New fitness-based migration operator for evolutionary programming

  • Authors:
  • Doo-Hyun Choi

  • Affiliations:
  • School of Electrical Engineering and Computer Science, Seoul National University, Seoul, Korea

  • Venue:
  • Neural, Parallel & Scientific Computations
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

A new fitness-based individual migration operator after the model of a migration phenomenon in a natural ecosystem is presented. The relative fitness variations are evaluated for each individual and those are used to determine whether an individual migrate to a new solution candidate which is far from its original position. This migration guarantees the search diversity by the uniformly dispersed individuals on a search space at the initial phase of evolution and fast convergence by the migrated individuals as the evolution progresses. The usefulness of the migration operator has been tested using evolutionary programming (EP) which adopts the operator. The performance of the proposed EP has been compared with those of other well-known EP algorithms through benchmark problems on continuous function optimization.