The 1/5-th rule with rollbacks: on self-adjustment of the population size in the (1 + (λ, λ)) GA

A. O. Bassin, Maxim Buzdalov

Research output: Chapter in Book/Report/Conference proceedingConference Proceeding (Non-Journal item)

Abstract

Self-adjustment of parameters can significantly improve the performance of evolutionary algorithms. A notable example is the (1 + (λ, λ)) genetic algorithm, where the adaptation of the population size helps to achieve the linear runtime on the OneMax problem. However, on problems which interact badly with the self-adjustment procedure, its usage can lead to performance degradation compared to static parameter choices. In particular, the one fifth rule is able to raise the population size too fast on problems which are too far away from the perfect fitness-distance correlation.

We propose a modification of the one fifth rule in order to have less negative impact in scenarios when the original rule reduces the performance. Our modification, while still having a good performance on OneMax, both theoretically and in practice, also shows better results on linear functions with random weights and on random satisfiable MAX-SAT instances.
Original languageEnglish
Title of host publicationGECCO '19
Subtitle of host publicationProceedings of the Genetic and Evolutionary Computation Conference Companion
EditorsManuel López-Ibáñez
PublisherAssociation for Computing Machinery
Pages277-278
Number of pages2
ISBN (Print)978-1-4503-6748-6, 1450367488
DOIs
Publication statusPublished - 13 Jul 2019
Externally publishedYes
EventGECCO 2019: The Genetic and Evolutionary Computation Conference - Prague, Czech Republic
Duration: 13 Jul 201917 Jul 2019
https://gecco-2019.sigevo.org

Conference

ConferenceGECCO 2019: The Genetic and Evolutionary Computation Conference
Country/TerritoryCzech Republic
CityPrague
Period13 Jul 201917 Jul 2019
Internet address

Keywords

  • (1 + (λ,λ)) ga
  • linear functions
  • max-sat
  • parameter adaptation

Fingerprint

Dive into the research topics of 'The 1/5-th rule with rollbacks: on self-adjustment of the population size in the (1 + (λ, λ)) GA'. Together they form a unique fingerprint.

Cite this