Intelligent Hospital Logistics

The coronavirus pandemic has shown that exceptional medical situations, coupled with staff shortages, can push entire hospital systems to their limits. The research project ”MediCar 4.0” seeks to create an advanced transport logistics platform for self-driving vehicles on clinic premises to avoid future bottlenecks and enhance operational efficiency. However, the dynamic nature of emergencies, failures, and other unpredictable events poses challenges. Many conflicts, such as congestion and deadlocks, can only be adequately resolved with expert knowledge, causing classical optimization algorithms to fall short. This challenge is compounded by the transportation of critical goods, each with unique requirements and time-sensitive needs. An example is blood transfusion samples requiring temperature control and shock-free deliveries within only a few minutes. Hereby, ensuring the safety and optimal treatment of patients necessitates maintaining reliable transportation at all times.

This team project aims to develop robust AI-based routing strategies that integrate world-knowledge from large language models to meet the challenges of clinical processes and heterogeneous supplies. In this context, the students will build a simulation environment of the University Hospital in Freiburg. Subsequently, the team implements routing algorithms of their choice and tests them in various scenarios. As an optional add-on, the system should be able to handle varying demands and vehicles on the hospital premises. Participants will become familiar with multi-agent simulation, routing algorithms, reinforcement learning, and large language models. Thus, the team should adopt Python as the main programming language, but is free to use any framework or opensource project.

The students from Mannheim, with a strong interest in machine learning, will work together with students from Cluj-Napoca, which includes a reciprocal visit. This project contributes to the future of hospital logistics and improved (possibly) self-learning routing systems aligned with human values.