[Cbc] Segmentation fault 11 when providing initial solution in command line

[Joel Cofield]

John,

That fixed it! I must have forgotten to set a path because I had 2.9.7
already compiled on my system.

Also, I see that CBC may not be able to build a solution unless all integer
variables are provided (even if the omitted ones are easily
determined/auxiliary).

This is a very valuable feature for me, so thanks to those who included it
in the latest versions!

Regards,

Joel


On Tue, Dec 29, 2015 at 1:34 PM, John Forrest <john.forrest at fastercoin.com>
wrote:

> Joel,
>
> Unable to reproduce with latest stable.
>
> I get
>
> opening mipstart file ./Matching-pulp-simplify.sol.
> MIPStart values read for 32 variables.
> Continuous objective value is 745 - 0.00 seconds
> Cgl0008I 15 inequality constraints converted to equality constraints
> Cgl0004I processed model has 68 rows, 167 columns (162 integer (162 of
> which binary)) and 503 elements
> Cutoff increment increased from 1e-05 to 0.9999
> Cbc0045I MIPStart provided solution with cost 1195
>
> or
>
> opening mipstart file ./Matching-manual.sol.
> MIPStart values read for 15 variables.
> Continuous objective value is 745 - 0.00 seconds
> Cgl0008I 15 inequality constraints converted to equality constraints
> Cgl0004I processed model has 68 rows, 167 columns (162 integer (162 of
> which binary)) and 503 elements
> Cutoff increment increased from 1e-05 to 0.9999
> Cbc0045I Trying just fixing integer variables.
> Cbc0045I Warning: mipstart values could not be used to build a solution.
>
> So try latest version and see if that fixes.
>
> John Forrest
>
> Using On 27/12/15 19:54, Joel Cofield wrote:
>
> Hi All,
>
> I am trying to run cbc through the command line with an initial solution
> but it does not work. My .lp and three tries at a .sol file are attached.
>
> My first attempt uses a .sol file generated by an initial run of the same
> MIP but does not provide all of the variables for some reason:
> cbc Matching-pulp.lp mips Matching-init-pulp.sol solve
> ...
> Cbc0045I MIPStart solution provided values for 1 of 162 integer variables,
> 22 variables are still fractional.
>
> My second attempt was to simplify the solution file but I get a
> segmentation fault:
> cbc Matching-pulp.lp mips Matching-init-pulp_simpler.sol solve
>
> So then I tried just making my own solution file and also received a
> segmentation fault:
>
> cbc Matching-pulp.lp mips Matching-manual.sol solve
> Welcome to the CBC MILP Solver
> Version: 2.9.0
> Build Date: Feb 12 2015
>
> command line -
> /Library/Python/2.7/site-packages/PuLP-1.5.6-py2.7.egg/pulp/solverdir/cbc/osx/64/cbc
> Matching-pulp.lp mips Matching-manual.sol solve (default strategy 1)
>  CoinLpIO::readLp(): Maximization problem reformulated as minimization
> will open mipstart file ./Matching-manual.sol.
> mipstart values read for 15 variables.
> Continuous objective value is 745 - 0.00 seconds
> Cgl0008I 15 inequality constraints converted to equality constraints
> Cgl0004I processed model has 68 rows, 167 columns (162 integer (162 of
> which binary)) and 503 elements
> Cutoff increment increased from 1e-05 to 0.9999
> Segmentation fault: 11
>
>
> If I just run cbc Matching-pulp.lp solve then cbc finds the optimal
> solution.
>
> Why are these file causing memory issues? Are there mips file standards I
> can follow or is this a bug?
>
> Thanks,
>
> Joel
>
>
>
> _______________________________________________
> Cbc mailing listCbc at list.coin-or.orghttp://list.coin-or.org/mailman/listinfo/cbc
>
>
>
> _______________________________________________
> Cbc mailing list
> Cbc at list.coin-or.org
> http://list.coin-or.org/mailman/listinfo/cbc
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20151229/8c12b2fa/attachment.html>