网格服务中基于GEP的资源自学习选取算法

The Resource Selecting Algorithm Based on Gene Expression Programming in Grid Service

在网格服务中,为有效地在域内合理选取恰当的资源分配给"服务消费者",提出了基于基因表达式编程(GEP)的网格资源自学习选取算法。算法通过采样网格资源历史分配情况进行合理评估与挖掘后,达到对网格资源分配失效、网格资源能力等进行预测,故解决了原传统算法因需要依赖专业知识,而使资源的选取和分配过程具有主观性和盲目性的问题。最后,通过实验对基于传统遗传算法的选取方式和基于GEP的算法的选取方式予以比较,表明该算法确有其优越性与实用性。

In Grid Service, a new approach of the grid resource selecting based on Gene Expression Programming （GEP） was proposed for aiming at Grid resource domain split technique. This approach could predict the performance of the Grid computation node. The experiment results of algorithm showed that this method is more effective in some aspects than the traditional GP ones.