Can I read Sound Heuristic Search Value Iteration for Undiscounted POMDPs with Reachability Objectives on EtoBox?

Sound Heuristic Search Value Iteration for Undiscounted POMDPs with Reachability Objectives by Ho, Qi Heng; Feather, Martin S.; Rossi, Federico; Sunberg, Zachary N.; Lahijanian, Morteza is a scholarly article available to read on EtoBox.

What is Sound Heuristic Search Value Iteration for Undiscounted POMDPs with Reachability Objectives about?

Partially Observable Markov Decision Processes (POMDPs) are powerful models for sequential decision making under transition and observation uncertainties. This paper studies the challenging yet important problem in POMDPs known as the (indefinite-horizon) Maximal Reachability Probability Problem (MRPP), where the goal is to maximize the probability of reaching some target states. This is also a core problem in model checking with logical specifications and is naturally undiscounted (discount factor is one). Inspired by the success of point-based methods developed for discounted problems, we study their extensions to MRPP. Specifically, we focus on trial-based heuristic search value iteration techniques and present a novel algorithm that leverages the strengths of these techniques for efficient exploration of the belief space (informed search via value bounds) while addressing their drawbacks in handling loops for indefinite-horizon problems. The algorithm produces policies with two-sided bounds on optimal reachability probabilities. We prove convergence to an optimal policy from below under certain conditions. Experimental evaluations on a suite of benchmarks show that our algorith

Author

Ho, Qi Heng; Feather, Martin S.; Rossi, Federico; Sunberg, Zachary N.; Lahijanian, Morteza

Published

2024

Language

EN