Publications• Sorted by Date • Classified by Publication Type • Classified by Research Category • Decentralized MCTS via Learned Teammate ModelsAleksander Czechowski and Frans A. Oliehoek. Decentralized MCTS via Learned Teammate Models. In Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence (IJCAI), pp. 81–88, July 2020. DownloadAbstractDecentralized online planning can be an attractive paradigm for cooperative multi-agent systems, due to improved scalability and robustness. A key diffi- culty of such approach lies in making accurate pre- dictions about the decisions of other agents. In this paper, we present a trainable online decentralized planning algorithm based on decentralized Monte Carlo Tree Search, combined with models of team- mates learned from previous episodic runs. By only allowing one agent to adapt its models at a time, un- der the assumption of ideal policy approximation, successive iterations of our method are guaranteed to improve joint policies, and eventually lead to convergence to a Nash equilibrium. We test the effi- ciency of the algorithm by performing experiments in several scenarios of the spatial task allocation en- vironment introduced in [Claes et al., 2015]. We show that deep learning and convolutional neural networks can be employed to produce accurate pol- icy approximators which exploit the spatial features of the problem, and that the proposed algorithm im- proves over the baseline planning performance for particularly challenging domain configurations. BibTeX Entry@inproceedings{Czechowski20IJCAI, author = {{Czechowski}, Aleksander and {Oliehoek}, Frans A.}, title = {Decentralized {MCTS} via Learned Teammate Models}, booktitle= IJCAI20, year = 2020, month = jul, pages = {81--88}, doi = {10.24963/ijcai.2020/12}, url = {https://doi.org/10.24963/ijcai.2020/12}, keywords = {refereed}, abstract = { Decentralized online planning can be an attractive paradigm for cooperative multi-agent systems, due to improved scalability and robustness. A key diffi- culty of such approach lies in making accurate pre- dictions about the decisions of other agents. In this paper, we present a trainable online decentralized planning algorithm based on decentralized Monte Carlo Tree Search, combined with models of team- mates learned from previous episodic runs. By only allowing one agent to adapt its models at a time, un- der the assumption of ideal policy approximation, successive iterations of our method are guaranteed to improve joint policies, and eventually lead to convergence to a Nash equilibrium. We test the effi- ciency of the algorithm by performing experiments in several scenarios of the spatial task allocation en- vironment introduced in [Claes et al., 2015]. We show that deep learning and convolutional neural networks can be employed to produce accurate pol- icy approximators which exploit the spatial features of the problem, and that the proposed algorithm im- proves over the baseline planning performance for particularly challenging domain configurations. } }
Generated by
bib2html.pl
(written by Patrick Riley) on
Mon Oct 07, 2024 14:17:04 UTC |