Design And Implementation Of Truck Logistics Distribution Prediction System

As one of the most famous e-commerce company, Amazon.cn has a complex network of warehouse, customer’s orders are trucked from fulfill center from station,then express delivery man would ship package to customer, the trucking part is outsourced to carriers.Different models of trucks, transportation costs are not the same, transportation costs is proportional to the truck’s volume, namely, the bigger volume, the higher trucking costs. Business Department staff would determine model and quantity of truck based on the historical experience, if the worker oversized estimated, then the truck space utilization is low, resulting in an unnecessary waste of resources, if estimated volume is too small, would delay part of the delivery of goods, customers would not receive their order in a timely manner, reduced customer satisfaction, the premise is that customers receive their orders in a timely manner, using truck space as high as possible, saving on shipping costs, the development of this system is very necessary.There is already has a system to forecast volume, but the prediction accuracy not very good. The new system would improve predict accuracy and give reasonable volume based on the old system. In this article, the author did requirement analyze, which include non-functional requirement, business classification and dividing function model, choose the system architecture based on business requirement at the same time, achieved summery design. Using decision tree to decide box type by asin category and order quantity, and then reduce noise by Random Forest to reduce noise during the choose box process. Testing forecast result, detailed design for all function module, focus on business process and module class design, described the meaning of mathematics model which applied in this system, make clear description for relationship between function points.The author take parts in the whole process of requirement analyze, design, implement and publish for this project. The work focus on design and implement of lane predict function, truck recommendation, business information extract, volume plot functions.After the system online, according to contrast the actual data collected, this system provides powerful digital credentials, improved the utilization successfully, save the cost of transportation, reducing truck space waste.