面向电力通信网动态适配的光网络资源智能分配方法OA

With the increasing volatility of renewable energy output,the rising real-time requirements of dispatching services,and the sustained concurrent growth of diverse fine-grained services in smart grids,power communication networks are facing multiple challenges,including highly dynamic traffic,fine granularity,and high reliability.Conventional optical network resource configuration is rigid and cannot adapt to rapid fluctuations in traffic load and network conditions,while its limited support for fine-grained services lead to communication risks during resource scheduling.To address these challenges,this paper proposes self-adaptive high-reliability dynamic fine-grained resource allocation,a highly reliable and dynamically adaptive fine-grained optical network resource allocation method for power communication networks with parameter fluctuation adaptability.Leveraging the fine-grain optical transport network architecture,a multi-dimensional state feature space is constructed and integrated into an asynchronous actor-critic reinforcement learning model,enabling the system to perceive the intricate correlation between power communication network state parameters and optical resource allocation,thereby achieving the coordinated optimization of reduced blocking probability and improved resource utilization under highly dynamic scenarios.