Fitted Q-iteration in continuous action-space MDPs

Andras Antos 1 Rémi Munos 2 Csaba Szepesvari 1
2 SEQUEL - Sequential Learning
LIFL - Laboratoire d'Informatique Fondamentale de Lille, Inria Lille - Nord Europe, LAGIS - Laboratoire d'Automatique, Génie Informatique et Signal
Abstract : We consider continuous state, continuous action batch reinforcement learning where the goal is to learn a good policy from a sufficiently rich trajectory generated by some policy. We study a variant of fitted Q-iteration, where the greedy action selection is replaced by searching for a policy in a restricted set of candidate policies by maximizing the average action values. We provide a rigorous analysis of this algorithm, proving what we believe is the first finite-time bound for value-function based algorithms for continuous state and action problems.
Document type :
Conference papers
Complete list of metadatas

Cited literature [16 references]  Display  Hide  Download
Contributor : Rémi Munos <>
Submitted on : Wednesday, January 9, 2008 - 5:08:45 PM
Last modification on : Thursday, February 21, 2019 - 10:52:49 AM
Long-term archiving on : Thursday, September 27, 2012 - 2:00:51 PM


Files produced by the author(s)


  • HAL Id : inria-00203359, version 1



Andras Antos, Rémi Munos, Csaba Szepesvari. Fitted Q-iteration in continuous action-space MDPs. Neural Information Processing Systems, 2007, Vancouver, Canada. ⟨inria-00203359⟩



Record views


Files downloads