Speeding up Online POMDP Planning - Unification of Observation Branches by Belief-state Compression Via Expected Feature Values

Gavin Rens

2015

Abstract

A novel algorithm to speed up online planning in partially observable Markov decision processes (POMDPs) is introduced. I propose a method for compressing nodes in belief-decision-trees while planning occurs. Whereas belief-decision-trees branch on actions and observations, with my method, they branch only on actions. This is achieved by unifying the branches required due to the nondeterminism of observations. The method is based on the expected values of domain features. The new algorithm is experimentally compared to three other online POMDP algorithms, outperforming them on the given test domain.

Paper Citation

in Harvard Style

Rens G. (2015). Speeding up Online POMDP Planning - Unification of Observation Branches by Belief-state Compression Via Expected Feature Values . In Proceedings of the International Conference on Agents and Artificial Intelligence - Volume 2: ICAART, ISBN 978-989-758-074-1, pages 241-246. DOI: 10.5220/0005165802410246

in Bibtex Style

@conference{icaart15,
author={Gavin Rens},
title={Speeding up Online POMDP Planning - Unification of Observation Branches by Belief-state Compression Via Expected Feature Values},
booktitle={Proceedings of the International Conference on Agents and Artificial Intelligence - Volume 2: ICAART,},
year={2015},
pages={241-246},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0005165802410246},
isbn={978-989-758-074-1},
}

in EndNote Style

TY - CONF
JO - Proceedings of the International Conference on Agents and Artificial Intelligence - Volume 2: ICAART,
TI - Speeding up Online POMDP Planning - Unification of Observation Branches by Belief-state Compression Via Expected Feature Values
SN - 978-989-758-074-1
AU - Rens G.
PY - 2015
SP - 241
EP - 246
DO - 10.5220/0005165802410246