灰色多目标光学寻优算法  

Gray optics multi-objective optimization

作　　者：王侃[1] 沈继红[2] 李焱[1] 胡波[1] 

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001 [2]哈尔滨工程大学理学院,黑龙江哈尔滨150001

出　　处：《华中科技大学学报（自然科学版）》2013年第1期56-61,共6页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：黑龙江省自然科学基金资助项目(F200931)

摘　　要：结合光学原理以及灰色系统思想提出了一种求解多目标问题的新方法.此方法利用自然界中最基本的光学现象设计迭代机制,同时与应用灰色系统原理维护外部档案的机制有机地结合在一起形成多目标优化方法.在数值实验中设计多种组合实验,分别从算法性能、Pareto最优集效果、收敛性多样性指标方面与经典算法进行比较.结果表明:灰色理论的引入大大加强了非劣解的均匀性,同时光学的迭代原理有效地增强了收敛性,为快速解决高维多目标优化问题提供新思路.By optics principle combined with gray system theory,a new intelligent multi-objective optimization was proposed.In this method,the iterative mechanisms was designed by using the most basic phenomenon of optics.At the same time,the new gray system theory principle of external files can effectively solve multi-objective problems.In the numerical experiments,many kinds of experiments were designed to compare the directions of optimization performance Pareto solution sets and convergence diversity indexes with the classical algorithm.The results show that the new optimization has better effects and can provide a new idea for solving the high dimension multi-objective problems.

关 键 词：多目标优化 费马原理 PARETO最优解 外部档案 灰色系统 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]