Research On The Resource Allocation Algorithm In One Novel Multiple Access

With the continuous development of wireless communication,the fifth generation(5G)mobile communication networks will face unprecedented challenges such as ultra-high data traffic,large-scale device connectivity,ultra-low latency and ultra-high reliability.As one of the 5G key technologies,new multiple access(NMA)can achieve high system capacity,low latency and large-scale connectivity to solve the above challenges of 5G.With the limited radio resources,reasonable resource allocation is essential to achieve NMA performance improvement.At present,the resource allocation of NMA was widely studied by researchers,and a large number of resource allocation algorithms were proposed.As a typical multi-carrier NMA scheme,pattern division multiple access(PDMA)becomes the most promising air interface technology for the challenges of enhanced mobile broadband and massive machine type communications in the 5G application scenarios.Although current research on PDMA has attracted great attention from researchers,the research on its resource allocation algorithm is still scarce.Therefore,research on resource allocation algorithm of PDMA is of great significance.Based on the above analysis,this thesis carries out the following research:Through the introduction and analysis of the existing resource allocation algorithms,it is found that the key elements of the resource allocation algorithm under various NMA schemes are mostly similar.For example,in user pairing,most algorithms are based on channel gains of users.In terms of power allocation,the closed-form solution of power allocation can be derived by the Karush-KuhnTucker(KKT)condition.Therefore,based on the existing user pairing algorithm and PDMA user pairing characteristics,this thesis proposes corresponding user pairing algorithms for two different scenarios of 5G.In the research of user pairing algorithm,this thesis first proposes the design criteria of the PDMA pattern matrix,then gives an example of the preferred pattern matrix,and simulates the performance of the preferred pattern matrix.The simulation results demonstrate which pattern matrix can provide better performance.In the research of power allocation algorithm,this thesis considers that the base station has two different types of channel state information(CSI)and proposes an iterative power allocation algorithm to maximize the system sum throughput.For the perfect CSI case,the closed-form solution of the optimal power allocation is derived by the KKT condition.Then,the Lagrangian multipliers are updated by the one-dimensional search algorithm and the sub-gradient method,and the optimal power allocation is obtained.For the imperfect CSI case,since the outage throughput of the system is selected as the metric,the optimization problem becomes a probabilistic hybrid problem.To solve this problem effectively,this thesis first converts the probabilistic mixture problem into a non-probability problem and then obtains the optimal power allocation strategy based on the proposed iterative power allocation algorithm.The simulation results show that compared with the existing power allocation algorithm,the proposed iterative power allocation algorithm can obtain better system sum throughput.