Within the growing area of data-aware processes, Data Petri nets (DPNs) with arithmetic data have recently gained popularity thanks to their ability to balance simplicity with expressiveness. DPNs can be automatically mined from event data, but these process discovery techniques typically come without any correctness guarantees. In particular, the generated models may violate the crucial property of data-aware soundness. While data-aware soundness can be checked automatically for a large class of models, nothing is known about how to repair such processes once a violation is detected. In this paper we are concerned with repairing DPNs so that the refined model satisfies the desired soundness properties. Our approach is based on conservative behavioural changes, which are minimally invasive in the sense that the behaviour of the repaired model coincides with that of the original model except for (prefixes of) traces that caused the violation. We show experimentally that the approach can be used to repair unsound DPNs from the literature.

Conformance checking is a key process mining task to confront the normative behavior imposed by a process model with the actual behavior recorded in a log. While this problem has been extensively studied for pure control-flow processes, data-aware conformance checking has received comparatively little attention. In this paper, we tackle the conformance checking problem for the challenging scenario of processes that combine data and control-flow dimensions. Concretely, we adopt the formalism of data Petri nets (DPNs) and show how solid, well-established automated reasoning techniques from the area of Satisfiability Modulo Theories (SMT) can be effectively harnessed to compute conformance metrics and optimal data-aware alignments. To this end, we introduce the CoCoMoT (Computing Conformance Modulo Theories) framework, with a fourfold contribution. First, we show how SMT allows to leverage SAT-based encodings for the pure control-flow setting to the data-aware case. Second, we introduce a novel preprocessing technique based on a notion of property-preserving clustering, to speed up the computation of conformance checking outputs. Third, we show how our approach extends seamlessly to the more comprehensive conformance checking artifacts of multi- and anti-alignments. Fourth, we describe a proof-of-concept implementation based on state-of-the-art SMT solvers, and report on experiments. Finally, we discuss how CoCoMoT directly lends itself to further process mining tasks like log analysis by clustering and model repair, and the use of SMT facilitates the support of even richer multi-perspective models, where, for example, more expressive DPN guards languages are considered or generic datatypes (other than integers or reals) are employed.