Research On Task Scheduling Strategy In Cloud Computing Based On Improved Genetic Algorithm

Task scheduling issue has always been a hot topic and also a difficult one in this cloud computing field.The strategy of task scheduling is the key point that affects the efficiency of computing task;it's aimed at efficient and rational allocation of resources which will assign to cloud tasks,so as to realize the optimization of scheduling goals.In recent years,some of artificial intelligence methods' applications on scheduling problem have raised widespread concern,among that;the genetic algorithms because of its strong robustness and global search fast convergence characteristics has been applied very efficiently to solve the scheduling problem of cloud problem.In the thesis,deeply analysis cloud computing task scheduling and then after a profound study of tasks scheduling algorithms in the cloud computing,aiming at the flaws and backwards,proposed an adaptive genetic algorithms to solve the scheduling problems of different kinds of task models.The main work are listed as below:(1)The theoretical knowledge and scheduling architecture of cloud computing task scheduling is studied.Several classical scheduling algorithms are analyzed and studied.According to the shortcomings of these algorithms,a task scheduling algorithm is proposed in this thesis.(2)For the independent task scheduling problems in the cloud computing,the tasks all unrelated to each other,so the process of task assign to the virtual machine can be seen as a multi-objective optimization problem.Then,an approximate optimal solution of this problem is presented by using the improved genetic algorithm in this paper.The algorithm uses the weighted round robin algorithm to initialize the population adaptive genitive algorithm can be used to search the global approximate optimal solution through its iteration.The algorithm uses a weight polling algorithm to initialize the chromosome population,which can ensure that the algorithm has a high quality initial population and to improve the efficiency of the algorithm.The re-selection strategy is added to speed up the optimization convergence speed of the algorithm and improve the efficiency of the algorithm.(3)As for the associated task scheduling problem in the cloud computing,there will be some correlation between tasks,so the DAG diagrams are used to describe this task.An improved adaptive genetic algorithm is proposed for the defects of the existing genetic algorithm in the initial population.The initial population is constructed by using the priority of task which makes full use of the parameter information of the task,so that the initial population is effective.(4)On the CloudSim platform,the performance of the two scheduling algorithms proposed in this paper is verified.Through comparison and analysis of experimental results,the proposed scheduling algorithm can solve the task scheduling problem,and improve task scheduling efficiency.(5)The research results of this thesis are applied to a project of explosion protection equipment testing service cloud platform cooperated with a research institute in Tianjin to verify that they are effective in the practical computing environment.