SWO:基于小世界效应的快速搜索算法

SWO:A Fast Search Algorithm Based on Small World Effect

借鉴小世界网络理论中层次树模型和多分类标准建模的理论,设计了一种基于小世界效应的快速搜索算法SWO。采用掩码规则将解空间构造为层次树网络模型,并提出采用相映射的空间与原解空间共同组成双分层标准的建模理论。SWO算法通过对两种空间网络中长短邻居关系的查询访问,将实验信封推送到目的地,从而找到解空间中的最优值。实验证明,引入映射空间搜索机制可避免算法陷入局部最优,而长邻居关系的访问又加快了算法的收敛速度。通过与遗传算法(GA)、粒子群优化算法(PSO)和差分算法(DE)的对比,SWO算法表现出较强的搜索能力和较高的搜索效率。

This paper designed a fast search algorithm called Small World Optimization（SWO） which was inspired by the hierarchical categorization tree model and multi categories method based on small world theory.The solution space can be divided into the hierarchical categorization tree model using mask rule in binary coding.Two bijective mapping solution space were adopted to establish multi categories method.SWO can find the optimal solution in the designed small world network by short and long distance neighbor relationship as pushing the mail to target.SWO was tested via a benchmark test functions in a simulation and the corresponding results show that two bijective mapping space can avoid the algorithm falling into early maturity and the long distance neighbor relationship can accelerate the convergence rate.Compared with the most popular optimization algorithm as genetic algorithm（GA）,Particle Swarm Optimization（PSO） and Difference Algorithm（DE）,the SWO algorithm is endowed with faster convergence ability to solve complex optimization problems.