一种模糊偏好排序的多目标粒子群算法  

Fuzzy preference ranking of multi-objective particle swarm optimization algorithm

作　　者：李世威 王建强[1] 曾俊伟[1] 

机构地区：[1]兰州交通大学交通运输学院,兰州730070

出　　处：《计算机应用研究》2011年第2期477-480,共4页Application Research of Computers

基　　金：国家社会科学基金资助项目(08XTQ010);甘肃省自然科学基金资助项目(096RJZA088)

摘　　要：为了将决策者对各目标属性的模糊评判信息转换为目标偏好信息,首先将模糊语义转换为三角模糊数,利用模糊数的广义加法、近似乘法和标量乘法进行计算,从而将决策者对目标属性的离散意见转换为对各目标的综合意见;然后采用多指标模糊排序法确定决策者权重,通过定义一种模糊综合排序指标来确定各目标偏好权重,依据目标权重构建判断多目标Pareto解的适应度函数,并采用粒子群算法对多目标问题进行求解;最后通过一个算例来说明该算法的实用性和有效性。In order to transfer fuzzy evaluation information of objective＇s attributes into objective preference information,this paper transferred fuzzy semantics into triangular fuzzy number,and used generalized additive function,an approximate multiplication and a scalar multiplication between fuzzy numbers to compute.Then transferred discrete levels of objective＇s attributes to integrated levels,and used fuzzy sorting multi-indexes to define weights of method makers.After that,defined a fuzzy integrated indexes to determine each objective sorting weight,defined a fitness function of multi-objective Pareto solutions based on these objective weights,and used particle swarm algorithm to solve the multi-objective problems.Finally,used a case to illustrate the algorithm＇s feasibility.

关 键 词：模糊效用 模糊评价 多属性 多目标 粒子群 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]