Intelligent Scheduling Method of Optical Transmission Network Based on Digital Twin

With the construction of optical transmission network(OTN), the huge number of optical cables brings great challenges to the maintenance of optical cable equipment. Due to the lack of a perfect intelligent management platform, much of the complex work is still manual. Dedicated to reducing the cost of network operation and maintenance and promote the evolution of optical transmission network to full automation, it is increasingly urgent to schedule and maintain the dumb resources automatically and intelligently by means of digital twin technology. In this paper, an automatic scheduling method based on digital twin for optical transmission networks is proposed. First, a digital twin layer of physical network (DTN) is constructed. In DTN, we construct and update network topology model according to the collected data, and realize topology dynamic generation function in combination with business model, which makes up the defects that the traditional scheduling method based on static model does not apply to dynamic network. In order to solve the model of topology generation optimization problem better, a topology optimization algorithm based on modified DDPG(Deep Deterministic Policy Gradient) is designed, which can effectively adapt to the dynamic process of network and make optimal decisions, reduce maintenance cost and improve resource utilization.