Master Thesis – Combining Process Mining and Machine Learning for Lufthansa’s Airport Operations

 

Description

Lufthansa is using process mining to analyze and improve several of its processes. Next to standard processes such as Accounts Payable, Lufthansa uses process mining to improve flight punctuality, reduce delays, prioritize flights, identify and alert flights potentially causing delays, etc. The Celonis EMS is already used for this. In this thesis project, we want to discover better process models and diagnostics for the following operational process at airports. Each airplane landing and taking off is seen as a case. In between landing and taking off, many events take place, e.g., start/end boarding, start/end fueling, start/end cleaning, start/end pushback plane, start/end de-icing, start/end loading, etc. The activities are mostly concurrent, but one problem will cause a delayed flight having a cascading effect at the airport and also on later flights. What happens in the time between when a plane lands and when it takes off again (known as turnaround time) is complex and affects customers, personnel, and Lufthansa's bottom line. Therefore, it is extremely important to ensure that the actual turnaround time meets the planned turnaround time (e.g., 60 minutes).

The combination of data availability and the need to ensure short turnaround time, calls for advanced analytics. Around 75 unique activities related to an airplane landing and taking off are available in the Celonis EMS. Through process mining, interesting and valuable machine-learning problems can be generated. The goal of this thesis project is to go beyond the analysis already done by the Lufthansa and Celonis teams. The goal is to create so-called situation tables from event data for the most relevant questions. The tables are used to train machine-learning models and answer a variety of questions. Next to machine learning, also data-driven optimization techniques will be used. A challenge is that some events are recorded manually, and sensors may be unreliable, leading to data quality problems. Also, domain knowledge needs to be exploited to parameterize the machine-learning models. Example questions include:

• How to dynamically reallocate staff to gates in case of delays?
• What are the main root causes of delayed flights?
• What are the factors that can be influenced to improve flight punctuality?
• What subsets of activities should be considered?
• How to create high-level event data about delays and high-loads from low-level event data and discover cascades of problems?
• How to deal with data quality problems (e.g., missing events)?

The research question is how the available event data, process mining, and machine learning can be combined to make better predictions and take action. The assignment will involve using the existing Celonis infrastructure at Lufthansa and creating novel prototype software to demonstrate and evaluate ideas. The developed solutions need to be tested on actual data from Frankfurt airport and/or München airport.

Prerequisites

• Excellent data science, machine learning, software engineering, and conceptualization / formalization skills.
• Process mining expertise as demonstrated by good grades for the respective courses.
• Programming / scripting languages
  • Java, Python, etc.
  • SQL and Celonis PQL knowledge is a plus
• Excellent communication skills (preferably in German and English due to the nature of the project).

Pointers

• W.M.P. van der Aalst. Process Mining: Data Science in Action. Springer-Verlag, Berlin, 2016.
• W.M.P. van der Aalst and J. Carmona, editors. Process Mining Handbook, volume 448 of Lecture Notes in Business Information Processing. Springer-Verlag, Berlin, 2022.
• M. de Leoni, W.M.P. van der Aalst, and M. Dees. A General Process Mining Framework for Correlating, Predicting and Clustering Dynamic Behavior Based on Event Logs. Information Systems, 56:235-257, 2016.
• W.M.P. van der Aalst and A. Berti. Discovering Object-Centric Petri Nets. Fundamenta Informaticae, 175(1-4):1-40, 2020.
• Böhm, M., Rott, J., Eggers, J., Grindemann, P., Nakladal, J., Hoffmann, M., & Krcmar, H. (2022). Process mining at Lufthansa CityLine: The path to process excellence. Journal of Information Technology Teaching Cases, 12(2), 135–145. https://doi.org/10.1177/20438869211022369
• https://www.celonis.com/customer-success-stories/lufthansa-celosphere-session-process-mining/

Supervisor

prof.dr.ir. Wil van der Aalst (formal thesis supervisor)

Please contact Mahsa Pourbafrani thesis@pads.rwth-aachen.de first. External theses like these are reserved for only the best students due to the added complexity of interacting with multiple stakeholders and the expectation that something usable comes out.

Daily advisor(s)

To be decided on the RWTH side. From Lufhansa, Maximilian Hoffmann and Patrick Schroeder are involved.

Important!

This a joint thesis project in collaboration with Lufthansa & Celonis. Do not approach people in Lufthansa or Celonis directly without first discussing this with Mahsa Pourbafrani (thesis@pads.rwth‐aachen.de) and later with prof. Wil van der Aalst. It is NOT allowed to organize your own thesis project in a company! We only do joint thesis projects with companies we already work with (in this case Lufthansa & Celonis). The project needs to be defined in the context of this collaboration to ensure good supervision and avoid later confusion.

Since external assignments are particularly challenging, we will only allow good students to do this. Concurrently meeting the expectations of Lufthansa & Celonis and ensuring a good academic level will be demanding. Hence, we only consider students with an average of 2.0 or better for such assignments.

If you are interested, make sure to include detailed information about your background (including a detailed CV), scores for completed courses, and your motivation to do this project. After an initial screening from the RWTH side, we will connect you to the person responsible on the Celonis side.

About Lufthansa and Celonis

The Lufthansa Group is an aviation group with operations worldwide. With 100.000+ employees, the Lufthansa Group generated revenue of EUR 32,770m in the financial year 2022. The Passenger Airlines segment includes, on the one hand, the network airlines Lufthansa German Airlines, SWISS, Austrian Airlines, and Brussels Airlines. As part of the multi-hub strategy, they offer their passengers a broad range of flights from their global hubs in Frankfurt, Munich, and Zurich as well as their national hubs in Vienna and Brussels. Lufthansa German Airlines also includes the regional airlines Lufthansa CityLine, Air Dolomiti and Eurowings Discover, the Lufthansa Group’s holiday airline. Lufthansa is highly innovative when it comes to using event data for improving operational excellence. This is reflected by the extensive use of process mining.

   

Celonis is the global leader in process mining and execution management. The Celonis Execution Management System provides companies with a modern way to run their business processes entirely on data and intelligence. Celonis has thousands of customers, including Lufthansa, ABB, AstraZeneca, Bosch, Coca-Cola, Citibank, Danaher Corporation, Dell, GSK, John Deere, L’Oréal, Siemens, Uber, Vodafone and Whirlpool. Celonis is headquartered in Munich, Germany and New York City, USA and has offices all over the globe, including the Celonis Engineering and Innovation Lab in Aachen (Melaten campus).