基于相似关系的多属性决策问题研究

属性权重确定是多属性决策问题中一个重要的研究内容。针对属性权重未知的多属性决策问题：首先,指出决策对象与理想对象之间的相似度与决策对象的优势度之间存在直接的关系；其次,借鉴离差最大化算法,提出用属性值的相似度来确定属性权重；再次,利用各个决策对象与理想对象的相似度对各决策对象进行排序并择优；最后,对属性值为区间数的多属性决策实例进行了分析,同时对该算法与离差最大化算法的异同点进行对比。

Research on multiple-attribute decision making problems based on similarity relationship

The definiteness of the attribute weight is an important research topic for multiple attribute decision making problems. Firstly, according to the uncertainty of the attribute weight problem for multi attribute decision making problems, the relationship of the similarity between the decision making subject and the ideal standard and the superiority of all decision making subject is analyzed. Secondly, with the help of maximizing deviation algorithm rules, the attribute weight based on the similarity of attribute values is proposed, then the similarity between decision making subject and the ideal standard is utilized to rank and pick over all decision making subjects. 〖JP2〗Finally, this paper takes the probabilistic multi attribute decision making problem that the attribute value is chosen as interval numbers as an example to state the applicability and rationality of the method. Meanwhile the characters between maximizing deviations and the algorithm are compared.