Dynamic Hierarchical Optimization for Train-to-Train Communication System

To enhance the operational efficiency of high-speed trains (HSTs), Train-to-Train (T2T) communication has received considerable attention. This paper introduces a T2T cooperative communication model that allows direct information exchange between HSTs, enhancing communication efficiency and system performance. The model incorporates a mix of dynamic and static nodes, and within this framework, we have developed a novel Dynamic Hierarchical Algorithm (DHA) to optimize communication paths. The DHA combines the stability of traditional algorithms with the flexibility of machine learning to adapt to changing network topologies. Furthermore, a communication link quality assessment function is proposed based on stochastic network calculus, which accounts for channel randomness, allowing for a more precise adaptation to the actual channel environment. Simulation results demonstrate that DHA has superior performance in terms of optimization time and effect, particularly in large-scale and highly dynamic network environments. The algorithm’s effectiveness is validated through comparative analysis with traditional and machine learning-based approaches, showing significant improvements in optimization efficiency as the network size and dynamics increase.