This project is concerned with the mathematical and algorithmic foundations of learning problems arising in the field of 3D printing.

In that field, a particularly relevant problem is the one of finding the right instructions to give to a 3D printer for a given design. This task is challenging, and it is currently done by trial and error. The goal of the PhD is to build, and analyze algorithms that automate this procedure. This will contribute to produce a final product with minimum waste, improved efficiency, and sustainability of production.

Mathematically, we will view the task as a learning problem: assuming that we have a database of available designs with their corresponding successful printing sequences, we want to construct an algorithm that automatically gives an admissible sequence of printing actions for a geometry that is not in the database. Building such a tool requires combining different mathematical fields such as optimization, machine learning, inverse problems, shape analysis. Having a good physical model of the mechanics of 3D printing and its numerical discretization comes also into the picture.

The candidate will work within the new research group on Data-Driven Scientific Computing led by Olga Mula, located at CASA, the Center for Analysis, Scientific Computing and Applications of TU Eindhoven. The work will take place in collaboration with researchers from the Mechanical Engineering and Built Environment department from TU/e, as part of an EAISI Exploratory Multidisciplinary AI Research project. EAISI is the Eindhoven Artificial Intelligence Systems Institute which brings together all AI activities of the TU/e. In this project, two PhD candidates will work together. The second PhD candidate will provide data from real experiments and expertise in mechanical modeling.