Effect of crossover operators under multirecombination: weighted tardiness, a test case

Abstract

Phone/Fax : +54 0297 4854888Abstron: I n evolutionary algorithms based on genetics, the crossover operation creates individuals by exchange of genes. Seleclion mechanisms propitiate reproduction of better individuals imposing P direction on the search process: copier of better individuals replace worst ones. Consequently, part of the genetic material contained in these worst individuals vanisher forever. This loss of divemi@ can lead to B premature convergence. To prevent an early convergence to a local optimum under the same selection mechanism then, either a larger population sue or adequate genetic operators m e needed. Multirecombination daws multiple crossover operations on two or more parents each time a new individual is created.I n this work we show the influence on genetic diversity, quality of ICEUIS and required computational effort, when applying different crossover methods to a set of hard instances, selected as a test case, ofthe weighted tardiness scheduling problem in Single machine environments under multire-combined approaches. A description of the multirecombination variant used, experiments and preliminary results me reported. 0-7803-85 15-2/04/$20.00 0 2 0 0 4 IEEE