Genetic network programming with reinforcement learning and its application to making mobile robot behavior

Shingo Mabu, Hiroyuki Hatakeyamay, Moe Thu Thu, Kotaro Hirasawa, Takayuki Furuzuki

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

A new graph-based evolutionary algorithm called "Genetic Network Programming, GNP" has been proposed. The solutions of GNP are represented as graph structures, which can improve the expression ability and performance. In addition, GNP with Reinforcement Learning (GNP-RL) has been proposed to search for solutions efficiently. GNP-RL can use current information (state and reward) and change its programs during task execution. Thus, it has an advantage over evolution-based algorithms in case much information can be obtained during task execution. The GNP we proposed in the previous research deals with discrete information, but in this paper, we extend the conventional GNP-RL which can deal with numerical information. The proposed method is applied to the controller of Khepera simulator and its performance is evaluated.

Original languageEnglish
Pages (from-to)1009-1015
Number of pages7
JournalIEEJ Transactions on Electronics, Information and Systems
Volume126
Issue number8
DOIs
Publication statusPublished - 2006

Fingerprint

Reinforcement learning
Mobile robots
Evolutionary algorithms
Simulators
Controllers

Keywords

  • Evolutionary Computation
  • Genetic Network Programming
  • Khepera robot
  • Reinforcement Learning

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Genetic network programming with reinforcement learning and its application to making mobile robot behavior. / Mabu, Shingo; Hatakeyamay, Hiroyuki; Thu, Moe Thu; Hirasawa, Kotaro; Furuzuki, Takayuki.

In: IEEJ Transactions on Electronics, Information and Systems, Vol. 126, No. 8, 2006, p. 1009-1015.

Research output: Contribution to journalArticle

@article{df9c11c40b8f43b1b53f7e568a01ed02,
title = "Genetic network programming with reinforcement learning and its application to making mobile robot behavior",
abstract = "A new graph-based evolutionary algorithm called {"}Genetic Network Programming, GNP{"} has been proposed. The solutions of GNP are represented as graph structures, which can improve the expression ability and performance. In addition, GNP with Reinforcement Learning (GNP-RL) has been proposed to search for solutions efficiently. GNP-RL can use current information (state and reward) and change its programs during task execution. Thus, it has an advantage over evolution-based algorithms in case much information can be obtained during task execution. The GNP we proposed in the previous research deals with discrete information, but in this paper, we extend the conventional GNP-RL which can deal with numerical information. The proposed method is applied to the controller of Khepera simulator and its performance is evaluated.",
keywords = "Evolutionary Computation, Genetic Network Programming, Khepera robot, Reinforcement Learning",
author = "Shingo Mabu and Hiroyuki Hatakeyamay and Thu, {Moe Thu} and Kotaro Hirasawa and Takayuki Furuzuki",
year = "2006",
doi = "10.1541/ieejeiss.126.1009",
language = "English",
volume = "126",
pages = "1009--1015",
journal = "IEEJ Transactions on Electronics, Information and Systems",
issn = "0385-4221",
publisher = "The Institute of Electrical Engineers of Japan",
number = "8",

}

TY - JOUR

T1 - Genetic network programming with reinforcement learning and its application to making mobile robot behavior

AU - Mabu, Shingo

AU - Hatakeyamay, Hiroyuki

AU - Thu, Moe Thu

AU - Hirasawa, Kotaro

AU - Furuzuki, Takayuki

PY - 2006

Y1 - 2006

N2 - A new graph-based evolutionary algorithm called "Genetic Network Programming, GNP" has been proposed. The solutions of GNP are represented as graph structures, which can improve the expression ability and performance. In addition, GNP with Reinforcement Learning (GNP-RL) has been proposed to search for solutions efficiently. GNP-RL can use current information (state and reward) and change its programs during task execution. Thus, it has an advantage over evolution-based algorithms in case much information can be obtained during task execution. The GNP we proposed in the previous research deals with discrete information, but in this paper, we extend the conventional GNP-RL which can deal with numerical information. The proposed method is applied to the controller of Khepera simulator and its performance is evaluated.

AB - A new graph-based evolutionary algorithm called "Genetic Network Programming, GNP" has been proposed. The solutions of GNP are represented as graph structures, which can improve the expression ability and performance. In addition, GNP with Reinforcement Learning (GNP-RL) has been proposed to search for solutions efficiently. GNP-RL can use current information (state and reward) and change its programs during task execution. Thus, it has an advantage over evolution-based algorithms in case much information can be obtained during task execution. The GNP we proposed in the previous research deals with discrete information, but in this paper, we extend the conventional GNP-RL which can deal with numerical information. The proposed method is applied to the controller of Khepera simulator and its performance is evaluated.

KW - Evolutionary Computation

KW - Genetic Network Programming

KW - Khepera robot

KW - Reinforcement Learning

UR - http://www.scopus.com/inward/record.url?scp=33746867121&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33746867121&partnerID=8YFLogxK

U2 - 10.1541/ieejeiss.126.1009

DO - 10.1541/ieejeiss.126.1009

M3 - Article

AN - SCOPUS:33746867121

VL - 126

SP - 1009

EP - 1015

JO - IEEJ Transactions on Electronics, Information and Systems

JF - IEEJ Transactions on Electronics, Information and Systems

SN - 0385-4221

IS - 8

ER -