While the advantages of actor-only algorithms, the ones that search directly the policy without the use of the value function, are clear (possibility of having a continuous action space, a stochastic policy...), I can't figure out the disadvantages, if not general statements about less stability or bigger variance with respect to critic-only methods (with which I refer to methods that are based on the value function).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.