 ##  [Policy Iteration](/policy-iteration) 

  ##  [Policy Iteration](https://mathlogic.quantumdictionary.io/policy-iteration-0) 

  
 [![Mathematics & Logic Dictionary](/sites/default/files/styles/large/public/2026-01/Mathematics%20%26%20Logic.png.webp?itok=UhtTRPnp)](/topic-specific-dictionaries/natural-formal-sciences/mathematics-logic)

- Natural &amp; Formal Sciences -

**Mathematics &amp; Logic Dictionary**

 
Definition

An algorithm for solving sequential decision problems (typically Markov decision processes) that alternates between evaluating the value of a current policy and improving the policy greedily with respect to that value until convergence to an optimal policy.