Systèmes d'aide à la décision - Une approche méthodologique intégrée

The paper puts forward a specification model for Decision Support Knowledge Bases. That model is intended for end-users as well as for professional developers. Specifications in this model must be easily mapped into widespread decision support tools such as spreadsheet processors, data managers and small rule-based systems. The knowledge base is split into a data base and a rule base. The data base is described through its E/R conceptual schema, while the rule base consists of a set of variables and a set of rules that tell how to derive the values of the variables from external data (e.g. from the data base). The paper concentrates on expressing the rule base. It develops simple variables and their defining rules, then dimensioned variables, which are indexed with simple variable(s), and finally variables which take their values in the data base. A specification language is suggested as well. The text is not an axiomatic presentation of abstract concepts but follows a pedagogical structure that best matches the experience of end-users.