Use of infeasible individuals in Probabilistic Model Building Genetic Network Programming

Xianneng Li, Shingo Mabu, Kotaro Hirasawa

Research output: Chapter in Book/Report/Conference proceedingConference contribution

15 Citations (Scopus)

Abstract

Classical EDAs generally use truncation selection to estimate the distribution of the feasible (good) individuals while ignoring the infeasible (bad) ones. However, various research in EAs reported that the infeasible individuals may affect and help the problem solving. This paper proposed a new method to use the infeasible individuals by studying the sub-structures rather than the entire individual structures to solve Reinforcement Learning (RL) problems, which generally factorize their entire solutions to the sequences of state-action pairs. This work was studied in a recent graph-based EDA named Probabilistic Model Building Genetic Network Programming (PMBGNP) which can solve RL problems successfully. The effectiveness of this work is verified in a RL problem, i.e., robot control, comparing with some other related work.

Original languageEnglish
Title of host publicationGenetic and Evolutionary Computation Conference, GECCO'11
Pages601-608
Number of pages8
DOIs
Publication statusPublished - 2011
Event13th Annual Genetic and Evolutionary Computation Conference, GECCO'11 - Dublin
Duration: 2011 Jul 122011 Jul 16

Other

Other13th Annual Genetic and Evolutionary Computation Conference, GECCO'11
CityDublin
Period11/7/1211/7/16

Fingerprint

Network Programming
Genetic Network
Reinforcement learning
Reinforcement Learning
Genetic Programming
Probabilistic Model
Factorise
Entire Solution
Robot Control
Substructure
Truncation
Entire
Robots
Graph in graph theory
Estimate
Statistical Models

Keywords

  • EDA
  • Infeasible individuals
  • Probabilistic Model Building Genetic Network Programming
  • Reinforcement learning

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Theoretical Computer Science

Cite this

Li, X., Mabu, S., & Hirasawa, K. (2011). Use of infeasible individuals in Probabilistic Model Building Genetic Network Programming. In Genetic and Evolutionary Computation Conference, GECCO'11 (pp. 601-608) https://doi.org/10.1145/2001576.2001659

Use of infeasible individuals in Probabilistic Model Building Genetic Network Programming. / Li, Xianneng; Mabu, Shingo; Hirasawa, Kotaro.

Genetic and Evolutionary Computation Conference, GECCO'11. 2011. p. 601-608.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Li, X, Mabu, S & Hirasawa, K 2011, Use of infeasible individuals in Probabilistic Model Building Genetic Network Programming. in Genetic and Evolutionary Computation Conference, GECCO'11. pp. 601-608, 13th Annual Genetic and Evolutionary Computation Conference, GECCO'11, Dublin, 11/7/12. https://doi.org/10.1145/2001576.2001659
Li X, Mabu S, Hirasawa K. Use of infeasible individuals in Probabilistic Model Building Genetic Network Programming. In Genetic and Evolutionary Computation Conference, GECCO'11. 2011. p. 601-608 https://doi.org/10.1145/2001576.2001659
Li, Xianneng ; Mabu, Shingo ; Hirasawa, Kotaro. / Use of infeasible individuals in Probabilistic Model Building Genetic Network Programming. Genetic and Evolutionary Computation Conference, GECCO'11. 2011. pp. 601-608
@inproceedings{47c72c965b7c48f98b67c74a97c1b1fa,
title = "Use of infeasible individuals in Probabilistic Model Building Genetic Network Programming",
abstract = "Classical EDAs generally use truncation selection to estimate the distribution of the feasible (good) individuals while ignoring the infeasible (bad) ones. However, various research in EAs reported that the infeasible individuals may affect and help the problem solving. This paper proposed a new method to use the infeasible individuals by studying the sub-structures rather than the entire individual structures to solve Reinforcement Learning (RL) problems, which generally factorize their entire solutions to the sequences of state-action pairs. This work was studied in a recent graph-based EDA named Probabilistic Model Building Genetic Network Programming (PMBGNP) which can solve RL problems successfully. The effectiveness of this work is verified in a RL problem, i.e., robot control, comparing with some other related work.",
keywords = "EDA, Infeasible individuals, Probabilistic Model Building Genetic Network Programming, Reinforcement learning",
author = "Xianneng Li and Shingo Mabu and Kotaro Hirasawa",
year = "2011",
doi = "10.1145/2001576.2001659",
language = "English",
isbn = "9781450305570",
pages = "601--608",
booktitle = "Genetic and Evolutionary Computation Conference, GECCO'11",

}

TY - GEN

T1 - Use of infeasible individuals in Probabilistic Model Building Genetic Network Programming

AU - Li, Xianneng

AU - Mabu, Shingo

AU - Hirasawa, Kotaro

PY - 2011

Y1 - 2011

N2 - Classical EDAs generally use truncation selection to estimate the distribution of the feasible (good) individuals while ignoring the infeasible (bad) ones. However, various research in EAs reported that the infeasible individuals may affect and help the problem solving. This paper proposed a new method to use the infeasible individuals by studying the sub-structures rather than the entire individual structures to solve Reinforcement Learning (RL) problems, which generally factorize their entire solutions to the sequences of state-action pairs. This work was studied in a recent graph-based EDA named Probabilistic Model Building Genetic Network Programming (PMBGNP) which can solve RL problems successfully. The effectiveness of this work is verified in a RL problem, i.e., robot control, comparing with some other related work.

AB - Classical EDAs generally use truncation selection to estimate the distribution of the feasible (good) individuals while ignoring the infeasible (bad) ones. However, various research in EAs reported that the infeasible individuals may affect and help the problem solving. This paper proposed a new method to use the infeasible individuals by studying the sub-structures rather than the entire individual structures to solve Reinforcement Learning (RL) problems, which generally factorize their entire solutions to the sequences of state-action pairs. This work was studied in a recent graph-based EDA named Probabilistic Model Building Genetic Network Programming (PMBGNP) which can solve RL problems successfully. The effectiveness of this work is verified in a RL problem, i.e., robot control, comparing with some other related work.

KW - EDA

KW - Infeasible individuals

KW - Probabilistic Model Building Genetic Network Programming

KW - Reinforcement learning

UR - http://www.scopus.com/inward/record.url?scp=84860415718&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84860415718&partnerID=8YFLogxK

U2 - 10.1145/2001576.2001659

DO - 10.1145/2001576.2001659

M3 - Conference contribution

AN - SCOPUS:84860415718

SN - 9781450305570

SP - 601

EP - 608

BT - Genetic and Evolutionary Computation Conference, GECCO'11

ER -