知识进化策略.pdfVIP

第 29 卷第6 期 系统工程与电子技术 Vol. 29 No.6 2007 年 6 月 Systems Engineering and Electronics Jun. 2007 :lt置信编号 :1001唰506X(2007)06-1017-04 知识进化策略 刘纯青1 ，杨莘元1 ，张颖2 (1.哈尔滨工程大学信息与通信学院，黑龙江哈尔旗 150001 ， 2. 哈尔演理工大学经济管理学院，黑龙江哈尔演 150080) 摘 要:根据卡尔·波普尔的知识进化论原缀，建立了知识进化策略的.1.本糕桨，并给出了一种用于求解无 约束非线性优化问题的具体实现步碟。知识进化策略的核心思想，就是侬说集与知识集的协闯进化，二者之间通 过猜测与反驳法联系起来，其进化给采最终逼近真缀，即待求解问题的最优解.对 12 个级典测试函数进行性能 仿真实验，结果表明该算法收敛速度有很大提高，并且在一定程度上抑制了-f熟现象. 关键词:进化计算p 知识进化策略自知识进化论p 侬说集p 知识集p 猜测与反驳法 中困分类号: TP18 文献标识码:A Knowledge evolution strategies LlU Chun-qing1 , Y ANG Shen-yuan1 , ZHANG YingZ (1. Schoolol Inlorrrwtíon and Communícation Engineering , Harbin Engineeríng Univ. , Haγbin 150001 • Chi时5 2. School 01 Economic Manag町nent. Harbin Univ. 01 Science and Technology , Harbin 150080. China) Abstract: Based on evolutionary epistemology proposed by Sir Karl Raimund Popper , a computational framework of Knowledge Evolution Strategies (K缸S) is proposed at first. The key idea behind Knowledge Evo幡 lution Strategies is that both of hypotheses set (H) and knowledge set (K) , which are connected with conjecture and refutation method (CRM) , evolve into the truth cooperatively , i. e. the best solutions of the problem to be solved. Then a specific implementation of Knowledge Evolution Strategies for nonlínear unconstraint optimiza帽 tion problems is produced. And 12 nonlinear multidimensional unconstraint function optimization problems are investigated. The computational experiment shows that Knowledge Evolution Strategies can produce substantial performance improvements as expressed in terms of high speed for convergence and a