diff --git a/course/lecture/lecture2.md b/course/lecture/lecture2.md index d337c4eff7051b845bb19d12340ac36567d451e5..71c30f9d0340aedbba3e5863fed4a2582d9f0a2b 100644 --- a/course/lecture/lecture2.md +++ b/course/lecture/lecture2.md @@ -81,7 +81,7 @@ and we need a [Newton-like](https://en.wikipedia.org/wiki/Newton%27s_method) sol It is important to understand that if $`(p_0^*, \lambda^*)`$ is solution of $`S=0`$, then the control $`u(\cdot) := u[z(\cdot, x_0, p_0^*)]`$ is a candidate as a solution of the optimal control problem. It is only a candidate and not a solution of the OCP since the PMP gives necessary conditions of optimality. We would have to go further to check whether the control is locally or globally optimal. <div align="center"> - <img src="attachment:21f9cc6f-5061-4fbe-ab9b-b905c2389dc9.png" width="400"> + <img src="../images/simple_shooting.png" width="400"> </div> <div align="center"> <i> @@ -89,3 +89,25 @@ Figure: Illustration of the shooting method in the state-time space. The blue tr </i> </div> +<div align="center"> + <img src="../images/simple_shooting_2.png" width="420"> +</div> +<div align="center"> +<i> +Figure: Illustration of the shooting method in the cotangent space. The blue extremal reaches the target in red. +</i> +</div> + +[Additional comments (video)](http://www.youtube.com/watch?v=YVG2Z_TEkBQ) + +### b) The iteration of the Newton solver and the Jacobian of the shooting function + +[Additional comments (video)](http://www.youtube.com/watch?v=hVbi9kShR90) + +### c) A word on the Lagrange multiplier + +[Additional comments (video)](http://www.youtube.com/watch?v=3VEi-UHAS6w) + +## IV) Numerical resolution of the shooting equations with the nutopy package + +This part is there: [notebook lecture](lecture.ipynb). Please follow the procedure given [here](../../README.md) to install all packages and use `jupyter-lab` to execute the code and do the exercices.