摘要： 任务调度是网格研究中的NP问题。在研究现有任务调度算法的基础上，分析小生境遗传算法的优点，引入种群排序三原则，优化种群结构，提出基于改进小生境遗传算法的网格任务调度方法，并运用GridSim工具进行仿真实验验证。实验结果表明，该方法结果精度和收敛速度优于普通遗传算法，能降低资源占用率。

关键词: 任务调度, 网格, 小生境遗传算法, 种群排序, 多峰值优化

Abstract: Task schedule is a NP problem on grid research. In this paper, Niche Genetic Algorithm(NGA) is used to be the grid schedule algorithm combined with its advantages after anglicizing current schedule algorithms and introducing the method of arraying species with three principles. Experimental results with software GridSim show that NGA has better result, convergence rate and takes less resource than genetic algorithm.

Key words: task schedule, grid, Niche Genetic Algorithm(NGA), species arraying, multi-peak optimization

中图分类号: 

• TP393.01