A straight priority-based genetic algorithm for a logistics network

Closed-loop logistics (forward and reverse logistics) has received increased attention of late due to customer expectations, greater environmental concerns, and economic aspects. Unlike previous works, which consider single products or single periods in multi-objective function problems, this paper considers a multi-product multi-period closed-loop logistics network with regard to facility expansion as a facility location-allocation problem, which is closer to real-world scenarios. A multi-objective mixed integer nonlinear programming formulation is developed to minimize the total cost, the product delivery time, and the used product collection time. The model is linearized by defining new variables and adding new constraints to the model. Then, to solve the model, a priority-based genetic algorithm is proposed that uses straight encoding and decoding methods. To assess the performance of the above algorithm, its final solutions and CPU times are compared to those generated by an initial priority-based genetic algorithm from the recent literature and the lower bound obtained by CPLEX. The numerical results show that the straight priority-based genetic algorithm outperforms the initial priority-based genetic algorithm at least in terms of obtaining a reasonable quality of final solutions for closed-loop logistics problems.