Batch Policy Iteration Algorithms for Continuous Domains

Bilal Piot (1) , Matthieu Geist (2, 3) , Olivier Pietquin

1 DeepMind [London]
2 CentraleSupélec
3 Georgia Tech Lorraine [Metz]

This paper establishes the link between an adaptation of the policy iteration method for Markov decision processes with continuous state and action spaces and the policy gradient method when the differentiation of the mean value is directly done over the policy without parameterization. This approach allows deriving sound and practical batch Reinforcement Learning algorithms for continuous state and action spaces.

Domaines: Machine Learning [stat.ML]

Bilal Piot, Matthieu Geist, Olivier Pietquin. Batch Policy Iteration Algorithms for Continuous Domains. European Workshop on Reinforcement Learning (EWRL), 2016, Barcelone, Spain. ⟨hal-01629651⟩