Non-Perturbative Functional Renormalization Group for Random Field Models and Related Disordered Systems. I: Effective Average Action Formalism

TL;DR

This paper introduces a nonperturbative functional renormalization group method for disordered systems with many metastable states, successfully unifying perturbative results and providing a comprehensive description across dimensions and component numbers.

Contribution

It develops a nonperturbative RG approach combining an exact flow equation with a cumulant-based approximation for disordered systems, applicable to the random field $O(N)$ model.

Findings

Reproduces known perturbative results near critical dimensions and at large N.

Provides a unified nonperturbative description across the (N,d) plane.

Applicable to models with many metastable states where perturbation theory fails.

Abstract

We have developed a nonperturbative functional renormalization group approach for random field models and related disordered systems for which, due to the existence of many metastable states, conventional perturbation theory often fails. The approach combines an exact renormalization group equation for the effective average action with a nonperturbative approximation scheme based on a description of the probability distribution of the renormalized disorder through its cumulants. For the random field $O(N)$ model, the minimal truncation within this scheme is shown to reproduce the known perturbative results in the appropriate limits, near the upper and lower critical dimensions and at large number $N$ of components, while providing a unified nonperturbative description of the full $(N,d)$ plane, where $d$ is the spatial dimension.