基于离散粒子群和禁忌搜索的网格资源调度算法

Grid Resources Scheduling Algorithm Based on Discrete Particle Swarm and Tabu Search

提出一种基于智能的网格资源分配和任务调度算法,由于网格资源调度是个离散问题,所以采用基于惯性权重的离散粒子群算法,针对离散粒子群算法在后期容易陷入局部最优的缺陷,将其与禁忌搜索算法相结合,在算法的前期采用离散粒子群算法进行搜索,在后期采用禁忌搜索算法进行局部搜索。并利用网格模拟工具Gridsim Toolkit开发一个网格模拟系统进行实验。实验结果表明,该算法具有较高的寻优能力。

A grid resource allocation and task scheduling algorithm based on intelligent is proposed,because the grid resources scheduling is a discrete problem,the discrete particle swarm algorithm based on self-adapting inertia is adopted.Aiming at the later of algorithm,the discrete particle swarm algorithm is easy to fall in local optimum,it is combined with tabu search algorithm.At prophase,the discrete particle swarm algorithm is presented,in the late tabu search algorithm is using to local search.And by using the grid simulation tools Gridsim Toolkit,a grid simulation system for experiments is developed.Experimental results show that the algorithm is of higher optimization ability.