Presented By: Applied Interdisciplinary Mathematics (AIM) Seminar - Department of Mathematics
AIM Seminar: Quantitative Homogenization of Hamilton–Jacobi Equations: An Optimal Control Approach
Yuxi Han (Purdue University)
Abstract: Homogenization describes how systems with rapidly oscillating microscopic structure give rise to effective macroscopic behavior at larger scales. For Hamilton–Jacobi equations arising in optimal control, a fundamental question is quantitative: how fast do solutions of the oscillatory problems converge to the solution of the effective equation? In this talk, I will present an optimal-control approach to this question. Representing solutions through optimal trajectories, and analyzing how these trajectories interact with the microscopic structure, yields sharp convergence rates in settings that go beyond the classical periodic framework. These settings include problems with multiple scales, domains perforated by small periodic obstacles under state-constraint or Dirichlet boundary conditions, and, more recently, Hamilton–Jacobi equations on the Wasserstein space of probability measures, which arise in mean field control.
Contact: Ibrahim Ekren
Contact: Ibrahim Ekren