基于风险加权的同构化区间型多属性决策新算法

发布时间：2018-06-05 19:00

本文选题：区间型决策 + 同构化　；参考：《计算机科学》2009年01期

【摘要】：多属性决策问题的复杂性、决策因素影响的不确定和传统评判方法的局限性,使不确定决策因素的属性测度常常难以精确量化,往往只能用区间数进行大致估量。为了精确量化表征属性决策因素测度值不确定性,根据同构化基本原理与相似性科学相关理论及相关思想,针对区间型多属性决策问题提出了一种基于同构化多属性决策新方法的新算法。该新算法的主要特点是:1)提出了决策者风险偏好权重;2)采用了同构化风险测度三元组(拟下限相似度,风险程度,风险偏好值),来精确量化决策过程中存在的风险程度以及决策者对此风险程度的偏好;3)生成了可描述各属性与决策目标关系的标杆方案;4)定义了方案相似度新概念;5)构造了风险加权相似度量算子(RWSMO),来度量各决策方案与标杆方案之间风险加权相似度的大小;6)挑选出风险加权相似度最大的方案作为最优或满意方案。
[Abstract]:The complexity of multi-attribute decision making problem, the uncertainty of decision factors and the limitation of traditional evaluation methods make the attribute measure of uncertain decision factors difficult to quantify accurately and can only be estimated roughly by interval number. In order to accurately quantify the uncertainty of measurement value of attribute decision factors, according to the basic principle of isomorphism and the theory of similarity science, A new algorithm based on isomorphism for interval multi-attribute decision making is proposed. The main feature of the new algorithm is that the weight of decision maker's risk preference is proposed by 1: 1) and isomorphic triplet of risk measure (quasi lower bound similarity, risk degree) is used. Risk preference value, to accurately quantify the degree of risk that exists in the decision-making process and the decision maker's preference for this risk degree 3) A benchmarking scheme that can describe the relationship between each attribute and the decision objective is generated. (4) A new probability of scheme similarity is defined. The RWSMOO operator is constructed to measure the size of the risk-weighted similarity between each decision scheme and the benchmarking scheme. The scheme with the largest risk weighted similarity is selected as the optimal or satisfactory scheme.
【作者单位】：西南财经大学信息技术应用研究所;西南财经大学经济信息工程学院;西南财经大学中国支付体系研究中心;
【分类号】：C934

【参考文献】