混合云计算环境下利用AMPL建模的应用程序计算开销最小化

On Computation Overhead Minimization of Application Program by AMPL Modeling in Hybrid Cloud Computing Environment

针对基于多云的任务规划问题,为了最小化应用程序计算开销,提出了一种基于AMPL语言的建模方法.分析了混合整数非线性规划问题,为应用程序最小化定义了输入数据、辅助参数和变量,利用AMPL语言建模将问题公式化,从而最小化任务规划的计算开销.最后通过仿真实验验证了本文方法的有效性及可靠性.

For the problem of task planning on multiple clouds, a method based on AMPL modeling lan- guage has been proposed to minimize computation overhead of application. Firstly, the mixed integer non- linear programming problem has been analyzed. Then, the input data, auxiliary parameters and variables have been defined. Finally, AMPL modeling language has been used to formulize the problem so as to minimize the computation overhead of task planning. The relationship between the time and cost has been calculated for data samples of data-intensive tasks and computationally intensive tasks in experiments, and sensitivity, impact of dynamic environment is analyzed, experimental results show the effectiveness and reliability of proposed method.