Optimizing Location-Routing Problem using Iterative Combination of GA and VNS

The location routing problem (LRP), which involves the decision of depot locations and vehicle routings, is one of the main factors for the success of logistic system. This paper focuses on solving LRP with multiple capacitated depots and multiple capacitated vehicles per depot. To generate a solution with approximate minimum system's total cost, a hybridization method is proposed to make facility location and delivery car routing decision simultaneously. The two main components of this proposed hybridization (GA-VNS) are Genetic Algorithm (GA) and Variable Neighborhood Search (VNS). With this hybrid algorithm, local optimal in GA is avoided by switching the searching space into VNS, and vice versa for avoiding local optimal in VNS. This process also makes the solution converges faster. The proposed method is evaluated on the classical benchmark instances. It can be seen that the performance of the proposed method is very promising, when compared with the individual using of GA or VNS.