405.405. Policy Iteration

Policy Iteration

===============

Policy iteration is a method for finding the optimal policy in a Markov decision process. It involves iterating between policy evaluation and policy improvement until convergence.

**Example:** We have a simple grid world with four states (A, B, C, D) and two actions (left, right). The reward function is as follows:

| State | Reward |

| --- | --- |

| A | 0 |

| B | 0 |

| C | 10 |

| D | 0 |

The goal is to find the optimal policy using policy iteration.

**Constraints:** The policy is stationary, and the reward function is deterministic.