为解决当前基于效用的网络资源调度中模型通用性不强、求解算法效果差、速度慢等问题,提出了基于边际效用函数的效用最优化资源调度方法.它根据边际效用函数的特点,将网络应用分为弹性和非弹性应用,并通过每个应用的边际效用函数求出其效用函数.将上述效用函数应用于网络资源调度问题中,给出了问题求解的高效算法.仿真实验表明,与目前最新的同类算法及经典优化求解工具Ling09.0中的算法相比,本文算法求得的总效用值平均提高5%和4%,而所用时间仅为上述算法的0.2%和0.003%.
For the poor scalability,effectiveness and efficiency of the up-to-date utility based scheduling solutions,in this pa- per, we proposed a marginal utility function based scheduling method. In accordance with the feature of the marginal utility func- tions, the networking applications were divided into two classes, e. g. elastic and inelastic class. And then we built a utility based net- work resource scheduling model with the utility functions obtained from the corresponding marginal utility functions. The simulation experiments show that compare to the latest similar method and the corresponding approach in Lingo9.0, our algorithm increased the total utility by average of 5 and 4 percent with only 0.2% and 0.003% in the elapsed time.