基于SAC的低轨卫星光网络负载均衡路由优化算法OA

To address the issues of network load imbalance and communication delay caused by satellite dynamics and complex environmental factors,a load-balancing routing optimization algorithm for low earth orbit Satellite optical network based on Soft Actor-Critic(SAC)in deep reinforcement learning(DRL)is proposed.By simulating the satellite network environment and dynamically adjusting decision-making strategies,the algorithm achieves dynamic optimization of routing paths,aiming to ra-tionally allocate network resources and improve network performance.Simulation results demonstrate that the proposed algo-rithm exhibits significant advantages in optimizing data flow distribution and enhancing system reliability.Under traffic intensi-ties of 50%and 100%,the average link load rate is optimized by 29.9%and 42.0%,respectively,while routing delay is also ef-fectively reduced.