Research on computer wireless communication technology based on machine learning algorithm

Different services in electric power communication network have different network requirements. Traditional centralized network architecture is difficult to provide communication network services that meet all business needs. Therefore, the design of dedicated power network slicing and the construction of slice-based communication network architecture will be the future development trend of the power industry. For different power grid service scenarios, network resources need to be allocated to different service slices on demand. And guarantee the quality of service. At the same time, in the network environment, the interference of pulse noise is the key factor affecting the quality of signal transmission, and the reduction of pulse noise before signal demodulation can improve the communication quality of electric power communication network. Based on Q learning algorithm in machine learning, this paper studies resource allocation and impulse noise reduction in network slice of electric power communication network. Aiming at the low efficiency of resource allocation method in power communication network slicing, this paper proposes a method of resource allocation based on end-to-end architecture combining experience quality and resource utilization. First, the resource allocation problem is modeled as a Markov process. Then, Q learning in reinforcement learning and deep Q network algorithm are used to solve the optimal allocation scheme. The simulation results show that the resource allocation method based on Q learning and DON can meet the requirements of multi-service experience quality and improve resource utilization. In addition, the resource allocation algorithm based on DON has faster convergence speed than the resource allocation algorithm based on Q learning.