Error using NLOPT and the AMPL interface

47 views
Skip to first unread message

Walton P. Coutinho

unread,
Nov 28, 2018, 3:35:13 PM11/28/18
to OPTI Toolbox Forum
Hi,

I am running into the error message below whenever I try to solve the attached 
AMPL model using NLOPT (calling GN_ISRES).

solveProb is my matlab wrapper function for calling different solvers.

>> [x,fval,exitflag,info] = solveProb('dubins.mod', 'NLOPT');

------------------------------------------------------------------
 This is NLOPT v2.4.2
 Author: Steven G. Johnson

 Algorithm: ISRES evolutionary constrained optimization (global, no-derivative)

 Problem Properties:
 # Decision Variables:      395
 # Inequality Constraints:  596
 # Equality Constraints:      0
------------------------------------------------------------------

 *** ERROR: Invalid Arguments ***
------------------------------------------------------------------

The problem is that I can't figure out what are the invalid arguments.
dubins.mod
solveProb.m

Jonathan Currie

unread,
Dec 3, 2018, 3:15:32 AM12/3/18
to opti-tool...@googlegroups.com
Hi,

Can you please supply the .nl version of your AMPL model, I no longer have access to the full version of AMPL.

This can be achieved by calling:

>> amplConvert('dubins.mod')

And sending the resulting .nl.

Jonathan

--
You received this message because you are subscribed to the Google Groups "OPTI Toolbox Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to opti-toolbox-fo...@googlegroups.com.
To post to this group, send email to opti-tool...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Walton P. Coutinho

unread,
Dec 3, 2018, 3:06:49 PM12/3/18
to OPTI Toolbox Forum
Dear Jonathan,

Please find the .nl file attached.

Best,
To unsubscribe from this group and stop receiving emails from it, send an email to opti-toolbox-forum+unsub...@googlegroups.com.
dubins.nl

Jonathan Currie

unread,
Dec 4, 2018, 3:02:46 AM12/4/18
to OPTI Toolbox Forum
Hi,

I can only guess that your chosen NLOPT algorithm (GN_ISRES) does not like infinite bounds, as your problem has. Changing these to finite seems to allow it to run.

In addition, NLOPT algorithm LN_COBYLA is my favourite and seems to do quite well with your problem.

Jonathan

To unsubscribe from this group and stop receiving emails from it, send an email to opti-toolbox-fo...@googlegroups.com.

To post to this group, send email to opti-tool...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "OPTI Toolbox Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to opti-toolbox-fo...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages