Service creation and provision is a major growth area in telecommunications, but feature interaction presents a major obstacle to the rapid introduction of services. Our research has shown that neither the state-of-the-art on-line nor off-line approaches are sufficient to tackle this problem, particularly in the presence of new and legacy service, and open party provision.

This project will develop a novel solution to the problem, aiming to develop the automated analysis and support which are needed to accelerate the speed at which services can be introduced.

Our main objectives are:

• to develop a novel, hybrid, adaptive approach to detecting and resolving feature interactions between telecommunications services, including those between new and existing services. This will be based on a synthesis of off-line and on-line techniques.
• to apply and evaluate the new techniques to a real switching enviroment, presenting both evolving and legacy aspects, provided by an industrial collaborator. We will demonstrate the effectiveness of our techniques through experimentation and reasoning about them in a formal way.

To achieve our objectives, we advocate a major adaptation of our existing on-line detection and resolution methods by forming an off-line/on-line hybrid to allow off-line service information to direct the on-line search. The on-line method is based on a run-time behaviour tree which gives a number of alternative behaviours as the call progresses; in the hybrid approach, information exchange will take the form of behaviour trees. The major research challenges in this project will be to build upon our experience of forml models to develop the appropriate off-line models and analysis techniques, and to develop the analysis. A further challenge will be that while a close relationship between the on-line system and the off-line models is required, it is essential that the models and the entire approach are not platform specfic.