Reinforcement learning with hierarchical decision-making

Shahar Cohen*, Oded Maimon, Evgeni Khmlenitsky

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Scopus citations

Abstract

This paper proposes a simple, hierarchical decision-making approach to reinforcement learning, under the framework of Markov decision processes. According to the approach, the choice of an action, in every time stage, is made through a successive elimination of actions and sets of actions from the underlined action-space, until a single action is decided upon. Based on the approach, the paper defines a hierarchical Q-function, and shows that this function can be the basis for an optimal policy. A hierarchical reinforcement learning algorithm is then proposed. The algorithm, which can be shown to converge to the hierarchical Q-function, provides new opportunities for state abstraction.

Original languageEnglish
Title of host publicationProceedings - ISDA 2006
Subtitle of host publicationSixth International Conference on Intelligent Systems Design and Applications
Pages177-182
Number of pages6
DOIs
StatePublished - 2006
EventISDA 2006: Sixth International Conference on Intelligent Systems Design and Applications - Jinan, China
Duration: 16 Oct 200618 Oct 2006

Publication series

NameProceedings - ISDA 2006: Sixth International Conference on Intelligent Systems Design and Applications
Volume3

Conference

ConferenceISDA 2006: Sixth International Conference on Intelligent Systems Design and Applications
Country/TerritoryChina
CityJinan
Period16/10/0618/10/06

Fingerprint

Dive into the research topics of 'Reinforcement learning with hierarchical decision-making'. Together they form a unique fingerprint.

Cite this