Closed joseph-gergaud closed 6 months ago
@joseph-gergaud You mean not transforming the Lagrangian part of the cost (could be Bolza) as an additional state integrated with the dynamics? For instance by using a specific numerical scheme for the quadrature of the cost?
Yes
And one more to the profiling pile
Well, I guess this one goes to low priority.
Do a version of CTDirect with the conservation of the Lagrangian cost and do comparaisons with the Mayer formulation (actual version)