Studies on Fuzzy Preference Evaluation

本文研究了群体决策评价中模糊评价的处理方法。首先，运用模糊理论对多属性风险分析进行了扩展。从隶属函数的角度考虑了风险评价的聚合方法。特别地，打算用与多属性效用函数一致的风险态度来构造模糊多属性风险函数。从可能性分布函数的角度对风险评价的模糊性进行了处理。这项研究的延伸是在两个方面进行的。一是基于可能性分布的模糊对策的构建及其在国际谈判过程中的应用。引入“外部联盟”是为了评估国际联盟的备选形式，并根据它们的评估对“内部联盟”的原始评估进行修改。二是开发群体决策智能支持系统（IDSS）。在这个系统中，强调了将决策支助系统与诸如决策分析等学科领域的理论基础相结合的必要性。的结构。考虑了IDSS的主框架。针对特定的Shell，提出了一种基于模糊隶属度函数的模糊效用函数的模糊效用函数推导方法。构造在可能性分布函数上。本文还对n人合作博弈和效用函数的联合利用进行了基础研究，在数学规划的框架下，研究了多层次系统中区域政策的评估方法，并结合经验数据进行了应用。

In this research, methods for treating ambiguous evaluation in group decision evaluation are developed. First, multiattribute risk analysis is extended with fuzzy theory. Aggregation methods for risk evaluation is considered on membership functions. In particular, it is intended that the fuzzy multiattribute risk function is constructed with a risk attitude consistent with the multiattribute utility function. The ambiguousness of risk evaluation is treated also in terms of possibility distribution functions.The extension of this research is performed in two ways. One is a construction of fuzzy game assessed on the possibility distributions and its application to international negotiation processes. The *external coalitions" are introduced for evaluating alternative formation of international coalitions and with their evaluation, the original assessment by the "internal coalitions" is modified.Another is to develop an intelligent decision support systems(IDSS)for group decision making. In this system, the necessity for combining an DSS with a theoretical basis in disciplinary fields, such as decision analysis, is stressed. A structure of the. main frame for the IDSS is considered. As a particular Shell, a method for deriving a group utility function as a fuzzy utility function is developed, which is based on anevaluation of fuzzy membership functions for deriving a possibility distributions and. constructed-on the possibility distribution functions.A basic research for combined utilization of n-person cooperative games and utility functions is also performed, in which a method to evaluate a regional policy is examined in a multilevel system and applied with empirical data in a framework of mathematical programming.