Behaviour Modelling and Transformations for Context-Aware Mobile Applications

Today's panorama of service offerings is characterised by the widespread diffusion of the Internet and Web-based technologies everywhere in society. We are surrounded by devices that can support us in several tasks of our every-day life, like, for example, e-readers to access books and magazines, or mobile phones with extremely intuitive user interfaces for browsing, checking emails, keeping in touch with colleagues and friends through social networks, finding maps and locations, and so on. Moreover, this daily experience would not be possible without ultra-fast networks and wireless technologies that allow us to exchange any kind of data, anywhere, in real time and at low-cost. In this panorama, it becomes of vital importance for service providers to offer services that are innovative and distinctive. On one hand, service providers have to preserve current customers and attract new ones in order to survive in an ever growing arena of competitors. On the other hand, service users become more and more aware of the opportunities offered by the evolving technologies and, consequently, more demanding and with stronger expectations than in the past. Therefore, users expect a profusion of services wherever they are, to support whatever they are doing, and according to their personal preferences and needs, while providers have to create a wide range of enriched services in a rapid, low-cost and user-centric way. This thesis proposes a layered methodology based on behaviour modelling and transformations for the development of context-aware mobile applications, which are distributed applications that can provide advanced and personalised services to their users. Currently available approaches, such as Service-Oriented Architecture (SOA) and Model-Driven Architecture (MDA), are used in this thesis to support such a methodology. The main objective is to progress the state-of-the-art in model-driven development of context-aware mobile applications by taking into account the behaviour of these applications already in early stages of the development process. In order to achieve this, we refine the application behaviour in several steps, from abstract specifications to final implementations, and develop automated model transformations throughout these refinement steps to generate executable models and reason about their behavioural correctness.