Paperlane

no. 15 · Monday 16 November 2026

Transformation of Stochastic Recursions and Critical Phenomena in the Analysis of the Athreya-Aldous-Shields Cascade and Related Mean Flow Equations

Radu Dascaliuc, Tuan N. Pham, Enrique Thomann and Edward C. Waymire

math.PR · v3 · 62 pp

Abstract

We study the discounted branching random walk process introduced by Athreya (1985) and Aldous and Shields (1988), referred to as the Athreya-Aldous-Shields (AAS) cascade. This process can be viewed as an idealized version of the stochastic model introduced by Le Jan and Sznitman in 1997 for the Navier-Stokes equations. The AAS cascade exhibits phase transitions at the critical branching rates $α=1$ and $α=2$. These transitions are reflected in the behavior of the associated mean flow equation, a nonlocal ODE referred to as the $α$-Riccati equation. It is known that if $α>1$, the AAS cascade is explosive in the sense of infinitely many splittings in finite time, and the $α$-Riccati equation has at least two solutions. In this paper, we establish finer structural properties of an explosive AAS cascade and obtain a stronger non-uniqueness result. Namely, for each initial data in a certain range, the $α$-Riccati equation has an infinite family of distinct solutions parametrized by their convergence rates at infinity. Our approach exploits a simple transformation between the AAS cascade and its "linearized" counterpart, whose mean flow equation is precisely the well-studied pantograph equation. This transformation demonstrates how a relationship at the level of stochastic processes can reveal an otherwise hidden relationship at the level of mean flow equations.

comments
62 pages, 7 figures
fields
Probability, Analysis of PDEs
pdf ↗ page arXiv:2411.00629 ↗