Adaptive scheduling for multimedia and text traffic in ocean networks through queuing theory integrated reinforcement learning

The Offshore Communication Network (OCN) is an Internet of Vehicles system that connects fishing vessels at sea, enabling communication and internet access to support sustainable growth in the blue economy. The dynamic maritime environment and limited infrastructure make it challenging to ensure low-latency and high-reliability transmission, especially for multimedia messages with strict quality demands. To address these challenges, this paper proposes an adaptive transmission queue scheduling framework for multimedia and text messages in the OCN. The framework combines a multidimensional quasi-birth-and-death queuing model to analyze queue states and loss rates with Deep Reinforcement Learning (DRL) to dynamically adjust queue priorities in response to real-time network conditions. In addition to enhancing lower layer performance, the proposed solution is designed to maintain application-level quality of service for critical offshore operations, including emergency response coordination, real-time multimedia exchange for maritime safety, and mission-critical reporting. Experimental results demonstrate that the intelligent transmission queue scheduling scheme reduces queue delay and loss rate, thus improving operational efficiency and enabling IoT-driven sustainable marine resource management.