Analysis of Logistics Linkage by Digital Twins Technology and Lightweight Deep Learning.

Abstract

The present work expects to meet the personalized needs of the continuous development of various products and improve the joint operation of the intraenterprise Production and Distribution (P-D) process. Specifically, this paper studies the enterprise's P-D optimization. Firstly, the P-D linkage operation is analyzed under dynamic interference. Secondly, following a literature review on the difficulties and problems existing in the current P-D logistics linkage, the P-D logistics linkage-oriented decision-making information architecture is established based on Digital Twins. Digital Twins technology is mainly used to accurately map the P-D logistics linkage process's real-time data and dynamic virtual simulation. In addition, the information support foundation is constructed for P-D logistics linkage decision-making and collaborative operation. Thirdly, a Digital Twins-enabled P-D logistics linkage-oriented decision-making mechanism is designed and verified under the dynamic interference in the linkage process. Meanwhile, the lightweight deep learning algorithm is used to optimize the proposed P-D logistics linkage-oriented decision-making model, namely, the Collaborative Optimization (CO) method. Finally, the proposed P-D logistics linkage-oriented decision-making model is applied to a domestic Enterprise H. It is simulated by the Matlab platform using sensitivity analysis. The results show that the production, storage, distribution, punishment, and total costs of linkage operation are 24,943 RMB, 3,393 RMB, 2,167 RMB, 0 RMB, and 30,503 RMB, respectively. The results are 3.7% lower than the nonlinkage operation. The results of sensitivity analysis provide a high reference value for the scientific management of enterprises.