Reinforcement learning agents applied to a class of control system problems

Michael Helm, Daniel Cooke, Klaus Becker, Larry Pyeatt, Nelson Rushton

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

Abstract

Some control system problems are sufficiently complex that it is difficult to define all of the decision weighting up-front. Reinforcement Learning (RL) can be applied to "tune" system performance. The Centipede Game from Economic Game Theory is used to simulate a control system problem with competing subsystem goals. RL agents are applied to the Centipede Game, which pits agents against each other in a game with increasing payoffs if cooperation is developed between the players. The Centipede Game has been studied both in theory and empirically using human players by other researchers. Human players are more inclined to cooperate and achieve the longer-term payoff than economic game theory would predict. This paper focuses on an experimental study of cooperation between RL players (agents) without explicit communications between the players. This work has application in control systems problems where communications is constrained.

Original languageEnglish
Title of host publication2006 IEEE Region 5 Conference
PublisherIEEE Computer Society
Pages35-40
Number of pages6
ISBN (Print)1424403596, 9781424403592
DOIs
StatePublished - Jan 1 2006
Event2006 IEEE Region 5 Conference - San Antonio, TX, United States
Duration: Apr 7 2006Apr 8 2006

Publication series

Name2006 IEEE Region 5 Conference

Conference

Conference2006 IEEE Region 5 Conference
CountryUnited States
CitySan Antonio, TX
Period04/7/0604/8/06

Keywords

  • Control system
  • Cooperative behavior
  • Game theory
  • Reinforcement learning

Fingerprint Dive into the research topics of 'Reinforcement learning agents applied to a class of control system problems'. Together they form a unique fingerprint.

Cite this