NOMA-based Wireless Networks Resource Allocation Technology

In today's network communication environment,the number of various smart devices and mobile users that access the network has increased dramatically,and the problem of scarcity of spectrum resources has become more and more prominent.Compared with traditional Orthogonal Multiple Access(OMA),Non-Orthogonal Multiple Access(NOMA)technology has the advantages of high system capacity,low transmission delay,and access to more users,which can better satisfy wireless networks Transmission needs.NOMA can provide the resources of one channel to several users in a non-orthogonal manner.This article focuses on the user allocation and power control issues of the NOMA system,and optimizes the sum rate of the NOMA system.First,the theoretical basis of NOMA is introduced,and the superposition coding and serial interference cancellation(SIC)technology of the NOMA system is briefly described.Superposition coding is to encode the user signal at the transmitting end and then superimpose and send;while SIC can decode gradually,reconstruct and eliminate the superimposed signal.Then,the user grouping algorithm and power control algorithm commonly used in the NOMA system are briefly introduced.Secondly,for the application scenarios of the NOMA system downlink,a system modeling was carried out.In order to maximize the system and rate,the IPF-CCCP algorithm is proposed.The algorithm is divided into two stages.First,an improved proportional fair user grouping algorithm(IPF)is adopted.In the user selection process,the forgetting factor is used to accelerate the update of user scheduling priority and balance the long-term and short-term fairness of user scheduling.Then,an improved power distribution algorithm based on the concave-convex process(CCCP)is proposed,which converts the non-convex problem into two sub-convex problems for solution,and discusses the classification of the solution space in detail.Simulation shows that the proposed algorithm has faster convergence speed and low computational complexity,and can effectively give local optimal solutions to non-convex problems.Compared with pre-grouping algorithm and random algorithm,the fairness of the proposed user grouping algorithm is better;compared with fixed power allocation and hierarchical power allocation,the complexity of the proposed power allocation algorithm is lower,and the system and rate are different in the number of users.The number of sub-channels is better than the comparison algorithm.Finally,for scenarios with a large number of connections,for more sub-channels and cell users,the Deep Q Network(DQN)algorithm is used to implement the resource allocation of the NOMA system.When the DQN algorithm is designed,by adding Mask prior rules before action selection,more user access requirements are met.Then,by adjusting the hyperparameter settings of the DQN algorithm,the algorithm performance is better.Among them,the calculation of the system and rate follows the improved power allocation algorithm based on CCCP with lower complexity.Finally,simulations are performed based on the Python development environment.The numerical results show that the DQN-based NOMA system resource scheduling algorithm has lower computational complexity,has a higher system and rate than traditional algorithms,and can better adapt to multi-connection scenarios.