Customizable Declarative Process Specification Scenarios via Automated Planning

Declarative specifications define the behaviour of processes by means of temporal rules exerting dependencies between the occurrence of activities. Those rules can constitute the system requirements and normative boundaries within which the workflow is framed. The multitude of rules and their interrelations, though, make it reportedly difficult for users to linearly figure out unfoldings that depict the admitted behaviour in the form of explicative scenarios. While simulation engines can produce example traces of customizable length and in varying numerosity, these state-of-the-art tools adopt random approaches that do not permit the selection of a subset of runs exhibiting characteristic traits. In other words, simulations could produce traces that show common behavior, leaving out significant parts of the possible execution paths, and thus substantially hindering a proper validation of specifications. To mitigate this issue, we propose a principled solution that leverages automated planning to produce execution scenarios depicting the specified behavior according to desirable execution costs, diversifying the enacted behavior, or matching a given subsequence of steps. We demonstrate its efficiency and efficacy through experiments demonstrating its scalability and superiority in terms of behavioral coverage with respect to the state-of-the-art simulation tool for declarative process specifications.