Solving Multi-Model MDPs by Coordinate Ascent and Dynamic ProgrammingDownload PDF

Published: 08 May 2023, Last Modified: 26 Jun 2023UAI 2023Readers: Everyone
Keywords: Markov decision process, policy gradient, dynamic programming
TL;DR: An efficient dynamic programming algorithm for solving multi-model MDPs.
Abstract: Multi-model Markov decision process (MMDP) is a promising framework for computing policies that are robust to parameter uncertainty in MDPs. MMDPs aim to find a policy that maximizes the expected return over a \emph{distribution} of MDP models. Because MMDPs are NP-hard to solve, most methods resort to approximations. In this paper, we derive the policy gradient of MMDPs and propose CADP, which combines a coordinate ascent method and a dynamic programming algorithm for solving MMDPs. The main innovation of CADP compared with earlier algorithms is to take the coordinate ascent perspective to adjust model weights iteratively to guarantee monotone policy improvements to a local maximum. A theoretical analysis of CADP proves that it never performs worse than previous dynamic programming algorithms like WSU. Our numerical results indicate that CADP substantially outperforms existing methods on several benchmark problems.
Supplementary Material: pdf
Other Supplementary Material: zip
0 Replies