带子群自组织蠕虫算法及其在多模态问题中的应用

Subgroup-self-organizing Worm Algorithm and Application in Multi-modal Function Optimization

提出了一种全新的多模态遗传算法——带子群的自组织蠕虫算法(SSOMA)。该算法的基本思想是:通过初始群体的前期寻优找到峰值点的所在邻域;随后分别在这些邻域内选择少量的个体组成子群,在这些子群中再利用自组织蠕虫算法进行后期寻优,从而找到所有的峰值点,该算法极大地降低了计算的复杂度、提高了收敛速度。最后,用经典测试函数对该算法进行了仿真实验,并进行了计算复杂度分析,结果表明该算法在多模态函数优化方面具有较为理想的应用前景。

In this paper, subgroup-self-organizing worm algorithm （SSOMA） is presented based on the emergence method in complexity research and the classical searching methods, The main idea of this algorithm can be described as follow： search the neighboring regions of peak points through the prophase optimization; select a small quantity of units to build up subgroup in every region, and process the anaphase optimization in these subgroups, by which the peak points will be found in these subgroups. The computation complexity can be lowered obviously and the convergence rate can also be improved efficiently by this method. At last, experiments are given to solve several typical multi-modal function optimization problems. The analysis results on the precision and computation complexity show that SSOMA is perfect for the multi-modal optimization,