【作者】 李少波； 胡建军； 陈茜；

【Author】 Li Shaobo 1,Hu Jianjun 2,Chen Xi 1 1.Key Laboratory of Advanced Manufacturing technology,Guizhou University,Guiyang 550003,P.R.China 2.Department of Computer Science and Engineering University of South Carolina,Columbia SC 29208

【机构】 贵州大学省部共建教育部现代制造技术重点实验室； 美国南卡罗莱纳州大学计算机科学与工程学院；

【摘要】 本文对HFC(Hierarchical Fair Competition)模型的搜索方式进行了相应的改进,引入基于遗传编程的结构参数混合优化方法。同时在HFC模型中加入了一些自适应机制,形成了三种衍生的HFC模型。第一种是静态的HFC模型(SHFC),它在原始HFC模型的基础上添加了遗传编程的开放式拓扑搜索能力;第二种是自适应输入阀值的HFC模型(HFC-ADM),它能够自适应的调整HFC模型中各层的输入阀值;第三种是自适应迁移的HFC模型(HFC-ATP),它允许HFC模型中各层间个体的动态移动。三种衍生的HFC模型在理论上保证了搜索无限的空间,实现了可持续的进化。论文最后结合旅行商问题实现了三种基于遗传编程的HFC模型(SHFC、HFC-ADM、HFC-ATP),并通过与标准遗传算法的对比实验,验证了三种基于遗传编程的算法模型不仅实现了可持续的进化,而且有效降低了算法的随机性。更多还原

【Abstract】 This paper improved the search manner of HFC(Hierarchical Fair Competition) model,and introduced an optimization method which was mixed with structure parameters based on genetic programming(GP) .Meanwhile,some adaptive mechanisms were added to HFC model,and then generated three derived HFC models:static generational HFC model(SHFC) which added an opened topology search to the initial HFC model;the generational HFC model with admission threshold adaptation mechanism(HFC-ADM) which set the admission thresholds of all ranks adaptively,and the generational HFC model with adaptive migration topology(HFC-ATP) which permits the individuals moved to different ranks dynamically.At last,we take the traveling salesman problem as benchmark;use the three models SHFC,HFC-ADM and HFC-ATP based on GP to solute this famous problem.We also use GA in this benchmark problem as a comparison.The result indicated these three improved HFC models based on GP achieved sustainable evolution,and reduced the randomicity of algorithms efficiently.更多还原