Heuristic Algorithms For Task Scheduling In Cloud Computing Environment

Due to the cloud computing was distributed calculation, distributed storage,virtual resource pool, and has some novel advantages of highly expansible, high availability, low cost, the scope of applying cloud computing has been broadening. Cloud computing involves a variety of fields from healthcare to network security. As cloud computing has the tremendous impact to the performance of our lives, scholars have a in-depth study to it. Cloud computing technologies include cloud data storage, cloud data management, task scheduling, task scheduling, programming model as well as cloud computing security. In particular, task scheduling is the one of key technologies of it. Task scheduling is mainly focused on two aspects:task scheduling model and task scheduling algorithm. In view of the characteristics of cloud computing tasks such as large amount of tasks and complex scheduling objectives, many researchers conducted wide and in-depth investigations about this issue and made great achievements, using heuristic algorithm for task scheduling. Derived from goal of task scheduling, this paper proposes two task.soheduling algorithms based on improved heuristic algorithm in cloud computing environment.Firstly, the knowledge of the concept of cloud computing, system services, architecture and key technologies was introduced. Then, the concept, characteristics and objectives of cloud computing task scheduling are introduced, and the advantages and disadvantages of the traditional algorithm are analyzed. A heuristic algorithm is proposed to solve the advantages of this kind of problem. Based on the analysis of several heuristic algorithms, CS and SFLA was chosen as the as special cases. After analyzing the background, the thought and the process of the two algorithms, the advantages and disadvantages of the algorithm are found. Considering the shortcoming of cuckoo algorithm search, which is easy to fall into the local optimal solution, and can not search the global optimal solution in the late stage of evolution, an improved cuckoo search algorithm is proposed, by using the Cauchy mutation operator with strong disturbance ability to improve its late search problem. In the analysis of the basic SFLA local search some shortcomings, an improved algorithm is proposed, in which the double factor learning algorithm is introduced and the local search update strategy is used, improving the searching capability of the algorithm. Then the standard function test is performed based on two improved algorithms. Finally, combined with task scheduling problem description, the coding mode and the fitness function are designed. and two task schedulling algorithms based on improved heuristic algorithm in cloud computing environment are designed deliberately by taking into consideration the characteristic of this problem.In this paper, the proposed task scheduling scheme is simulated on the CloudSim simulation platform, for evaluating the performance of cloud computing task scheduling, made an objective evaluation for the scientific performance of the algorithm. Comparing the experimental results, the ACCS and the DLSFLA perform well in the time span and load balancing, which can meet the design requirements.