Research On The Task Scheduling Strategy Based On Multi-Objective Optimization In Cloud Environment

With the development of communication technology and Internet technology,cloud computing came into being.It connects low-cost computing units over the network to form cloud computing systems that provide computing and storage services for businesses and individuals.And it saves the cost of resources for the whole society while providing unlimited storage and computing services.In the cloud computing technology,task scheduling is an inevitable and very important part.The cloud computing task scheduling determines the system operating efficiency and customer satisfaction.So the cloud computing task scheduling strategy has been the hot spot is difficult to study.At present,the task scheduling mechanism of cloud computing has not yet formed a unified standard and specification.Many scholars,according to the characteristics of cloud computing task scheduling,from one of the indicators such as makespan,optimal span,cost,reliability,energy consumption Optimize the goal to do the study.It is not mature enough to study the scheduling strategy with multiple targets as the index.Especially the makespan,load balancing,service costs together as a constraint target research less.In addition,cloud computing task scheduling problem is a NP-Hard problem,and ant colony algorithm in dealing with the problem has a very good effect.So this thesis,based on the study of the ant colony algorithm and the task scheduling model,proposes a scheduling strategy which takes into account the makespan,cost and system load balancing.This thesis first studies the important technologies related to cloud computing,analyzes the basic task scheduling model,and learns and the existing task scheduling algorithms.According to the characteristics of cloud computing task scheduling and the current situation of the current study,the makespan,cost and system load are selected as the target conditions for scheduling optimization,and the related objective functions are defined for each target.Then,the two objective optimization problems of task completion time and cost are transformed into single objective optimization problem.Secondly,we study the intelligent algorithm,especially the ant colony algorithm.The heuristic function,pheromone initialization and the pheromone updating rules of the basic ant colony algorithm are improved,and the pseudo-random transition probability rule in the ant colony system is adopted in the probability transfer strategy.Then,the objective functionis integrated into the improved ant colony algorithm.In order to balance the system load,the load factor is defined and integrated into the heuristic function.Finally,a multi-objective optimization cloud computing task scheduling strategy based on improved ant colony algorithm is proposed.In addition,the parameters in the algorithm were studied and analyzed to determine the approximate range of parameters.Finally,on the Cloudsim platform,a simulation experiment was carried out.The validity of the proposed algorithm is verified.