This dissertation presents a new approach for handling decision-making problems when (1) the information obtained is fuzzy, (2) no voluminous statistical data are available, and (3) the decision process relies on expert's knowledge. The proposed approach is based on fuzzy multicriteria decision-making paradigm and is designed to assist managerial and strategic level decision making in an organization.;Three scenarios in the decision-making paradigm are studied: (1) single decision maker with multiple criteria, (2) single decision maker with higher-order criteria, and (3) multiple decision makers with multiple criteria. These studies employ fuzzy sets to handle incomplete and uncertain information. The decision reasoning is carried out using fuzzy logic. The framework to model a decision-making problem is first presented. The relationship among the criteria, not considered in the traditional decision-making theory, is studied and represented by using if-then rules and higher-order structures. The if-then rules are utilized to describe the criteria relationship as well as to represent expert's knowledge, whereas higher-order structures are used to distinguish prioritized criterion and important criterion.;It is noted that one of the primary issues in group decision making is the consensus technique. This research proposes a new aggregating technique that maps the confidence value or belief value to a consensus space to obtain a consensus degree of the concerned criterion. The consensus term for each criterion is thereby generated by utilizing the Dempster's Rule of Combination. The overall decision outcome yields a linguistic solution and a corresponding conformity degree.;Based on the proposed model and techniques, a fuzzy decision support system, XPROS, is constructed to demonstrate the applicability and the reliability of the approach. The input and the output of this system can be linguistic terms, real numbers, or unknown data. Several real-world test cases are used to manifest the performance of the system.
展开▼
