A genetic algorithm with an elitism replacement method for solving the nonfunctional web service composition under fuzzy QoS parameters

The nonfunctional (QoS-aware) web Service Composition (QSC) problem, which is a strong NP-hard optimization one, is widely addressed by considering the advertised Quality of Service (QoS) values of web services as non-ambiguous. However, in real world environments, and due to some of their unconditional factors like network architectures changes, communications congestion and economic policies, the QoS values ambiguity should be undertaken in formulating the QSC problem. In this paper, we present a genetic algorithm that integrates an elitism replacement method for solving the QoS problem under fuzzy QoS parameters, which have been expressed as generalized trapezoidal fuzzy numbers. The addressed QSC problem is formulated as a fuzzy nonlinear integer constrained single-objective optimization model through adapting the well-known simple additive weighting method. To illustrate the performance and the efficiency of the proposed algorithm, we present the experimental comparisons to a fuzzy approach of an existing Particle Swarm Optimization (PSO)-based web service selection algorithm over a fuzzy extended version of the real-world QWS dataset.