July
2004
Approximate Solutions for Partially Observable Stochastic Games with Common Payoffs
Authors:
Abstract:
Partially observable decentralized decision making in robot teams is fundamentally different from decision making in fully observable problems. Team members cannot simply apply single-agent solution techniques in parallel. Instead, we must turn to game theoretic frameworks to correctly model the problem. While partially observable stochastic games (POSGs) provide a solution model for decentralized robot teams, this model quickly becomes intractable. We propose an algorithm that approximates POSGs as a series of smaller, related Bayesian games, using heuristics such as QMDP to provide the future discounted value of actions. This algorithm trades off limited look-ahead in uncertainty for computational feasibility, and results in policies that are locally optimal with respect to the selected heuristic. Empirical results are provided for both a simple problem for which the full POSG can also be constructed, as well as more complex, robot-inspired, problems.
copied = false, 2000);
">
@conference{Emery-Montemerlo-2004-119829,
author = {R. Emery-Montemerlo And G. Gordon And J. Schneider And S. Thrun},
title = {Approximate Solutions for Partially Observable Stochastic Games with Common Payoffs},
booktitle = {Proceedings of 3rd International Joint Conference on Autonomous Agents and MultiAgent Systems (AAMAS '04)},
year = {2004},
month = {July},
volume = {1},
pages = {136 - 143},
}
author = {R. Emery-Montemerlo And G. Gordon And J. Schneider And S. Thrun},
title = {Approximate Solutions for Partially Observable Stochastic Games with Common Payoffs},
booktitle = {Proceedings of 3rd International Joint Conference on Autonomous Agents and MultiAgent Systems (AAMAS '04)},
year = {2004},
month = {July},
volume = {1},
pages = {136 - 143},
}