机构地区: 铜陵学院数学与计算机科学系
出 处: 《计算机仿真》 2009年第4期240-243,共4页
摘 要: 传统的资源调度算法在网格环境下存在一定缺陷,如不能很好地平衡资源节点的负载,不能很好满足用户服务质量需求等。为了提高网络质量,应用遗传算法全局快速收敛的优点,将遗传算法融入到蚁群算法的每一次迭代中,使之具有很强的全局搜索能力,以加快算法的收敛速度,提出了在价格机制驱动下,应用蚁群遗传算法进行网格资源调度的算法。仿真实验结果表明,在价格机制驱动下,应用蚁群遗传算法进行网格资源调度可以减少系统总执行时间和任务完成时间,系统负载均衡度好,提高了资源调度效率,在网格环境下,算法具有稳定性和高效性。 Traditional resource scheduling algorithm in the grid environment has some defects, such as the balance of resource node load, can't be achieved, and the demand for quality of service can't be met and so on. A new hybrid algorithm combining ant colony algorithm with genetic algorithm is proposed. The global fast convergence of genetic algorithm is utilized to combine ant colony algorithm with genetic algorithm in each generation, which makes the new algorithm have a strong global searching capacity, enhances the convergence rate and improves the efficiency. The paper proposes methods of grid resource scheduling, based on ant colony - genetic algorithm driven by the price mechanism, and the emulation experiment indicated that driven by the price mechanism, the application of ant colony -genetic algorithm in grid resource scheduling can reduce the executing time and task completion time, and further improve the scalability of the resource scheduling model. This algorithm has stability and high efficiency in grid environment.
关 键 词: 网格计算 蚁群遗传算法 资源调度 计算经济 融合
领 域: [自动化与计算机技术] [自动化与计算机技术]