Within the mathematical programming community, it is commonly accepted that a universal optimization strategy is impossible, and that the only way one strategy can outperform another is if it is specialized to the structure of the considered optimization problem. It is also well-known that some problem structures are numerically more tractable than others, where probably the most important distinction is the one between convex programs (nonlinear programs of which every local optimum is also globally optimal) and nonconvex programs.
Within the engineering community, the importance of obtaining and exploiting favorable (preferably convex) structure is far less well-known, that is, fairly standard knowledge within specific application areas (automatic control, signal processing, system identification, shape optimization of mechanical parts) but virtually unknown in many others. The main rationale underlying the creation of OPTEC is that applying these insights to a large variety of engineering applications has great potential to shift the state of the art in these research areas. More details concerning the research program are provided below.
All research at OPTEC is guided by the idea that method development and real world engineering applications should go closely hand in hand. This principle ensures that:
- Challenging engineering applications are addressed by the most modern optimization methods, and the problems are formulated already with efficient solution methods in mind.
- Optimization algorithm development is guided and inspired by real world engineering practice, so that new interesting problem classes can be quickly discovered and structure exploiting software be developed for these classes.
Since 2009, the research activities inside OPTEC are organized in working groups. These groups consist of self-chosen members from at least three departments which are interested in the same methodological area. The six working groups are:
- Dynamic and Embedded Optimization;
- Data-driven modelling;
- Parameter and State Estimation;
- Shape and Topology Optimization;
- Advanced Linear Systems Analysis and Control;
- PDE-constrained Optimization.
Experimental validation of the developed optimization methodologies and algorithms is an essential element for the success of OPTEC’s mission. Besides the many, often simplified, academic test setups that will be used in the different working groups, it is also essential to consider real-world applications with high industrial and/or societal impact, and to demonstrate feasibility and improved performance. For this reason, OPTEC research will focus on the following application projects:
- Optimal Human-Robot Interaction;
- Bioreactor Optimization;
- Kite Power;
- Assessment of the Response of a Footbridge to Pedestrian Excitation.
A description of the previous OPTEC organization is avilable here.
