Research On Algorithms For Solving Task Scheduling Problems In Cloud Computing

Cloud computing,as a new computing model,has the advantage of convenient,fast and on-demand network access,which also greatly meets the increasing data computing needs of enterprises and individuals in the digital economy.Task scheduling,which is the act of allocating tasks submitted by users to virtual machine resources,is a crucial issue in the cloud computing environment.The results of task scheduling have a direct impact on customer satisfaction with cloud computing services,in addition to having an impact on the profitability of cloud computing service providers.Given the NP property of the cloud computing task scheduling problem,the mainstream task scheduling methods generally apply heuristic intelligent algorithms to solve.Due to the relatively weak mathematical foundation of heuristic intelligent algorithms,in the process of solving this problem,they often tend to show: prematurity,in the later stage convergence slows down,and solution accuracy reduces.It is worth studying how to obtain a reasonable and efficient task scheduling scheme.This paper takes the application of heuristic intelligent algorithms to solve task scheduling problems and improve the performance of the algorithms as the research entry point,with the following main research work.(1)A single-objective cloud computing task scheduling dual-population TLBO algorithm.The algorithm has two populations and evolves in separate mechanisms,with ecologically meaningful competition implemented between the populations.The ’teaching’operator is transformed into a non-linear adaptive behavior and the ’learning’ operator is combined with brainstorming ideas to improve the accuracy and speed of the algorithm.It also enables the optimal individual to have the behavior of inversion mutation,giving the algorithm the ability to jump out of the local optimum constraint.With task completion time as the optimization objective,the algorithm is applied to solve the task scheduling problem.Experimental results show that the algorithm has a fast convergence speed and good solution accuracy,and obtains a task scheduling scheme with a short task completion time.It is trustworthy.(2)A multi-objective cloud computing task scheduling Hybrid HHO-TLBO algorithm.The algorithm blends the ideas of TLBO and HHO to give individuals the behavior of learning from the optimal individual and the population mean.And based on probability through t-distribution hybrid variation to obtain new individuals to get out of local distress.All phase replaces the prey in the standard algorithm with a multi-elite weighting mechanism.Apply it to the cloud computing task scheduling problem.Considering task completion time,task completion cost,and virtual machine load balance,to design the fitness function.Experimental results show that this algorithm not only reduces the cloud task completion time and has a more balanced workload of virtual machines,but also has an appropriate cost consumption.