On a pinning model in correlated Gaussian random environments
Zian Li, Jian Song, Ran Wei, Hang Zhang
TL;DR
This work analyzes a pinning model in a correlated Gaussian environment with covariance $\gamma(n) \sim |n|^{2H-2}$ and renewal exponent $\alpha$, identifying intermediate-disorder scaling limits. The authors construct Wick-ordered and non-Wick partition functions, prove Slkorohod (Wiener) chaos limits for the Wick-ordered case and Stratonovich chaos limits for the original model, under precise conditions on $H$ and $\alpha$ (notably $H\in(\tfrac12,1)$ and $\alpha>\tfrac12$, with stronger $\alpha+H>\tfrac32$ for Stratonovich). The limits are given by continuum chaos expansions with kernels $\phi_k$ and $\psi_{m,k}$ against Gaussian noise with covariance $|t-s|^{2H-2}$, aligning with the Weinrib–Halperin disorder-relevance predictions in the correlated setting. The paper also establishes uniform integrability and $L^1$-boundedness in stronger regimes, highlighting a new feature where full $L^2$-boundedness can fail in the disorder-relevant region due to environmental correlations, and discusses potential $L^1$ limits when $\alpha\le\tfrac12$. Overall, the results provide a rigorous link between discrete disordered pinning models and their continuum limits in correlated environments, enriching the understanding of disorder relevance beyond the classical Harris criterion.
Abstract
We consider a pinning model in correlated Gaussian random environments. For the model that is disorder relevant, we study its intermediate disorder regime and show that the rescaled partition functions converge to a non-trivial continuum limit in the Skorohod setting and in the Stratonovich setting, respectively. Our results partially confirm the prediction of Weinrib and Halperin for disorder relevance/irrelevance.
