Abstract
Declarative business process discovery aims at identifying sets of constraints, from a given formal language, that characterise a workflow by using pre-recorded activity logs. Since the provided logs represent a fraction of all the consistent evolution of a process, and the fact that many sets of constraints covering those examples can be selected, empirical criteria should be employed to identify the “best” candidates. In our work we frame the process discovery as an optimisation problem, where we want to identify optimal sets of constraints according to preference criteria. Declarative constraints for processes are usually characterised via temporal logics, so different solutions can be semantically equivalent. For this reason, it is difficult to use an arbitrary finite domain constraints solvers for the optimisation. The use of Answer Set Programming enables the combination of deduction rules within the optimisation algorithm, in order to take into account not only the user preferences but also the implicit semantics of the formal language. In this paper we show how we encoded the process discovery problem using the ASPrin framework for qualitative and quantitative optimisation in ASP, and the results of our experiments.