A Study On Task Scheduling In A Cloud-Computing Platform Based On Simulated Annealing Genetic Algorithm

In recent years, with the development of Internet technology, cloud computing has become the research focus in the academic field and commercial field. The center idea of cloud computing is to distribute a large number of computing tasks on the virtual resources pool which is in composed of cheap computer cluster. When a large number of users request for the service of cloud computing resources, cloud service providers need to develop an efficient resource allocation scheme that can make the user has a satisfying experience. So in this huge cloud computing system, how to carry out the reasonable resources organization and task scheduling is one of the key technology of cloud computing. But today's study of cloud computing task resource allocation problem was still in the stage of grope and development, the existing relevant theory on the science and methodology is not too perfect. If the distribution scheduling method is unreasonable, that woulde cause some problems such as the user's needs are not being met and so on. In order to design a reasonable task scheduling scheme to bring efficient experience to users, the task scheduling problem under the cloud computing environment for mathematical modeling in this paper, established a mathematical model based on simulated annealing genetic algorithm (ga). The single thread and independent task scheduling problems in full analysis of the research.In this paper, on the basis of reference to domestic and foreign related research, cloud environment order thread independent task scheduling problems are analyzed. First of all, this paper introduces the concept of cloud computing and the process and characteristics of cloud environment task scheduling. Then, as the goal, with the minimum task completion time scheduling of mathematical programming model is established, and use the calculation time needed for priority task scheduling algorithm of the least Min-Min algorithm simulation analysis, the time needed for completion of a task which is 572 seconds. Finally, according to the advantages and disadvantages of genetic algorithm and simulated annealing algorithm, combine two kinds of algorithms, and the encoding and crossover and mutation operators of genetic algorithm in the traditional genetic algorithm made great improvement, on the basis of the established model of task scheduling based on simulated annealing genetic algorithm, simulated annealing genetic algorithm for task scheduling scheme is finally the time required for completion of the task in 334 seconds. Compared with Min-Min algorithm, simulated annealing genetic algorithm greatly shortens the time of task need, the result confirm the validity of the algorithm. At the same time, the algorithm can also be applied to different targets in task scheduling problem under the cloud computing environment.In the process of concrete simulated annealing genetic algorithm, in this paper, the genetic algorithm the chromosome is divided into two separate substring coding, they represent the task allocation and task scheduling sequence. In the selection, mutation and crossover operation two molecules is operated separately, and then connect the chromosome substring become a chromosome representing an individual or a solution. In addition, in the mutation and crossover operations, in order to make excellent individual gene values are not damaged and make less excellent individuals toward the excellent individual evolution, this paper adopts adaptive crossover probability Pc and adaptive mutation probability Pm: fmax and favg,respectively, for a maximum and average of population fitness.f, for any individual in the population fitness value. k1,k2,k3 and k4 are parameters. After genetic operation to get the new solution, the simulated annealing algorithm, the Metropolis criterion to judge the data processing. Set time to the current solution, newtime to the new solution, T to the current temperature, so the Metropolis criterion is If newtime< time, with probability 1 to accept the new; Otherwise, the probability of accepting the new, abandon the old solution.