Have you ever wondered how the meal plan in the university's cafeterias is created? How do they ensure there's always a meal for every dietary need and there's not too much repetition? Could you even optimize which meal is offered when to help lower food waste, improve customer satisfaction or reduce climate impact?

Join us in this Bachelor's thesis to have a real impact on sustainability in institutional kitchens.

This thesis is offered in cooperation with better bites, a Münster-based startup in the field of institutional kitchen software.

The Problem

Kitchen managers balance a wide range of requirements every week: How many vegetarian dishes should be on the menu? Which meals fit the season and the available kitchen equipment? How do you avoid repetition without losing sight of margins? On top of that, several objectives need to be optimized simultaneously: guest demand, contribution margin, CO₂ footprint, and nutritional balance. There is no single "optimal" solution, the goal is to provide kitchen teams with a set of Pareto-optimal plans from which they can choose based on their own priorities.

The Research Task

Before this problem can be solved algorithmically, it needs to be modeled rigorously. That is exactly what this thesis is about.

How can constraints like "a dish may not be repeated within 30 days" or "the kitchen can run at most three oven dishes per day simultaneously" be translated into mathematical language? Are these hard requirements or soft preferences? Are the objective functions linear, or do non-linearities emerge for instance because dishes influence each other's demand? These modeling decisions are far from trivial and have significant consequences since they determine which class of solution methods is even applicable.

In this thesis, you will formally model the optimization problem, analyze its structural properties, and based on this analysis identify, compare, and empirically validate a suitable algorithmic approach using real production data provided by better bites.

Requirements

Prior knowledge or coursework in operations research, mathematical optimization, or algorithm design is expected. Programming experience, ideally Python, is beneficial for the empirical validation component. The thesis can be written in German or English.

Note: This thesis is suited for students with a strong interest in operations research, combinatorial optimization, and mathematical modeling.

In case you are interested, please send a short motivational statement to Dr. Lennart Schäpermeier (lennart.schaepermeier@uni-muenster.de).

About better bites

better bites is a young startup based in Münster, founded a year ago by Information System graduates from the University of Münster. We build software that helps large institutional kitchens in student canteens and corporate cafeterias planning their production and ordering quantities using a machine learning forecast software. Our next big step is a module for optimized meal plan generation. More to read about us: https://www.wiwi.uni-muenster.de/fakultaet/de/news/5380

We provide real production data from live kitchen operations, as well as direct access to partner institutions such as Studierendenwerk Münster. We are a small team and look forward to close, ongoing collaboration. In the context of this thesis, you are welcome to work together with us at REACH Start-up Center in Münster.