A mixed problem for the infinity laplacian via Tug-of-War games

TL;DR

This paper establishes that the value of a Tug-of-War game corresponds exactly to the unique viscosity solution of a mixed boundary value problem for the infinity Laplacian, linking game theory with PDE solutions.

Contribution

It proves the equivalence between the Tug-of-War game value and the viscosity solution of the infinity Laplacian with mixed boundary conditions, extending previous results.

Findings

The game value matches the unique viscosity solution.

The PDE has a unique solution as the absolutely minimizing Lipschitz extension.

The results connect stochastic game theory with nonlinear PDE analysis.

Abstract

In this paper we prove that a function $ u\in\mathcal{C}(\bar{\Omega})$ is the continuous value of the Tug-of-War game described in \cite{PSSW} if and only if it is the unique viscosity solution to the infinity laplacian with mixed boundary conditions {-\Delta_{\infty}u(x)=0\quad & \text{in} \Omega, \frac{\partial u}{\partial n}(x)=0\quad & \text{on} \Gamma_N, u(x)=F(x)\quad & \text{on} \Gamma_D. By using the results in \cite{PSSW}, it follows that this viscous PDE problem has a unique solution, which is the unique {\it absolutely minimizing Lipschitz extension} to the whole $\bar{\Omega}$ (in the sense of \cite{Aronsson} and \cite{PSSW}) of the boundary data $ F:\Gamma_D\to\R $.