An explanation subsystem to support decision problem structuring

This paper develops the principles underlying the development of an explanation subsystem, a component of a knowledge-based exploratory analysis system (KEAS). A KEAS is a decision support system (DSS) which supports decision problem structuring (DPS) through an interactive process of consultation between a single decision maker (DM) and the DSS. By DPS, we refer to a DM´s cognitive process (1) of identifying decision elements (e.g. objectives, alternatives, and evaluation criteria) and their interrelationships, which characterize a decision situation and (2) of specifying the DM´s expectations in dealing with that situation. Based on these expectations and an initial specification, however incomplete, of the decision problem structure, a DSS can assist the DM in conducting exploratory analyses. Exploratory analysis begins with the specification of a decision problem structure based largely on decision analysis concepts. As a result of exploratory analysis, a decision problem structure may be modified using a mechanism based on concepts of knowledge-based systems. Hence, a KEAS exploits the natural synergy between the concepts of decision analysis and the incremental modifiability of knowledge structures. This paper develops the principles underlying the explanation subsystem component of a KEAS