一种正交量子进化算法  

An orthogonal quantum-inspired evolutionary algorithm

作　　者：钱洁[1,2] 郑建国[1] 

机构地区：[1]东华大学智能系统研究中心,上海200051 [2]湖北汽车工业学院信息系,湖北十堰442002

出　　处：《华中科技大学学报（自然科学版）》2012年第11期106-110,共5页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(70971020);湖北省教育厅资助项目(2007q040;B20092306;2012Y051;Q20122302)

摘　　要：提出一种基于正交试验设计的量子进化算法.通过正交试验设计方法找到最有效的优秀个体知识组合,利用基于正交试验因素分析的约束修补算子对知识组合进行修补,并将获得的知识组合作为量子群体的吸引子.同时,采用正交试验结果动态调节量子更新转角大小,使得量子更新更简洁且具有自适应学习能力.Rastrigin和Ackley函数优化和0-1背包组合优化实验表明:该算法缩短了量子寻找到最佳知识组合的时间,降低了计算成本,提高了算法精度.A novel quantum-inspired evolutionary algorithm based on orthogonal experimental design method was proposed. By using orthogonal experiment on different excellent individuals, the algo- rithm can get a better knowledge combination. A new repair operator based on orthogonal experimen- tal factor analysis was proposed for the knowledge combination which does not satisfy the constraint conditions. The result of the knowledge combination was used as a new attractor for quantum individ- uals. In addition, the quantum individuals were updated by new rotation gates base on the result of the orthogonal experiment. The proposed updating of quantum can simply self-adaptively adjust length during evolution. This algorithm was tested with Rastrigin, Ackley benchmark functions and 0-1 knapsack problem. The results show that the algorithm shortens the time to find the best knowl- edge combination and improves the accuracy of the algorithm.

关 键 词：量子进化算法 正交试验设计 因素分析 函数优化 组合优化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]