Research On Offloading Strategy Of Mobile Edge Computing Task For Industrial Internet Of Things

With the growth in the amount of data in the industrial internet of things(IIoT)and the increase in the intelligence of terminal equipment,the traditional centralized cloud computing has been unable to provide low-latency,low-energy,and high-efficiency services.Mobile edge computing(MEC)can provide computing services closer to terminal devices,thereby reducing task transmission delays,reducing network congestion,and improving the service quality.However,there is a performance bottleneck in the computing resources of a single MEC server,and how to efficiently offload numerous tasks to different MEC servers for processing is an urgent problem.Thus,this thesis investigates task offloading strategies in line with industrial application scenarios based on the state of industrial devices in the IIoT environment.The main work is as follows:1.In the IIoT big data processing scenario combined with artificial intelligence,aiming at the problem that the limited computational resources of MEC servers cannot meet the demand of task models for all industrial devices,a Q-learning based mobile edge computing offloading strategy for IIoT is proposed.Firstly,an IIoT-oriented MEC framework is built to deploy a single artificial intelligence model in a single MEC server,and a software defined network is used to monitor MEC server information.Secondly,by assigning different weight coefficients to delay,energy consumption,and load balancing,an offloading utility function is constructed to transform the task unloading problem into a utility minimization problem.Finally,Q-learning is used to select the offloading location and offloading path adaptively according to the task type and utility function.The experimental results show that,compared with all cloud offloading and the benchmark offloading strategies,the proposed offloading strategy reduces latency and energy consumption,and achieves load balancing.2.In the scenario of large-scale real-time tasks generated by industrial equipment,aiming at the problem of task processing latency and energy consumption growth caused by limited resource competition,a priority task based mobile edge computing offload strategy for IIoT is investigated.Firstly,the analytic hierarchy process is used to prioritize tasks based on latency requirements,task types,data size,and computational resource requirements for determining the offloading platform of tasks.Secondly,by assigning different weights to offloading latency and energy consumption,an optimization function is constructed to transform the task offloading and resource allocation problem of MEC servers into a total cost minimization problem.Finally,the optimization function is solved using Q-learning methods for task offloading and resource allocation.The experimental results demonstrate that the proposed strategy achieves better results in the system of latency,energy consumption,and task completion rate compared with the fully local,fully MEC,and no priority consideration offloading strategies.