为了根据公告服务质量随机性和实际服务质量随机性之间的一致性来计算服务质量的信誉,首先利用分布函数表示服务质量随机性,并给出一种服务质量经验分布函数的增量式更新方法。通过将服务质量公告值和实际观察值视为随机变量,并以随机优势理论为基础,设计结合随机服务质量公告及实际观察值的信誉度量模型。模型考虑了公告服务质量随机性和实际服务质量随机性之间不同的优势关系,以及服务质量观察值波动对信誉的影响。理论分析与实验验证了模型的合理性及其在服务选择中应用的有效性。
To measure the Quality of Service(QoS) reputation based on the consistency between advertised QoS randomness and actual QoS randomness,distribution function was used to represent the stochastic nature of QoS and an incremental update method of empirical distribution function was designed firstly.Based on the stochastic dominance theory,a reputation measurement model combined with advertised QoS and actual QoS was proposed by modeling advertised QoS and actual QoS as random variables.The different dominance relationship between advertised QoS and actual QoS with the influence on reputation by QoS observation wave were considered.Theoretical analysis and experiments showed that the proposed model was feasible and effective in service selection.