Multiple-Attribute Decision Making Based on Attribute Preference

摘要

Multiple-attribute decision making in the presence of independence between attributes, the shortage of quantitative value of one attribute can be compensated with that of other attributes. But, many a times, this is not true. On another side, the quantitative value of one attribute is treated as equally important in different intervals. However, this violates the law of diminishing marginal rate of substitution and destroys the additive of attributes. So, the valuations of alternatives based on weighed aggregation operator are failure. In this paper, two styles of attribute substitution are proposed, named by, complete-substitution and incomplete-substitution. For the complete-substitution, firstly, based on the law of diminishing marginal rate of substitution, the quantitative value of attributes is divided into many intervals with different weight levels. On the same level, the substitution rate is 100% and the additive of different attributes is effective. Secondly, the multiple-attribute decision making is transformed into single-attribute decision making and the valuations of alternatives are calculated based on single attribute with different levels. Thirdly, the comprehensive value of each alternative is calculated through by accumulating different levels and all alternatives are ranked by the comprehensive value of alternative. For the incomplete substitution, with the help of shortest path method and one-vote-down system, it is transformed into complete-substitution and calculated.