Flexible Event Log Generation from Object-centric Process Models

Universität Ulm

BA Abschlussvortrag, Erwin Ricardo Pentz, Ort: O27/5202, Datum: 18.12.2023, Zeit: 16:30 Uhr

Process mining is crucial for organizations as it leverages event data to uncover, analyze, and optimize business processes, providing actionable insights for enhanced operational efficiency and informed decision-making. This results in the continuous necessity of event logs suitable for testing process mining techniques. However, event logs based on real-world data are often limited, sensitive, or unavailable.
A different way of obtaining suitable event logs is their synthetic generation using an event log generator. This thesis compares existing approaches and proposes a new tool for generating event logs in the context of object-centric processes. The tool is able to generate guided, tolerated and deviating behaviour on state and step level. Moreover, it can produce event logs comprising multiple different object lifecycles and actors. A PHILharmonic Flows process model serves as input for the event log generation. The tool converts the object lifecycles into Petri nets and uses a token-based simulation approach to generate the event logs. To validate the generated event logs, an exhaustive conformance check was performed.
It was shown that the tool accepts generic input scenarios by testing it on two different object-centric scenarios (modelled in PHILharmonic Flows).