Under-Approximating Expected Total Rewards in POMDPs
Abstract
We consider the problem: is the optimal expected total reward to reach a goal state in a partially observable Markov decision process (POMDP) below a given threshold? We tackle this -- generally undecidable -- problem by computing under-approximations on these total expected rewards. This is done by abstracting finite unfoldings of the infinite belief MDP of the POMDP. The key issue is to find a suitable under-approximation of the value function. We provide two techniques: a simple (cut-off) technique that uses a good policy on the POMDP, and a more advanced technique (belief clipping) that uses minimal shifts of probabilities between beliefs. We use mixed-integer linear programming (MILP) to find such minimal probability shifts and experimentally show that our techniques scale quite well while providing tight lower bounds on the expected total reward.
Cite
@article{arxiv.2201.08772,
title = {Under-Approximating Expected Total Rewards in POMDPs},
author = {Alexander Bork and Joost-Pieter Katoen and Tim Quatmann},
journal= {arXiv preprint arXiv:2201.08772},
year = {2022}
}
Comments
Technical report for TACAS 2022 paper with the same title