基于遗传编程和效用最优的网格资源调度及仿真  

Grid Resource Scheduling Algorithm and Simulation Based on Genetic Programming and Utility Optimization

作　　者：王忠群[1] 李钧[1] 刘涛[1] 王勇[1] 

机构地区：[1]安徽工程科技学院计算机与科学工程系,安徽芜湖241000

出　　处：《系统仿真学报》2008年第16期4442-4445,共4页Journal of System Simulation

基　　金：安徽省自然科学基金支持项目(070412058);安徽教育厅自然科学基金重点项目(2006KJ016A)(2005KJ065)。

摘　　要：由于因特网的开放、动态性,传统的计算资源调度方法已不再适用网格计算。基于经济模型的网格资源管理和调度成为研究热点。针对计算市场模型中非线性效用最优化问题,提出了一种基于遗传编程改进的效用最优的网格资源调度算法。该算法使用遗传编程构造计算市场模型中的效用函数,使得计算复杂度控制为O(n)。仿真结果表明该算法可以提高网格计算中的资源调度性能。Because of openness and dynamics of Intemet traditional resource management and scheduling algorithms have no longer been valid in grid computing, and now grid resource management and scheduling based on economic model is a research hotspot. Aiming at the optimization of non-linear utility functions in computational market model, a computing resource scheduling algorithm based on genetic programming and utility optimization for grid computing was proposed. This algorithm constructs utility functions in computational market model by means of genetic programming and its computational complexity is O（n）. The experimental simulations show that this algorithm can improve the performance of grid resource scheduling.

关 键 词：网格 效用最优 遗传编程 资源调度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]