An Energy-Aware Network Routing Algorithm Based On Q-Learning

With the increasing network scale and network device number, reducing network energy consumption has become one of the research focuses in network optimization field. In this paper, we design an energy-aware network routing algorithm based on Q-Learning. By aggregating network traffic to construct minimum topology subset, we can effectively reduce the number of occupied devices during transmission process. Considering the real-time traffic requirements and residual bandwidth, we set the reward value function to classify the candidate links into three different levels for path selection. We aim to make full use of the link bandwidth to reduce the number of occupied devices, while satisfying the delay requirement and balancing load. To improve the convergence rate, we design heuristic factor to control the solution space search and enhance the guiding effect of the existing optimal solution. Simulation results show that the proposed algorithm outperforms existing algorithms in traffic aggregation and bandwidth utilization while meeting the delay and load-balance requirements.