Intelligent Logistics Optimization for Carbon Emission Reduction in Cement Industry: A Sensor-Driven Dynamic Order Matching Approach

Efficient logistics management is critical for reducing carbon emissions in the cement industry. This work proposes an intelligent logistics optimization framework integrating sensor-driven vehicle state detection, reinforcement learning-enhanced order matching, and adaptive path optimization. The system employs IMU-based vehicle monitoring, a Deep Double Q-Network (DDQN)-Hungarian matching algorithm for dynamic order allocation, and a reinforcement learning-guided shortest path strategy combining Dijkstra and A*. Experimental results on real-world logistics data demonstrate reduced empty mileage, improved order matching efficiency, and lower carbon emissions. The proposed approach outperforms traditional methods, offering a scalable and adaptive solution for sustainable cement transportation.