COUENNE with Julia

Specific Domains Optimization (Mathematical)
21

Sorry I am not following what is happening.
You are trying to use Baron now?
If it is not about Counne anymore I recommend a separate topic.

Also, you should start by making sure the Baron executable works fine and then try to use from JuMP.
You should probably follow Baron installation manual for that. I have no experience at all with Baron.
You might be talking about this: https://github.com/joehuchette/BARON.jl
Note that the latest version of JuMP does no support this.

Finally, what kind of problem class do you want to solve?
The solvers you are looking usually do not scale much.

22

I have a non convex nlp problem @joaquimg

23

Interesting, its pretty broad though. Feel free to open a new topic to continue the discussion!

1 Like
24

@joaquimg
can you provide me with useful links to integrate the couenne with julia
I downloaded the visual studio necessary file ā€¦ What next?

25

Hey,
The couenne .exe that you downloaded from AMPL webpage still does not work?
even after the Visual Studio files?

26

I tried to uninstall it and delete it, then download and install it again but now when I run the exe file, I do not get anything ā€¦ @joaquimg

27

What have you downloaded ? the binary or the source code and did you use this link for installation ?
https://github.com/ampl/coin/blob/master/Couenne/INSTALL

28

Open a cmd session. And call couenne from it.
It should display something.

29

@joaquimg
like that ?

cmd
cmd754Ɨ454 10 KB

30

Use the full path to the executbale.

31

sd
@joaquimg
this is what I get

32

Looks good. This should work thenā€¦

1 Like
33

o k
I will try now and tell you what happened

34

@joaquimg
I get error ā€œOptimizer not definedā€

run1

35

You need AmplNLWriter version 0.6. Try ] add AmplNLWriter@0.6.

1 Like
36

@odow
I tried this

dd
dd719Ɨ459 71.6 KB

37

Run ] rm CouenneNL Then run ] up.

1 Like
38

rm means to remove CouenneNL right ?
I ran the program now
however I got ā€œOTHER_ERRORā€ as a termination status

39

Yes.

the screen log should give you more info. My experience is that couenne is not super robust though.

40

so what do you recommend as robust solvers? @joaquimg