基于愿景滿意度函數(shù)的多屬性群決策方法

摘要： 針對定性和定量屬性相混合的多屬性群決策問題,提出基于愿景滿意度函數(shù)的多屬性群決策方法。該方法從損失和收益的角度考慮問題,通過計算標準化的愿景值和屬性值之間的距離,建立考慮專家不同風險態(tài)度的愿景滿意度函數(shù),求得專家各自的滿意度,最后應用簡單加權算子集結專家的排序結果。通過案例分析和算法對比說明了該算法可適用于不同風險態(tài)度的專家進行決策,而風險態(tài)度和愿景值的分析結果表明當屬性不確定性較大時,專家看重損失或者收益取決于專家的風險態(tài)度和愿景值的高低,該方法可為參與決策的雙方提供有價值的參考意見。
For multi-attribute group decision-making problems composed of qualitative and quantitative attributes, aspiration is an important factor in decision making for both individuals and organizations. The existing researches mainly combine the aspiration and Prospect Theory to get the prospect value for selective preference. However, Prospect Theory assumes that most decision makers display a risk preference when facing the prospect of loss, and display a risk aversion when facing the prospect of gain. The theory sets fixed risk attitude coefficient to solve prospect value, and it cannot be fully applicable to solve the multi-attribute decision making problems with aspiration satisfaction. Hence scientific decisions should be made according to the decision-makers' psychological aspiration satisfaction. In this paper, a new method called two-stage aspiration satisfaction function is proposed based on the combination of attribute aspiration, risk attitude, utility function and satisfaction. In this method, the distances between the standardized aspiration values is calculated and the attribute values is calculated by using the distance computational formula and the differentiation method on the cost and benefit of attribute. Then the as piration satisfaction function is built which integrates varied risk attitudes and aspiration of experts into the multi-attribute decision-making method to get satisfaction. Finally, the weighted algorithm is used to get the sorting results of aggregation experts. In the case analysis, comparison with the existing algorithms and risk attitude analysis show the superiority that this algorithm can be used for decision makings of different risk attitudes in investors. The analysis results of risk attitude and aspiration represent that investor's value loss or gain depends on the risk attitude and the level of aspiration while the attribute uncertainty is large. Computational results also have indicated that the proposed method is robust with respect to the variance of experts' different risk attitudes and the level of aspirations. When companies' attributes vary a lot, an interesting result reveals that the risk averse specialists value losses more than gains,and the risk appetite experts consider gains to be more important. What's more, the proposed method could be a useful tool for investors with different risk attitude who tries to make a profit, and it is also useful for companies to be the winner by sharpening their attributes. In particular, a company, who wants to be an investee, needs to improve the gain attributes, if the investor's aspiration is high; or to improve the loss attributes, if the investor's aspiration is low.