Approximate Policy Iteration with a Policy Language Bias

DMCA

Approximate Policy Iteration with a Policy Language Bias (2003)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Alan Fern , Sungwook Yoon , Robert Givan

Venue:	Journal of Artificial Intelligence Research
Citations:	139 - 18 self

BibTeX

@INPROCEEDINGS{Fern03approximatepolicy,
    author = {Alan Fern and Sungwook Yoon and Robert Givan},
    title = {Approximate Policy Iteration with a Policy Language Bias},
    booktitle = {Journal of Artificial Intelligence Research},
    year = {2003},
    publisher = {MIT Press}
}

OpenURL

Abstract

We explore approximate policy iteration (API), replacing the usual costfunction learning step with a learning step in policy space. We give policy-language biases that enable solution of very large relational Markov decision processes (MDPs) that no previous technique can solve.

Keyphrases

approximate policy iteration policy language bias enable solution previous technique policy space learning step policy-language bias usual costfunction large relational markov decision process

DMCA