Periodic FPU system: Continuum limit to KdV via regularization and Fourier analysis

TL;DR

This paper investigates the continuum limit of the periodic FPU system, connecting it to the KdV equation through advanced Fourier analysis and regularization techniques, addressing challenges posed by the lack of smoothing effects.

Contribution

It extends the continuum limit analysis of the periodic FPU system to broader initial data classes using L4-Strichartz estimates and normal form regularization.

Findings

Established uniform well-posedness for rough data

Derived continuum limit to KdV for periodic FPU

Developed L4-Strichartz estimates for FPU solutions

Abstract

The Fermi-Pasta-Ulam (FPU) system, initially introduced by Fermi for numerical simulations, models vibrating chains with fixed endpoints, where particles interact weakly, nonlinearly with their nearest neighbors. Contrary to the anticipated ergodic behavior, the simulation revealed nearly periodic (quasi-periodic) motion of the solutions, a phenomenon later referred to as the FPU paradox. A partial but remarkable explanation was provided by Zabusky and Kruskal [36], who formally derived the continuum limit of the FPU system, connecting it to the Korteweg-de Vries (KdV) equation. This formal derivation was later rigorously justified by Bambusi and Ponno [4]. In this paper, we revisit the problem studied in [4], specifically focusing on the continuum limit of the periodic FPU system for a broader class of initial data, as the number of particles N tends to infinity within a fixed domain. Unlike the non-periodic case discussed in [15], periodic FPU solutions lack a (local) smoothing effect, posing a significant challenge in controlling one derivative in the nonlinearity. This control is crucial not only for proving the (uniform in N) well-posedness for rough data but also for deriving the continuum limit. The main strategies to resolve this issue involve deriving L4-Strichartz estimates for FPU solutions, analogous to those previously derived for KdV solutions in [7], and regularizing the system via the normal form method introduced in [1].