Policy evaluation
= Policy evaluation
Estimating the expected return under a specified control policy. It is distinct from improving or optimizing that policy. A <Bellman expectation equation> describes the fixed-policy value, whereas a Bellman optimality equation compares admissible actions.