一种支持混合 QoS 属性的服务选择方法

摘要

Current Web service selection methods mostly assume that all the QoS attribute values of Web services have to be the definite real numbers;but do not take into account the fuzziness property of QoS.This hypothesis has certain limitation in practical application requirement;what’s more,it will also cause the loss of a large amount of data information.To enhance the representation capability of QoS, we described the QoS attributes as the accurate numerical type,the interval numerical type and the fuzzy numerical type.At the same time, we used the order relations vector to represent the demand preference of user on different QoS attributes and converted them into user’s subjective weights on QoS attributes,as well as calculated the objective weights of QoS attributes with entropy weight.On this basis,we gave Web service selection process of hybrid QoS attributes by using relative superiority degree algorithm.At last we proved the effectiveness and rationality of the relative superiority degree algorithm through simulation and experiment verification.%当前的服务选择方法大都假设所有服务的 QoS（Quality of Service）属性值必须均为确定的实数，并未考虑 QoS 属性的模糊性。这个假设在实际应用需求中具有一定的局限性，还会丢失大量的数据信息。为增强 QoS 表达能力，将 QoS 属性值描述成精确数值型、区间数值型、模糊数值型。同时，用序关系向量表示用户对不同 QoS 属性的需求偏好，将其转换成用户对 QoS 属性的主观权重，并采用熵权法计算 QoS 属性的客观权重。在此基础上，采用相对优势度算法给出混合 QoS 属性的服务选择过程。最后通过模拟和实验验证，证明相对优势度算法的有效性与合理性。