QoS-Aware Semantic Web Service Composition with Uncertainties

Choosing the optimal execution plan is an important process of semantic web service composition. However, Quality of service (QoS) based plan selection is still challenging because the uncertain QoS values and uncertain user preference make data aggregation and decision making hard. This paper presents a novel Uncertain multiattribute decision making-based semantic web service composition algorithm (UMC) to solve the above difficulties for the first time. UMC takes all possible QoS expression types (real number, interval and linguistic) into consideration and includes three main steps: defuzzifying linguistic data, normalizing the decision matrix and evaluating alternatives synthetically. Further, the strategies of uncertain preference conversion are also discussed in UMC. Other contributions of the paper include an extensible QoS ontology to express the relative concepts, a QoS aggregation algorithm and a set of experiments that demonstrate the effectiveness and benefits of our approach.