Lasse Hjuler Christiansen: Optimal Control of PDE-constrained Systems

Many engineering and scientific problems in design, control, and parameter estimation can be formulated as optimization problems that are constrained by partial differential equations (PDEs). Applications comprise diverse areas such as geoscience, chemical process industry, aerodynamics, quantum systems in physics and chemistry, medicine, and financial engineering. As a result, science and industry express an increasing demand for numerical methods and software to efficiently solve PDE-constrained optimization problems. To this end, this thesis proposes and investigates new methods and algorithms that contribute to ensure fast, efficient and robust PDE-constrained optimization solvers of practical relevance. In this regard, the thesis addresses a selection of computational challenges that are central to largescale non-linear model predictive control (NMPC). The associated contributions fall into two main parts:

New methods for efficient solution of multi-objective optimization problems that arise in oil reservoir management.

Customized iterative solvers tailored for a class of PDE-constrained optimization problems that are central to applications in physics, biology and chemical engineering.

In a range of simulation-based case studies, the new methods and solvers demonstrate a potential to improve current state-of-the-art methodologies and/or conventional industrial practices. In this way, the contributions of this thesis can be seen as new steps towards NMPC for large-scale processes that can be described by PDEs.


Mon 25 Mar 19
13:00 - 16:00


DTU Compute


Building 341, Auditorium 22

Files for download