Research On Dynamic Priority Task Scheduling Of Smart Grid Based On Cloud And Fog Computing

With the continuous expansion of the scale of the power grid and the continuous improvement of the intelligence of the power grid,the automatic task scheduling architecture of the power grid that only considers cloud computing can no longer meet the needs of the smart grid for fast task data processing.Cloud-fog collaborative computing technology can sink the server layer closer to the terminal,which can effectively meet the requirements of smart grid for fast task data processing.Considering the difference in computing performance of different servers,how to make full use of each server resource and make the server network load distribution more balanced under the premise of meeting the task deadline has become an urgent problem to be solved in the cloud-fog collaborative computing environment.Therefore,this paper comprehensively considers the task processing delay,load balancing and total task set completion value in the cloud computing environment,and studies the unloading strategies of independent static priority tasks and dynamic priority tasks with deadline constraints.Firstly,combined with the respective advantages of cloud and fog computing,the cloud and fog computing technology is introduced into the smart grid dispatching automation system,a three-layer distributed grid dispatching automation architecture based on cloud-fog-end interaction is designed,and network resource management is studied from the aspects of scalable network topology and flexible task orchestration,and two fog computing network resource scheduling schemes are proposed for different types of power tasks.That is,a static task priority scheduling scheme considering resource prediction and a real-time task scheduling scheme considering task deadline based on dynamic priority based on load balancing.Secondly,aiming at the problem of high scheduling delay of static priority tasks in smart grids in fog computing environment,according to the task set information monitored by the fog layer management node at the initial time and the performance parameters of the server,a greedy algorithm is used to assign the initial task set to the task set to be processed.On this basis,considering the changes of server performance parameters,the quadratic smoothing index model is used to predict them,and the task unloading decision is updated according to the predicted results.The simulation results show that compared with the advanced first-out algorithm and greedy algorithm,the virtual resource prediction algorithm based on quadratic smoothing index can obtain better task offloading decisions and have better performance in terms of delay.Finally,aiming at the problem of complex task environment and frequent changes in task arrival rate of smart grid under cloud-fog-device three-layer architecture,a task offloading strategy under dynamic task priority considering task deadline is proposed.Considering that the computing performance of different server nodes is different and the task execution effect is obviously different,in order to avoid the waste of server resources,a dynamic task unloading optimization model considering load balancing,with the minimum and maximum task completion value of the task set comprehensive scheduling delay as the optimization goal,considering the task deadline,according to the dynamic priority of the current server computing resources and the changes of each task,the tasks are reasonably allocated in real time,and the improved ant colony optimization algorithm is used to solve the optimization problem.The simulation results show that the dynamic task offloading strategy proposed in this paper can effectively improve the load balancing degree and on-time task completion rate of server network.