Against the backdrop of rapid e-commerce development, the efficiency of terminal distribution systems is crucial for customer satisfaction and operational cost management, especially amid the increasing demand for instant delivery. This study addresses the "cold start" problem for new users in logistics by optimizing delivery routes through mixed-integer linear programming (MILP) and an improved genetic algorithm (IGA), aiming to enhance route efficiency for new users lacking substantial historical interaction data. Simulation experiments involving 300 e-commerce platform users verified the model's effectiveness, demonstrating that the optimized delivery routes reduced average delivery time by 15% and costs by 20%. Performance comparisons revealed that MILP + IGA outperformed traditional methods in runtime, convergence accuracy, and iteration efficiency, exhibiting higher solving efficiency and adaptability. Furthermore, analysis of typical community case studies showed that the optimized route design achieved station consolidation and route reconfiguration, leading to significant reductions in distribution costs and improved resource allocation. The results highlight the practicality of user profiling in tailoring logistics strategies to meet individual user needs without relying on extensive historical data. This study makes an important contribution to logistics management theory by demonstrating how user behavior prediction can be integrated into delivery route optimization to enhance operational efficiency and reduce e-commerce logistics costs. This approach not only helps overcome initial barriers faced by new users but also lays a foundation for future research aimed at incorporating real-time data to further personalize delivery services.
Zhang et al. (Tue,) studied this question.