The DDP implementation #1071 is preliminary and could be made more efficient.

Testing:

• So far DDP is only tested for explicit integrators. For implicit integrators additional precautions need to be taken in the line search. In particular, this requires the possibility to stash the internal state of the implicit integrator.
• Adding to the previous point: problems with algebraic variables should be tested.
• The procedure for finding a feasible initial guess needs to be tested. When the problem dynamics are unstable it can happen that the DDP forward sweep gives a bad initial guess, and the step size should be reduced in this case.

Other:

• In the QP tolerances setter: The accuracy should be set to the minimum of the current QP accuracy and the one specified.
• Early termination: The nonlinear forward sweep should be terminated as soon as (a) NaNs are detected in one of the function evaluations or (b) the sufficient decrease condition is violated.
• Evaluate how costly the forward sweep of DDP is for the overall solve process.