A prototype DSS for structuring and diagnosing managerial problems

A methodology for managerial problem recognition, structuring, and diagnosis is presented. The system integrates diagnostic concepts from artificial intelligence with database management and structural modeling techniques. Given an organizational database, a user specifies control bounds for selected data items and causal relationships between variables in the database. Structural modeling techniques are used to represent knowledge of causal relations. If a control variable´s value is out of bounds, structural models are used in an attempt to diagnose (explain) the discrepancy. If the discrepancy is explained, a diagnosis is at hand. If not, the user´s model is somewhat deficient and what-if analysis can be used to construct an improved model. This methodology has been implemented in a prototype decision support system (DSS) that is described, along with an example of its use