Examples of Macdonald polynomials

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

Last update: 25 September 2012

Type A2

The Weyl group W0= s1,s2 s12=s22=1, s1s2s1= s2s1s2 acts on the lattices

𝔥=ω1+ ω2and 𝔥*=ω1+ ω2, (4.7)

where s1 and s2 are the reflections in the hyperplanes determined by

α1=2ω1 -ω2, α2=-ω1 +2ω2 ,α1= 2ω1-ω2, andα2=-ω1 +2ω2, (4.8)

with ωi,αj =δij, and ωi,αj =δij. In this case,

φ=α1+ α2,and φ=α1+α2. (4.9)

The double affine braid group is generated by T0,T1,T2,g, Xω1,Xω2, and q1/3, with relations

TiTjTi= TjTiTj, forij, XμXλ= Xμ+λ= XλXμ, forμ,λ 𝔥*, T1Xω2= Xω2T1, T2Xω1= Xω1T2, T1Xω1T1 =X-ω1+ω2, T2Xω2 T2= Xω1-ω2, g3=1,g Xω1= q1/3 X-ω1+ω2g, gXω2= q2/3 X-ω1g, gT0g-1=T1, gT1g-1=T2, gT2g-1=T0. (4.10)

The formula (2.27) gives

g=Yω1 T1-1 T2-1, g2=Yω2 T2-1T1-1, T0=Yφ T1-1T2-1 T1-1, (4.11) g= Xω1T1 T2, (g)2= Xω2T2 T1, (T0)-1 =XφT1 T2T1. (4.12)

At this point, the following Proposition, which is the type A2 case of Theorem 2.1, is easily proved by direct computation.

(Duality). Let Yd=q-1. The double affine braid group is generated by T0, T1, T2, g, Yω1, Yω2, and q1/3, with relations

(T1)-1 Yω1 (T1)-1= Y -ω1+ ω2 , (T2)-1 Yω2 (T2)-1= Y ω1- ω2 , (T1)-1 Yω2= Yω2 (T1)-1, (T2)-1 Yω1= Yω1 (T2)-1, (g)3=1, gYω1= q-1/3 Y-ω1+ω2 g,g Yω2=q-2/3 Y-ω1g, gT0 (g)-1= T1,g T1 (g)-1= T2and gT2 (g)-1= T0.

To give a concrete example of Theorem 3.4 let us compute the symmetric Macdonald polynomial Pρ where ρ=α1+α2. Since

10= Xs1s2s1+ t-1/2Xs1s2 +t-1/2Xs2s1 +t-2/2Xs1+ t-2/2Xs2+ t-3/2,

and Xρm=s0 is the minimal length element of the coset XρW0,

Pρ = 10Eρ=10 τ01 = ( Xs1s2s1+ t-1/2 Xs1s2+ t-1/2 Xs2s1 ) ( T0+ t-1/2(1-t) 1-Y-α0 ) 1 + ( t-2/2Xs1+ t-2/2Xs2+ t-3/2 ) ( (T0)-1+ t-1/2(1-t) Y-α0 1-Y-α0 ) 1 = ( Xs1s2s1s0 +t-1/2 Xs1s2s0+ t-1/2 Xs2s1s0+ t-2/2Xs1s0 +t-2/2Xs2s0 +t-3/2Xs0 ) 1 + ( Xs1s2s1+ t-1/2Xs1s2 +t-1/2Xs2s1 ) t-1/2 (1-t) 1-Y-α0 1 + ( t-2/2Xs1+ t-2/2Xs2+ t-3/2 ) t-1/2 (1-t) Y-α0 1-Y-α0 1.

Since Y-α01= Yφ-d1=q Yα1+α2 1=t2q1,

Pρ = ( Xw0ρ+ t-1/2 Xs1s2ρ T2+ t-1/2 Xs2s1ρ T1 +t-2/2 Xs1ρ T2T1+ t-2/2 Xs2ρ T1T2+ t-3/2Xρ T1T2 T1 ) 1 + ( T1T2 T1+t-1/2 T1T2+ t-1/2T2 T1 ) t-1/2 (1-t) 1-t2q 1 + ( t-2/2T1+ t-2/2T2+ t-3/2 ) t-1/2 (1-t) t2q 1-t2q 1 = ( Xw0ρ+ Xs1s2ρ+ Xs2s1ρ+ Xs1ρ+ Xs2ρ+ Xρ ) 1 + ( t3/2+ t1/2+t1/2 ) t-1/2 (1-t) 1-t2q 1+ ( t-1/2+ t-1/2+ t-3/2 ) t-1/2 (1-t)t2q 1-t2q 1 = ( Xw0ρ+ Xs1s2ρ+ Xs2s1ρ+ Xs1ρ+ Xs2ρ+ Xρ+ ( t+2+2tq+q ) 1-t 1-t2q ) 1.

The set 𝒫(ρ) contains 12 alcove walks,

and

The Hall-Littlewood polynomial and the Weyl character are

Pρ(0,t)=mρ+ (2+t)(1-t) andsρ=Pρ (0,0)=mρ+2,

where mρ=Xw0ρ= Xs1s2ρ+ Xs2s1ρ+ Xs1ρ+ Xs2ρ+ Xρ.

The expression Xs1s2ρs2 =s1s2 s0 is a reduced word for the minimal length element in the coset Xs1s2ρW0 and Theorem 2.2 is illustrated by

τ1τ2τ0 = ( T1+ t-1/2 (1-t) 1-Y-α1 ) τ1τ0= ( T1τ2+ τ2 t-1/2 (1-t) 1-Y-s2α1 ) τ0 = ( T1T2+ T1 t-1/2 (1-t) 1-Y-α2 +T2 t-1/2 (1-t) 1-Y-s2α1 + t-1/2 (1-t) 1-Y-α2 t-1/2 (1-t) 1-Y-s2α1 ) τ0 = T1T2τ0 +T1τ0 t-1/2 (1-t) 1-Y-s0α2 +T2τ0 t-1/2 (1-t) 1-Y-s0s2α1 +τ0 t-1/2 (1-t) 1-Y-s0α2 t-1/2 (1-t) 1-Y-s0s2α1 = T1T2T0 +T1T2 t-1/2 (1-t) 1-Y-α0 +T1 (T0)-1 t-1/2 (1-t) 1-Y-s0α2 + T1 t-1/2 (1-t) Y-α0 1-Y-α0 t-1/2 (1-t) 1-Y-s0α2 +T2 (T0)-1 t-1/2 (1-t) 1-Y-s0s2α1 + T2 t-1/2 (1-t) Y-α0 1-Y-α0 t-1/2 (1-t) 1-Y-s0s2α1 +(T0)-1 t-1/2 (1-t) 1-Y-s0α2 t-1/2 (1-t) 1-Y-s0s2α1 + t-1/2 (1-t) Y-α0 1-Y-α0 t-1/2 (1-t) 1-Y-s0α2 t-1/2 (1-t) 1-Y-s0s2α1 ,

where the eight terms in this expansion correspond to the eight alcove walks in (1,s1s2s0) =(s1s2ρ) pictured below. Applying the expansion of τ1τ2τ0 to 1 and using

Y-α01= Yφ-d1= t2q1, Y-s0α2 1=Yα1-d 1=tq1, and Y-s0s2α1 1=Yφ-2d1 =t2q21, (4.13)

computes

Es1s2ρ = ( Xs1s2ρ t1/2+t t-1/2 (1-t) 1-t2q +Xs1ρt t-1/2 (1-t) 1-tq +t1/2 t-1/2 (1-t) t2q 1-t2q t-1/2 (1-t) 1-tq + Xs2ρt t-1/2 (1-t) 1-t2q2 +t1/2 t-1/2 (1-t) t2q 1-t2q t-1/2 (1-t) 1-t2q2 + Xρt3/2 t-1/2 (1-t) 1-tq t-1/2 (1-t) 1-t2q2 + t-1/2 (1-t) t2q 1-t2q t-1/2 (1-t) 1-tq t-1/2 (1-t) 1-t2q2 ) 1. = t1/2 ( Xs1s2ρ+ (1-t) 1-t2q +Xs1ρ (1-t) 1-tq +t (1-t)q 1-t2q (1-t) 1-tq +Xs2ρ (1-t) 1-t2q2 +t (1-t)q 1-t2q (1-t) 1-t2q2 +Xρ (1-t) 1-tq (1-t) 1-t2q2 + (1-t)q 1-t2q (1-t) 1-tq (1-t) 1-t2q2 ) 1. Xs1s2ρ t1/2 t1/2 (1-t) 1-t2q Xs1ρt1/2 (1-t) 1-tq t3/2 (1-t) 1-tq (1-t)q 1-t2q Xs2ρt1/2 (1-t) 1-t2q2 t3/2 (1-t)q 1-t2q (1-t) 1-t2q2 Xρt1/2 (1-t) 1-tq (1-t) 1-t2q2 t3/2 (1-t) 1-tq (1-t)q 1-t2q (1-t) 1-t2q2

Notes and References

This page is taken from a paper entitled A combinatorial formula for Macdonald Polynomials by Arun Ram and Martha Yip.

page history