Flexible QoS-aware services composition for service computing environments

The rapid development of service computing environments has led to an increase in the number of functionally equivalent services that differ according to their QoS (Quality of Service) attributes values. Accordingly, a challenging issue in the context of services composition is how to select the most appropriate candidate services for each user's requirements such that the obtained composite service satisfies the user's non-functional requirements in terms of global QoS constraints. In this paper, a Flexible QoS-aware Services Composition (FQSC) algorithm is proposed to increase the composition feasibility and reduce the composition time, while the composition optimality is maintained as high as possible. The proposed algorithm deals with the limitations of Pareto dominance-based and QoS decomposition-based services composition approaches. Indeed, the number of compositions increases exponentially with the number of services and QoS attributes in the case of Pareto dominance-based approaches, whereas the QoS decomposition-based approaches can lead to a decrease in the composition feasibility when the user's global QoS constraints are highly restrictive. The simulation results show that the proposed approach is effective compared to other services composition approaches in terms of composition time, optimality and search space.