基于稳定匹配机制的网格任务调度算法

Stable Matching Mechanism-based Grid Computing Task Scheduling Algorithm

任务调度和资源分配是网格计算的关键技术。提出了一种基于稳定匹配机制的网格任务调度算法(SMM:StableMatchingMechanism),以及根据各条件因素值及其所占比率进行量化匹配的思想。用户和资源可以根据自身特点动态地设置各条件因素的基准值,通过与实际提供的条件因素值进行比较,得到匹配度,按照匹配度的高低顺序循环匹配最终达到稳定。算法综合考虑用户和资源双方的需求而得出相应的稳定匹配调度策略。给出了算法的数学模型和描述,在Matlab语言环境下仿真表明,SMM算法可以有效地实现网格任务调度,满足任务和资源双方的需求,得到稳定匹配。

Task scheduling and resource allocation are the key of Grid Computing. A new grid task scheduling algorithm （SMM： Stable Matching Mechanism） based on matching problem was proposed and the quantification matching method founded on each condition value and its proportion value was put forward. The user and the resource can dynamically preset the benchmark values of each condition follow along with its character and obtain the matching degree via the comparison with the actual values. According to the rank of the matching degree, SMM can circularly match until stabilization. The demand of both the users and the resource was considered in SMM and the relevant stable matching scheduling strategy could be obtained. The mathematical model and the relevant description were proposed. The simulation in Matlab demonstrates that SMM can implement grid task scheduling effectively. This algorithm can meet the standard of both the works and the resource; further more stable matching will be achieved.