DEVELOPMENT OF A DECISION SUPPORT SYSTEMS GENERATOR FOR ILL-STRUCTURED MULTIPLE CRITERIA DECISION PROBLEMS

Decision support systems (DSS) are interactive, computer-based management information systems (MIS) for facilitating decision processes by utilizing data and decision models for ill-structured and semi-structured problems. The ultimate goal of DSS is to support the decision maker by structuring ill-structured problems. Thus, the key feature of a DSS is its modeling ability, which serves as a support to the decision-making process. Generally, from an information system standpoint, operations research or management science is concerned with models and model building. However, models have been applied primarily to well-structured problems. In fact, modeling the decision process and formulating proper models are the critical aspects of the MS/OR implementation process. A promising approach to building a DSS is using a DSS generator as the facilitator to formulate a problem and configure the DSS. Currently, there is no software system which is generalized, powerful, user-friendly, and provides a full-range of capabilities for easily formulating problems and building specific DSS in the field of ill-structured multiple criteria decision making (MCDM) tasks. The development of DSS generators is a complicated task. To integrate data and decision models with modeling capability into a well-designed DSS generator, a conceptual framework must be established as a foundation for developing such software systems. The objectives of this dissertation are: (1) to establish a comprehensive conceptual framework for developing DSS generators; (2) to design and develop a prototype DSS generator for multiple criteria decision problems based on the conceptual framework; and (3) to apply the prototype DSS generator to a real-world problem to validate the system. The conceptual framework employs the concept of Simon\u27s decision process to extend the conventional meaning of problem formulation in a DSS environment. The real case version of the prototype DSS generator is implemented on a personal computer. This workable DSS generator validates the design of the prototype DSS generator and the conceptual framework established for developing DSS generators