Self-Adaptive Genotype-Phenotype Maps: Neural Networks as a Meta-Representation

Luís F. Simões, Dario Izzo, Evert Haasdijk, Agoston Endre Eiben

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

In this work we investigate the usage of feedforward neural networks for defining the genotype-phenotype maps of arbitrary continuous optimization problems. A study is carried out over the neural network parameters space, aimed at understanding their impact on the locality and redundancy of representations thus defined. Driving such an approach is the goal of placing problems' genetic representations under automated adaptation. We therefore conclude with a proof-of-concept, showing genotype-phenotype maps being successfully self-adapted, concurrently with the evolution of solutions for hard real-world problems.
Original languageEnglish
Pages (from-to)110-119
Number of pages10
JournalLecture Notes in Computer Science
Volume8672
DOIs
Publication statusPublished - 2014
EventParallel Problem Solving from Nature (PPSN) -
Duration: 13 Sep 201417 Sep 2014

Bibliographical note

Winner of the Best Paper Award at PPSN 2014.
Proceedings title: Parallel Problem Solving from Nature – PPSN XIII
Publisher: Springer
ISBN: 978-3-319-10761-5
Editors: T. Bartz-Beielstein, J. Branke, B. Filipič, J. Smith

Keywords

  • Adaptive representations
  • Genotype-Phenotype map
  • Neuroevolution
  • Redundant representations
  • Self-adaptation

Fingerprint Dive into the research topics of 'Self-Adaptive Genotype-Phenotype Maps: Neural Networks as a Meta-Representation'. Together they form a unique fingerprint.

  • Cite this