2026/03/19 by Stefano Galatolo, Valerio Lucarini
#math.DS #cond-mat.stat-mech #nlin.CD
Linear response theory aims to predict how an additional forcing alters the statistical properties of a reference system. Such questions have been studied predominantly for autonomous dynamical systems, although many systems in the physical, natural, and social sciences are inherently nonautonomous and evolve under time-dependent external forcings. In this setting, one would like to understand how the system's time-dependent statistical properties change when an additional infinitesimal forcing is applied. Despite its practical relevance, this question has received a rigorous mathematical treatment only for a limited number of systems and perturbations. We develop a rigorous linear response theory for a broad class of deterministic and random nonautonomous systems under uniform assumptions extending those commonly used in the autonomous setting. A central ingredient is rapid loss of memory, namely sufficiently fast forgetting of initial conditions along the nonautonomous evolution. Our main strategy is to reformulate the sequential dynamics as a fixed-point problem for a global transfer operator acting on a sequence space of measures. This yields explicit causal response formulas for predicting the effect of small perturbations on time-dependent statistical states. We illustrate the theory for sequential compositions of expanding maps and for sequential compositions of random maps with additive noise, where uniform positivity of the noise implies exponential loss of memory. We also prove linear response for compact reflected Euler-Maruyama discretizations of dissipative nonautonomous stochastic differential equations. Finally, we apply the framework to a finite-dimensional stochastic discretization of the Ghil-Sellers energy balance model and study its response to a time-dependent perturbation of the greenhouse parameter.