DMCA
Influence-based abstraction for multiagent systems (2012)
| Venue: | In AAAI |
| Citations: | 9 - 4 self |
BibTeX
@INPROCEEDINGS{Oliehoek12influence-basedabstraction,
author = {Frans A. Oliehoek and Stefan J. Witwicki and Leslie P. Kaelbling},
title = {Influence-based abstraction for multiagent systems},
booktitle = {In AAAI},
year = {2012}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
This paper presents a theoretical advance by which factored POSGs can be decomposed into local models. We formal-ize the interface between such local models as the influ-ence agents can exert on one another; and we prove that this interface is sufficient for decoupling them. The result-ing influence-based abstraction substantially generalizes pre-vious work on exploiting weakly-coupled agent interaction structures. Therein lie several important contributions. First, our general formulation sheds new light on the theoretical re-lationships among previous approaches, and promotes future empirical comparisons that could come by extending them beyond the more specific problem contexts for which they were developed. More importantly, the influence-based ap-proaches that we generalize have shown promising improve-ments in the scalability of planning for more restrictive mod-els. Thus, our theoretical result here serves as the foundation for practical algorithms that we anticipate will bring similar improvements to more general planning contexts, and also into other domains such as approximate planning, decision-making in adversarial domains, and online learning. 1
Keyphrases
multiagent system influence-based abstraction local model several important contribution new light promising improve-ments weakly-coupled agent interaction structure theoretical re-lationships specific problem context approximate planning influence-based ap-proaches restrictive mod-els influ-ence agent general formulation adversarial domain practical algorithm online learning pre-vious work theoretical advance result-ing influence-based abstraction general planning context similar improvement future empirical comparison previous approach theoretical result
