Research on Urban Distribution Optimization Under Point-based Billing on Simulated Annealing with Variable Neighborhood

Abstract

In this paper, we study the optimization of urban distribution networks with leased third-party logistics fleets under point-based billing and construct a mathematical model of multi-model vehicle path optimization (MVRPTW) with time windows aiming at the lowest logistics expenditure under point-based billing mode. For the problem that the simulated annealing solution to the VRPTW problem tends to fall into slow convergence, a variable neighborhood structure and adaptive temperature control are introduced to improve the simulated annealing. Finally, using the order data of a Beijing’s third-party logistics company Z, the model is solved by first clustering the customer nodes through the K-Means algorithm to generate the initial solution and then optimizing it with the improved simulated annealing algorithm. Compared with the simulated manual empirical scheduling, algorithmic scheduling can reduce the logistics cost by 88.1%.