CASE-BASED SELECTION OF ALLOCATION METHODS

THE OBJECTIVE IS TO DEVELOP A CASE-BASED REASONING (CBR) SYSTEM FOR DETERMINING THE ALLOCATION METHODS USED BY FIRE DIRECTION OFFICERS. CASE-BASED REASONING IS A NEW APPROACH TO EXPERT SYSTEMS; COMPARED TO RULE-BASED SYSTEMS, CBR APPEARS TO BE A MORE EFFICIENT, SENSITIVE, AND EFFECTIVE STRATEGY TO DERIVE AND REPRESENT CRITICAL KNOWLEDGE. PHASE I WILL DEMONSTRATE THE FEASIBILITY OF A CBR APPROACH THAT USES TACTICAL FIRE CONTROL DECISION DATA FROM PREVIOUS EXERCISES AS THE BASIS FOR PREDICTIONS. STANDARD ARTIFICIAL INTELLIGENCE (AI) APPROACHES RUN INTO THE DIFFICULTY OF HAVING TO MODEL THE COMPLEXITIES OF THE HARDWARE/SOFTWARE ELEMENTS ALONG WITH THE COMPLEXITIES OF THE HUMAN FACTORS. A CBR STRATEGY AVOID THIS PROBLEM BY RELYING ON MATCHES TO SIMILAR PROJECTS SO THAT IT IS NOT NECESSARY TO REPRESENT WORLD KNOWLEDGE. ADDITIONALLY, A CBR APPROACH OFFERS THE POSSIBILITY OF IMPROVING REAL-TIME PERFORMANCE CONSIDERABLY, SINCE A SIMILAR CASE REPRESENTING A NEAR SOLUTION CAN BE RETRIEVED RAPIDLY.