Research On Maximum Sum Rate Of Multi-carrier NOMA System

With the rapid development of mobile communication and the continuous evolution of technical standards,the fourth-generation mobile communication technology has been unable to meet the increasing needs of people.In order to meet the challenges of spectrum resources,mobile access technology and network architecture,the fifth-generation mobile Communication technology emerged at the historic moment.Non-orthogonal multiple access technology(NOMA)is one of the key 5G technologies.This technology uses channel coding and power allocation technology at the transmitting end to allow multiple users to share time-frequency resources,and the receiving end uses serial interference.The cancellation technology demodulates the received signal,so a reasonable and reliable channel and power allocation algorithm can ensure effective decoding at the receiving end and increase the total rate of the system.Under the premise of ensuring user fairness,research how to maximize the sum rate of the multi-carrier NOMA system.The main work of this paper is as follows:1.Based on the traditional channel and power allocation algorithm,a step-by-step optimization idea is proposed.First,a stable matching grouping algorithm is proposed.A preliminary grouping scheme is selected according to the equivalent channel gain,and then the final grouping scheme is determined according to the sum rate.User grouping scheme,the second step is to propose a power allocation algorithm based on genetic algorithm.A certain number of users form a population.The power is composed of chromosomes.The total speed of the NOMA system is used as a fitness function.Through continuous selection and crossover,Mutation obtains the optimal power allocation matrix under this condition.Simulation results show.On the basis of ensuring that the complexity is not increased,the algorithm effectively improves the total system rate.2.Based on the popular deep reinforcement learning framework,an attention-based deep reinforcement learning resource allocation algorithm is proposed.First,the optimization objective function is established to maximize the sum rate and maximize the minimum rate,and first derive the given channel conditions,The closed solution of optimal power allocation,and then a deep reinforcement learning framework based on attention neural network is proposed to solve the channel allocation problem.The neural network uses an attention frame of encoder-decoder structure,and the encoder output In all state spaces,the decoder outputs the probability distributions of all states,combined with the derived power allocation strategy,to obtain the optimal channel and power joint allocation strategy.The simulation results show that the proposed joint channel and power allocation algorithm has high system performance.