Research On Computation Offloading Of Edge Computing For Internet Of Things

With the explosive development of the Internet of Things(Io T),enormous devices are connected through Io T techniques,and these Io T devices will generate massive amounts of data and demand.As resource-constrained Io T devices cannot meet the increasingly complex computing requirements,the traditional cloud computing paradigm can offload computing tasks to a cloud center for execution.However,cloud centers are mostly far from Io T devices,and the resulting high latency is unacceptable for some latency-sensitive Io T services.As an emerging and promising computing paradigm,edge computing is considered a key technology in the Internet of Things.Io T devices can achieve low-latency requirements by offloading tasks to edge computing nodes near the edge of the network.Most of the existing computation offloading works assume that computing tasks are independent.That is,computation offloading with inter-task dependency relationships,especially the task dependency among various devices,have seldom been considered and addressed.This kind of inter-task dependency is ubiquitous in the Io T environment,and this kind of task dependency relationship between devices will have a great impact on the strategy of computation offloading.In view of the above problem,this thesis considers the scenario where exist inter-task dependency relationship among various Io T device and studies the task computation offloading problem in this scenario.The main research work of this thesis is summarized as follows:1.Taking inter-task dependency and service completion time constraint into consideration,this thesis formulates the computation offloading strategy problem as a mixed integer optimization problem on the cloud-edge integrated computing framework.Since energy consumption is critical for most Io T devices,the goal of the optimization problem is to minimize the total energy consumption of Io T devices.Then offloading algorithms are designed to solve this problem in two different scenarios.2.In the static scenario,this thesis proposes an energy-efficient collaborative task computation offloading algorithm(ECTCO)to solve the optimization problem.The algorithm obtains computation offloading decisions through semidefinite relaxation approach and probability-based stochastic mapping method.Simulation results show that in the inter-task dependency scenario,the proposed ECTCO algorithm performs well in terms of energy consumption.Also,the performance evaluations verify the effectiveness and the adaptability of the proposed algorithm under different system parameters.3.In the dynamic network scenario,due to the dynamic change of wireless transmission rate,the algorithm proposed in the static scenario cannot obtain an effective offloading strategy.Therefore,this thesis proposes a dynamic offloading scheduling algorithm based on reinforcement learning(RDOS).Finally,by analyzing the results of simulation,the effectiveness and adaptability of RDOS in reducing energy consumption in dynamic networks is verified.