基于遗传编程的网格资源调度算法

网格资源调度是一个非常重要的研究课题.由于因特网的开放、动态性,传统的资源调度和分配方法已经不再适用网格计算,基于经济模型的资源管理和调度成为研究热点.在计算市场模型中,构造有效的效益函数又是提高算法性能的关键.有关文献中采用的是线性效益函数,虽然降低了复杂度,但不能很好地反映用户的效益.文中提出了基于遗传编程来寻找和构造非线性效益函数的方法,并将其应用到网格调度算法中.实验结果表明该算法可以提高网格中的资源调度性能.

Grid Resource Scheduling Algorithm Based on Genetic Programming

Management and scheduling is a very important problem in grid computing. Because of openness and dynamics of Intemet trsditional resource management and scheduling algorithms have no longer been valid in grid computing, and now resource management and .scheduling based on economic model is a research hotspot. In computational market model, constitution of benefit function is key to improvement of scheduling algorithm. The relevant literatures present linear benefit function, it reduces complexity, but it can not reflect benefit of users. This paper proposes a method for finding and constructing benefit function, based on genetic programming, and it is used in grid scheduling algorithm. The experimental results show that this algorithm can improve performance of grid resource scheduling.