i+1; (2.41) τ02= (1-t012tk12W1-1) 1-W1-1 (1+t012tk-12W1-1) 1+W1-1 (1+t0-12tk12W1-1) 1+W1-1 (1-t0-12tk-12W1-1) 1-W1-1 ; (2.42) τi2= (t12-t-12Wi-1Wi+1) (t12-t-12Wi+1-1Wi) (1-Wi-1Wi+1) (1-Wi+1-1Wi) ,fori=1,,k-1. (2.43)

Proof.

Notes and References

This is an excerpt from the paper Two boundary Hecke Algebras and the combinatorics of type C Zajj Daugherty (Department of Mathematics, The City College of New York, NAC 8/133, Convent Ave at 138th Street, New York, NY 10031) and Arun Ram (Department of Mathematics and Statistics, University of Melbourne, Parkville VIC 3010, Australia).

page history