Rollout Sampling Policy Iteration for Decentralized POMDPs

TL;DR

DecRSPI is a scalable Monte Carlo-based algorithm for multi-agent decision-making in DEC-POMDPs, capable of solving larger problems efficiently with good solution quality.

Contribution

Introduces DecRSPI, a novel decentralized rollout sampling policy iteration algorithm that improves scalability and handles large DEC-POMDPs without explicit models.

Findings

Linear time complexity over number of agents

Bounded memory usage

Effective on large, intractable problems

Abstract

We present decentralized rollout sampling policy iteration (DecRSPI) - a new algorithm for multi-agent decision problems formalized as DEC-POMDPs. DecRSPI is designed to improve scalability and tackle problems that lack an explicit model. The algorithm uses Monte- Carlo methods to generate a sample of reachable belief states. Then it computes a joint policy for each belief state based on the rollout estimations. A new policy representation allows us to represent solutions compactly. The key benefits of the algorithm are its linear time complexity over the number of agents, its bounded memory usage and good solution quality. It can solve larger problems that are intractable for existing planning algorithms. Experimental results confirm the effectiveness and scalability of the approach.