Topological Aspects in Genetic Algorithms

We investigate topological aspects in Genetic Algorithms (GAs). Two dimensional combinatorial optimization problems, cell placement problems, are concerned. We construct topological recombination and uniform recombination based on exchange operations, and compare their performance. Simulation results show that the contributions of these recombinations are qualitatively different. 1 Introduction Though GAs are used in wide variety of optimization problems, they contains several factors to be tuned according to each problem. Among those factors, design of coding and operators has strong effects on performance, since it makes GA different from local nor random searches [1, 2]. In this paper we examine topological aspects of GAs. Under this category, one natural question arises: how are three dimensional structures of proteins are coded effectively on strings of bases in natural genetics? Another problem is more pragmatic: how can we use intrinsic topology of problems in optimization ta..