Joel Andersson
Joel Andersson
That would make sense, though I wonder why it was not made `int` to begin with. In principle, you can also get an `int` overflow if you call it enough...
If you solve the optimal control problem by writing it as an NLP, you can get the Lagrange multipliers from the solver. These can be used to construct costate trajectories.
It's not clear to me what you exactly mean by the "adjoint state" here. Do you define the necessary conditions for optimality for the continuous time system using (Pontryagin's)? If...
Iirc, ```Function::call(const std::vector &arg, std::vector &res, ...)``` will not necessarily avoid the non-thread issues, for example if sparsity has to be projected. I think the easiest overload to use for...
ok, noted
In any case, I don't believe any new coloring algorithm should be implemented before the current coloring implementation is cleaned up and code duplication is removed.
Postponed to 2.1.
The best way to solve this issue appears to be to unpack the FMU into a directory with a unique name.
A pythonic way to unzip into a temporary directory and then load is something like: ```python import zipfile, tempfile with zipfile.ZipFile(fmu_file, 'r') as zf: with tempfile.TemporaryDirectory() as tempdir: zf.extractall(tempdir) dae...
Workaround solution: ```python def unpack_fmu(fmu_file, **kwargs): """ Unpack the contents of an FMU file """ # Read options unpacked_fmu = kwargs.pop('unpacked_fmu', None) verbose = kwargs.pop('verbose', True) # By default, strip...