Model-Based Dependable Composition of Self-Adaptive Systems

Building mobile and pervasive systems as a selection, composition, adaptation and evolution of pre-existing software entities may arise dynamically and continuously different issues related to inconsistencies, changes or faults. We pro-pose an approach to detect and handle these issues with the appropriate methodology in every case. This is performed by tackling three great challenges in software engineering related to self-adaptive systems: (i) their formalisation, by using model-based SOA, which bridge the business and software processes, (ii) their development and maintenance, by performing adaptation and/or evolution when inconsistencies or changes occur, and (iii) their monitoring to handle faults, by using error recovery techniques. We use an example based on an intelligent transportation system to validate our proposal.