The Research Of Multi-parent Genetic Algorithm And Its Application

Multi-parent Genetic Algorithm (MGA) is a heuristic search algorithm. MGA has not been developed for a long time, but it has been preliminarily applied in many function optimization problem and many combinatorical optimization problem including Nondeterministic Polynomial completeness. MGA is a special kind of traditional Genetic Algorithm (TGA), but if differs from TGA in a point that in MGA, the crossover operation is among many individuals(more than two) while in TGA it is among one or two, that is to say crossover operation is improved. At present, MGA absorbs more and more attentions from experts and scholars at home and abroad, moreover, they have valuable achievements.In this paper, first, we analyze the research works that experts and scholars are going along within the aspect of Multi-parent Genetic Algorithm. Second, in order to improve the convergent speed, we put forward the Multi-parent Genetic algorithm based on shared memory. In this algorithm, shared-memory plays two important roles: first it can preserve the excellent individuals gained from the algorithm implementing process; Second, it provides excellent individuals to crossover operation, and in some case it can improve the performance of the Multi-parent Genetic Algorithm. Meanwhile, we have studied and discussed the parallelization of the multi-parent Genetic Algorithmbased on shared-memory. Finally, we apply the multi-parent Genetic Algorithm to multi-objective optimization problem, and it is shown through experiments that the algorithm has excellent qualities.