一种基于神经网络的网格实时调度算法

Grid real-time scheduling algorithm based on neural network

提出一种在指定的最终期限内,利用队列技术来模拟调度动态资源,构建一个数学神经模型调度应用的子任务,使用GridSim工具测试的调度算法,通过大约90%的模拟实验说明了模型调度任务是成功和高效的。

Within specified deadlines, this paper modeled performance dynamism of resources with the help of queuing techniques. Employed a mathematical neural model afterwards, to schedule subtasksof the application. Implemented the.proposed model on GridSim toolkit to evaluate the performance of scheduling algorithms. Simulation experiments show that in approximately 90% of cases,this model schedules tasks successfully and efficiently.