Research on TTE Scheduling Algorithm Based on Hybrid Method

Time-Triggered Ethernet (TTE) is a strongly real-time deterministic network. It supports various types of traffic, including Time-Triggered (TT) and Rate-Constrained (RC) traffic, ensuring good support for mixed real-time applications. Existing scheduling methods for different types of TTE traffic mainly focus on optimizing TT traffic, while insufficiently considering the delay optimization for RC traffic. To address this issue, this paper proposes a hybrid reinforcement learning algorithm that transforms the TTE scheduling problem into a two-dimensional bin packing problem, utilizing heuristic algorithms for packing and combining reinforcement learning algorithms for optimizing scheduling strategies. Experimental validation shows that the proposed hybrid reinforcement learning algorithm (HYRL) demonstrates significant advantages in scheduling efficiency and delay optimization. The experimental results indicate that HYRL can generate high-quality scheduling tables at a speed tens of times faster than traditional Satisfiability Modulo Theories (SMT) and Genetic Algorithms (GA) when handling TT and RC messages. Additionally, the waiting time for TT messages and the transmission delay for RC messages are improved by over 10%. This method exhibits higher efficiency and flexibility in complex scheduling problems, providing valuable insights for the future development of TTE network scheduling technologies.