Policy Search by Dynamic Programming (2003)

Drew Bagnell, Sham Kakade, Andrew Ng, Jeff Schneider

Tags

Markov Decision Processes, Reinforcement Learning

Abstract

Abstract We consider the policy search approach to reinforcement learning. We show that if a ``baseline distribution'' is given (indicating roughly how often we expect a good policy to visit each state), then we can derive a policy search algorithm that terminates in a finite number of steps, and for which we can provide non-trivial performance guarantees. We also demonstrate this algorithm on several grid-world POMDPs, a planar biped walking robot, and a double-pole balancing problem.

Full text

Download (application/pdf, 157.4 kB)

Approximate BibTeX Entry

@inproceedings{bagnellPSDP,
    Year = {2003},
    Booktitle = {Proceedings of Neural Information Processing Systems},
    Author = { Drew Bagnell, Sham Kakade, Andrew Ng, Jeff Schneider },
    Title = {Policy Search by Dynamic Programming}
}