Invariance principle for non-homogeneous random walks

TL;DR

This paper establishes an invariance principle for non-homogeneous zero-drift random walks in multiple dimensions, characterizing their scaling limits as elliptic martingale diffusions with detailed excursion and angular component analysis.

Contribution

It introduces a novel Riemannian metric-based approach to characterize the limit diffusion and develops the excursion theory without relying on the strong Markov property.

Findings

Limit process is an elliptic martingale diffusion.

Established skew-product decomposition of excursions.

Identified conditions for time-reversibility of angular components.

Abstract

We prove an invariance principle for a class of zero-drift spatially non-homogeneous random walks in $\mathbb{R}^d$, which may be recurrent in any dimension. The limit $\mathcal{X}$ is an elliptic martingale diffusion, which may be point-recurrent at the origin for any $d\geq2$. To characterise $\mathcal{X}$, we introduce a (non-Euclidean) Riemannian metric on the unit sphere in $\mathbb{R}^d$ and use it to express a related spherical diffusion as a Brownian motion with drift. This representation allows us to establish the skew-product decomposition of the excursions of $\mathcal{X}$ and thus develop the excursion theory of $\mathcal{X}$ without appealing to the strong Markov property. This leads to the uniqueness in law of the stochastic differential equation for $\mathcal{X}$ in $\mathbb{R}^d$, whose coefficients are discontinuous at the origin. Using the Riemannian metric we can also detect whether the angular component of the excursions of $\mathcal{X}$ is time-reversible. If so, the excursions of $\mathcal{X}$ in $\mathbb{R}^d$ generalise the classical Pitman-Yor splitting-at-the-maximum property of Bessel excursions.