Solving Deterministic Policy (PO)MDPs using

The viewpoint of solving Markov Decision Processes and their partially observable extension refers to nding policies that max- imise the expected reward. We follow the rephrasing of this problem as