PERFORMANCE OF HYBRID GENETIC ALGORITHM FOR THE GREY PATTERN PROBLEM

Abstract

Recently, genetic algorithms (GAs) are quite popular by solving combinatorial optimization problems. In this paper, we discuss a hybrid genetic algorithm that uses a new kind of solution recombination operators − a so-called multiple parent crossover. We examined this innovative crossover operator on the grey pattern problem, which is as special case of the well-known problem, the quadratic assignment problem. The results obtained during the experimentation with the set of 62 instances of the grey pattern problem demonstrate promising efficiency of the multiple parent crossover. All the instances tested were solved to pseudo-optimality within surprisingly small computation times.