Research On Logistics Scheduling Based On Machine Learning And Heuristic Algorithms

In recent years,with the improvement of people's living standards,online shopping has become an indispensable part of people's lives.The rapid development of e-commerce has brought unprecedented opportunities to the express delivery industry,but the huge order volume of the double eleven is likely to lead to the burst of express delivery,and greatly exceeds the carrying capacity of the logistics system,and the huge workload in the short term gives Companies and distributors also bring great pressure,and slow delivery also brings a poor service experience to users.This paper uses machine learning to analyze and model the operation data of e-commerce platform users,and formulate strategies for pre-delivery of users who predict purchases,and reasonably reduce the peak of transportation and distribution.The work of this paper is carried out in the following aspects:(1)Analyze the original data,construct the sample features,and the model information is less likely to lead to the model over-fitting problem.The Pearson correlation coefficient and PCA feature selection method are designed.The algorithm is used to select the features and retain the data.The core information and a solution to the integrated learning approach to the sample disequilibrium of the data in this paper.(2)Using the basic learning methods such as support vector machine,GBDT and random forest in machine learning to construct the sample data,a hybrid model based on bagging is designed to improve the accuracy of purchase prediction and the generalization ability of the model.And predict the user's purchases.(3)Using the stored user geographic data of the e-commerce platform,a logistics distribution model is established to realize an early delivery strategy for the purchased user.For the classic logistics distribution model,the objectives and constraints are not comprehensive enough,and there are certain defects in the practical application.Based on the classic model,the author constructs a logistics distribution path optimization model that considers the integration of time window and fuel consumption.A particle swarm optimization algorithm based on k-medoids dynamic clustering hybr-id topology is designed to solve the problem that the classical particle swarm optimization algorithm is easy to fall into the local optimal solution when solving such models.The simulation results show that the improved particle swarm optimization algorithm can jump out of the local optimal solution and converge to the global optimal solution quickly,which can effectively solve the logistics distribution path optimization problem.