Rate of convergence for homogenization of nonlinear weakly coupled Hamilton-Jacobi systems
Hiroyoshi Mitake, Panrui Ni
TL;DR
The paper addresses periodic homogenization for nonlinear weakly coupled Hamilton-Jacobi systems with convex Hamiltonians and nonlinear coupling, introducing an effective Hamiltonian via cell problems. It develops a fixed-point iteration framework to control the multiscale error and proves a sharp convergence rate $|\mathbf{u}^\varepsilon-\bar{\mathbf{u}}|\le C\sqrt{\varepsilon}$ for times away from initial, along with a small-time bound, using quantitative homogenization estimates. The results extend to a stationary discounted problem under a strict monotonicity condition $\lambda>\Theta$, yielding the same $O(\sqrt{\varepsilon})$ rate. The methods combine iteration techniques with recent quantitative homogenization theory for Hamilton-Jacobi equations, establishing the first rate-of-convergence result for nonlinear, non-monotone, weakly coupled HJ systems and broadening applicability beyond single-equation settings.
Abstract
Here, we study the periodic homogenization problem of nonlinear weakly coupled systems of Hamilton-Jacobi equations in the convex setting. We establish a rate of convergence $O(\sqrt{\varepsilon})$ which is sharp.
