<html><head><meta http-equiv="content-type" content="text/html; charset=utf-8"></head><body dir="auto"><div dir="ltr"><span></span></div><div dir="ltr">Hi Gerald,<div><br></div><div>I went for solution 2) to disable the pricer and to restart scip. The problem is that the pricer can only be disabled when scip is in stage problem or in stage solved and I could not manage to reach these stages from the pricer. In scip, time limit is reached but even though after the SCIPsolve() is finished, scip is in stage solving. How could scip reach stage solved? Is it necessary that the problem is solved to optimality? Could we go back to stage problem? <br><br><div dir="ltr">Best,</div><div dir="ltr">Jan</div><div dir="ltr"><br>Am 25.07.2019 um 21:29 schrieb Gerald Gamrath <<a href="mailto:gamrath@zib.de">gamrath@zib.de</a>>:<br><br></div><blockquote type="cite"><div dir="ltr">
  
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  
  
    Hi Jan,<br>
    <br>
    there are two options that you already mentioned which I would
    suggest as well, but I did not try them myself:<br>
    <br>
    1) copy the problem using SCIPcopy(), where you can set copypricers
    to FALSE. Additionally, you need to loop over all constraints in the
    subscip once and set them to be not modifiable via
    SCIPsetConsModifiable(). SCIP can only perform very few presolving,
    propagation, and separation steps if constraints are marked to be
    modifiable, because in that case, SCIP needs to assume that new
    variables will be added to the constraint later which might conflict
    with the reductions performed now.<br>
    <br>
    2) Again, set all constraint to be non-modifiable, disable the
    pricer, and then perform a restart and solve the problem.<br>
    <br>
    In both cases, you potentially need to re-enable features that you
    disabled for the column generation phase.<br>
    <br>
    I would probably go for the second option first, because it should
    be easier to implement, but I am not sure if perhaps some flags can
    still be set incorrectly in that case. If you still observe
    considerable differences, please write again.<br>
    <br>
    Best,<br>
    Gerald<br>
    <br>
    <div class="moz-cite-prefix">Am 24.07.19 um 18:54 schrieb Jan
      Berling:<br>
    </div>
    <blockquote type="cite" cite="mid:CAPeMmq0TTGyT3s6s-WSUTqYX6GAbWyzh5rNYmELFUxrkrLDK8g@mail.gmail.com">
      <meta http-equiv="content-type" content="text/html; charset=UTF-8">
      <div dir="ltr">
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif"">Dear
          SCIP community,<span></span></p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif""><span> </span></p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif"">I
          have a binary optimization problem which uses column
          generation, i. a. because adding
          all variables is not possible due to memory limitations.
          However,
          solving takes a lot of time and so pricing is stopped<span> </span>when
          a timelimit is exceeded, by setting the
          result pointer to SCIP_SUCCESS<span style="font-size:9.5pt;font-family:Consolas;color:black">.</span><span></span></p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif""><span> </span></p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif"">SCIP
          continues solving the “final restricted master problem”
          but this takes a lot of time.<span></span></p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif""><span> </span></p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif"">I
          made an experiment giving the final transformed lp file to
          out-of-the-box-SCIP and the problem is solved more quickly,
          due to presolvers
          and cutting planes.<span></span></p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif""><span> </span></p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif"">Is
          it possible to restart the solving of the transformed
          problem after pricing is stopped and to enable
          presolving/cutting-planes (e.g. by copying the problem,
          enabling presolvers, restarting, setting SCIP stage...)?</p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif""><br>
        </p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif"">Kind
          regards,</p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif"">Jan<br>
        </p>
        <p class="MsoNormal" style="margin:0cm 0cm
0.0001pt;font-size:11pt;font-family:"Calibri","sans-serif""><span></span></p>
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <pre class="moz-quote-pre" wrap="">_______________________________________________
Scip mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Scip@zib.de">Scip@zib.de</a>
<a class="moz-txt-link-freetext" href="https://listserv.zib.de/mailman/listinfo/scip">https://listserv.zib.de/mailman/listinfo/scip</a>
</pre>
    </blockquote>
    <br>
  

</div></blockquote></div></div></body></html>