多属性决策的支持向量机方法  

The Support Vector Machine Method for Multiple Attribute Decision Making

作　　者：王强[1] 沈永平[1] 陈英武[1] 

机构地区：[1]国防科技大学信息系统与管理学院,湖南长沙410073

出　　处：《系统工程理论与实践》2006年第6期54-58,123,共6页Systems Engineering-Theory & Practice

摘　　要：为获取和表达决策者的偏好,实现对决策方案的选择,提出了基于支持向量机的多属性决策方法.首先,分析了多属性决策支持向量机方法的原理.其次,提出了基于属性效用函数估计的学习样本构造方法,从决策问题本身抽取学习样本.最后给出了一个算例.In order to capture and represent the decision maker＇s preferences and then to select the most desirable alternative, a method for solving multiple attribute decision making （MADM） problems is proposed based on support vector machine （SVM）. Firstly, the principle of MADM based on SVM is discussed. Secondly, to extract learning samples from the MADM problem, an approach to estimate the utility functions for attributes is presented. Finally, an example is used to illustrate the proposed method.

关 键 词：多属性决策 支持向量机 效用函数 学习样本 

分 类 号：C934[经济管理—管理学]