周期变异概率的免疫克隆算法

在传统免疫克隆算法的基础上提出了一种新的基于周期变异概率的免疫克隆算法,该算法进一步提高了收敛速度,有效地克服了早熟现象,很好地解决了类似高维函数优化等复杂问题.通过对比计算实验表明:种群的初始分布对该算法的性能影响很小,且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自适应遗传算法等优化算法.

An immune clonal algorithm based on the probability of cyclic mutation

Based on the traditional immune clonal algorithm, we suggest a new immune clonal algorithm with cycle mutation. This algorithm can enhance the convergence speed, effectively overcome the premature convergence, and well resolve similar complex problems of multidimensional function optimization. Its performances exhibit little effect from the initial distribution of population. In comparison with the simple clone algorithm and the self-adapting genetic algorithm, this algorithm shows a higher speed in computation and a superior ability and stability in the global and local search for optimal solutions in space.