基于QoS约束的启发式网格资源分配算法研究  

Grid Resource Allocation Heuristic Algorism Based on Quality of Service Constraint

作　　者：穆瑞辉[1] 苗国义[2] 

机构地区：[1]新乡学院计算机与信息工程学院,河南新乡453003 [2]河南科技学院信息工程学院,河南新乡453003

出　　处：《计算机测量与控制》2012年第12期3357-3360,共4页Computer Measurement &Control

摘　　要：由于网格环境的资源分配问题属于NP难题而难以通过精确算法获得最优解,提出一种针对具有用户QoS需求的资源分配问题进行求解的启发式方法;首先,在资源分配模型目标函数中考虑QoS约束,然后采用改进的蚁群算法对资源分配方案进行寻优,在信息素的初始化、局部信息素和全局信息素的更新以及资源选择机制中,都加入了QoS约束,使得通过改进蚁群算法求解出的最优解能满足用户QoS需求和实现资源负载平衡;实验结果表明:文中算法得到的平均调度时间较QoS-F和QoS-GA分别减少了42.96%和24.49%,是一种符合网格环境的有QoS约束的有效资源分配方法。For the grid resource allocation problem is the NP problem, so it is difficult to get the precise solution, a heuristic solution aiming at the grid resource allocation with the QoS （Quality of Service）. Firstly, the QoS constraint is added to the resource allocation model goal function, and then the improved ant colony algorism is used to find the optimal solution. The QoS constraint is also added the process of the initialization of ＂pheromone, the update of the local pheromone and the global pheromone, and the selection mechanism of resource selection. The improved ant colony algorism can satisfy the optimal solution and realize the load balance of resource. The experiment shows the method in the paper reduced 42.96 % and 24. 49 % respectively in the average scheduling time, and it is an effective resource allocation meth od under the grid environment with QoS constraint.

关 键 词：网格 任务 资源分配 蚁群算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]