Comparison of direct collocation methods for optimal control

Jennifer Hofmann

University of Freiburg

Tuesday, February 06, 2018, 11:00

Room 02-012, Georges-Köhler Allee 102, Freiburg 79110, Germany

After giving a short introduction into optimal control, the direct collocation method will be discussed in details. Such a method divides the time horizon into subintervals and on each of them the state and control trajectories are approximated by polynomials. The performance in terms of accuracy and CPU-time associated with the solution of optimal control problems using different collocation methods will be assessed. In particular, different combinations of discretization intervals and polynomials degrees for both state and control trajectories will be considered. The benchmark shows that the performance largely depends on several factors. In this presentation, we will try to describe and discuss different scenarios with a focus on a comparison between local and global collocation methods.