Research On Task Execution Efficiency Optimization Strategy In Mobile Edge Computing Scenario

With the explosive growth of massive low-latency applications on the mobile side,Mobile Edge Computing(MEC),which sinks computing and storage to the edge side,has become a research hotspot.In the MEC scenario,the execution efficiency of a user task depends on the data transfer rate of the task and the computational rate of the task during processing,and the ability to edge cache the required data resources of the task and computationally offload the task is an important element of MEC.Currently,the main problems in the MEC scenario are:(1)limited cache space and processing power of edge nodes;(2)limited computing resources of user devices.Therefore,this paper takes advantage of the large number of edge nodes or user devices in the MEC scenario and adds an inter-device communication(Device-to-Device,D2D)approach to the existing research to propose a multi-node hierarchical collaborative caching strategy and a distributed computational offload strategy to fully exploit the features of MEC.The main areas of work are as follows.First,a multi-node tiered collaborative caching strategy is proposed to address the problem of limited cache space and processing power of edge nodes and underutilized cache space of user devices.Based on the existing MEC collaborative caching framework,this paper divides the edge collaborative cache into local cache domain and collaborative cache domain,taking into account the cache space of MEC edge servers and user devices,and the mobility of user devices.Next,this caching strategy problem that minimizes the average transmission delay of the system is transformed into a nonlinear 0-1 backpack problem,where the proposed objective function is optimized by using an improved genetic algorithm.Finally,detailed parameters and cache effects are verified by setting up simulation experiments.The results of the simulation experiments show that this caching strategy based on the improved genetic algorithm effectively reduces the average transmission latency of acquired data resources relative to the current caching strategy.Second,a distributed computation offloading strategy is proposed to address the problem of long delays in processing computation-intensive tasks due to limited compute resources of user devices,and the current compute offload strategy for multi-user devices in a single cell under the MEC scenario is studied.First,the communication and computational resources in this scenario are modeled separately,and computational cost models are built that take into account both the task processing delay and energy consumption of the user device.Next,by offloading the computational tasks generated by user devices in a single cell to the MEC edge servers and D2 D collaboration devices to which they are connected,an optimization problem with the goal of minimizing the total computational cost of user devices in a single cell is formed,and the optimization problem reaches Nash equilibrium through a game theory approach.Finally,through simulation experiments,the computational offloading strategies of full local computational processing,full MEC offloading processing and random offloading processing are compared respectively,which proves that the proposed strategy has better performance compared with other strategies.