Deep learning methods for high dimensional PDEs

Abstract

In this work we study numerical methods for partial differential equations in high dimensions using approximations with neural networks. We apply those methods to solve optimal control problems and N-agent games problems.