Add FeasibilityDrivenDDPSolver (FDDP)

[wxmerkt]

This PR adapts and adds the Crocoddyl implementation of FDDP into Exotica.

The following differences have to be noted:

• Crocoddyl uses T running models and a final model after. This leads to differences in T lengths and on top of that an additional model. Care needs to be taken to adapt all T and iterations
• The StateDelta in Pinocchio works the other way around in the argument list.
• I did not yet add an Integrate as the Pinocchio integrate assumes unit time, while we usually use dt_ in the dynamics.

This now works well - better on some dynamics systems than on others.

This version currently has numerical issues - particularly with the state feedback gain matrix K_. It works better without in the forward-pass. This could be due

a) t/T arguments being wrong in a place, b) the differences in the cost computation with StateDelta (arguments are swapped) in the DynamicTimeIndexedShootingProblem, c) the calculation of the defects fs_ being wrong d) or something else.

The convergence is also poor compared with VanillaDDP.

[wxmerkt]

After some fixes to make them comparable (FDDP didn't use the dt_ running cost scaling), I now get the following on Cart-Pole (example 7):

The flat line at the beginning are the infeasible iterations in FDDP. Not sure why there are so many in the cart-pole example.