Theoretical aspects of genetic algorithms in dynamic environments

Abstract

In this project, Genetic Algorithms (GAs) in dynamic optimization problems, which are characterized by changes in the surface of fitness occurring during the execution of the optimization algorithm, will be theoretically investigated. The area of evolutionary algorithms in dynamic environment has received great attention from researchers in recent years. Unfortunately, few studies have studied GAs on dynamic optimization problems from a theoretical point of view. The main objective of the work to be done at the post-doctoral stage is the theoretical analysis of fitness landscapes, mainly elementary fitness landscapes, in GAs in dynamic environments. One of the researchers with important works in the area of theory of GAs, with several articles on elementary fitness landscapes, is Prof. Darrell Whitley, who is the supervisor of this post-doctoral stage at Colorado State University. The technique of elementary fitness landscapes has not yet been applied to dynamic optimization problems, despite the great potential of this area. The researcher should investigate the applicability of theoretical analysis tools based on elementary fitness landscapes in dynamic optimization problems of the real world and those produced by benchmark problems generators. New problems generators should be developed based on the analysis done. (AU)