Increasing genetic programming robustness using simulated dunning-kruger effect

Thomas D. Griffiths, Anikó Ekárt

Research output: Chapter in Book/Published conference outputConference publication

Abstract

Robustness is a key characteristic of genetic programming candidate solutions, providing the ability to maintain a satisfactory level of performance under dynamic and uncertain environments. In this paper we perform experiments on Tartarus problem instances[2] exploring the hypothesis that the introduction of a fitness distribution bias, inspired by the Dunning-Kruger effect [5], will lead to an increase in the diversity and robustness of candidate solutions.

Original languageEnglish
Title of host publicationGECCO 2019 Companion - Proceedings of the 2019 Genetic and Evolutionary Computation Conference Companion
PublisherACM
Pages340-341
Number of pages2
ISBN (Electronic)9781450367486
ISBN (Print)978-1-4503-6748-6
DOIs
Publication statusPublished - 13 Jul 2019
Eventthe Genetic and Evolutionary Computation Conference Companion - Prague, Czech Republic
Duration: 13 Jul 201917 Jul 2019

Conference

Conferencethe Genetic and Evolutionary Computation Conference Companion
Period13/07/1917/07/19

Bibliographical note

© 2019 Copyright held by the owner/author(s). Publication rights licensed to ACM.

Keywords

  • Diversity
  • Dunning-Kruger
  • Robustness
  • Tartarus Problem

Fingerprint

Dive into the research topics of 'Increasing genetic programming robustness using simulated dunning-kruger effect'. Together they form a unique fingerprint.

Cite this