Optimization Model of Multicenter Distribution Logistics Network Based on NSGA-II

With the rapid development of globalization and e-commerce, traditional logistics models are facing issues such as rising costs and low delivery efficiency. This study innovatively proposes a multicenter delivery logistics network optimization model based on an improved nondominated sorting genetic algorithm II. By constructing a virtual delivery network and introducing adaptive scaling factors and variable neighborhood operators, the model enhances the convergence and diversity of solutions in multi-objective optimization. The model aims to minimize transportation distance costs, vehicle fuel consumption costs, personnel scheduling costs, delivery time, and pollution emissions. It demonstrates significant advantages in multicenter complex scenarios. Experimental results show that the Pareto front coverage of the optimization model reaches up to 0.98, and the number of optimal solutions increases to 7.4 after integrating the improved operators, an increase of 4.4 compared with the traditional NSGA-II. In tests involving 10 service recipients in Chengdu, the delivery time cost was only 3 h, the maximum difference in transportation costs compared with the MODE model was 5,800 yuan, the minimum single-person scheduling cost was 1,165.78 yuan, the minimum pollution emissions were 40.22 ppm, and the maximum user satisfaction reached 92.15%. These results clearly demonstrate that the proposed model can achieve a multi-objective balance in reducing delivery costs, improving system efficiency, and optimizing environmental impact, providing effective new insights and technical support for the design and optimization of future multicenter logistics delivery networks.