Research On Task Offloading Strategy And Computation Resource Allocation Based On Mobile Edge Computing

With the rapid growth of the internet,applications such as data stream processing,augmented reality,and other applications with high requirements for computation resources as well as latency and energy consumption continue to emerge.User Equipment(UE)cannot run these applications well locally due to its physical resource limitations.The emergence of Mobile Edge Computing(MEC)not only provides low-lantency computing services for UE,but also extends the battery life of UEs.However,MEC server usually have limited computation resources,so how to efficiently allocate offloading strategy of UE and computation resource allocation of MEC server has become a hot research topic in MEC.Therefore,this thesis investigates the offloading strategy of task and the computation resource allocation of MEC server under different MEC scenarios,as follows:1.For the single MEC server multi-UE scenario,this thesis proposes a task offloading and computation resource allocation algorithm based on an improved arithmetic optimization algorithm.The task offloading decision of UE and the computation resource allocation of MEC server are jointly optimized.The weighted cost of lantency and energy consumption of task is modeled as the system cost,and the goal is to minimize the system cost while satisfying the time-delayed energy consumption constraint of the UE task.Simulation results show that the algorithm proposed in this thesis can effectively reduce the system cost and bring up to 20% performance improvement compared with the benchmark algorithm.2.For the problem that a single MEC server cannot handle many tasks due to limited computation resources,cooperative computing offloading scenario between multiple MEC servers and cloud server is considered in this thesis.The scenario considers not only the vertical cooperation among UE,MEC servers and cloud server but also the horizontal cooperation among MEC servers.In addition,in order to reduce the delay and energy consumption generated during task transmission,the data compression strategy is considered in it and a three-tier cooperative computing offloading scheme combining data compression strategy is proposed in this thesis.To solve the problem,a task offloading and computation resource allocation algorithm based on an improved genetic algorithm is proposed,with the goal of minimizing the system cost while ensuring the latency and energy consumption constraints of UE tasks.Simulation results show that the proposed scheme can effectively reduce the system cost and bring up to 14% performance improvement compared with the benchmark scheme.