Research On Task And Resource Scheduling Method Based On Ant Colony Algorithm Of Cloud Computing

Cloud computing can be seen as a new service model that provides computing service to the customers by putting all the applications and data files stored in the cloud side through a collecting of tens of thousands of physical machine resources working together.As an emerging distributed system,it follows the "on-demand pays" business model.From the user level,they are more concerned about the total costing time from submitting an assignment to the results returning.,From the point of the cloud service provider,how to trade off the energy consuming and the task completing efficiency is what they care of.The former task is the mapping of virtual machine,while the latter one is a mapping virtual machine to physical one problem.Two algorithms will be proposed respectively to solve the two problems in this paper.According to the mapping problem from tasks to virtual machine we introduce a load balance task scheduling method based on the ant colony algorithm,the main idea is to let the cloud environment of virtual machine reach a load balance condition to maximize output.Against disadvantages of lack of initial pheromone of ant colony algorithm,the introduction of MIN-MIN scheduling pheromone initialization,thought to solve the problem of long search time.Secondly by introducing LBF load balance factor,combined with global and local pheromone update coordination,make ant colony can real-time when search the solution space based on iteration load,through the pheromone update mechanism affect the choice of the next round of decisions.Algorithm is introduced in the late volatilization coefficient adjustment mechanism,in order to improve the global search ability of algorithm.In this paper,the algorithm and the new task scheduling algorithm is put forward in recent years,as well as the scheduling algorithm commonly used a large amount of experimental contrast.The experimental results show that the algorithm in addition execution time,task queue response time,and a cloud environment virtual machine load balancing degree aspects have significant advantages compared with other algorithms.According to the virtual machine to the physical machine mapping problem,this paper puts forward the virtual machine migration algorithm based on improvedant colony system.On the one hand,through the multidimensional measure of the objective function to regulate and optimize the process of ant colony search,the load forecasting mechanism are introduced to build the migration queue,and build the independent distributed monitoring system to collect and deal with the load information,cloud platform to trigger and perform the migration strategy.On the other hand,the introduction of random variables and search the sliding window mechanism,improve the efficiency of the algorithm in the solution space search and quality.Finally from the two dimensions of resource scheduling and task allocation,to avoid a lot of algorithms focus on only one dimension of the scheduling results respectively.In this paper,the reinforcement and the algorithm and the commonly used virtual machine migration algorithm made a lot of experimental contrast.The experimental results show that the algorithm in the total resource consumption,the SLA service hierarchy levels and virtual machine migration of total number of aspects have significant advantages compared with other algorithms.The experimental results show that through the scheduling algorithm with two levels of the benign interaction,the algorithm greatly improve the efficiency of the cloud platform,the user task response speed,completion time and the energy consumption and so on several aspects,weigh the sex is better than other algorithms,this algorithm has good performance.