Most companies evolve in competitive markets by quickly adapting their business processes. Their performance depends on their capability to adopt continuous improvement techniques for their organization through the development of durable Information Systems (IS) for automating the processes. In this sense, Service-Oriented Architectures (SOA) enable the development of flexible IS with a predominant architectural paradigm of service compositions. However, aligning the SOA to the business processes evolutions remains an overriding concern. In this thesis, we study business process management and evolution from their design until their automation in SOA and their adaptation. We propose a model-driven engineering approach which is integrated to a tool for business process modeling and SOA development. It begins with the design of a model with Business Process Modeling Notation (BPMN) describing process interactions. With an automated transformation chain, we produce a model with Service-Component Architecture (SCA) describing service compositions solution based on a SOA. This ensures the architectural traceability between the business level of the IS and its technical level, since generated services describe an application logic that implements the BPMN model. Then, we introduce an incremental change propagation method between business process and SOA founded on BPMN and SCA models synchronization. It is integrated to a change impact simulation and analysis tool that is based on graph rewriting technique to avoid the inconsistencies caused by change propagation. Finally, we propose metrics for measuring the impact of business process evolutions on SOA maintenance and ensuring better change governance.
