Dissemin is shutting down on January 1st, 2025

Links

Tools

Export citation

Search in Google Scholar

Heuristically Accelerated Reinforcement Learning: Theoretical and Experimental Results

Journal article published in 2012 by Reinaldo A. C. Bianchi ORCID, Carlos H. C. Ribeiro, Anna H. R. Costa
This paper is available in a repository.
This paper is available in a repository.

Full text: Download

Question mark in circle
Preprint: policy unknown
Question mark in circle
Postprint: policy unknown
Question mark in circle
Published version: policy unknown

Abstract

Since finding control policies using Reinforcement Learning (RL) can be very time consuming, in recent years several authors have investigated how to speed up RL algorithms by mak-ing improved action selections based on heuristics. In this work we present new theoretical results – convergence and a superior limit for value estimation errors – for the class that encompasses all heuristics-based algorithms, called Heuristically Accelerated Reinforcement Learning. We also expand this new class by proposing three new al-gorithms, the Heuristically Accelerated Q(λ), SARSA(λ) and TD(λ), the first algorithms that uses both heuristics and eligibility traces. Empirical evaluations were conducted in traditional control problems and results show that using heuristics significantly enhances the per-formance of the learning process.