A Parallel Immune Genetic Algorithm for Multimodal Function Optimization Problem

For the simple genetic algorithm is difficult to find all the optimal solutions in solving multimodal function optimization problem, a parallel immune genetic algorithm was proposed in this paper. Multi-parent crossover operator of GuoTao algorithm and niche technique was introduced to improve the algorithm. The optimization procedure of the algorithm was divided into two stages. In the early stage of the algorithm, antibody concentration depression of immune algorithm was referenced to increase the population diversity. The concentration of superior similar individual in a large population was reduced by mutation operator to enlarge the search space, thereby determining every modal region. The population was divided into a number of sub-populations in the late algorithm. And immune memory bank was introduced into the sub-populations to record the current optimization of each sub-population and instructionally help sub-population to converge to each modal quickly. Algorithm convergence was ensured by immune memory keeping the elite individual of generations. The experimental results show that the proposed algorithm takes good performance on the multi-modal function optimization problem.