Integrating optimization heuristic techniques and GIS for solving the districts determination problem in urban areas

Abstract

Recent research in Geographic Information Systems (GISs) has created the kind of systems capable of modeling a number of interesting real world phenomena. Some applications where GISs have made significant contribution include routing, scheduling, dispatching, transportation logistics, vehicle tracking, market research, construction Planning, facility management and resource distribution. However, there seems to exist little connection between GIS solutions and much related research being carried out in other disciplines. This thesis studies the advantages of integrating recent research in heuristic combinatorial algorithms with GIS to solve real world problems. In particular, we investigate the use of four different heuristic algorithms to solve the problem of mail distribution in urban areas where districts determination is a priority goal for automation. A variation to the local search phase of the GRASP (Greedy Randomized Adaptive Search Procedures) algorithm is proposed and implemented which improved the workload balance between mailmen.