Thank you. <div>I'll try printing the function evaluations on the screen to trace the program. But as for increasing the print level of ipopt, I don't know how to do it.<div>In the tutorial on the official website, it's said that after the compile of the ipopt matlab interface, I can get the usage instruction of it by typing 'help ipopt' in the matlab prompt like the other matlab original functions if I set the ipopt.mexw32 directory in the 'set path' window. But after those procedure including compile and setting path, I actually get 'ipopt not found' instead of any usage instruction when I typed 'help ipopt'. So the only way to learn how to use ipopt matlab interface is the four examples in the ipopt source code.</div>
<div>So if there's any detail usage document of the ipopt matlab interface, please send me a copy.</div><div>Thank you very much.<br clear="all"><div>------------------------------------------------------------<br>Regards<br>
Neo Ma 马平川 | Graduate<br><br>T:+86-10-6278 2545 | <a href="mailto:E%3Achambertinofn@gmail.com" target="_blank">E:chambertinofn@gmail.com</a><br>------------------------------------------------------------<br>Dept. of Electrical Engineering, Tsinghua Univ.<br>
BLDG. 28# RM.312 Tsinghua Univ. Beijing 100084 P.R.CHINA<br>------------------------------------------------------------<br>Remember what should be remembered, and forget what should be forgotten. Alter what is changeable, and accept what is immutable.</div>
<br>
<br><br><div class="gmail_quote">On Wed, Nov 7, 2012 at 9:27 PM, Stefan Vigerske <span dir="ltr"><<a href="mailto:stefan@math.hu-berlin.de" target="_blank">stefan@math.hu-berlin.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi,<br>
<br>
Easiest should be if you print your values in your implementation of the function and gradient evaluation routines. If its not in the function values, then maybe in a gradient (e.g., sqrt(x) has an infinite derivative at 0).<br>
Check your starting point.<br>
Increase the print level of Ipopt to get more output.<br>
<br>
Stefan<div><div class="h5"><br>
<br>
On 11/07/2012 01:46 PM, 马平川 wrote:<br>
</div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5">
I'm very grateful for your last reply. I had gone over my own function<br>
evaluations. There's no logarithm function or a divide by zero.<br>
My test suite have 311 decision variable and 253 equality constraints, let<br>
alone 311 variable bound constraints. What scale does ipopt could deal<br>
with. But I thought my test suite wasn't big enough.<br>
Is there any other reason that could lead to the same error message "EXIT:<br>
Invalid number in NLP function or derivative detected. "<br>
Expecting your reply.<br>
Thank you.<br>
------------------------------<u></u>------------------------------<br>
Regards<br>
Neo Ma 马平川 | Graduate<br>
<br>
T:+86-10-6278 2545 | <a href="mailto:E%3Achambertinofn@gmail.com" target="_blank">E:chambertinofn@gmail.com</a><br>
------------------------------<u></u>------------------------------<br>
Dept. of Electrical Engineering, Tsinghua Univ.<br>
BLDG. 28# RM.312 Tsinghua Univ. Beijing 100084 P.R.CHINA<br>
------------------------------<u></u>------------------------------<br>
Remember what should be remembered, and forget what should be forgotten.<br>
Alter what is changeable, and accept what is immutable.<br>
<br>
<br>
<br>
On Tue, Nov 6, 2012 at 9:29 PM, Miao Li <<a href="mailto:miao.li@epfl.ch" target="_blank">miao.li@epfl.ch</a>> wrote:<br>
<br>
</div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5">
" EXIT: Invalid number in NLP function or derivative detected. ":The<br>
error message means that Ipopt detected an error in function evaluations,<br>
typically a divide by zero or a logarithm of a negative number. You'd<br>
better check you model about this issue.<br>
For the second, you can change the number of maximum iteration if I<br>
understand your problem correctly.<br>
<br>
Best,<br>
Miao<br>
<br>
<br>
2012/11/6 马平川 <<a href="mailto:chambertinofn@gmail.com" target="_blank">chambertinofn@gmail.com</a>><br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
First, this is the only way to post my problems that I can find on the<br>
Ipopt home page. Maybe this message would be send to all the members on the<br>
mailing list. I'm sorry for bothering you with this mail. But if you could<br>
solve my problems, I would appriciate you for your reply.<br>
<br>
I'm coding in Matlab for my master degree issue. But there was some<br>
unknown bugs in my program. I worried about that mistakes in my Primal-dual<br>
Inner Point Method codes had led these bugs. So I compiled the Ipopt matlab<br>
interface on Windows (ipopt.mexw32) on my PC. And then all the test which<br>
was from the Ipopt code (\build\Ipopt\contrib\<u></u>MatlabInterface\examples)<br>
using my own compiled Ipopt matlab interface(ipopt.mexw32) has passed.<br>
<br>
But the first running result after inputting 'clear all' in the matlab<br>
prompt is below.<br>
<br>
******************************<u></u>******************************<u></u>******************<br>
This program contains Ipopt, a library for large-scale nonlinear<br>
optimization.<br>
Ipopt is released as open source code under the Eclipse Public License<br>
(EPL).<br>
For more information visit <a href="http://projects.coin-or.org/Ipopt" target="_blank">http://projects.coin-or.org/<u></u>Ipopt</a><br>
******************************<u></u>******************************<u></u>******************<br>
<br>
This is Ipopt version 3.10.0, running with linear solver ma27.<br>
<br>
Number of nonzeros in equality constraint Jacobian...: 1336<br>
Number of nonzeros in inequality constraint Jacobian.: 0<br>
Number of nonzeros in Lagrangian Hessian.............: 3236<br>
<br>
<br>
Number of Iterations....: 0<br>
<br>
Number of objective function evaluations = 0<br>
Number of objective gradient evaluations = 1<br>
Number of equality constraint evaluations = 0<br>
Number of inequality constraint evaluations = 0<br>
Number of equality constraint Jacobian evaluations = 1<br>
Number of inequality constraint Jacobian evaluations = 0<br>
Number of Lagrangian Hessian evaluations = 0<br>
Total CPU secs in IPOPT (w/o function evaluations) = 0.082<br>
Total CPU secs in NLP function evaluations = 0.006<br>
<br>
EXIT: Invalid number in NLP function or derivative detected.<br>
<br>
And if I rerun this program without doing anything at once. Usually I can<br>
get this result.<br>
This is Ipopt version 3.10.0, running with linear solver ma27.<br>
<br>
Number of nonzeros in equality constraint Jacobian...: 1336<br>
Number of nonzeros in inequality constraint Jacobian.: 0<br>
Number of nonzeros in Lagrangian Hessian.............: 3236<br>
<br>
Total number of variables.....................<u></u>.......: 310<br>
variables with only lower bounds: 39<br>
variables with lower and upper bounds: 38<br>
variables with only upper bounds: 0<br>
Total number of equality constraints.................: 253<br>
Total number of inequality constraints...............: 0<br>
inequality constraints with only lower bounds: 0<br>
inequality constraints with lower and upper bounds: 0<br>
inequality constraints with only upper bounds: 0<br>
<br>
iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du<br>
alpha_pr ls<br>
0 0.0000000e+000 8.31e+000 4.49e+299 0.0 0.00e+000 - 0.00e+000<br>
0.00e+000 0<br>
1 8.4037298e+005 8.30e+000 1.59e+000 -1.5 1.84e+301 - 6.87e-300<br>
6.50e-301h 5<br>
2 2.0336766e+008 6.34e+002 1.64e+002 0.3 9.20e+001 - 8.39e-001<br>
1.00e+000f 1<br>
...<br>
...<br>
...<br>
Number of Iterations....: 100<br>
<br>
(scaled) (unscaled)<br>
Objective...............: 8.5002463604944314e+003<br>
8.5002463604944312e+011<br>
Dual infeasibility......: 1.8701445737198796e+003<br>
1.8701445737198798e+011<br>
Constraint violation....: 3.9817380411393437e+001<br>
7.1727917826186584e+001<br>
Complementarity.........: 1.8222872952405129e+002<br>
1.8222872952405128e+010<br>
Overall NLP error.......: 1.8701445737198796e+003<br>
1.8701445737198798e+011<br>
<br>
<br>
Number of objective function evaluations = 1016<br>
Number of objective gradient evaluations = 101<br>
Number of equality constraint evaluations = 1216<br>
Number of inequality constraint evaluations = 0<br>
Number of equality constraint Jacobian evaluations = 101<br>
Number of inequality constraint Jacobian evaluations = 0<br>
Number of Lagrangian Hessian evaluations = 100<br>
Total CPU secs in IPOPT (w/o function evaluations) = 0.391<br>
Total CPU secs in NLP function evaluations = 6.904<br>
<br>
EXIT: Maximum Number of Iterations Exceeded.<br>
<br>
Please neglect the unconverged problem because my own faults may cause<br>
this problems. My problem is that the running result of exceeding the<br>
maximum number iterations or of jumping out without any iteration happened<br>
without any regular patterns. I wonder what had led this problem and how to<br>
avoid this.<br>
My PC configuration:<br>
OS: windows vista home premium X86<br>
Matlab 7.10.0(R2010a)<br>
Ipopt 3.10.0<br>
<br>
Thank you for your reading. Hoping your reply.<br>
------------------------------<u></u>------------------------------<br>
Regards<br>
Neo Ma 马平川 | Graduate<br>
<br>
T:+86-10-6278 2545 | <a href="mailto:E%3Achambertinofn@gmail.com" target="_blank">E:chambertinofn@gmail.com</a><br>
------------------------------<u></u>------------------------------<br>
Dept. of Electrical Engineering, Tsinghua Univ.<br>
BLDG. 28# RM.312 Tsinghua Univ. Beijing 100084 P.R.CHINA<br>
------------------------------<u></u>------------------------------<br>
Remember what should be remembered, and forget what should be forgotten.<br>
Alter what is changeable, and accept what is immutable.<br>
<br>
<br>
______________________________<u></u>_________________<br>
Ipopt mailing list<br>
<a href="mailto:Ipopt@list.coin-or.org" target="_blank">Ipopt@list.coin-or.org</a><br>
<a href="http://list.coin-or.org/mailman/listinfo/ipopt" target="_blank">http://list.coin-or.org/<u></u>mailman/listinfo/ipopt</a><br>
<br>
<br>
</blockquote>
<br>
<br>
--<br></div></div>
*<br>
Miao Li*<br>
* Learning Algorithms and Systems Laboratory** (LASA)*<br>
*École polytechnique fédérale de Lausanne (EPFL)<br>
<br>
*<br>
<br>
</blockquote><div class="im">
<br>
<br>
<br>
______________________________<u></u>_________________<br>
Ipopt mailing list<br>
<a href="mailto:Ipopt@list.coin-or.org" target="_blank">Ipopt@list.coin-or.org</a><br>
<a href="http://list.coin-or.org/mailman/listinfo/ipopt" target="_blank">http://list.coin-or.org/<u></u>mailman/listinfo/ipopt</a><br>
<br>
</div></blockquote>
<br>
</blockquote></div><br></div></div>