On Some Issues With Model In Decision Support System

Decision Support System is one kind of man-machine interaction system that helps decision-makers solve structured and half structured problems using various of data, information and knowledge, especially the model technique. Model management is the kernel of DSS. Taking model techniques as research background, this thesis makes a focal research on the implementation and application technologies of the model normalization and model automatic selection.The primary works of this thesis are as follows:1. This thesis analyzes the problems of the existing Decision Support Systems and point out the new requirement of DSS along with the development of information technique and Web technique.2. With respect to the research on the normalization, model coding normalization, model data normalization, and model encapsulation have been achieved.3. This thesis designs the interfaces of model component for DSS, which include model normalization interfaces, model management interfaces, component based DSS model modules and so on. Creation, information retrieval, and operation of model components have been implemented with COM.4. The Ontology technique is used in natural language understanding to improve the semantic analysis capacity. The method of automatic model class selection based on Ontology and natural language understanding is proposed in the dissertation. Model class selection based on rules, and variables extraction using morphological analysis is discussed. These approaches improve the performance, efficiency and intelligence of model selection greatly.5. An interactive method of model structure selection is proposed, which has been applied efficiently in the developed model base system.