<div dir="auto">Dear Dr. Galati,</div><div dir="auto"><br></div><div dir="auto">Thank you so much for your informative comments. </div><div dir="auto"><br></div><div dir="auto">All the best</div><div dir="auto">Abbas</div><div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Feb 9, 2024 at 02:13 Matthew Galati <<a href="mailto:matthew.galati@gmail.com">matthew.galati@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)"><div dir="ltr">Cplex has classical benders.<div><br></div><div>Gurobi has something where they take advantage of fully disconnected matrices; and I think they try to do some reductions when you have the case of "almost fully disconnected". Gu gave a talk on this a few years ago. </div><div><br></div><div>But, in general, the answer is "no". No commercial vendors (except for SAS) do automated branch-cut-and-price like what GCG does. See: <a href="https://documentation.sas.com/doc/en/pgmsascdc/v_047/casmopt/casmopt_decomp_overview.htm" target="_blank">https://documentation.sas.com/doc/en/pgmsascdc/v_047/casmopt/casmopt_decomp_overview.htm</a><br><br>Matt</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, Feb 8, 2024 at 5:40 PM Abbas Omidi <<a href="mailto:abb.omidi@gmail.com" target="_blank">abb.omidi@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)"><div dir="ltr">Dear Dr. Galati,<div><br></div><div>Thank you so much for pointing this out. </div><div>I have often used SCIP-GCG or CPLEX and do not have any experience with SAS. </div><div>Do you know if there are any implementations with CPLEX or Gurobi under a hood?</div><div><br></div><div>Kind regards</div><div>Abbas</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, Feb 8, 2024 at 3:51 PM Matthew Galati <<a href="mailto:matthew.galati@gmail.com" target="_blank">matthew.galati@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)"><div dir="auto">The only commercial solver that has these methods is SAS. Look for the DECOMP option. <div><br></div><div>Cheers, </div><div>Matt </div><div><br id="m_6266390413499941418m_-1200776570926013048m_168131694313725158lineBreakAtBeginningOfSignature"><div dir="ltr">Sent from my iPhone</div><div dir="ltr"><br><blockquote type="cite">On Feb 8, 2024, at 4:26 AM, Katrin Halbig <<a href="mailto:katrin.halbig@fau.de" target="_blank">katrin.halbig@fau.de</a>> wrote:<br><br></blockquote></div><blockquote type="cite"><div dir="ltr">
Dear Abbas,<br>
<br>
To use decompositions, it is best to first write a *.dec file and
then read it into Scip. You can write the *.dec file manually or use
gcg to automatically create a decomposition and write it out to a
*.dec file.<br>
This file is read in the same way as the problem itself. On the
command line enter "read path/to/problem read
path/to/decomposition.dec optimize".<br>
In pyscipopt:<br>
m = Model()<br>
m.readProblem("path/to/problem")<br>
m.readProblem("path/to/decomposition.dec")<br>
<br>
There are several plugins that (can) use such a decomposition, for
example, heur_gins, heur_dps, heur_padm, and the benders framework.
Switch on the heuristics by setting the corresponding parameters,
for example, m.setParam('heuristics/dps/freq', 0).<br>
It's not always efficiently to use these methods, but just try it.<br>
<br>
Independent from a decomposition is the cons_component presolver.
This method detects on its own whether the problem is split into
independent components and treats them separately. But since you
mentioned that you have linking constraints and variables, your
problem is probably connected.<br>
I don't know what commercial solvers do, but I would strongly assume
that they also use methods like cons_components.<br>
<br>
Best,<br>
Katrin<br>
<br>
<div>Am 05.02.24 um 12:48 schrieb Abbas
Omidi:<br>
</div>
<blockquote type="cite">
<div dir="ltr">
<div dir="ltr"><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">Dear Alexander,</font>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif"><br>
</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">Thank you so much
for your quick response. </font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">Please, let me
explain more about what I meant.</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif"><br>
</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">I am currently
working on a scheduling problem and its structures, indeed
from the modeling point of view, to investigate how a
standard MIP solver like SCIP or Gurobi can solve this
problem efficiently.</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">For that, I am
using GCG, actually PyGCGopt, to analyze these problems.
In my previous email, I asked about some parameters at GCG
that seed up the solving process, and the comment by
Prof. Lübbecke was really helpful. </font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif"><br>
</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">Now, at
the moment I do not need to use/implement this method
manually and I think everything can be done automatically.
Also, on the other hand, I am not aware of that the SCIP
has a tool that applies <a href="https://www.scipopt.org/doc-8.0.2/html/cons__components_8c.php" style="box-sizing:border-box;text-decoration-line:underline;outline:0px;font-family:arial,sans-serif;background-position:0px 0px" target="_blank">cons_components</a> as a
successful presolving technique. By that, my questions
are:</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif"><br>
</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">1) Is it possible
to employ the mentioned decomposition by using the
command line? (e.g. read problem -> set this method
-> optimize)</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">2) Is there any
Python implementation, PySCIPopt, that shows how we can
use this method? (unfortunately, I do not have more
experience with C/C++)</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">3) Based on the
mentioned "successful presolving technique", can we infer
that commercial solvers like Gurobi and CPLEX have gained
from such techniques under a hood?</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif"><br>
</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">Best regards</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">Abbas</font></div>
<div><br>
</div>
</div>
<br>
<div class="gmail_quote">
<div dir="ltr" class="gmail_attr">On Mon, Feb 5, 2024 at
2:09 PM Hoen, Alexander <<a href="mailto:hoen@zib.de" target="_blank">hoen@zib.de</a>> wrote:<br>
</div>
<blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)">
<div>
<div id="m_6266390413499941418m_-1200776570926013048m_168131694313725158m_-89375102679366714m_2434439602759906148divtagdefaultwrapper" style="font-size:12pt;font-family:Calibri,Helvetica,sans-serif;color:rgb(0,0,0)" dir="ltr">
<p style="font-family:Calibri,Helvetica,sans-serif">Hi Abbas,<br>
</p>
<p style="font-family:Calibri,Helvetica,sans-serif"><br>
</p>
<p style="font-family:Calibri,Helvetica,sans-serif">SCIP allows the user to provide a problem
decomposition. Section 4.2 in the SCIP 7 Release
report describes how SCIP uses the decomposition.<br>
<br>
On <a href="https://www.scipopt.org/doc-8.0.2/html/DECOMP.php" id="m_6266390413499941418m_-1200776570926013048m_168131694313725158m_-89375102679366714m_2434439602759906148LPlnk74537" target="_blank" style="font-family:Calibri,Helvetica,sans-serif">https://www.scipopt.org/doc-8.0.2/html/DECOMP.php</a> there
is an explanation of how to provide such a
decomposition.<br>
<br>
Best,</p>
<p style="font-family:Calibri,Helvetica,sans-serif">Alex</p>
</div>
<hr style="display:inline-block;width:98%">
<div id="m_6266390413499941418m_-1200776570926013048m_168131694313725158m_-89375102679366714m_2434439602759906148divRplyFwdMsg" dir="ltr"><font style="font-size:11pt;font-family:Calibri,sans-serif;color:rgb(0,0,0)" face="Calibri,
sans-serif"><b style="font-family:Calibri,sans-serif">From:</b> Scip <<a href="mailto:scip-bounces@zib.de" target="_blank" style="font-family:Calibri,sans-serif">scip-bounces@zib.de</a>>
on behalf of Abbas Omidi <<a href="mailto:abb.omidi@gmail.com" target="_blank" style="font-family:Calibri,sans-serif">abb.omidi@gmail.com</a>><br>
<b style="font-family:Calibri,sans-serif">Sent:</b> Monday, February 5, 2024 10:49:50 AM<br>
<b style="font-family:Calibri,sans-serif">To:</b> <a href="mailto:scip@zib.de" target="_blank" style="font-family:Calibri,sans-serif">scip@zib.de</a><br>
<b style="font-family:Calibri,sans-serif">Subject:</b> [SCIP] Dealing with a problem that
contains a specific structure with a standard MILP
solver</font>
<div> </div>
</div>
<div>
<div dir="ltr"><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">Dear
support team,<br>
</font>
<div><br>
</div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">Suppose
we have a specific formulation which contains a
particular structure. For example, the formulation
contains a block diagonal format with some linking
constraints or variables. </font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif"><br>
</font></div>
<div><font style="font-family:arial,sans-serif;background-color:rgb(255,255,255);color:rgb(0,0,0)" face="arial, sans-serif">As far as
I know, this format would be suitable for the
decomposition scheme. Now, I am interested to know
if is there any benefit to using this special
structure with a standard MILP solver line SCIP or
Gurobi.<br>
</font></div>
<div><span style="background-color:rgb(255,255,255)"><font face="arial, sans-serif" style="font-family:arial,sans-serif;color:rgb(0,0,0)"><br>
</font></span></div>
<div><span style="background-color:rgb(255,255,255)"><font face="arial, sans-serif" style="font-family:arial,sans-serif;color:rgb(0,0,0)">All
the best</font></span></div>
<div><span style="background-color:rgb(255,255,255)"><font face="arial, sans-serif" style="font-family:arial,sans-serif;color:rgb(0,0,0)">Abbas</font></span></div>
</div>
</div>
</div>
</blockquote>
</div>
</div>
<br>
<fieldset></fieldset>
<pre style="font-family:monospace">_______________________________________________
Scip mailing list
<a href="mailto:Scip@zib.de" target="_blank" style="font-family:monospace">Scip@zib.de</a>
<a href="https://listserv.zib.de/mailman/listinfo/scip" target="_blank" style="font-family:monospace">https://listserv.zib.de/mailman/listinfo/scip</a>
</pre>
</blockquote>
<br>
<pre cols="72" style="font-family:monospace">--
Katrin Halbig, FAU Erlangen-Nürnberg, Analytics & Mixed-Integer Optimization, Tel. +49 9131 85-67179</pre>
<span>_______________________________________________</span><br><span>Scip mailing list</span><br><span><a href="mailto:Scip@zib.de" target="_blank">Scip@zib.de</a></span><br><span><a href="https://listserv.zib.de/mailman/listinfo/scip" target="_blank">https://listserv.zib.de/mailman/listinfo/scip</a></span><br></div></blockquote></div></div></blockquote></div>
</blockquote></div>
</blockquote></div></div>