Measures for Performance Evaluation of Genetic Algorithms (Extended Abstract)

) Kazuo Sugihara Dept. of ICS, Univ. of Hawaii at Manoa 1 Introduction In recent years, genetic algorithms (GAs) [1] have been widely recognized as an effective solving technique for complex problems in the real world. GAs can be regarded as a paradigm of algorithms in the sense that the GAs are parameterized and applicable to a variety of problems by instantiating the paradigm. There are three major components to be designed in GAs. The first one is the coding which is a mapping scheme from a problem to the GA paradigm and represents potential solutions. The second one is a fitness function which quantifies quality of solutions and enables us to differentiate "good" solutions from "bad" solutions. The third one is a set of parameters including population size, population structure, a sequence of genetic operators, the operators' parameters, termination condition, etc. As GAs are getting a larger spectrum of applications, it becomes more crucial to develop a methodology for the design..