Hi,<br><br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
However, with a current Bonmin 1.5, it seems to work correctly.<br>
So I guess this bug has been fixed in Bonmin. (Pierre?)<br></blockquote><div>I was not aware of the bug but I am glad if it was fixed :)</div><div>Best,</div><div>Pierre</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
It will also be available in the next GAMS release, which I hope will be out soon.<span class="HOEnZb"><font color="#888888"><br>
<br>
Stefan</font></span><div class="HOEnZb"><div class="h5"><br>
<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Hi,<br>
<br>
I'm sending you a slightly modified model. I'm solving it with the<br>
following options:<br>
<br>
bonmin.milp_solver Cplex<br>
bonmin.allowable_fraction_gap 0.1<br>
bonmin.algorithm B-QG<br>
<br>
- This is the output I get:<br>
<br>
Bonmin finished. Found feasible point. Objective function value = -9.40927.<br>
Resolve with fixed discrete variables to get dual values.<br>
NLP0014I 4 OPT -9.4093 20 1.963<br>
Warning: Optimal value of NLP subproblem differ from best MINLP value<br>
reported b<br>
y Bonmin.<br>
Will not replace solution, dual values will not be available.<br>
Best solution: -1.859887e+001 (41 nodes, 12.074 seconds)<br>
Best possible: 9.409270e+000 (only reliable for convex models)<br>
<br>
Absolute gap: 2.800814e+001 (absolute tolerance optca: 0)<br>
Relative gap: 297.665415% (relative tolerance optcr: 0.1%)<br>
<br>
<br>
- If I solve with B-OA, I get the following:<br>
<br>
Bonmin finished. Found feasible point. Objective function value = -9.40927.<br>
Resolve with fixed discrete variables to get dual values.<br>
NLP0014I 3 OPT -9.4093 25 2.488<br>
<br>
Best solution: 9.409269e+000 (0 nodes, 16.239 seconds)<br>
Best possible: 9.409268e+000 (only reliable for convex models)<br>
<br>
Absolute gap: 1.460884e-006 (absolute tolerance optca: 0)<br>
Relative gap: 0.000016% (relative tolerance optcr: 0.1%)<br>
<br>
<br>
Hope this helps,<br>
Juan<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Hi,<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi,<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
This is an example output I'm getting from a model solved by bonmin:<br>
<br>
Bonmin finished. Found feasible point. Objective function value =<br>
-8.477.<br>
Resolve with fixed discrete variables to get dual values.<br>
NLP0014I 4 OPT -8.477 18 1.887<br>
Warning: Optimal value of NLP subproblem differ from best MINLP value<br>
reported by Bonmin.<br>
Will not replace solution, dual values will not be available.<br>
<br>
<br>
Best solution: -9.701475e+000 (36 nodes, 12.637 seconds)<br>
Best possible: 8.477002e+000 (only reliable for convex models)<br>
<br>
Absolute gap: 1.817848e+001 (absolute tolerance optca: 0)<br>
Relative gap: 214.444651% (relative tolerance optcr: 0.05%)<br>
<br>
However, the primal solution given to GAMS corresponds to objective<br>
value -9.701475, and not 8.477002.<br>
</blockquote>
<br>
The best feasible solution that Bonmin found has value -9.701475.<br>
The 8.477002 is only an upper bound that tells you that there exists<br>
no solution with a value above 8.477002.<br>
</blockquote>
<br>
<br>
I thought the output line:<br>
"Bonmin finished. Found feasible point. Objective function value =<br>
-8.477."<br>
<br>
meant that bonmin has found a feasible point with objective value<br>
8.477. I don't understand why it subsequently says that the best<br>
solution is -9.701475, which is much worse. In fact, I thought bonmin<br>
is exact for convex models. The solution reported, however, is way<br>
outside the required tolerance levels, as you can see in the output.<br>
On the other hand, if I try with a different algorithm (OA), the best<br>
solution found *is* 8.477.<br>
<br>
I'll send you a model later to see if you can reproduce the problem.<br>
<br>
Juan<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Stefan<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Juan<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi,<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hello,<br>
<br>
I'm trying to solver a number of convex MINLP instances of a problem<br>
using Bonmin through GAMS and am frequently getting this error<br>
message<br>
when bonmin finishes:<br>
<br>
Warning: Optimal value of NLP subproblem differ from best MINLP value<br>
reported by Bonmin.<br>
Will not replace solution, dual values will not be available.<br>
<br>
When this happens, GAMS does not receive the last feasible point (and<br>
supposedly optimal solution) found by bonmin. So first of all I'm not<br>
sure if bonmin is correctly calculating the solution in these<br>
cases, and<br>
if it is, it's not sending the solution to GAMS. Is there any way to<br>
solve this or get around it?<br>
</blockquote>
<br>
When Bonmin finishes and has found a feasible solution, then the<br>
Gams/Bonmin link solves the NLP obtained from the MINLP by fixing the<br>
discrete variables to the values in the solution. It's doing this to<br>
compute dual values.<br>
The warning says, that this final NLP solve ended with a solution that<br>
has a worse objective function value than the solution value reported<br>
by Bonmin.<br>
However, in this case, you should still get the primal solution that<br>
Bonmin reported, only a dual solution will be unavailable.<br>
<br>
If you don't get Bonmin's solution, it would be great if you could<br>
send me a model to reproduce this behaviour.<br>
<br>
In the next GAMS release, there will be an option to disable the final<br>
solve in the Gams/Bonmin link.<br>
<br>
Stefan<br>
<br>
<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Regards,<br>
Juan<br>
<br>
<br>
______________________________<u></u>_________________<br>
Bonmin mailing list<br>
<a href="mailto:Bonmin@list.coin-or.org" target="_blank">Bonmin@list.coin-or.org</a><br>
<a href="http://list.coin-or.org/mailman/listinfo/bonmin" target="_blank">http://list.coin-or.org/<u></u>mailman/listinfo/bonmin</a><br>
<br>
</blockquote>
<br>
<br>
</blockquote>
<br>
<br>
<br>
</blockquote>
<br>
<br>
</blockquote>
<br>
<br>
<br>
--<br>
Juan José Gálvez García, Ph.D.<br>
Dept. Ingeniería de la Información y las Comunicaciones<br>
Facultad de Informática<br>
Universidad de Murcia<br>
Campus de Espinardo, 30100 E-mail: <a href="mailto:jjgalvez@um.es" target="_blank">jjgalvez@um.es</a><br>
Murcia, SPAIN Phone: <a href="tel:%2B34868887882" value="+34868887882" target="_blank">+34868887882</a><br>
<br>
<br>
______________________________<u></u>_________________<br>
Bonmin mailing list<br>
<a href="mailto:Bonmin@list.coin-or.org" target="_blank">Bonmin@list.coin-or.org</a><br>
<a href="http://list.coin-or.org/mailman/listinfo/bonmin" target="_blank">http://list.coin-or.org/<u></u>mailman/listinfo/bonmin</a><br>
<br>
</blockquote>
<br>
</blockquote>
<br>
______________________________<u></u>_________________<br>
Bonmin mailing list<br>
<a href="mailto:Bonmin@list.coin-or.org" target="_blank">Bonmin@list.coin-or.org</a><br>
<a href="http://list.coin-or.org/mailman/listinfo/bonmin" target="_blank">http://list.coin-or.org/<u></u>mailman/listinfo/bonmin</a><br>
</div></div></blockquote></div><br>