Institutional Repository [SANDBOX]
Technical University of Crete
EN  |  EL

Search

Browse

My Space

Influence of state-variable constraints on partially observable Monte Carlo planning

Castellini Alberto, Chalkiadakis Georgios, Farinelli Alessandro

Full record


URI: http://purl.tuc.gr/dl/dias/CC802758-9A21-4D99-A2B7-E8EF862C8D47
Year 2019
Type of Item Conference Full Paper
License
Details
Bibliographic Citation A. Castellini, G. Chalkiadakis and A. Farinelli, "Influence of state-variable constraints on partially observable Monte Carlo planning," in 28th International Joint Conference on Artificial Intelligence, 2019, pp. 5540-5546. doi: 10.24963/ijcai.2019/769 https://doi.org/10.24963/ijcai.2019/769
Appears in Collections

Summary

Online planning methods for partially observable Markov decision processes (POMDPs) have recently gained much interest. In this paper, we propose the introduction of prior knowledge in the form of (probabilistic) relationships among discrete state-variables, for online planning based on the well-known POMCP algorithm. In particular, we propose the use of hard constraint networks and probabilistic Markov random fields to formalize state-variable constraints and we extend the POMCP algorithm to take advantage of these constraints. Results on a case study based on Rocksample show that the usage of this knowledge provides significant improvements to the performance of the algorithm. The extent of this improvement depends on the amount of knowledge encoded in the constraints and reaches the 50% of the average discounted return in the most favorable cases that we analyzed.

Available Files

Services

Statistics