论文部分内容阅读
为解决基于服务质量(QoS)的web服务选择中只考虑QoS的收益(均值),不考虑风险(即动态变化)的Skyline查询问题,提出均值标准差描述QoS(均值描述QoS的收益,标准差描述QoS的风险)的服务Skyline计算方法和4种服务Skyline查询算法,即基于均值标准差的BNL算法(BNL_MS)、基于均值标准差的D&C算法(DC_MS)、基于均值标准差的NN算法(NN_MS)、基于均值标准差的BBS算法(BBS_MS).该方法能剔除被支配服务,给出QoS属性全优且稳定的服务集,有效缩减备选服务集.实例和实验结果表明:1)均值标准差较均值、区间数、模糊数、随机数不仅能很好地刻画QoS的收益还能刻画风险;2)服务Skyline计算能有效计算服务最优集;3)BBS_MS算法较BNL_MS、D&C_MS和NN_MS算法具有更好的性能.
In order to solve the problem of QoS queries (mean value) only based on quality of service (QoS) and Skyline query without considering the risk (ie, dynamic change), an average standard deviation description QoS (average value describes the return of QoS, standard deviation Service risk (QoS risk) QoS Skyline algorithm and four service Skyline query algorithms, namely BNL algorithm based on mean standard deviation (BNL_MS), D & C algorithm based on mean standard deviation (DC_MS), NN algorithm based on mean standard deviation ), BBS algorithm based on mean standard deviation (BBS_MS). This method can remove the service under control and give a service set with excellent and stable QoS attributes and reduce the service set effectively.Examples and experimental results show that: 1) Difference mean, interval number, fuzzy number and random number can not only depict the benefits of QoS well but also describe the risk; 2) Service Skyline calculation can effectively calculate service optimal set; 3) BBS_MS algorithm is better than BNL_MS, D & C_MS and NN_MS algorithm With better performance.