User Grouping And Energy Consumption Optimization Methods In NOMA Based MEC Systems

The increasingly diversified multimedia services bring challenges to wireless communication networks in terms of capacity and number of access users,especially traditional wireless access networks,which is limited by the orthogonal multiple access method,restricting the number of user accesses at a single base station and making it difficult to meet the performance requirements of the Internet of Everything.Non-orthogonal Multiple Access(NOMA)technology based on power domain exploits the power difference between mobile devices as a non-orthogonal resource to differentiate users,which increases the number of access users while ensuring quality of service.In addition,mobile devices have two distinctive features:firstly,they are energy constrained.Mobile devices are very sensitive to power consumption,especially for wireless sensors,which require periodic forced energy replenishment to maintain normal operation.Secondly,the computing power of mobile devices is limited.Limited by the size and cost of the device,the computing power is weak and the power consumption is limited.Mobile Edge Computing(MEC)technology can improve the computing ability of edge devices,reduce user processing latency,and increase access network capacity by deploying high-performance computing servers at the edge of mobile networks.NOMA-based MEC system(NOMA-MEC)combines NOMA and MEC to increase the number of access users,improve network capacity,reduce end-to-end latency,and improve the energy efficiency of wireless access networks.Considering that user grouping has a significant impact on NOMA performance,and how to optimize the overall network energy efficiency is the focus problem of NOMA-MEC,this thesis analyzes the impact of grouping strategy and offloading decision on performance and investigates the selection of user access modes in the network to optimize network energy efficiency.The main work and innovation points are summarized as follows:1.A user grouping method based on greedy algorithm is proposed.For the problems of high complexity of traditional exhaustive search algorithm and low performance of random grouping algorithm,a user grouping algorithm based on greedy algorithm NOMA-MEC is proposed considering the channel variability among large scale different users.Simulation results show that the proposed greedy algorithm-based grouping strategy can achieve about 12%increase in system capacity when the maximum user transmit power is higher than 20dBm compared with the exhaustive and random grouping strategies.2.The theory and algorithm of energy consumption minimization of NOMA-MEC are studied.Based on the above user grouping method,an optimization model for minimizing system energy consumption under power,offloading task,and delay constraints is constructed for the resource-constrained scenario.Considering that the model is a non-convex problem and it is difficult to obtain the optimal solution directly,by using the augmented Lagrange method,the algorithm of finding suboptimal solution is analyzed and given.Meanwhile,for the problem of high complexity of the Augmented Lagrange Method in the large-scale user scenario,the original non-convex problem is decoupled into two subproblems by using the Stackelberg game principle,and a resource allocation method based on many-to-one matching is proposed to reduce the algorithmic complexity of the solution scheme.Simulation results show that the energy consumption of the system based on the proposed algorithm is about half of that of the conventional system based on OFDMA-MEC when the user task volume is higher than 1.5 Mb.This thesis will provide significant theoretical research value and application prospect for supporting multi-user low-power access under wide coverage of wireless communication,and it will also promote the development and maturity of NOMA-MEC.