Managing variability in process-aware information systems

Configurable process models are integrated representations of multiple variants of a process model in a given domain, e.g. multiple variants of a shipment-to-delivery process in the logistics domain. Configurable process models provide a basis for managing variability and for enabling reuse of process models in Process-Aware Information Systems. Rather than designing process models from scratch, analysts can derive process models by configuring existing ones, thereby reusing proven practices. This thesis starts with the observation that existing approaches for capturing and managing configurable process models suffer from three shortcomings that affect their usability in practice. Firstly, configuration in existing approaches is performed manually and as such it is error-prone. In particular, analysts are left with the burden of ensuring the correctness of the individualized models. Secondly, existing approaches suffer from a lack of decision support for the selection of configuration alternatives. Consequently, stakeholders involved in the configuration of process models need to possess expertise both in the application domain and in the modeling language employed. This assumption represents an adoption obstacle in domains where users are unfamiliar with modeling notations. Finally, existing approaches for configurable process modeling are limited in scope to control-flow aspects, ignoring other equally important aspects of process models such as object flow and resource management. Following a design science research method, this thesis addresses the above shortcomings by proposing an integrated framework to manage the configuration of process models. The framework is grounded on three original and interrelated contributions: (i) a conceptual foundation for correctness-preserving configuration of process models; (ii) a questionnaire-driven approach for process model configuration, providing decision support and abstraction from modeling notations; (iii) a meta-model for configurable process models covering control-flow, data objects and resources. While the framework is language-independent, an embodiment of the framework in the context of a process modeling language used in practice is also developed in this thesis. The framework was formally defined and validated using four scenarios taken from different domains. Moreover, a comprehensive toolset was implemented to support the validation of the framework.