针对传统遗传算法解决大规模复杂问题效率比较低的问题,提出了一种并行隔代映射遗传算法。该算法采用的是多种群并行进化的方法,在各种群之间引入竞争,既能较好地丰富和保持种群的多样性,有效地避免早熟收敛,又能大大提高求解大规模复杂问题的效率。将该算法应用于冷加工金属板和正交各向异性复合板材料参数的反演问题中,将计算结果与并行化前的算法进行比较,验证了该并行算法具有高效率解决大规模复杂问题的能力。
A multiple-deme parallel genetic algorithm based on the intergeneration projection genetic algorithm was suggested to solve the complex optimal problems. In this method, competitive migrations were applied among the multiple demes to keep the diversity of each deme. The applications to determine material parameters of cold-working metal plates and anisotropic plates demonstrate the efficiency and better performance of the suggested method, according to the comparison of results to that of the intergeneration projection genetic algorithm.