Research On Computation Offloading Strategy Based On Cooperative Computing In Mobile Edge Computing Networks

The real-time communication and computation of massive wireless devices(WDs)promotes the rapid growth of a series of emerging applications.In practice,the real-time computation tasks to be executed may be quite intensive,but WDs are generally of small size and only have limited computation capacity.Therefore,how to improve their computation capabilities and reduce the computation latency is a crucial but challenging problem to be solved for realizing these applications.Fortunately,such limitations can be tackled in mobile edge computing(MEC),where strict latency requirements can be guaranteed by deploying MEC servers at the network edge.However,the future wireless networks are expected to serve massive devices.It may not be feasible if all data are offloaded to MEC server.To tackle this problem,cooperative computing has been proposed as a promising solution.However,considering that the helpers have no computing task,the research works involving cooperative computing usually ignore computation frequency resource optimization and results downloading,which is unfavorable for the practical engineering implementation.And in engineering practice,the helpers usually have their own computing tasks.At present,only a few research works have considered how to simultaneously meet the computing requirements of the user and helpers.However,they ignore the user can simultaneously exploit the computation resource of multiple helpers,which indicates that the MEC performance can be further improved.Hence,this paper mainly focuses on the computation offloading strategy based on cooperative computing in the two cases of whether the helpers have computing tasks.The specific contents are as follows:1)Computation offloading strategy based on cooperative computing when the helpers have no computing tasksConsidering that the helpers have no computing task,aiming at how to effectively coordinate the cooperative computing between the user and the helpers,this paper proposes a novel cooperative computing approach in the multi-user cooperative MEC system,in which multiple nearby helpers share the computation and communication resources to actively help the user.By considering an orthogonal frequency-division multiple access(OFDMA)-aided three-phase transmission protocol,we design an energy-efficient cooperative computing framework.Under the user’s computation latency constraint,we optimize the user’s task partition,jointly with the communication and computation resources allocation for computation offloading and results downloading,so as to minimize the system energy consumption.Based on convex optimization method,we propose an effective algorithm to obtain the globally optimal solution.Simulation results show that the proposed joint task partition and resource allocation(JTPRA)scheme has advantages over other benchmark schemes.2)Computation offloading strategy based on cooperative computing when the helpers have their own computing tasksConsidering that the helpers have computing task,aiming at how to efficiently meet the computing requirements of the user and helpers at the same time,this paper proposes a novel joint communication and computing resource sharing approach in the base station(BS)-aided multi-user cooperative MEC system,in which the helpers and the BS integrating with a MEC server share the computation and communication resources to actively help the user and helpers,respectively.By considering an OFDMA-aided partial offloading protocol,we design an energy-efficient cooperative computing framework.Under the common latency constraint,we jointly optimize the task partition and resources allocation of the user and helpers,so as to minimize the system energy consumption.Based on convex optimization method,we propose an effective algorithm to obtain the globally optimal solution.Simulation results show that the proposed cooperative computing scheme with BS has advantages over other benchmark schemes without joint consideration.