Descrição
Routing optimization has long been a problem in the networking field. With the rapid development of user applications, network traffic is continuously increasing in dynamicity, making optimization of the routing problem NP-hard. Traditional routing algorithms cannot ensure both accuracy and efficiency. The emergence of SDN enables network traffic to be steered from a global view of the network, which can provide more information for routing algorithms. However, the traffic distribution in the network is complicated and dynamically changes, making it difficult to design a good algorithm or policy to control the traffic throughout the whole network. Deep reinforcement learning (DRL) has recently shown great potential in solving networking problems. However, existing DRL-based routing solutions cannot process the graph-like information in the network topology and do not generalize well when the topology changes. In this paper, we propose AutoGNN, which combines a GNN and DRL for the automatic generation of routing policies.