Convergence properties of genetic algorithms

Genetic algorithms are regarded as robust and efficient search and optimization techniques which can be applied to many applications including control. Several genetic algorithm parameters may affect convergence properties. To guarantee that these algorithms converge to the global optimum, these algorithm parameters should be analyzed and selected properly. In this paper, we analyzed the effects of selection operator, crossover operator and mutation operator on the convergence properties of genetic algorithms using the hamming distance and the schema theorem. Rules of thumb are provided to improve algorithm convergence performance.