Solving large-scale uncapacitated facility location problems with evolutionary simulated annealing

Uncapacitated Facility Location (UFL) Problems are, in general, modelled as mixed integer programming problems, which are known as NP-hard problems. In recent years, a few publications have appeared on the metaheuristics for solving UFL problems, discussing the performance of particular implementations of metaheuristics for small and middle size UFL benchmarks. The large-scale problems remain untouched. The approach presented in this paper attempts to tackle them with a metaheuristics combining two well-known approaches. The idea is to enable algorithm searching through solution space by taking advantage of both underlying approaches in order to avoid local minima. The power of simulated annealing (SA) in local search and that of the evolutionary approach in global search have been brought together to obtain the desired solution quality within a shorter time.