自组织迁移算法(SOMA)是一种新型的群体智能算法。在对原始自组织迁移算法分析的基础上,针对基于随机变异步长的自组织迁移算法存在的不足,提出了线性递减步长策略,即有针对性地以线性方式动态调整步长,以满足群体迭代在不同阶段的需求,从而加速群体在多峰复杂空间中收敛速度的同时提高算法的局部搜索能力。实验结果表明,该算法优于原始自组织迁移算法和基于随机变异步长的自组织迁移算法。
Self-Organizing Migrating Algorithm(SOMA) is a kind of new swarm intelligent algorithm.After analyzing the basic self-organizing migrating algorithm and modified self-organizing migrating algorithm with random mutation step,this paper proposes a improved self-organizing migrating algorithm with linear-digress step.The algorithm accelerates the convergence ratio of swarm in multi-modal complex space and enhances the ability of local search by adjusting step length dynamic and linear under some directions.Experiments reveal that the proposed algorithm behaves better than the self-organizing migrating algorithm with random mutation step and the basic one.