<div dir="ltr"><div><div><div>Hi Brian,<br></div>Thanks for the quick response. I actually generated the problem file in a C code using CSDP routine write_prob().<br></div>When I called SDPT3 from MATLAB to solve this test problem I got <br>
gap := trace(XZ) = 3.08e-07<br> relative gap = 6.86e-11<br> actual relative gap = 4.47e-11<br><br></div><div>You are right that Sedumi seems to run into numerical problem too. But the returned value agrees with the optimal solution given by SDPT3 so I consider it solved successfully.<br>
</div><div>FYI, Here are the commands I used for SDPT3:<br>startup<br>[blk,At,C,b] = read_sdpa(’./sdplib/mcp250-1.dat-s’);<br>[obj,X,y,Z,info] = sqlp(blk,At,C,b);<br>x1=X{1};<br>x1(1:35)-x1(36:70)<br><br>Here are the commands I used for Sedumi:<br>
<font color="#000000">[At,b,c,K]=readsdpa('arch0.dat-s');<br>[x,y,info] = sedumi(At,b,c,K);<br>x(1:35)-x(36:70)</font><br><br><br></div><div><font color="#000000">x(1:35)-x(36:70) is the return result of interest. The CSDP solution is far from those obtained from SDPT3 and Sedumi and I found that it is actually not feasible.<br>
</font></div><div><table cols="1" border="0" cellspacing="0"><tbody><tr><td align="LEFT" valign="BOTTOM" height="129"><br></td>
</tr>
</tbody></table>
<br></div><div><br></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Mon, Jan 27, 2014 at 12:59 AM, Brian Borchers <span dir="ltr"><<a href="mailto:borchers@nmt.edu" target="_blank">borchers@nmt.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">When I ran your test case through SeDuMi (with pars.errors=1 to get the DIMACS error measures), I got:<div>
<br></div><div><div> 47 : 2.24E+03 9.80E-12 0.000 0.3687 0.9000 0.9000 0.93 6 6 2.7E-06</div>
<div> 48 : 2.24E+03 6.68E-12 0.000 0.6815 0.9000 0.9000 0.98 11 10 1.9E-06</div><div>Run into numerical problems.</div><div><br></div><div>iter seconds digits c*x b*y</div><div> 48 2.3 6.5 2.2390568439e+03 2.2390575678e+03</div>
<div>|Ax-b| = 3.1e-04, [Ay-c]_+ = 1.7E-09, |x|= 1.8e+04, |y|= 4.3e+04</div><div><br></div><div>Detailed timing (sec)</div><div> Pre IPM Post</div><div>3.022E-02 2.354E+00 3.420E-03 </div>
<div>Max-norms: ||b||=1.800647e+01, ||c|| = 1,</div><div>Cholesky |add|=4, |skip| = 0, ||L.L|| = 516.211.</div><div><br></div><div>DIMACS error measures</div><div> PInf PConInf DInf DConInf RelGap RelComp</div>
<div>3.92E-08 0.00E+00 0.00E+00 2.83E-12 -1.62E-07 1.06E-04 </div></div><div><br></div><div>SeDuMi wasn't able to get the relative complentarity gap below 1.0e-4, so it clearly had difficulty with this problem. </div>
<div> </div><div>Looking at your data file, it appears that the coefficients were rounded to single precision before being written out. This is a bad idea in this case since the problem is badly conditioned. </div><div>
<br>
</div><div>In your testing with SeDuMi, did you start with your problem in MATLAB, write it out in sparse SDPA format, and then read it back in before solving the problem, or did you solve it directly with SeDuMi without going through the data file? </div>
<div><br></div></div><div class="gmail_extra"><br><br><div class="gmail_quote"><div><div class="h5">On Sun, Jan 26, 2014 at 7:35 PM, Feng Nan <span dir="ltr"><<a href="mailto:fnan@bu.edu" target="_blank">fnan@bu.edu</a>></span> wrote:<br>
</div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div class="h5"><div dir="ltr"><div><div>Hi CSDP users/developers,<br></div>I am trying to solve a SDP problem inside a code in C. The structure of the SDP formulation I have is such that I can always guarantee a feasible solution no matter how big sample size is. CSDP successfully solves the problem with small sample sizes; but when the sample size grows it keeps giving up saying lack of progress or stuck on the edge of primal feasibility. <br>
<br>I generated such a moderately large problem in SDPA sparse format (attached here, it has 661 constraints, 2 blocks in the X,Z matrices with sizes 12 -671) and fed it into MATLAB solvers such as SDPT3 and Sedumi and both of them gave the correct solution. Using the same input file I just couldn't get even a feasible solution from CSDP. <br>
<br></div><div>I've also played with the parameters to no avail... <br></div><div>Do you have any suggestions? Is there a setting to make it behave the same as those MATLAB solvers?<br></div><div>Thank you very much!<span><font color="#888888"><br>
</font></span></div><span><font color="#888888"><div>Feng<br></div><div><br></div><br></font></span></div>
<br></div></div>_______________________________________________<br>
Csdp mailing list<br>
<a href="mailto:Csdp@list.coin-or.org" target="_blank">Csdp@list.coin-or.org</a><br>
<a href="http://list.coin-or.org/mailman/listinfo/csdp" target="_blank">http://list.coin-or.org/mailman/listinfo/csdp</a><br>
<br></blockquote></div><span class="HOEnZb"><font color="#888888"><br></font></span></div><span class="HOEnZb"><font color="#888888"><br clear="all"><div><br></div>-- <br>Brian Borchers <a href="mailto:borchers@nmt.edu" target="_blank">borchers@nmt.edu</a><br>
Department of Mathematics <a href="http://www.nmt.edu/~borchers/" target="_blank">http://www.nmt.edu/~borchers/</a><br>
New Mexico Tech Phone: (575) 322-2592<br>Socorro, NM 87801 FAX: (575) 835-5366
</font></span></blockquote></div><br></div>