Macdonald processes and measures
macdonald_processes macdonald_measures macdonald_polynomials
Macdonald measures can be found in [{forrester-rains02}].
Macdonald measures
Let m,n be positive integers, a∈[0,1]m, b∈[0,1]n and 0≤q,t≤1. Let Pλ and Qλ be the Macdonald polynomials, we define the Macdonald measure by
μMac(λ)=Pλ(a)Qλ(b)∏ij(ajbi)∞(ajbit)∞.It is a probability measure due to
- the Cauchy-Littlewood identity (4.13) in macdonald_polynomials.
- and the fact that Pλ(a),Qλ(b)≥0 if we plug in the explicit formulas of the Macdonald polynomials.
This is our default Macdonald measure, which we also call the usual Macdonald measure since it involves the usual specialisation of the Macdonald polynomials.
By contrast, here is the dual Macdonald measure, involving the dual specialisation of the Macdonald polynomials.
μdMac(λ)=Pλ(a;q,t)Pλ′(b;t,q)∏ij11+ajbi.TODO: Exponential specialisation case.
For now we only care about the usual case μMac.
Macdonald processes
Let ϕ be the quantity defined in the Pieri rule of Macdonald polynomials. For u∈[0,1], define kernel Au by
Au(λ,μ)=ϕμ/λu|μ|−|λ|∏i(aiu)∞(taiu)∞Claim. For any u, the Macdonald polynomial Pλ(b) is Au-harmonic.
Proof. Recall by (2.8)(5.5) of macdonald_polynomials,
∑nQ(n)(a)un=∏i(taiu)∞(aiu)∞and the Pieri rule (i) of macdonald_polynomials
Pλ(a)Q(r)(a)=∑μ:λ≺μ,|μ|−|λ|=rϕμ/λPμ(a).Combining these two we have
Pλ(a)∏i(taiu)∞(aiu)∞=∑r≥0Pλ(a)Q(r)(a)ur=∑μ:λ≺μϕμ/λPμ(a)u|μ|−|λ|and we are done. ◻
Definition. Let h(λ)=Pλ(a). The Macdonald process is a Markov process X with transition kernel R that is the [[doob_transform|Doob-h transform]] of kernel Abn at time n:
P(Xn=μ|Xn−1=λ,X0:n−2)=Abn(λ,μ)Pμ(a)Pλ(a).The Macdonald process induces the Macdonald measure:
Claim. Let X be a Macdonald process with empty initial condition: X0≡∅ is an empty partition. Then for any fixed n, Xn distributes according to the Macdonald measure.
Proof. Use the branching formula (7.13) of macdonald_polynomials, we have
P(Xn=λ)=∑∅=λ0≺λ1≺...≺λn(∏iϕλi/λi−1b|λi|−|λi−1|i)Pλ(a)∏ij(biaj)∞(biajt)∞=Pλ(a)Qλ(b)∏ij(biaj)∞(biajt)∞.◻
References
- [forrester-rains02] Interpretations of some parameter dependent generalizations of classical matrix ensembles, , Probability Theory and Related Fields, Vol. 131, No. 1 2002. [ link ]