考虑风险偏好和属性约简的熵可靠性决策模型

摘要

针对属性值为区间数的大量属性决策问题，提出考虑决策者风险偏好和属性约简的熵可靠性决策模型。首先引入风险偏好因子将决策者进行分类，然后提出基于熵可靠性的属性约简方法并提取有效属性，匹配相应的信息熵权重确定方法和排序方法进行方案排序和择优，通过案例分析验证该方法的科学合理性；并通过选取不同的风险偏好因子，对决策对象排序结果进行灵敏度分析。结果表明随着决策者风险偏好程度的增加，决策属性保留个数递减；最后经算法对比表明该算法考虑决策信息的可靠性，减少信息损失且计算简便。%For a large number of attribute decision making problems of attribute values as interval number,a new multiple attribute decision making model is proposed based on entropy reliability of attribute reduction method,considering the risk preferences.First,the risk preference parameter is used to classify the decision makers.Then,the paper puts forward the attribute reduction method based on entropy reliability and extracts the effective properties,matches the corresponding infor-mation entropy weight method,ranking method for alternative ranking,and selects the best.Real examples with interval value attribute values are presented to prove the scientific rationality of the new method.And then through choosing differ-ent risk preference parameters,the sensitivity analysis is carried out on the ranking results of decision -making objects. The results show that the method proposed is more applicable in the actual situation.Finally by comparison with algorithm, the paper concludes that because the algorithm takes into account the reliability of decision -making information,it reduces the information loss and it is simple and convenient.