A combined GA-TS algorithm for two-echelon dynamic vehicle routing with proactive satellite stations

Abstract

The make-to-stock strategy can improve the operational efficiency of urban freight delivery, especially in a dynamic transportation system with idle storage resources. This study proposes a two-echelon dynamic vehicle routing problem with proactive satellite stations (2E-DVRP-PSSs), which converts customers with available idle storage to satellite stations and optimizes the operating cost and make-to-stock cost. In 2E-DVRP-PSSs, heavy trucks depart from the depot, serve original static customers and proactive satellite stations (PSSs), and then return to the depot in the first-echelon network. In the second-echelon network, light vehicles depart from the PSSs to serve the dynamic customer demands. To compute the delivery routes efficiently, a hybrid algorithm integrating the cutting plane and improved genetic algorithm-tabu search (GA-TS) algorithm is proposed and implemented. An exact method using Gurobi solver and an improved GA-TS algorithm with multiply optimization strategies are also incorporated. Furthermore, the effectiveness of the 2E-DVRP-PSS formulation and the applicability of the hybrid algorithm for various instances are experimentally evaluated. An empirical case study of a two-echelon dynamic network in Dalian, China indicates that the proposed make-to-stock strategy with the PSS network can reduce costs and improve the efficiency of operations in urban transportation networks.