Research On The Cloud Computing Task Scheduling Strategy Based On Multidimensional QoS Constraints

Cloud computing task scheduling strategy directly affects the efficiency of the whole cloud platform resource usage and cloud platform user satisfaction,meanwhile,cloud computing task scheduling is a completely NP problem,so the task scheduling algorithm has been a research difficulty and hot spot in the cloud computing field.Currently,cloud computing task scheduling algorithms are mainly divided into traditional algorithms and intelligent heuristics.Traditional algorithms are more inclined to single-indicator optimization.Intelligent heuristic algorithms,from the cloud platform perspective,optimize task completion time,task completion cost,etc.;from the user perspective,improve user service quality;however,these algorithms do not consider user service quality and cloud platform load in a comprehensive manner.Ant colony algorithm has good effect in dealing with complex problems.By conducting an in-depth study on ant colony algorithm and cloud computing task scheduling model,and considering both user QoS demand and cloud platform load balancing,we propose an improved cloud computing task scheduling strategy based on QoS constraint for ant colony algorithm.First,the task completion time model and execution cost model are constructed,and the QoS integrated benefit function is created by the completion time model and cost model,and different weight factors are used to indicate the preference degree of user QoS demand on task completion time and cost,respectively,and the cloud platform load balancing model is defined.Secondly,the ant colony algorithm is improved to dynamically change the pheromone intensity so that the pheromones at different stages are different,thus enabling the algorithm to achieve a dynamic balance between exploration and merit seeking.Using the ant colony system path selection formula,a random factor is added to the path selection to effectively avoid the ant colony algorithm from falling into local optimum.The expectation heuristic function is improved to give priority to the virtual machine with fast computation speed and low load to perform the task,which can reduce the task completion time and ensure the load balance of the system.The pheromone update method is improved to update the pheromone locally to ensure the path pheromone in real time,and to update the pheromone globally to increase the pheromone difference between the optimal path and other paths,so as to improve the convergence speed of the algorithm.Finally,a reasonable parameter setting range is found by simulating the relevant parameters in the improved ant colony algorithm.Simulation experiments are conducted through the Cloud Sim platform to compare four aspects: task completion time,task completion cost,system load averaging and user satisfaction.The experimental results show that the scheduling strategy in this paper improves user satisfaction and improves the cloud platform load situation.