DRL-Driven Service Routing in Software-defined IP+Optical Smart Grid Communication Networks

The rapid rise of Multimedia Broadband Services(MBS) makes the network traffic grow rapidly, which brings great challenges to the construction of smart grid. IP + optical network architecture based on SDN is an effective solution to this problem. However, the coordination ability and resource utilization of each layer of the network are often low. Therefore, for MBS, this paper proposes a DRL-Driven Service Routing method in Software-defined IP+Optical Smart Grid Communication Networks. Aiming at delay, network risk and primary/alternate routes similarity, this paper a reasonable routes allocation model. Then, deep reinforcement learning algorithm is used to solve the model. The simulation results show that this method can not only make full use of the resources of IP + optical networks, but also reduce the network risk, which provides a theoretical guidance for the routing method of MBS in smart grid communication network.