Steven G. Johnson
Steven G. Johnson
No, the C API doesn't provide this. Note that, given a local optimum (i.e. the NLopt result), you can easily compute Lagrange multipliers simply by solving a small least-squares problem....
The type of solver doesn't matter. For any solver that returns a local optimum, the KKT equations tell you how to compute Lagrange multipliers.
I've renamed this issue and marked it as an "enhancement", as it would be nice to have a built-in API for this.
> It is using the GUILE_CONFIG_EXECUTABLE=guile-config command that in turn is a wrapper over pkg-config. No, `guile-config` is *not* a wrapper around `pkg-config`. It is installed as a part of...
You can print anything you want from your objective function. e.g. I typically add a few lines to my objective function to print a status output, e.g. maybe the value...
Note that we don't implement the original MMA algorithm, only the CCSA variant from 2001, so the parameters are a bit different. We could certainly add new algorithm-specific parameters (via...
In principle, yes — we could update [the code that evaluates f in the inner iterations](https://github.com/stevengj/nlopt/blob/496be736b8b249273838b891f4c8ca3669551127/src/algs/mma/mma.c#L292) to pass `NULL` for the gradient (indicating that it isn't needed). It would then...
> I can make a pull request to pass NULL for the gradients and remove the code that updates the gradient vector on inner iterations. Would that work? Yes, that...
What would this be for? (I'm not even sure how I would define "outer iteration" in general.) Maybe a callback that would be called each time a feasible point with...
I still don't understand how to define an "outer" iteration in a way that is meaningful for all algorithms, or why a GUI interface can't just update every time the...