Research On Multi User OFDM Resource Allocation Optimization Based On Evolutionary Algorithm And Neural Network

Orthogonal Frequency Division Multiplexing is the core technology in the next generation of wireless communication. In order to meeting the emerging demand in the next generation of wireless communication for diverse quality of service(QoS) more and more people began to focus on efficient and intelligent multi-user orthogonal frequency division multiplexing scheduling algorithm. Recently, scholars have expanded the scope of multi-user OFDM systems radio resource allocation problems: 1, ensure fairness criteria; 2, providing the user QoS guarantee; 3, user channel mobility and network dynamic. These three aspects of the multi-user OFDM resource allocation issues raise higher requirements, is also difficult the next generation of wireless communication.For best effort service, proposed a novel water-filling method that effectively and accurately achieves Joint subcarrier and power allocation, and maximizes weighted user rate in the multiuser OFDM channel resource allocation with the static or ergodic fading channel. for non-real-time QoS service, using particle swarm optimization to adjust rate threshold multipliers, and combined with the novel water-filling method obtain maximum weighted user rate under the users’ rate threshold limits.Faced multi-user OFDM resource allocation problem with the ergodic channel, we propose a multi-user OFDM Joint Two-Step Optimization that can solve a non-real-time QoS and a variety of utility functions Using a novel water-filling method to quickly and accurately obtain the Joint subcarrier and power optimal allocation scheme and complete the first step of the model and maximize weighted rate sum. Use combining multi-objective optimization with differential evolution(CMODE) algorithm to optimize the second step model. Maximize sum of the user rate utility value with various types of utility functions under the non-real-time QoS type request.For the traditional linear joint subcarrier and power optimization process, in the multi-user OFDM transmission under static fading channels and best-effort service, using chaotic neural network find the subcarrier allocation optimization. In each step carrier allocation optimization process, though generalized the novel water-filling method optimal power. Strictly in accordance with the optimal power allocation the sub-carrier allocation gets nonlinear neural network dynamics with rich gradient descent search, in order to take full advantage of multi-user diversity gain.