International Journal of Computers Communications & Control · 2013 · 10 citations · 8 references
ArchitectureIndustrial EngineeringOperations ManagementFacility LocationFacility ManagementOptimizationEvolutionary ComputationEvolutionary OptimizationGenetic AlgorithmsMetaheuristic OptimizationPopulation-based SearchFacility Location OptimizationGenetic Algorithm OptimizationLocal SearchLocal Search AlgorithmsLocal Search Optimization
In this paper, an improved genetic algorithm (GA) for solving the multi-level uncapacitated facility location problem (MLUFLP) is presented. First improvement is achieved by better implementation of dynamic programming, which speeds up the running time of the overall GA implementation. Second improvement is hybridization of the genetic algorithm with the fast local search procedure designed specially for MLUFLP. The experiments were carried out on instances proposed in the literature which are modied standard single level facility location problem instances. Improved genetic algorithm reaches all known optimal and the best solutions from literature, but in much shorter time. Hybridization with local search improves several best-known solutions for large-scale MLUFLP instances, in cases when the optimal is not known. Overall running time of both proposed GA methods is signicantly shorter compared to previous GA approach.
8