Rule Propagation: Adapting Procedural Process Models to Declarative Business Rules

The debate on advantages and disadvantages of declarative versus procedural process modeling languages for different usage scenarios has been intense. Procedural languages are more suited for describing operational processes while declarative ones for expressing regulations/guidelines and, in many situations, the need of combining the benefits of the two rises. Instead of forcing modelers to use a hybrid language, we envisage to keep the two specifications separate and propose a technique that automatically adapts procedural models so as to comply with sets of declarative rules. This not only fits scenarios where, e.g., company processes have to be modified according to changing external rules, but, more in general, it presents a way to take advantage of the flexibility of declarative while maintaining the high level of support provided by procedural languages. Furthermore, by comparing the original and the resulting procedural models, the impact of rules is clearly exposed. In this paper, we frame the problem above by providing its theoretical characterization and propose an automata-based solution, which is then evaluated against approaches leveraging state-of-the-art techniques for process discovery and model repair.