A multi-objective programming for web services composition

Web service plays an important role in implementing Service Oriented Architecture (SOA). It is becoming vital to provide an efficient service composition mechanism with respect to user's requirements. Measuring the Quality of Service (QoS) can help to offer appropriate functions. However, most of the previous QoS-based researches were unable to guide a service selection to achieve a rapid and effective composition. To address this issue, firstly, we present a method to measure QoS in a composition. Secondly, QoS risk is defined as the variance between actual value and the expectation of QoS. Thirdly, we propose a multi-objective programming model, aimed at composite service plan optimization. Finally, genetic algorithm is used with synthetic data, and a Weighted sum of risk values constitutes the fitness to lead the evolution of populations. Experimental results demonstrate that the presented model and algorithm provide effective and comprehensive information for decision makers to select appropriate web services for composition.