Research On Resource Allocation For Energy-Efficient Mobile Edge Computing

In recent years,with the rapid development of Mobile Communication and Internet of Things technology,the number of mobile devices has increased explosively.With the popularity of mobile devices,the types of computing intensive and latency sensitive mobile applications based on mobile devices are also increasing,such as online interactive games,face recognition and virtual reality,which,undoubtedly,brings huge challenges to mobile devices with limited battery capacity and computing power.By providing computing services at the edge of the network,mobile edge computing(MEC)has been considered to be a desirable technology in addressing the above challenges.This technology supports users to upload tasks to the MEC server with rich computing resource at the network edge for processing,after processing,mobile users can download the computing results of tasks from the network edge.Since the task computing process is completed at the edge of the network,and with the development and growth of the network scale,the future wireless network is bound to contain more complex mobile computing,which makes people pay attention to the energy consumption at the edge of the network caused by task offloading.Therefore,how to design an energy-efficient MEC system has become one of the current hot topics.The design of energy-efficient MEC system needs to jointly optimize the resources at the edge of the network.Based on this,aiming at reducing energy consumption,this paper deeply studies the joint optimization of the resources at the edge of the network,and formulates reasonable resource allocation schemes,so as to obtain an energy-efficient MEC system.The main research contents of this thesis can be summarized as follows:1.Resource allocation for energy-efficient parallel mobile edge computing.A multiuser MEC system is considered,which adopts frequency division duplex mode and considers orthogonal frequency division multiple access.Firstly,a parallel mechanism for communication and computation is established.The mechanism includes task input data upload,task execution,and task computing result download.It can support that the uplink/downlink transmission and execution processes of different computing tasks are carried out simultaneously,and capture non-negligible task computing result sizes.On this basis,a nonconvex mixed-integer nonlinear programming(MINLP)problem is established.The goal of this problem is to minimize the energy consumption of the parallel MEC systems.By using the Mc Cormick method and exploring the problem’s structural properties,the original problem is equivalent to a convex MINLP problem,and two low complexity resource allocation algorithms are proposed.Particularly,the first algorithm bases on continuous relaxation and the second one follows from penalty convex-concave procedure.Finally,simulation experiments verify the advantages of the two algorithms in reducing system energy consumption.2.Incentive mechanism and resource allocation for energy-efficient mobile edge computing.In a multiuser MEC system,firstly,in order to encourage idle users to participate in the collaborative offloading of tasks,a total cost minimization problem is established by considering the requirement of the amount of computing resource planned to be collected by the base station.On this basis,a computing resource sharing auction mechanism is proposed to select the winning bids and calculate the rewards for the winning collaborators.Secondly,in order to obtain an energy-efficient MEC system,a nonconvex MINLP problem with the goal of minimizing the system energy consumption is established.By using the classical Mc Cormick method,it can be equivalent to a convex MINLP problem,and a low complexity resource allocation algorithm is proposed.Finally,simulation experiments verify the effectiveness of the proposed mechanism and resource allocation algorithm.