Research
The Junior Professorship for AI Application in Production and Logistics investigates how artificial intelligence methods can contribute to solving complex planning, control, and optimization problems in industrial value chains. To this end, our research combines methods and models from machine learning, generative artificial intelligence, metaheuristics, as well as modeling and simulation for problems in production, logistics, supply chain management, robotics, and transportation.
Our research is structured around three focus areas.
AI Methods for Production and Logistics
At the core of our work is the question of how AI methods can be leveraged to make industrial decision-making processes more intelligent, adaptive, and efficient. In the area of production planning and control, we develop learning agents that autonomously optimize manufacturing processes and adapt to changing conditions. At the supply chain level, we investigate how AI-based approaches can dynamically manage networks and rapidly reconfigure them in the event of disruptions. A further focus lies on AI-driven robotics applications, where we explore adaptive handling strategies and learning-based control concepts for industrial robotics.
Modeling and Simulation in Production and Logistics
Modeling and simulation form an essential foundation for the successful application of AI methods. Simulation models enable the realistic representation of complex systems, the generation of training data for machine learning methods, and the testing of AI-based decision strategies in a risk-free environment. In particular, simulation environments play a critical role in training agentic systems – for instance via deep reinforcement learning – by enabling the necessary interaction between agent and environment without real-world costs or risks. Within the professorship, we develop fast and efficient simulation approaches – for instance based on mesoscopic models –, investigate the automated data-driven creation and parameterization of simulation models, and deploy simulation-based digital twins for operational decision support.
Generative AI for Industrial Applications
As a third focus area, we address the growing potential of generative AI (genAI) systems for industrial applications. This includes genAI-assisted factory planning, where generative models produce layout designs and planning alternatives, as well as the development of genAI-specific AutoML pipelines and MLOps infrastructures to ensure scalable and maintainable deployment of genAI systems in practice. Complementing the data-driven creation and parameterization of simulation models, the professorship further investigates how generative AI can simplify model building through natural language interaction – particularly where automated data-driven approaches reach their limits due to the complexity of real-world process logic.
The three areas are closely interlinked: simulation models provide training environments for AI agents, while generative methods accelerate and automate model generation.
