[Csdp] SDP is dual infeasible
Hui Wu
wuhuing at gmail.com
Thu Nov 14 17:33:16 EST 2013
Hi Brian, thanks for your answer. I attach my formulations process. Please
have a look.
My original formal if primal form with
Min <C * X)>
st. <Ai, X> = bi
X>=0
but csdp likes Max <C, X>, then I changed Min <C, X> to Max <-C, X>.
Any feedback is appreciated, thanks
Hui
On Thu, Nov 14, 2013 at 2:28 PM, Brian Borchers <borchers.brian at gmail.com>wrote:
> This output indicates that CSDP has determined that the problem doesn't
> have a primal-dual optimal pair of solutions. In particular, the code has
> found a primal improving direction D such that if X0 is any feasible
> solution than X0+a*D is feasible for all a>=0 and the objective function
> keeps improving as you move in this direction. Thus the primal problem is
> either infeasible (if there isn't actually a primal feasible solution X0),
> or the primal problem is feasible and unbounded (in which case the dual
> problem is infeasible.)
>
> You haven't told us anything about how you formulated this SDP, so there's
> no practical way to determine what mistake you made in formulating your SDP.
>
> How did you formulate this SDP?
>
> What did you expect the solution to this SDP to be?
>
>
>
> On Thu, Nov 14, 2013 at 10:17 AM, Hui Wu <wuhuing at gmail.com> wrote:
>
>> resend after joining the mail list :)
>> On Nov 13, 2013 9:51 PM, "Hui Wu" <wuhuing at gmail.com> wrote:
>>
>>> Hi Guys, I am using csdp's matlab interface for the following sdp
>>> problem, which is described in the attachment. While I run the script, I get
>>>
>>> *>>[x, y, z] = genSample();*
>>> *Number of constraints: 6 *
>>> *Number of SDP blocks: 1 *
>>> *Number of LP vars: 0 *
>>> *C block 1, blocksize, 6*
>>> *Checking constraint 1 *
>>> *Checking constraint 2 *
>>> *Checking constraint 3 *
>>> *Checking constraint 4 *
>>> *Checking constraint 5 *
>>> *Checking constraint 6 *
>>> *Iter: 0 Ap: 0.00e+00 Pobj: 0.0000000e+00 Ad: 0.00e+00 Dobj:
>>> 0.0000000e+00 *
>>> *Iter: 1 Ap: 1.00e+00 Pobj: 2.4901133e+04 Ad: 6.77e-01 Dobj:
>>> 3.0972415e+03 *
>>> *Iter: 2 Ap: 1.00e+00 Pobj: 1.7734229e+05 Ad: 5.18e-01 Dobj:
>>> 1.6715331e+02 *
>>> *Iter: 3 Ap: 2.10e-02 Pobj: 9.3394935e+05 Ad: 2.66e-02 Dobj:
>>> 1.2916856e+03 *
>>> *Iter: 4 Ap: 1.00e+00 Pobj: 5.7585163e+07 Ad: 1.05e-01 Dobj:
>>> 1.0092369e+03 *
>>> *Iter: 5 Ap: 2.82e-03 Pobj: 5.3050644e+08 Ad: 2.97e-03 Dobj:
>>> 8.2813968e+02 *
>>> *Iter: 6 Ap: 1.29e-03 Pobj: 2.8396651e+09 Ad: 2.90e-03 Dobj:
>>> 9.6682104e+02 *
>>> *Declaring dual infeasibility.*
>>> *Success: SDP is dual infeasible*
>>> *Certificate of dual infeasibility: tr(CX)=1.00000e+00,
>>> ||A(X)||=4.91757e-09*
>>> *Elements time: 0.000010 *
>>> *Factor time: 0.000045 *
>>> *Other time: 0.000882 *
>>> *Total time: 0.000937 *
>>>
>>> Can some body help tell me why my SDP is dual infeasible ? Is it because
>>> there is something wrong with my problem setting or it is normal to my SDP?
>>> Thanks
>>>
>>> Hui
>>>
>>>
>> _______________________________________________
>> Csdp mailing list
>> Csdp at list.coin-or.org
>> http://list.coin-or.org/mailman/listinfo/csdp
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/csdp/attachments/20131114/cee9d3cc/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: SDP Formulation.pdf
Type: application/pdf
Size: 81149 bytes
Desc: not available
URL: <http://list.coin-or.org/pipermail/csdp/attachments/20131114/cee9d3cc/attachment-0001.pdf>
More information about the Csdp
mailing list