UNIQUE APPLICATIONS FOR ARTIFICIAL NEURAL NETWORKS

THE INVESTIGATION HAS THE FOLLOWING OBJECTIVES: 1. ESTABLISH THAT MODULAR NEURAL NETWORKS, WORKING SYNERGISTICALLY WITH GENETIC SEARCH, PROVIDE A UNIQUELY POWERFUL MEANS OF INTELLIGENTLY CONTROLLING HEURISTIC MATHEMATICAL ALGORITHMS FOR LARGE-SCALE VEHICLE ROUTING AND SCHEDULING PROBLEMS. THE DESIGN LENDS ITSELF NATURALLY TO PARALLEL COMPUTING ON COMPUTING NETWORKS. 2. CODE A PROTOTYPE SYSTEM USING THE DESIGN, DEMONSTRATING THE VIABILITY AND STRENGTHS OF THE SYSTEM. 3. PREPARE A DETAILED REPORT THAT INCLUDES THE DESIGN, RESULTS OF COMPUTATIONAL EXPERIMENTS, AND FUTURE PLANS FOR THE TECHNOLOGY. THE METHODS BUILD UPON THE PROVEN STRENGTHS OF OPERATIONS RESEARCH TECHNIQUES (E.G., GENERALIZED ASSIGNMENT ALGORITHMS) FOR VEHICLE ROUTING, BUT SIGNIFICANTLY AMPLIFY THE PERFORMANCE OF THE TECHNIQUES BY USING NEURAL NETWORK AND GENETIC SEARCH PARADIGMS TO SET PARAMETERS AND ADAPTIVELY GUIDE THEIR COMPUTATIONS. THE METHODS ARE NEW TO THE VEHICLE ROUTING AND SCHEDULING PROBLEM SOLVING DOMAIN. PRELIMINARY TESTING SUGGESTS THAT SOLUTION QUALITY IS SIGNIFICANTLY BETTER THAN WHAT ANY OF THE UNDERLYING MATHEMATICAL METHODS WORKING INDIVIDUALLY CAN ACHIEVE. THE PRIMARY ISSUES TO BE RESOLVED INCLUDE THE TOPOLOGY OF THE NEURAL NETWORKS, THE PARAMETERS AND DEGREES OF FREEDOM OF THE GENETIC ALGORITHMS, AND SOME DETAILS OF THE UNDERLYING HEURISTIC BEING CONTROLLED. ANTICIPATED BENEFITS/POTENTIAL COMMERCIAL APPLICATIONS - TRANSPORTATION OF GOOD AND PEOPLE IS A PERVASIVE AND CONSTANT ACTIVITY WITHIN BOTH MILITARY AND COMMERCIAL OPERATIONS. GIVEN THE SCOPE OF THE ASSOCIATED VEHICLE ROUTING AND SCHEDULING PROBL