Incremental commitment in genetic algorithms

Successful recombination in the simple GA requires that interdependent genes be close to each other on the genome. Several methods have been proposed to reorder genes on the genome when the given ordering is unfavorable. The Messy GA (MGA) is one such 'moving-locus' scheme. However, gene reordering is only part of the Messy picture. The MGA uses another mechanism that is influential in enabling successful recombination. Specifically, the use of partial specification (or variable length genomes) allows the individuals themselves, rather than the ordering of genes within an individual, to represent which genes 'go together' during recombination. This paper examines this critical feature of the MGA and illustrates the impact that partial specification has on recombination. We formulate an Incremental Commitment GA that uses partially specified representations and recombination inspired by the MGA but separates these features from the moving-locus aspects and many of the other features of the existing algorithm.

[1]  Kalyanmoy Deb,et al.  Analyzing Deception in Trap Functions , 1992, FOGA.

[2]  Tim Jones.  Evolutionary Algorithms, Fitness Landscapes and Search , 1995 .

[3]  David E. Goldberg,et al.  Learning Linkage , 1996, FOGA.

[4]  Kalyanmoy Deb,et al.  Messy Genetic Algorithms: Motivation, Analysis, and First Results , 1989, Complex Syst..

[5]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[6]  David E. Goldberg,et al.  Genetic Algorithms in Search Optimization and Machine Learning , 1988 .

[7]  Melanie Mitchell,et al.  Relative Building-Block Fitness and the Building Block Hypothesis , 1992, FOGA.

[8]  Jordan B. Pollack,et al.  Modeling Building-Block Interdependency , 1998, PPSN.

[9]  K Vekaria,et al.  Selective crossover in genetic algorithms , 1998 .

[10]  Hillol Kargupta.  Gene expression: The missing link in evolutionary computation , 1997 .

[11]  Lee Altenberg,et al.  The Schema Theorem and Price's Theorem , 1994, FOGA.

[12]  Kalyanmoy Deb,et al.  RapidAccurate Optimization of Difficult Problems Using Fast Messy Genetic Algorithms , 1993, ICGA.

[13]  Kanta Premji Vekaria,et al.  Selective Crossover in Genetic Algorithms: An Empirical Study , 1998, PPSN.

[14]  G. Harik.  Learning gene linkage to efficiently solve problems of bounded difficulty using genetic algorithms , 1997 .

[15]  Jordan B. Pollack,et al.  How Symbiosis Can Guide Evolution , 1999, ECAL.