This project implements a model-based policy iteration algorithm for an agent navigating a discretized 2D terrain to find the global minimum.