Principal Investigators:
- Mathieu Besançon - Inria
- Sebastian Pokutta - Zuse Institute & TU Berlin
Members:
- Sébastien Designolle - Inria
- Deborah Hendrych - Zuse Institute & TU Berlin
- Gioni Mexi - Zuse Institute & TU Berlin
Computational Global Optimization with First-Order Methods
Mathematical optimization is a fundamental tool in various fields, including engineering, computer science, economics, and physics. The team aims to address key challenges in mixed-integer and convex optimization, which are crucial for solving real-world problems efficiently. The development of optimization algorithms, their analysis and convergence guarantees, and their efficient implementation are essential for handling large-scale and complex systems.
This associate team will be a joint effort between Inria (Grenoble and Lyon) and the Zuse Institute. We will tackle challenging optimization problems with applications in quantum information theory, learning, and transportation. The collaboration is inscribed within the long-term lines of research of the participants, designing and implementing solution methods for classes of increasing complexity. Some recent work from members of the team already showed the potential of first-order methods as a building block to find solutions to nonconvex quadratic problems.