提出了基于种群协同进化的并行免疫克隆算法,将种群中个体的亲和度计算并行在多个计算节点上同时进行。引入免疫记忆机制,使抗体种群的演化过程和记忆单元的演化过程并行进行,更好地实现了抗体间的相互协作,保证了解集从可行域内部和不可行域边缘向着最优解逼近。采用了克隆增殖变异和交叉算子的操作,增加了种群中优秀个体获得克隆增殖实现亲和度成熟的机会,提高抗体群分布的多样性,在深度搜索和广度寻优之间取得了平衡。从而保证了算法较强的收敛性以及搜索空间的多样性。利用标准问题库对算法进行测试,并分析算法参数对算法结果的影响,仿真结果表明,该算法对待寻优空间的全局搜索能力和局部搜索能力以及算法的稳定性与计算速率都要强于简单免疫克隆算法和遗传算法等优化算法。
Parallel immune clone algorithm is proposed based on population coevolution theory and parallel computing affinity of individual at multiple compute nodes.Introducing the immune memory mechanism,the evolution processes of antibody population and memory units are conducted simultaneously,meanwhile,it improves mutual cooperation among antibodies,and ensures solution set approaching optimal solution from the inside of feasible region or infeasible region border.Clone proliferation,high frequency variation and operation of crossover operators increase the chance that better individuals gain affinity maturation by the operation of clone expansion,improve diversity of antibody population distribution,achieve the balance of optimization between depth and range,and ensure the convergence of the algorithm and the diversity of the search range.A computational study for a standard data set is carried out to test the validity of the algorithm,and the effect of algorithm parameters on the results is analyzed.The simulation results show that the global search capability,local search capability,algorithm stability and computing speed of the algorithm are all superior to conventional optimization algorithms such as normal immune clone optimization algorithm,genetic algorithm,etc.