<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css" style="display:none;"><!-- P {margin-top:0;margin-bottom:0;} --></style>
</head>
<body dir="ltr">
<div id="divtagdefaultwrapper" style="font-size:12pt;color:#000000;font-family:Calibri,Helvetica,sans-serif;" dir="ltr">
<p>Hi Marc.</p>
<p><br>
</p>
<p>I did some further investigation and the log output of Gurobi indicates some numerical issues:</p>
<p><br>
</p>
<p><span style="font-family: Consolas, Courier, monospace;">...</span></p>
<p></p>
<div><span style="font-family: Consolas, Courier, monospace;">Gurobi Optimizer version 9.0.2 build v9.0.2rc0</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">Optimize a model with 32793 rows, 102025 columns and 2555610 nonzeros</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">Coefficient statistics:</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> Matrix range [1e+00, 2e+03]</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> Objective range [3e+01, 6e+05]</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> Bounds range [1e+00, 1e+00]</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> RHS range [1e+00, 5e+04]</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">Iteration Objective Primal Inf. Dual Inf. Time</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> 0 4.3786197e+07 0.000000e+00 1.667889e+04 0s</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> 1152 4.3786110e+07 0.000000e+00 0.000000e+00 1s</span></div>
<div><br>
</div>
<div><span style="font-family: Consolas, Courier, monospace;">Solved in 1152 iterations and 0.90 seconds</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">Optimal objective 4.378611043e+07</span></div>
<div><span style="font-family: Consolas, Courier, monospace; color: rgb(255, 0, 0);">Warning: unscaled dual violation = 1.07138e-06 and residual = 1.07138e-06</span></div>
<div><span style="font-family: Consolas, Courier, monospace; color: rgb(255, 0, 0);">(node 1) numerical troubles in LP 117 -- solve again with primal simplex without scaling</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">...</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"><span style="font-family: Consolas, Courier, monospace; font-size: 16px;">...</span><br>
</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">Gurobi Optimizer version 9.0.2 build v9.0.2rc0</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">Optimize a model with 32326 rows, 87534 columns and 2182273 nonzeros</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">Coefficient statistics:</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> Matrix range [1e+00, 2e+03]</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> Objective range [3e+01, 6e+05]</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> Bounds range [1e+00, 1e+00]</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> RHS range [1e+00, 5e+04]</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">Iteration Objective Primal Inf. Dual Inf. Time</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> 0 4.4043728e+07 0.000000e+00 4.354066e+04 0s</span></div>
<div><span style="font-family: Consolas, Courier, monospace; color: rgb(255, 0, 0);">Warning: 1 variables dropped from basis</span></div>
<div><span style="font-family: Consolas, Courier, monospace; color: rgb(255, 0, 0);">Warning: switch to quad precision</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> 2506 4.4034847e+07 0.000000e+00 1.764252e+04 5s</span></div>
<div><span style="font-family: Consolas, Courier, monospace;"> 3271 4.4034676e+07 0.000000e+00 0.000000e+00 8s</span></div>
<div><br>
</div>
<div><span style="font-family: Consolas, Courier, monospace;">Solved in 3271 iterations and 8.15 seconds</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">Optimal objective 4.403467558e+07</span></div>
<div><span style="font-family: Consolas, Courier, monospace;">...</span></div>
<br>
<p></p>
Do I have any chance, to get an LP file of the program with that numerical issues? I tried to output the LP file at the end of each pricing iteration within my pricer, but that does not output exactly the LP that is solved in above examples? Is there anything
else I can do to figure out the root cause of that issues?<br>
<br>
Regards
<div>Christian<br>
<br>
<div style="color: rgb(0, 0, 0);">
<hr style="display:inline-block; width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" color="#000000" style="font-size:11pt"><b>Von:</b> Scip <scip-bounces@zib.de> im Auftrag von Franzen, Christian <franzen@or.rwth-aachen.de><br>
<b>Gesendet:</b> Sonntag, 20. Dezember 2020 09:58<br>
<b>An:</b> Marc Pfetsch; scip@zib.de<br>
<b>Betreff:</b> Re: [SCIP] "Resolve instable" in SCIP statistics?</font>
<div> </div>
</div>
<div>
<div id="divtagdefaultwrapper" dir="ltr" style="font-size:12pt; color:#000000; font-family:Calibri,Helvetica,sans-serif">
<p>Hi Marc.</p>
<p><br>
</p>
<p>Thanks for your support. I am using Gurobi as LP-solver. Scale of the coefficients should be fine, they are all between 1e2 and 1e6, however I have some Big-M's in my objective function that I can not get rid of. I am wondering why I run into stability issues
only when I solve my LP for some time (e.g., for a day). It seems to me that after some time my pricer generates columns that imply that issues. I will activate LP solver logging and see if Gurobi is giving me some hints.</p>
<p><br>
</p>
<p>Regards</p>
<p>Christian</p>
<br>
<br>
<div style="color:rgb(0,0,0)">
<div>
<hr style="display:inline-block; width:98%" tabindex="-1">
<div id="x_divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" color="#000000" style="font-size:11pt"><b>Von:</b> Scip <scip-bounces@zib.de> im Auftrag von Marc Pfetsch <pfetsch@mathematik.tu-darmstadt.de><br>
<b>Gesendet:</b> Freitag, 18. Dezember 2020 20:58<br>
<b>An:</b> scip@zib.de<br>
<b>Betreff:</b> Re: [SCIP] "Resolve instable" in SCIP statistics?</font>
<div> </div>
</div>
</div>
<font size="2"><span style="font-size:10pt">
<div class="PlainText"><br>
<br>
Hi Christian,<br>
<br>
the "resolve instable" indeed refers to the work that SCIP and the<br>
LP-solver try to obtain a numerically correct result. SCIP checks<br>
whether the solution returned by the solver is actually feasible. If<br>
this is not the case, then it tries to switch the LP-solver from dual to<br>
primal, tries to tighten the feasibility tolerance, starts from scratch<br>
(no warm start) etc.<br>
<br>
In your case, the LP-solver seems to struggle with the LP-relaxation.<br>
Which solver are you using?<br>
<br>
You can try to improve your model, e.g., by scaling or checking that you<br>
do not have very large coefficients together with very small ones in<br>
constraints or the coefficients. You can also check whether you have<br>
big-Ms in the model that you might be able to reduce. But first should<br>
probably turn on the output of the LP-solver to maybe get an idea of<br>
what is going wrong (parameter display/lpinfo).<br>
<br>
Best<br>
<br>
Marc<br>
<br>
<br>
<br>
On 18/12/2020 13:56, Franzen, Christian wrote:<br>
> Hi SCIP Team.<br>
> <br>
> <br>
> I am implementing a branch-and-price solver and for one of my benchmark<br>
> instances solving is significantly slower than for other instances. In<br>
> the statistics of SCIP is see the following output:<br>
> <br>
> <br>
> LP : Time Calls Iterations Iter/call <br>
> Iter/sec Time-0-It Calls-0-It ItLimit<br>
> primal LP : 39822.63 1350 22117847 22431.89 <br>
> 555.41 80.20 364<br>
> dual LP : 903.64 9 622937 69215.22 <br>
> 689.36 0.00 0<br>
> lex dual LP : 0.00 0 0 0.00 -<br>
> barrier LP : 0.00 0 0 0.00 <br>
> - 0.00 0<br>
> resolve instable : 32994.26 583 21004525 36028.34 636.61<br>
> diving/probing LP: 8.31 27 28543 1057.15 3436.36<br>
> strong branching : 0.00 0 0 0.00 <br>
> - - - 0<br>
> (at root node) : - 0 0 0.00 -<br>
> conflict analysis: 0.00 0 0 0.00 -<br>
> <br>
> I am not sure what these statistics tell me. Most of the time the LP<br>
> solver is struggeling with resolving any instability. What kind of<br>
> instability is meant here? Is this something related to numerical<br>
> issues? What can I do to solve that "issue"?<br>
> <br>
> <br>
> Regards<br>
> <br>
> Christian<br>
> <br>
> <br>
> _______________________________________________<br>
> Scip mailing list<br>
> Scip@zib.de<br>
> <a href="https://listserv.zib.de/mailman/listinfo/scip">https://listserv.zib.de/mailman/listinfo/scip</a><br>
> <br>
_______________________________________________<br>
Scip mailing list<br>
Scip@zib.de<br>
<a href="https://listserv.zib.de/mailman/listinfo/scip">https://listserv.zib.de/mailman/listinfo/scip</a><br>
</div>
</span></font></div>
</div>
</div>
</div>
</div>
</div>
</body>
</html>