Artificial Intelligence–Driven Digital Twin Models for Optimizing Freight Flow Management in Dry Cargo Ports

Abstract

Efficient management of freight flows in dry cargo ports is critical for reducing vessel turnaround times, minimizing congestion, and improving overall supply chain performance. This paper proposes an AI-driven digital twin (DT) framework designed to optimize freight flow management in dry cargo ports. The framework integrates multi-source real-time data, including vessel schedules, terminal operating logs, and IoT sensor streams, with predictive machine learning models and an optimization engine. Through discrete-event simulation and reinforcement learning, the DT continuously learns and adapts to port dynamics, enabling smarter resource allocation and congestion mitigation. The proposed approach enhances operational visibility, allowing port managers to predict potential bottlenecks and proactively allocate resources. The results of a simulated case study demonstrate significant improvements in vessel turnaround time, crane utilization, and overall port throughput. These findings highlight the transformative potential of AI-driven DT models in modern port logistics and their contribution to sustainable maritime transport.