Fault Processing Algorithm of Power backbone Communication networks Based on Artificial Intelligence and State Perception

With the expansion of power backbone communication network, the topology structure is increasing complex, and the network operation and maintenance data are increasing exponentially. In order to effectively ensure that the large number of deep-level operation data which reflecting the essence of production and operation could be analysed, a fault recovery method for power backbone communication network based on artificial intelligence and state perception is proposed. Firstly, based on the state awareness technology, a fault management model of power backbone communication networks(PBCN) in virtualized network environment is established. Then a reinforcement learning algorithm for PBCN fault recovery is proposed, and a risk value evaluation model for service failure of PBCN is established with the aim at ensuring the quality of service (QoS) of PBCN. The goal of minimizing network loss caused by link failure is achieved. The simulation results show that the algorithm can recovery as much power communication services as possible and performance better in fault recovery of power communication services networks and the revenue of power backbone communication basic networks.