<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body>
<p dir="ltr">Which linear solver were you using? I'd usually experiment with problem scaling both at the modeling level and at the linear solver level. Were you seeing any warnings? Can you share some representative logs at the default print level, to see how
the primal vs dual infeasibilities, barrier parameter, and any regularization were behaving? I also assume you've enabled the constant jacobian and hessian options for a qp, which should at least make each iteration cheaper though should not change convergence
behavior.<br>
</p>
<br>
<br>
<br>
<div class="gmail_quote">On Tue, Oct 13, 2015 at 11:09 PM -0700, <span dir="ltr">
<<a href="mailto:jmogali@andrew.cmu.edu" target="_blank">jmogali@andrew.cmu.edu</a>></span> wrote:<br>
<br>
</div>
<div class="BodyFragment">
<div class="PlainText">Hi all,<br>
For my problem (~10000 constraints, ~12000 variables), IPOPT fails<br>
to converge on Convex Quadratic problem (positive semi definite<br>
though) within 3000 iterations. I am aware of some earlier posts<br>
on this issue which stated enabling Mehrotra's algorithm. Enabling<br>
that still did not give me satisfactory timings. I was wondering<br>
if there were any recent additions to the solver, methods such as<br>
block coordinate descent or BSUM or any other fancy method which<br>
although may not fall into an interior point framework. Anyways, I<br>
would still like to hear any suggestions that has helped people in<br>
particular with convex quadratic problems using IPOPT or any other<br>
solver which is very effective on these sort of problems.<br>
<br>
Thanks,<br>
Jayanth<br>
<br>
_______________________________________________<br>
Ipopt mailing list<br>
Ipopt@list.coin-or.org<br>
<a href="http://list.coin-or.org/mailman/listinfo/ipopt">http://list.coin-or.org/mailman/listinfo/ipopt</a><br>
</div>
</div>
</body>
</html>