Two meta-heuristics for a multi-period minisum location-relocation problem with line restriction

摘要

This paper investigates a multiperiod rectilinear distance minisum location problem, as a mixed-integer nonlinear programming (MINLP) model with a line-shaped barrier restriction, in which the starting point of the barrier uniformly distributed in the plane. The objective function of this model is to minimize the sum of the costs associated with the expected weighted barrier distance of the new facility from the existing facilities and the costs incurred by location-dependent relocation during the planning horizon. Then, a lower bound based on the forbidden region is presented. To show the validation of the presented model, a number of numerical examples are illustrated. The associated results show that the optimization software is effective for small-sized problems. However, the optimization software is unable to find an optimum solution for large-sized problems in a reasonable time. Thus, two meta-heuristics, namely genetic algorithm (GA) and imperialist competitive algorithm (ICA), are proposed. Finally, the associated results are compared and discussed.