基于meta-种群理论的免疫遗传算法

Immune genetic algorithm based on meta-population

自然meta-种群中局部种群之间存在相对隔离和种群个体冒险迁徙的机制,提出了一种基于meta-种群理论的免疫遗传算法。该算法模拟了自然meta-种群中局部种群克隆、生殖、变异和自然灭绝等自然过程。其主要步骤包括初始种群,种群及其个体适应度计算,选择,克隆变异、交叉生殖、含记忆B细胞个体克隆和局部种群灭绝及其最优个体的冒险迁徙等。该算法的特点是模拟了meta-种群的自然机制,具有并行性,能够产生高适应能力的个体并不断地更新,直到最优个体的出现。对两种问题进行了模拟实验,并与普通遗传算法IMA进行了比较,结果表明所提出的算法能以较少的迭代次数完成最优解的寻找。

There are the isolation correspondingly and the migration of risk between the populations in the meta-population. So this paper proposed a new optimization algorithm based on the meta-population theory. The algorithm simulated the natural process of the meta-population mechanism, such as initial the populations, the fitness of the population and the individual, selecting, cloning and mutating, crossover procreating, cloning of the mnemonic B ceils, the depopulation of the sub-population, and the migration of risk of the best individual. The advantages were that the algorithm simulated the meta-population＇ s natural mechanism, produced and updated the good optimum until the best one appeared, and had the parallelism. The method was utilized to optimize the two problems, the simulation results show that the algorithm can find the better optimum than the common IGA at the same iterative times.