Composition of web services through genetic programming

Abstract

Web Services are interfaces that describe a collection of operations that are network-accessible through standardized web protocols. When a required operation is not found, several services can be compounded to get a composite service that performs the desired task. To find this composite service a search process in a, generally, huge search space must be performed. The algorithm that composes the services must select the adequate atomic processes and, also, must choose the correct way to combine them using the different available control structures. In this paper a genetic programming algorithm for web services composition is presented. The algorithm has a context-free grammar to generate the valid structures of the composite services and, also, it includes a method to update the attributes of each node. Moreover, the proposal tries to minimize the number of services, and looks for compositions with the minimum execution path. A full experimental validation with four different repositories with up to 1,090 web services has been done, showing a great performance in all the tests as the algorithm finds a valid solution with a short execution path.

Genetic Improvement of Programs

“…There is already some work on web "mashups" and using GP to create new web services by combining web services defined by the OWL protocol [30].…”

“…Genetic programming has also been used to create distributed algorithms [35], mobile applications [5] and composing new webservice mashups from other web services [30].…”

Automated Web Service Composition Using Genetic Programming

“…In Majithia et al (2004), a framework is presented to facilitate automated service composition in Service-Oriented Architecture by using semantically-described heterogeneous services. In Rodriguez-Mier et al (2010), genetic programming is also used to compose web services and several public service repository are used in the experiments. However, the fitness function of the genetic programming depends too much on atomic services' input/output variable name and the composition they generated is short of a reliable accurate.…”

“…For the same parent candidates, we can get better children candidates if we take proper genetic operator based on the analysis of the input/output of the parent candidates. Fifth, comparing to Rodriguez-Mier et al (2010) which is too much rely on the semantic meaning of input/output parameter such as the variable name, our method is more stable when the available atomic services have the same input/output parameter name and type but have very different functionality actually.…”

“…Chapter 5 describes an analysis about the complexity about this algorithm. Chapter 6 show the results of experiments using our method in the service composition problem and compare our method with the existing method in Rodriguez-Mier et al (2010). Chapter 7 concludes this thesis and presents some future work.…”