<html><head><meta http-equiv="Content-Type" content="text/html charset=us-ascii"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class="">NO, if the QCQPs are completely binary you do not need a global solver. Look at my benchmarks again.<div class="">While SCIP is publicly available it is not free for use by the public!!! Only as member of an academic</div><div class="">institution. No commercial use! But, if you are at an academic institution you can also get CPLEX.</div><div class=""><br class=""><div><blockquote type="cite" class=""><div class="">On Aug 9, 2017, at 7:25 AM, Vyacheslav Kungurtsev <<a href="mailto:vyacheslav.kungurtsev@fel.cvut.cz" class="">vyacheslav.kungurtsev@fel.cvut.cz</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div dir="ltr" class="">QCQPs. CPLEX for some reason can handle QCQPs with indefinite objective, but not constraint quadratics. So yes, I would need a global solver, however, I suppose there should be something specifically tailored to QCQPs, as opposed to a generic nonlinear IP solver. It seems like scip is the only one I've found so far publicly available for this<br class=""></div><div class="gmail_extra"><br class=""><div class="gmail_quote">On Wed, Aug 9, 2017 at 3:52 PM, Hans Mittelmann <span dir="ltr" class=""><<a href="mailto:mittelma@asu.edu" target="_blank" class="">mittelma@asu.edu</a>></span> wrote:<br class=""><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div style="word-wrap:break-word" class="">Are you talking about QPs or QCPQs? CPLEX can handle non convex QPs and they do not even have <div class="">to be binary. If they are binary they can be linearized or convexified and other solvers can handle them</div><div class="">as well. See my QPLIB benchmarks on <a href="http://plato.asu.edu/bench.html" target="_blank" class="">http://plato.asu.edu/bench.<wbr class="">html</a></div><div class="">However, if you have quadratic constraints you need a global solver.</div><span class="HOEnZb"><font color="#888888" class=""><div class=""><br class=""></div><div class="">Hans Mittelmann</div></font></span><div class=""><br class=""><div class=""><blockquote type="cite" class=""><div class=""><div class="h5"><div class="">On Aug 9, 2017, at 3:33 AM, Vyacheslav Kungurtsev <<a href="mailto:vyacheslav.kungurtsev@fel.cvut.cz" target="_blank" class="">vyacheslav.kungurtsev@fel.<wbr class="">cvut.cz</a>> wrote:</div><br class="m_-6192705984354132962Apple-interchange-newline"></div></div><div class=""><div class=""><div class="h5"><div dir="ltr" class=""><div class=""><div class="">Hello,<br class=""><br class="">I am trying to find a solver (available via academic license at most)
for nonconvex binary QPs. CPLEX requires the constraint quadratics to be
positive semi-definite, so I stumbled on to SCIP after finding the OPTI
toolbox. But the OPTI toolbox is apparently only available for Windows,
so I figured directly installing scip to run in matlab should be
possible, but unlike opti there is no matlab interface, apparently. <br class=""><br class="">I haven't
heard of any of the formats specified in the SCIP documentation, are
there any you would recommend as being straightforward to write a script
to convert into the proper data format by printing appropriately from
MATLAB problem data? I have heard it is possible to access it through ampl, but how would you run scip through the ampl mex?<br class=""><br class=""></div>Thank you<br class=""></div></div></div></div><span class="">
______________________________<wbr class="">_________________<br class="">Scip mailing list<br class=""><a href="mailto:Scip@zib.de" target="_blank" class="">Scip@zib.de</a><br class=""><a href="https://listserv.zib.de/mailman/listinfo/scip" target="_blank" class="">https://listserv.zib.de/<wbr class="">mailman/listinfo/scip</a><br class=""></span></div></blockquote></div><br class=""></div></div></blockquote></div><br class=""></div>
</div></blockquote></div><br class=""></div></body></html>