Applying Process Mining on Highly Concurrent Processes with Continuous Data Footprints
Student: Sezin Maden
Title: Applying Process Mining on Highly Concurrent Processes with Continuous Data Footprints
Supervisor: Dr. Sebastiaan J. van Zelst
1st Examiner: Prof. Wil van der Aalst
2nd Examiner: Prof. Markus Strohmaier
Summary
Continuous production processes generate large amounts of data during their execution. A multitude of sensors collects data about the current state of the production plant. Process mining provides a wide variety of techniques for process-centric analysis of data. Most existing process mining techniques assume an event log as their starting point. Event logs are a discrete footprint of processes where ordered events are grouped into process instances. However, the footprints of many production processes are of a continuous form. In this thesis, we aim to bridge the gap between process mining and processes with continuous footprints, by introducing a 3-step framework to efficiently transform sensor data of continuous production processes to conventional event logs. We demonstrate a proof-of-concept implementation of a stand-alone process mining tool using the framework allowing for end-to-end process mining of raw continuous production processes data. Furthermore, we show the viability of the framework by applying it to two real-world data sets of continuous production processes provided by a leading manufacturer of production plants for the food manufacturing industry. We argue that different event logs can be generated from the same raw data and that these event logs can be used for various analytic purposes. Our framework makes the design decisions during log generation explicit.
Note
This work was performed in collaboration with an industrial partner. Therefore, the final thesis is not shared publically. For more information, contact Dr. Sebastiaan J. van Zelst