Error concerning ipopt when using ATOM fmincon() in Scilab 2026

Tools Atoms packages
1

I need to use the ATOM fmincon. I have installed it in Fedora 44 both in Scilab 2026.0.1 and the latest Scilab 2026.1.0 (installed with the official .tar.xz),with its two dependencies via the command atomsInstall("fmincon"):

--> atomsGetInstalled()
fmincon    1.0.9  allusers  SCI/contrib/fmincon/1.0.9    I
sci_ipopt  1.3.5  allusers  SCI/contrib/sci_ipopt/1.3.5  A
apifun     0.4.3  allusers  SCI/contrib/apifun/0.4.3     A

In order to learn fmincon(), I have downloaded the demos from its GitLab: demos · master · scilab / forge / fmincont · GitLab

--> exec('/home/blabla/fmincont/demos/fmincon_hesspattern.sce', -1)
Ă  la ligne   287 de la fonction fmincon_ipopt (/home/blabla/scilab-2026.1.0/share/scilab/contrib/fmincon/1.0.9/macros/fmincon/fmincon_ipopt.sci ligne 297 )
Ă  la ligne   406 de la fonction fmincon       ( /home/blabla/scilab-2026.1.0/share/scilab/contrib/fmincon/1.0.9/macros/fmincon/fmincon.sci ligne 418 )
à la ligne    45 du fichier exécuté /home/blabla/fmincont/demos/fmincon_hesspattern.sce

ipopt : Taille erronée de l'argument d'entrée #1 : une seule structure attendue.

I have tried several other demos, and obtain always the same error about ipopt.

2

Hello,

I updated sci_ipopt yesterday evening, you should be fine with

atomsRemove fmincon

quit and relaunch Scilab, then

atomsSystemUpdate
atomsInstall fmincon
atomsLoad fmincon

S.

1 Like
3

Thanks a lot @mottelet ,
I can confirm it now works in Scilab 2026.1.0! I have tested successfully several demos from the fmincon GitLab sources.
Thanks for your efforts maintaining it.
V.

4

Thanks for using fmincon. If you didn’t already run this example : https://scilab.discourse.group/t/large-scale-nonlinear-optimization-with-fmincon-in-scilab-vs-matlab, I encourage you to do so, as it shows that fmincon+IPopt outperforms Matlab’s fmincon.

S.

1 Like