Producing efficient error-bounded solutions for transition independent decentralized MDPs
Résumé
There has been substantial progress on algorithms for single-agent sequential decision making problems represented as partially observable Markov decision processes (POMDPs). A number of efficient algorithms for solving POMDPs share two desirable properties: error-bounds and fast convergence rates. Despite significant efforts, no algorithms for solving decentralized POMDPs benefit from these properties, leading to either poor solution quality or limited scalability. This paper presents the first approach for solving transition independent decentralized Markov decision processes (MDPs), that inherits these properties. Two related algorithms illustrate this approach. The first recasts the original problem as a finite-horizon deterministic and completely observable Markov decision process. In this form, the original problem is solved by combining heuristic search with constraint optimization to quickly converge into a near-optimal policy. This algorithm also provides the foundation for the first algorithm for solving infinite-horizon transition independent decentralized MDPs. We demonstrate that both methods outperform state-of-the-art algorithms by multiple orders of magnitude, and for infinite-horizon decentralized MDPs, the algorithm is able to construct more concise policies by searching cyclic policy graphs.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...