Research

The scientific programme builds the foundation of a long-term European research & training network in
the field of optimal control theory, with emphasis on: 

• Sensitivity analysis, which is concerned with the robustness of optimal control strategies to changes in theunderlying optimization problem
• Deterministic controller design based on real-time solution of optimal control problems.

The scientific activities of the project are split into 5 scientific workpackages:

• WP1: Necessary and sufficient conditions
• WP2: Hamilton-Jacobi theory
• WP3: Stabilization of nonlinear systems
• WP4: Perturbed and large-scale systems
• WP5: Differentential games
  

Optimal control concerns the determination of control strategies for complex, dynamic systems, to optimize some measure of best performance.

Optimal Control came into existence in the 1960's, driven by the 'space race' between the US and the former USSR, and the need of new theory and computational methods, for the calculation of flight trajectories in space exploration. But the field now has far wider scope than its first applications to aerospace engineering would suggest, and now embraces areas where the dynamical models describe chemical reactors (process control), vehicles (traffic flow control), wind generators and solar panels (power systems control) or might even describe 'virtual' economic systems relating management decisions to economic consequences over time (econometrics and resource economics).

Sensitivity analysis and optimization-based controller design, which are central to our programme, are two especially active themes of optimal control research:

• On the one hand, applied optimal control is often based on the use of idealized models providing only an approximate description of highly complicated phenomena and sensitivity analysis is needed to study the effects of these approximations.
• Optimization-based controller design is attracting increasing interest in industry, because of its versatility and constrainy-handling capability. The need for a coordinated programme of training in sensitivity analysis for controller design arises from a number of significant trends on modern day technology.