Region-Based Incremental Pruning for POMDPs

Zhengzhu Feng and Shlomo Zilberstein. Region-Based Incremental Pruning for POMDPs. Proceedings of the Twentieth Conference on Uncertainty in Artificial Intelligence (UAI), 146-153, Banff, Canada. 2004.

Abstract

We present a major improvement to the incremental pruning algorithm for solving partially observable Markov decision processes. Our technique targets the cross-sum step of the dynamic programming (DP) update, a key source of complexity in POMDP algorithms. Instead of reasoning about the whole belief space when pruning the cross-sums, our algorithm divides the belief space into smaller regions and performs independent pruning in each region. We evaluate the benefits of the new technique both analytically and experimentally, and show that it produces very significant performance gains. The results contribute to the scalability of POMDP algorithms to domains that cannot be handled by the best existing techniques.

Bibtex entry:

@inproceedings{FZuai04,
  author	= {Zhengzhu Feng and Shlomo Zilberstein},
  title		= {Region-Based Incremental Pruning for {POMDP}s},
  booktitle     = {Proceedings of the Twentieth Conference on Uncertainty in
                   Artificial Intelligence},
  year		= {2004},
  pages		= {146-153},
  address       = {Banff, Canada},
  url		= {http://rbr.cs.umass.edu/shlomo/papers/FZuai04.html}
}

shlomo@cs.umass.edu