On the convoy of the ASEP speed process
Yuan Tian
TL;DR
This work analyzes the convoy size in the ASEP speed process, connecting probabilistic convoy structure to an exact formula via $q$-Genocchi numbers and to a universal $\sqrt{n}$-scaling for fixed $q\in[0,1)$. It leverages Martin's convoy construction and an orthogonal-polynomial framework built from continuous big $q$-Hermite and Askey–Wilson polynomials, together with a birth-death–Karlin–McGregor representation, to obtain both explicit formulas and limiting results. A notable result is the universality of the normalized convoy size, with $\lim_{n\to\infty}\frac{1}{\sqrt{n}}\mathbb{E}_x[\#\mathcal C_0^n]=\sqrt{\frac{4x(1-x)}{\pi}}$, and a distributional limit in the special case $q=0$; a critical weakly asymmetric scaling $q_n=e^{-\gamma/\sqrt{n}}$ yields a nontrivial limiting regime described by a joint density $f^{\gamma}_{X,Y}$.
Abstract
We investigate the size of the convoy in the speed process in the multi-species asymmetric simple exclusion process (ASEP). Through a coupling argument, we obtain an exact formula for the expected convoy size by relating it to a combinatorial structure. We prove that the asymptotic expected convoy size is universal for all fixed jump rates $q \in [0,1)$. In the special case $q=0$, we upgrade this to full convergence in distribution. We further establish a critical scaling $q = 1 - γ/\sqrt{n}$ that yields a nontrivial limiting regime. Our analysis builds on Martin's construction of the convoy and makes use of an orthogonal-polynomial representation of random-walk transition probabilities.
