Dynamic services selection approach for the composition of complex services in the web of objects

The composition of services in the web of objects poses important challenges concerning resources limitation, adaptability, performance and scalability. In this paper, we focus on the optimization of service selection as the key operation in dynamic service composition by using multi-agent architecture that achieves an online generation and rescheduling of the optimal services execution workflow by using a heuristic planning technique. The validation of the proposed work has been carried out large-scale composition scenarios corresponding to realistic use cases, treated in the domain of smart cities infrastructure monitoring. The results of composition scenarios simulations show clearly the performance of the selection and how the composition is well adapted for Web of Objects in terms of a timely fashion handling in of the fluctuation and variation of user context and quality of service. The ongoing work concerns the implementation of the composition platform by using the middleware and the infrastructure of the Ubistruct living lab.