Vehicle Defleet Optimization

During my PhD, I acquired and executed a research collaboration between Sixt SE and the University of Göttingen.

The project focused on Sixt's vehicle defleet process. When rental cars reach a certain age or mileage, they're sold and need to be transported to auction sites or dealerships. I modeled this process mathematically to investigate transport optimization potential.

The approach combined machine learning (gradient boosting to predict vehicle values and demand patterns) with mathematical optimization (mixed-integer programming via Gurobi to optimize transport routes and timing). This hybrid ML + optimization approach is increasingly common in industry, where ML handles prediction and optimization handles decision-making.