THERE ARE FIVE MAJOR THEORIES* OF RELEVANCE TO DEVELOPMENT OF A "CONSENSUS THEORY" FOR HANDLING EVIDENCE IN EXPERT SYSTEMS.

THERE ARE FIVE MAJOR THEORIES* OF RELEVANCE TO DEVELOPMENT OF A "CONSENSUS THEORY" FOR HANDLING EVIDENCE IN EXPERT SYSTEMS. SINCE THEY HAVE EVOLVED FROM VERY DIFFERENT FIELDS, THEY DO NOT HAVE IDENTICAL ASSUMPTIONS NOR DO THEY SHARE A COMMON NOTATION, TERMINOLOGY, OR SEMANTICS. THE PROPOSED EFFORT SEEKS TO ELUCIDATE THE PRACTICAL IMPACT OF THESE THEORETICAL APPROACHES FOR BUILDERS OF EXPERT SYSTEMS IN THE MILITARY DOMAIN AND TO LAY THE GROUNDWORK REQUIRED FOR FORMULATION OF A "CONSENSUS THEORY" BASED ON CURRENT THEORIES AND METHODS. TO THIS END, THE PAHSE I EFFORT HAS THREE MAJOR TECHNICAL OBJECTIVES. FIRST, IT WILL DEVELOP A COHERENT AND SYSTEMATIC DESCRIPTION OF EACH THEORY USING A SINGLE, CONSISTENT NOTATION ACROSS ALL FIVE. SECOND, IT WILL FORMULATE A SET OF ILLUSTRATIVE PROBLEMS, SOLVE THEM USING THEORY IN TURN, AND COMPARE THE RESULTS. THIRD, IT WILL ASSESS THE IMPACT OF THESE RESULTS AND IDENTIFY THE MOST IMPORTANT RESEARCH TOPICS FOR THE PHASE II PROJECT.