What makes some POMDP problems easy to approximate?

Part of Advances in Neural Information Processing Systems 20 (NIPS 2007)

Bibtex Metadata Paper


Wee Lee, Nan Rong, David Hsu


Point-based algorithms have been surprisingly successful in computing approx- imately optimal solutions for partially observable Markov decision processes (POMDPs) in high dimensional belief spaces. In this work, we seek to understand the belief-space properties that allow some POMDP problems to be approximated efficiently and thus help to explain the point-based algorithms’ success often ob- served in the experiments. We show that an approximately optimal POMDP so- lution can be computed in time polynomial in the covering number of a reachable belief space, which is the subset of the belief space reachable from a given belief point. We also show that under the weaker condition of having a small covering number for an optimal reachable space, which is the subset of the belief space reachable under an optimal policy, computing an approximately optimal solution is NP-hard. However, given a suitable set of points that “cover” an optimal reach- able space well, an approximate solution can be computed in polynomial time. The covering number highlights several interesting properties that reduce the com- plexity of POMDP planning in practice, e.g., fully observed state variables, beliefs with sparse support, smooth beliefs, and circulant state-transition matrices.