Complex Systems

Locus-Shift Operator for Function Optimization in Genetic Algorithms Download PDF

Hiroshi Inazawa
Electronic mail addresses: genzoh@shoin.ac.jp, ihiroshi@cs.ucsd.edu
Center for Education in Information Systems,
Kobe Shoin Women's University,
1-2-1 Shinohara Obanoyama, Nada, Kobe 657-0015, Japan

Kazuhisa Kitakaze
Department of Information Systems,
Tokyo University of Information Sciences,
1200-2 Yato Wakaba, Chiba 265-8501, Japan

Abstract

Function optimization is the most important context for studying genetic algorithm (GA) operators. In this paper a new GA operator is introduced which greatly improves several well-known benchmark functions used in function optimization. The new operator cuts a circular chromosome at any locus selected randomly. By this operation, various types of linear chromosomes can be formed from a parent circular chromosome. The new operator is called locus-shift (LS) because the locus of the linear chromosome produced by LS almost always shifts. In this paper, we study the dynamics of the evolution of chromosomes by the LS in simulation and show the effects of LS by various benchmark functions.