A VNS-NPGA approach to multi-objective optimization of hub-and-spoke logistics network

To realize low-cost freight transport in the logistics network and improve the network operation efficiency, a multi-objective optimization model and the corresponding algorithm for a hub-and-spoke logistics network are proposed based on the multi-level location of hub points and channels layout. By considering the structure of the multi-level hub-and-spoke logistics network and the features of the connectivity between the hub and spoke points, the multi-objective optimization model is constructed with two objectives of minimizing the total network operation costs and the total network service time. By considering the characteristics of decision variables and models, a variable neighborhood search (VNS)-niched Pareto genetic algorithm (NPGA) approach with a three-stage encoding structure chromosome is proposed, where the VNS algorithm nested in NPGA is used for individual variable neighborhood search to optimize individual channel level genes, and NPGA is adopted to solve the multi-objective optimization model. To evaluate the performance of the proposed VNS-NPGA approach, a real-life case study based on a small-scale Australia Post data set was conducted, and 25 nodes of the Australia Post and 14 nodes of the Gansu Province 3-level hub-and-spoke logistics networks were established, respectively. The analysis results indicated that the network structure of multi-level hub points could avoid the detour problem existing in the traditional hub-and-spoke network, and showed better applicability in the narrow geographical structure. Compared to the traditional multi-objective evolutionary algorithms, VNS-NPGA can obtain better solutions through the distributed optimization of channel levels, avoiding the problem that a single algorithm cannot effectively deal with coupling relationships in genes.