Research On Task Offloading And Scheduling Strategy Supporting Device-edge-cloud Cooperation In Industrial Internet Of Things

In the Industrial Internet of Things(IIo Ts),single cloud computing or edge computing mode can no longer effectively meet the demands of high reliability,low delay and low cost of industrial tasks.The complex and diversified demand scenarios require the three layers of heterogeneous resources in the device-edge-cloud to cooperate closely,cooperate with each other,complement each other,and jointly provide services.Only in this way can each task be completed as required with high efficiency and quality,so as to maximize the application value of all kinds of resources.However,device-edge-cloud collaboration makes the research of task offloading and scheduling more complicated.This paper mainly studies the task offloading and scheduling problems that support device-edge-cloud collaboration in the IIo Ts.The specific work is as follows:Combined with the actual industrial scene,this paper studies a task offloading architecture,which supports the comprehensive collaborative offloading of the three layers of heterogeneous resources in the device-edge-cloud,and the tasks can be selected to be offloaded to other devices in the same factory,the edge or the cloud.Then,from the perspective of task and user,the offloading problem is modeled as a multi-objective optimization problem,aiming at maximizing degree of task completion,minimizing delay and cost.Aiming at the problem of multi-objective task offloading modeled,this paper proposes a BAS-based Individual Heuristic Offloading Algorithm(BIHA).This algorithm is based on the beetle antennae search algorithm from the two directions of position update rules and optimization methods,combined with other heuristic algorithm optimization ideas to improve,the purpose of improvement is to ensure the algorithm with low time complexity and improve the optimization effect.Experiments show that,compared with other heuristic algorithms,the improved BIHA algorithm improves the task completion by 48%,reduces the delay by 9% and reduces the cost by 31% on average.To solve the problem of task scheduling,this paper designs a Priority Quantizing-based Scheduling Policy(PQSP),which is divided into two parts: task priority division rule and task Scheduling rule.The priority assignment of tasks is divided into two stages: the first stage is roughly divided into the first priority and the second priority tasks;the second phase is precise prioritization,which is done only for the first priority tasks.Task scheduling rules can also be divided into two situations according to the classification of priority: the first priority task is scheduled in the form of linked list or Hash list,and the second priority task is scheduled in the form of queue.The results show that compared with the traditional scheduling policy,the PQSP policy can improve the task processing efficiency by 20% on average.