一种基于逆变异系数的云服务价值评价方法

A CLOUD SERVICE VALUE ASSESSMENT APPROACH BASED ON INVERSE COEFFICIENT OF VARIATION

当前云服务平台中存在大量功能相似、性能不同的服务,为有效地从中评估、选择各个服务,提出一种新的基于逆变异系数的服务价值评估方法,称为Inv-CV(Inverse Coefficient of Variation)。采用以Pareto支配为基础的快速非支配排序算法确定Pareto集,然后根据Pareto集计算各评价指标的变异系数,并得到权重值。与传统的基于CV的方法不同,在Inv-CV方法中,一个指标的变异系数越小,对应的权重越大,反之亦然。最后,使用加权综合法对所有服务进行价值评估。实验结果验证了Inv-CV方法的可行性和准确性。

There exist a large number of services with similar function and different performance in the cloud service platform. To well evaluating and selecting the right service, a novel assessment approach based on inverse coefficient of variation, called Inv-CV, is proposed. A fast non-dominated sort algorithm based on Pareto dominance was adopted to determine the Pareto set. The coefficient of variation of each criterion was calculated from the Pareto set, and the weight value was obtained. Different from the traditional CV based approach, in Inv-CV, the smaller the coefficient of a criteria, the larger the corresponding weight, and vice versa. The integrated weighting method was used to evaluate the value of all services. The experimental results verify that the proposed Inv-CV approach is feasible and accurate.