Abstract: The rapid development of the internet technology has greatly improved the service level of home pick-up in express industry, which also resulted in an unprecedented fierce competition. Considering high cost and low customer satisfaction in the process of home pick-up service, a dynamic routing problem model is proposed for home pick-up service based on the internet appointment platform. In this model, customers’ time windows are considered as fuzzy variables, and minimizing the total cost is taken as the goal. Then, a dynamic scheduling-based genetic algorithm (GA) is proposed to solve the model, in which initial coding based on the Clarke and Wright savings algorithm or the greedy insertion is designed, and a best-cost route crossover and an elite retention strategy are adopted. Based on the 21 classic benchmarks composed of Taillard, Christophides and Fisher, a comparative analysis is performed with other four published algorithms to evaluate the performance of the improved GA. The results show that the improved algorithm obtains 15 better solutions and has good convergence. Finally, based on the result analysis and sensitivity analysis about a simulation case of an express company in Tianjin, the validity and applicability of the model are verified, which provides strong decision support for home pick-up service.

Key words: ‘Internet+’, appointment platform, home pick-up service, dynamic routing, genetic algorithm