New instance of Space Shuttle ( Task #4)

[0Yassine0]

@frapac
As discussed, I added the model of space shuttle inspired from Constructive Methods for Initialization and Handling Mixed State-Input Constraints in Optimal Control.
I created the 2 model files : with JuMP space_Shuttle_JMP_V2 and OptimalControl space_Shuttle_OC_V2 and a notebook testSpaceShuttleV2.
With OptimalControl, I don't have a valid results yet. I can't run it on the notebook.
The execution is done in a separate terminal with

using OptimalControl
using NLPModelsIpopt
include("./Problems/OptimalControl/space_Shuttle_OC_V2.jl")
SpaceSV2OC = space_Shuttle_OC_V2();
import HSL_jll
nh = 502
init = space_Shuttle_init(;nh=nh)
sol = OptimalControl.solve(SpaceSV2OC, grid_size=nh, init =init,
              tol=1e-8, constr_viol_tol=1e-6, max_iter=1000,
              linear_solver="ma57",hsllib=HSL_jll.libhsl_path, print_timing_statistics="yes")

[frapac]

Can you try solving the problem with HSL ma27 instead of HSL ma57?

[0Yassine0]

Yes, the results below are with HSL ma27

Somehow when I run the solving a second time on the same terminal I get this error!

[frapac]

Is this with OC only, or both with OC and JuMP?

[0Yassine0]

Unfortunately only with OC.
I can run it normally with JuMP and the results for JuMP in the notebook are with ma27.

[frapac]

Perfect, thank you.
Looking at the log, OC is facing an issue with the linear solver. We should assess what's going on exactly (but we can do that later).

[frapac]

@0Yassine0 I had a closer look at the notebooks, and it looks like we get exactly the same solution between testSpaceShuttleJMP.ipynb and testSpaceShuttleV2.ipynb (both objective are equal, and primal solutions are very similar). I am not sure we are considering the state constraints there.

Would it be possible to reproduce and plot the Figure 6 in the initial paper to check that we are indeed constraining the heating on the shuttle?

[ocots]

@0Yassine0 Can we close this issue?

[frapac]

reopening this issue, as I think the space shuttle reentry instance with state constraints is not having the expected behavior, cf #22 (comment)