Investigating direct/indirect methods for solving optimal control problems

Moshe Ben Saadon
Work towards MSc degree under the supervision of Professor Joseph Ben Asher (Technion)
Department of Aerospace Engineering
Technion – Israel Institute of Technology

Optimal control problems can be classified into two cases: indirect or direct approach. The 1st class is based on the minimum principle and several key conditions to obtain a TPBVP (Two Point Boundary Value Problem), which is then solved numerically in order to find the optimal trajectories. The advantage of these methods is their high solution accuracy, and certainty in satisfying the optimality conditions. The direct methods are based on parametrization of the problems as dimensionally finite, therefore can be solved by specific algorithms for constrained parameter optimization. These methods are less prone to convergence problems, but are less accurate. In this research, the performance of several solvers (which employ these methods) is compared using predefined use cases. The abundance of generated data is then being collected and analyzed using self-developed data analysis tools, that are based on optimal control metrics in order to grade the behavior of the results – which in turn are translated into insights and recommendations concerning the usage of the aforementioned solvers.

The talk will be given in Hebrew

Wed, 21-09-2022, 13:30-14:30 (Gathering at 13:30)

Classroom 165, ground floor, Library, Aerospace Eng. & https://technion.zoom.us/j/2541346366

Light refreshments will be served before the lecture


Investigating direct/indirect methods for solving optimal control problems