2013arXiv (Cornell University)Open access

A Method for Speeding Up Value Iteration in Partially Observable Markov\n Decision Processes

Nevin L. Zhang, Stephen S. Lee, Weihong Zhang

Open full text 1 citations

Abstract

We present a technique for speeding up the convergence of value iteration for\npartially observable Markov decisions processes (POMDPs). The underlying idea\nis similar to that behind modified policy iteration for fully observable Markov\ndecision processes (MDPs). The technique can be easily incorporated into any\nexisting POMDP value iteration algorithms. Experiments have been conducted on\nseveral test problems with one POMDP value iteration algorithm called\nincremental pruning. We find that the technique can make incremental pruning\nrun several orders of magnitude faster.\n

Open-access reader

About this research paper

What this paper is about

We present a technique for speeding up the convergence of value iteration for\npartially observable Markov decisions processes (POMDPs). The underlying idea\nis similar to that behind modified policy iteration for fully observable Markov\ndecision processes (MDPs). The technique can be easily incorporated into any\nexisting POMDP value iteration algorithms. Experiments have been conducted on\nseveral test problems with one POMDP value iteration algorithm called\nincremental pruning. We find that the technique can make incremental pruning\nrun several orders of magnitude faster.\n

Why it matters

OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

We present a technique for speeding up the convergence of value iteration for\npartially observable Markov decisions processes (POMDPs). The underlying idea\nis similar to that behind modified policy iteration for fully observable Markov\ndecision processes (MDPs). The technique can be easily incorporated into any\nexisting POMDP value iteration algorithms. Experiments have been conducted on\nseveral test problems with one POMDP value iteration algorithm called\nincremental pruning. We find that the technique can make incremental pruning\nrun several orders of magnitude faster.\n

Key concepts: Observable, Markov decision process, Partially observable Markov decision process, Value (mathematics), Computer science, Markov process, Markov chain, Decision process

Related papers

Back to paper searchBrowse research topicsOriginal source
A Method for Speeding Up Value Iteration in Partially Observable Markov\n Decision Processes — Research Paper | ScholarLens