Optimizing Lifelong Multi-Agent Pathfinding Using Reinforcement and Imitation Learning
- Forschungsthema/Bereich
- Mobile Agents and Robotic Systems
- Typ der Abschlussarbeit
- Master
- Startzeitpunkt
- 01.11.2025
- Bewerbungsschluss
- 30.04.2026
- Dauer der Arbeit
- 6 Monate
Beschreibung
Field:Mobile robotics is one of the fastest-growing and most dynam-ic areas in intralogistics. As fleet sizes rapidly increase, the challenge shifts from single-robot navigation to large-scale, cooperative fleet coordination. Efficient and adaptive management of these fleets is essential to ensure high system throughput, reliability, and productivi-ty. At the IFL, we are at the forefront of this development, actively contributing to the industry standard VDA 5050, which defines com-munication between mobile robots and fleet management systems.Problem Statement:
Traditional Lifelong Multi-Agent Pathfinding (LMAPF) algorithms rely on fixed heuristics and lack adaptability to dynamic, real-world environments. In practice, robot fleets must op-erate continuously under uncertainty, delays, and changing task de-mands—conditions where static methods quickly reach their limits. This thesis explores learning-based approaches such as Reinforce-ment Learning, Imitation Learning, or hybrid methods to enable scal-able and adaptive coordination. The goal is to develop policies that learn cooperative behaviors, reduce congestion and deadlocks, and improve overall system throughput and robustness.
Voraussetzung
- Voraussetzungen an Studierende
-
- Experience with Python or a similar programming language.
- Familiarity with machine learning frameworks such as PyTorch or TensorFlow is an advantage.
- Knowledge of graph theory and pathfinding algorithms is a plus.
- Problem-solving mindset and an independent working style.
- Studiengangsbereiche
-
- Ingenieurwissenschaften
Informatik
Maschinenbau
Mechatronik & Informationstechnik
Mechanical Engineering - Wirtschafts- und Rechtswissenschaften
Wirtschaftsinformatik
Wirtschaftsingenieurwesen
- Ingenieurwissenschaften
Betreuung
- Titel, Vorname, Name
- M. Sc. Marvin Rüdt
- Organisationseinheit
- Institute for Material Handling and Logistics (IFL)
- E-Mail Adresse
- marvin.ruedt@kit.edu
- Link zur eigenen Homepage/Personenseite
- Website
Bewerbung per E-Mail
- Bewerbungsunterlagen
-
- Anschreiben
- Lebenslauf
- Notenauszug
E-Mail Adresse für die Bewerbung
Senden Sie die oben genannten Bewerbungsunterlagen bitte per Mail an marvin.ruedt@kit.edu
Zurück