LAUSR

The strong integration of technology in the physical world caused the emergence of smart environments. These environments are supposed to improve the quality of life of their users by providing them with customized services when needed and adapting to their changing needs. The dynamics of the users, the huge number of available services and the strong collaboration between the stakeholders, make traditional service-oriented approaches incapable of providing relevant services to the users in these environments. To deal with these issues, we propose a recommendation-based system for service composition targeting smart environments. The proposed system is able to capture the situation of the users through the analysis of their context information, which in turn allows the system to capture their requirements and select the appropriate service models to satisfy their needs. Then, based on the invocation log, the system implements two recommendation policies. First, it selects the best services in terms of QoS that satisfy the captured requirements. Second, the system recommends new tasks to be integrated in existing service models. The conducted experiments show the efficiency and effectiveness of the recommendation policies proposed. To illustrate the workings of the proposed system, we present a case study called SMARTROAD pertaining to the transport domain and road security.