MIXED-PARADIGM CASE-BASED REASONING USING A BLACKBOARD ARCHITECTURE

CASE-BASED REASONING (CBR) INVOLVES THE USE OF SUITABLE CASES OF PAST EXPERIENCE TO INTERPRET OR SOLVE A NEW PROBLEM. THIS PROPOSED WORK EXPLORES INCORPORATING CBR TECHNIQUES IN MIXED-PARADIGM KNOWLEDGE-BASED APPLICATIONS. IN MANY POTENTIAL CBR APPLICATIONS, CBR APPROACHES ALONE ARE INSUFFICIENT AND MUST BE COMBINED WITH OTHER PROBLEM-SOLVING TECHNIQUES IN A MIXED-PARADIGM CBR FRAMEWORK. OUR APPROACH TO MIXED-PARADIGM CBR IS TO USE A BLACKBOARD ARCHITECTURE AS AN INTEGRATION FRAMEWORK FOR CASE-BASED KNOWLEDGE SOURCES (KSS) OPERATING WITH KSS USING OTHER PROBLEM-SOLVING APPROACHES. THE OPPORTUNISTIC CONTROL CAPABILITIES OF THE BLACKBOARD PARADIGM ARE WELL SUITED TO MIXED-PARADIGM PROBLEM SOLVING. CBR AND THE BLACKBOARD PARADIGM ALSO SHARE FEATURES THAT FACILITATE AND EXTEND THE CAPABILITIES OF SUCH INTEGRATION BEYOND THE MERE SUPPORT OF KSS EMPLOYING CBR TECHNIQUES. THESE SIMILARITIES INCLUDE: INDEXING AND RETRIEVAL (CASE S AND BLACKBOARD OBJECTS), RELATIVELY LARGE-GRAINED CONTROL DECISIONS (CASE SELECTION AND KS EXECUTION) WE WILL BUILD ON THESE SIMILARITIES TO DEVELOP A FLEXIBLE, MIXED-PARADIGM CBR FRAMEWORK BASED ON THE BLACKBOARD ARCHITECTURE. IN THE PROCESS WE WILL ADDRESS ISSUES IN CBR INDEXING AND RETRIEVAL AND IN CONTROL OF MIXED-PARADIGM CBR APPLICATIONS. ANTICIPATED BENEFITS/POTENTIAL COMMERCIAL APPLICATIONS - THE PROPOSED WORK WILL RESULT IN A FLEXIBLE FRAMEWORK FOR MIXED-PARADIGM CBR APPLICATIONS BASED ON THE BLACKBOARD ARCHITECTURE. IT WILL EXPAND THE USE OF CBR TO APPLICATION AREAS THAT ARE NOT COMPLETELY SUITED TO CBR TECHNIQUES AND WILL INCREASE UNDERSTANDING OF THE SYNERGY BETWEEN CBR AND OTHER PROBLEM-SOLVING PARADIGMS. THE WORK WILL ALSO PROVIDE EXAMPLES OF MIXED-PARADIGM CBR APPLICATIONS. MIXED-PARADIGM CBR IS EXPECTED TO PLAY A SIGNIFICANT ROLE IN FUTURE GOVERNMENT AND COMMERCIAL APPLICATIONS.