A Renormalization Group Approach to Spontaneous Stochasticity

TL;DR

This paper introduces a renormalization group framework to understand spontaneous stochasticity in near-singular classical systems, revealing universal statistics and fixed points that explain turbulence phenomena.

Contribution

It develops a novel RG approach to characterize spontaneous stochasticity, providing exact solutions for a minimal model and insights into universal turbulence effects.

Findings

RG fixed points determine spontaneous statistics

Universal approach to fixed points via large-deviations scaling

Numerical simulations confirm analytical predictions

Abstract

We develop a theoretical approach to ``spontaneous stochasticity'' in classical dynamical systems that are nearly singular and weakly perturbed by noise. This phenomenon is associated to a breakdown in uniqueness of solutions for fixed initial data and underlies many fundamental effects of turbulence (unpredictability, anomalous dissipation, enhanced mixing). Based upon analogy with statistical-mechanical critical points at zero temperature, we elaborate a renormalization group (RG) theory that determines the universal statistics obtained for sufficiently long times after the precise initial data are ``forgotten''. We apply our RG method to solve exactly the ``minimal model'' of spontaneous stochasticity given by a 1D singular ODE. Generalizing prior results for the infinite-Reynolds limit of our model, we obtain the RG fixed points that characterize the spontaneous statistics in the near-singular, weak-noise limit, determine the exact domain of attraction of each fixed point, and derive the universal approach to the fixed points as a singular large-deviations scaling, distinct from that obtained by the standard saddle-point approximation to stochastic path-integrals in the zero-noise limit. We present also numerical simulation results that verify our analytical predictions, propose possible experimental realizations of the ``minimal model'', and discuss more generally current empirical evidence for ubiquitous spontaneous stochasticity in Nature. Our RG method can be applied to more complex, realistic systems and some future applications are briefly outlined.