不完全信息的离散型模糊随机多准则决策方法  

Discrete fuzzy-stochastic multi-criterion decision-making method with incomplete information

作　　者：任剑[1,2] 高阳[1] 

机构地区：[1]中南大学商学院,长沙410083 [2]湖南农业大学科学技术师范学院,长沙410128

出　　处：《系统工程理论与实践》2011年第1期122-130,共9页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(70771115);湖南省自然科学基金(10JJ6107);湖南省哲学社会科学基金(09YBA072);湖南省教育厅科学研究项目(10C0792);湖南农业大学青年科学基金(08QN61)

摘　　要：针对状态概率与准则权重均为区间数且准则值有缺失的离散型模糊随机多准则决策问题,设计了一种基于信息集结的新方法.该方法在区分准则类型和确定决策者的风险偏好水平后,首先根据准则类型利用最大、最小算子计算方案集在不同状态下有缺失值的准则上的正、负理想值,再通过决策者的风险偏好水平与正、负理想值补充方案集在不同状态下缺失的准则值,然后参照准则类型与决策者的风险偏好水平集结方案集在不同状态下赋值为梯形模糊数的准则值并进行规范化,接着利用熵与离差最大化分别构建规划模型以求解最优状态概率分布与最优准则权重向量,进而得到方案集的综合评价值并确定排序,最后给出具体算例.结果显示了该方法的有效性和可行性.To solve the discrete fuzzy-stochastic multi-criterion decision-making problems with states＇ probabilities and criterions＇ weights both evaluated interval numbers and some of alternatives＇ evaluations on criterions missing,a new method based on the information integration was designed.After distinguishing the types of criterions and determining the risk preference level of decision-makers,the following were these tasks.Firstly,according to the types of criterions,the positive ideal values and negative ideal values of alternatives＇ evaluations on criterions in different states were worked out by the maximum operator and minimum operator.Secondly,via the risk preference level of decision-makers,positive ideal values and negative ideal values,the alternatives＇ missing evaluations on criterions in different states were filled completely.Thirdly,in the light of the types of criterions and risk preference level of decision-makers,the alternatives＇ evaluations in form of trapezoidal fuzzy numbers on criterions in different states were integrated and normalized.Fourthly,with the help of maximizing the entropy and deviation,the programming models were constructed separately to find the optimal state probability distribution and weight vector of criterions.Fifthly,the integrated evaluations of alternatives were worked out.And then,the rank of alternatives was obtained.Finally,a numerical example was given.The result shows the validity and feasibility of the method.

关 键 词：离散型模糊随机多准则决策 不完全信息 信息集结 风险偏好水平 理想值 熵 离差 

分 类 号：C934[经济管理—管理学]