Memoryless Policy Iteration for Episodic POMDPs

TL;DR

This paper introduces a novel class of memoryless policy iteration algorithms for episodic POMDPs that improve computational efficiency and can be implemented in a model-free manner, outperforming existing methods.

Contribution

It proposes a new family of policy iteration algorithms for POMDPs that operate directly in output space and identifies optimal periodic patterns for policy improvement.

Findings

Achieves significant speedups over policy-gradient methods.

Develops a model-free variant that learns directly from data.

Demonstrates effectiveness across several POMDP examples.

Abstract

Memoryless and finite-memory policies offer a practical alternative for solving partially observable Markov decision processes (POMDPs), as they operate directly in the output space rather than in the high-dimensional belief space. However, extending classical methods such as policy iteration to this setting remains difficult; the output process is non-Markovian, making policy-improvement steps interdependent across stages. We introduce a new family of monotonically improving policy-iteration algorithms that alternate between single-stage output-based policy improvements and policy evaluations according to a prescribed periodic pattern. We show that this family admits optimal patterns that maximize a natural computational-efficiency index, and we identify the simplest pattern with minimal period. Building on this structure, we further develop a model-free variant that estimates values from data and learns memoryless policies directly. Across several POMDPs examples, our method achieves significant computational speedups over policy-gradient baselines and recent specialized algorithms in both model-based and model-free settings.