Research On Computation Offloading Methods For D2D-MEC Network

In recent years,more and more computationally intensive mobile applications with advanced functions have been developed,including interactive online games,object recognition and voice control.Due to the scarcity of resources and limited energy of mobile devices,mobile edge computing(MEC)is considered a promising technology to enhance the cloud computing capabilities of mobile devices at the edge of the network.MEC is given cloud computing capabilities,so that users can offload computing tasks to nearby edge servers for remote execution to improve energy efficiency and reduce latency.However,considering the limited computing power of the MEC server and possible resource competition issues,D2 D communication technology can be used to improve the offloading performance of the MEC network.However,the current research work rarely considers the energy consumption of multi-users equipment in the D2D-MEC network and the task execution cost.Therefore,this paper studies the computation offloading methods for D2D-MEC networks from the following two perspectives.(1)The computation offloading problem of minimizing energy consumption in D2D-MEC network is studied.A D2D-enabled MEC framework is considered,in which user equipment can offload its computing tasks to edge servers or nearby auxiliary devices.The goal is to minimize the total energy consumption of all user equipment tasks,while optimizing computing offloading decisions and resource allocation.First,the energy saving computation offloading problem is formulated as MINLP problem.And then,by combining the merit of genetic algorithm(GA)and particle swarm optimization algorithm(PSO),a suboptimal algorithm(LGAP)based on layered GA and PSO is designed to solve this problem.(2)The computation offloading problem of joint user decision-making and resource allocation in the D2D-MEC network is studied,and formulates the weighted sum of task excution delay and energy consumption as the problem of minimizing task execution cost.The goal is to minimize the users' task execution cost.Since this problem is a MINLP problem,it is inconvenient to solve.This paper proposes a heuristic algorithm to solve the computational offloading sub-problem and the MEC resource allocation sub-problem in turn by comparing the task execution costs of the three offloading modes and using the Lagrangian dual method.(3)In this paper,experimental verification and analysis of the proposed task execution cost model and minimizing energy consumption model are carried out.The simulation experiment corresponding to the task execution cost model verifies that the method can diminish the overall task execution cost compared with the traditional MEC computation offloading research work.In the meantime,results show that the proposed method improves the bandwidth and computing resource allocation efficiency in the MEC server.The energy minimization model studies the convergence of the algorithm through experimental simulations,and compares with other baseline algorithms to verify the performance of the algorithm.LGAP is proved that can lessen the energy consumption of excution tasks.