Research On Task Scheduling Based On Cross Entropy Method In Cloud Computing Environment

With the increasing development of Cloud Computing, task scheduling problem as a crucial aspect of Cloud Computing also get a lot of attention and development. However, there are still some problems existing in cloud scheduling tasks, such as, the convergence speed of the scheduling algorithm is slow, or the task scheduling target is single that is not taking into account the total execution time and total cost of tasks at the same time. In order to solve these problems, the cross entropy method, which is a method for solving combinatorial optimization problems, has been used in task scheduling of Cloud Computing. The following several aspects have mainly been done.(1) Built a hierarchical model of task scheduling in Cloud Computing by CPN tools, thereby the cloud scheduling process was clear, which can help to properly design objective functions which is used in task scheduling algorithm based on cross entropy method in Cloud Computing.(2) Established a mathematical model of Cloud Computing task scheduling problem, then designed three different kinds of objective functions:the shortest total execution time of tasks, the lowest total cost of tasks, and the relatively lowest of the total execution time and the total cost of tasks.(3) Solved the task scheduling problem with cross entropy method to meet three different objective functions in CloudSim simulation environment. Then solved the same cloud task instance respectively used FIFO method and greedy algorithm. According to the comparison of the experimental results, highlighted the advantage of solving task scheduling problem with the cross entropy method in Cloud Computing.