Quality‑aware web service composition using a hybrid summarization

Abstract

A web application contains diverse functions where a distinct collection of candidate web services exists on the web to perform each function. Each web service in a collection performs the same function but with a different quality of service (QoS). When a web service with a specific QoS is selected from each collection for each application function, a composition of web services is created. Therefore, many compositions can be made for a web application so that their near-optimal selection is a matter of concern. Two issues exist with the selection: (1) how to compute the composition’s QoS and (2) how to select the compositions based on their QoSs, which is an NP-hard problem. This paper aims to address resolving these issues. To show the effectiveness of our proposed method, we applied it to four web-based service- oriented applications with a dataset of 2507 web services in collections. Then, regarding a few performance indicators, we evaluated the solutions obtained using our method against those obtained using six related studies. Moreover, we exploited statistical tests to show the generality of the results in terms of the QoSs. The performance indicator of the coverage ratio showed our solutions dominate 79% of related studies’ solutions on average.