Design and Implementation of Regional Logistics Resource Intelligent Matching and Supply Chain Integration Control System Based on Multi-Source Data Fusion
Regional logistics serves as a critical pillar supporting the efficient circulation of industrial and supply chains. Aiming at prominent challenges in current regional logistics, including fragmented multi-source data, severe resource mismatches, and static & rigid scheduling modes, traditional approaches fail to cope with complex logistics scenarios compounded by order fluctuations, traffic disruptions and meteorological disturbances. This paper constructs a hierarchical fusion framework for multi-source heterogeneous regional logistics data, integrating Internet of Things (IoT) perception data, order data, geographic transportation data, meteorological data and supply chain ledger data. An entropy weight-attention improved fusion model is adopted to realize accurate feature reconstruction for logistics data with high noise and missing values. On this basis, a multi-constraint mathematical model for logistics resource supply-demand matching is established, and an improved Proximal Policy Optimization (PPO) reinforcement learning matching algorithm is designed. A multi-objective reward function balancing cost, timeliness, resource utilization rate and carbon emissions is constructed to complete dynamic intelligent matching of transportation capacity, warehousing and distribution nodes. Meanwhile, a closed-loop management and control mechanism of “Perception-Matching-Scheduling-Execution-Feedback” is built, and a regional logistics intelligent control system is developed based on cloud-edge collaboration. Simulation experiments are carried out using real logistics data of a domestic metropolitan area from 2024 to 2025. The results show that compared with traditional algorithms, the proposed method improves matching accuracy by 12.47%, reduces resource idle rate by 9.62%, shortens supply chain turnover time by 14.13%, and achieves excellent robustness under order peak and sudden disturbance scenarios. This research provides effective theoretical models and engineering practical schemes for integrated management & control of regional smart logistics and flexible coordination of supply chains. (Sun X & Wang H., 2025)