Memory-Efficient Symbolic Online Planning for Factored MDPs

Aswin Raghavan Oregon State University, Prasad Tadepalli Oregon State University, Alan Fern Oregon State University, Roni Khardon Tufts University
Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence, PMLR R13:800-809, 2015.

Abstract

Factored Markov Decision Processes (MDP) are a de facto standard for compactly modeling sequential decision making problems with uncertainty. Offline planning based on symbolic operators exploits the factored structure of MDPs, but is memory intensive. We present new memory-efficient symbolic operators for online planning that effectively generalize experience. The soundness of the operators and convergence of the planning algorithms are shown followed by experiments that demonstrate superior scalability in benchmark problems.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR13-university15r, title = {Memory-Efficient Symbolic Online Planning for Factored MDPs}, author = {University, Aswin Raghavan Oregon State and University, Prasad Tadepalli Oregon State and University, Alan Fern Oregon State and University, Roni Khardon Tufts}, booktitle = {Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence}, pages = {800--809}, year = {2015}, editor = {Meila, Marina and Heskes, Tom}, volume = {R13}, series = {Proceedings of Machine Learning Research}, month = {12--16 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r13/main/assets/university15r/university15r.pdf}, url = {https://proceedings.mlr.press/r13/university15r.html}, abstract = {Factored Markov Decision Processes (MDP) are a de facto standard for compactly modeling sequential decision making problems with uncertainty. Offline planning based on symbolic operators exploits the factored structure of MDPs, but is memory intensive. We present new memory-efficient symbolic operators for online planning that effectively generalize experience. The soundness of the operators and convergence of the planning algorithms are shown followed by experiments that demonstrate superior scalability in benchmark problems.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Memory-Efficient Symbolic Online Planning for Factored MDPs %A Aswin Raghavan Oregon State University %A Prasad Tadepalli Oregon State University %A Alan Fern Oregon State University %A Roni Khardon Tufts University %B Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2015 %E Marina Meila %E Tom Heskes %F pmlr-vR13-university15r %I PMLR %P 800--809 %U https://proceedings.mlr.press/r13/university15r.html %V R13 %X Factored Markov Decision Processes (MDP) are a de facto standard for compactly modeling sequential decision making problems with uncertainty. Offline planning based on symbolic operators exploits the factored structure of MDPs, but is memory intensive. We present new memory-efficient symbolic operators for online planning that effectively generalize experience. The soundness of the operators and convergence of the planning algorithms are shown followed by experiments that demonstrate superior scalability in benchmark problems. %Z Reissued by PMLR on 04 October 2026.
APA
University, A.R.O.S., University, P.T.O.S., University, A.F.O.S. & University, R.K.T.. (2015). Memory-Efficient Symbolic Online Planning for Factored MDPs. Proceedings of the 31st Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R13:800-809 Available from https://proceedings.mlr.press/r13/university15r.html. Reissued by PMLR on 04 October 2026.

Related Material