[Coin-discuss] OsiSolverInterfaceTest.cpp
JP Fasano
jpfasano at us.ibm.com
Sat May 3 20:18:31 EDT 2003
I also think it is great having OsiGlpk running wiht unitTest.
JP Fasano
jpfasano at us.ibm.com
(914)945-1324 (tie line 862-1324)
COIN www.coin-or.org
Brady Hunsaker
<hunsaker at isye.gatech.edu> To: coin-discuss at www-124.southbury.usf.ibm.com
Sent by: cc: coin-discuss at www-124.southbury.usf.ibm.com
coin-discuss-admin at www-124.southbury. Subject: Re: [Coin-discuss] OsiSolverInterfaceTest.cpp
usf.ibm.com
05/03/2003 06:31 PM
Please respond to coin-discuss
I forgot to mention that the testing in unitTest.cpp and
OsiSolverInterfaceTest.cpp really helped me improve OsiGlpk,
especially by helping me identify several significant errors.
Thanks to everyone who has spent time working on the test sets.
Brady
On Sat, May 03, 2003 at 03:01:26PM -0400, Brady Hunsaker wrote:
> I went ahead and implemented setColSolution and setRowPrice in OsiGlpk
> by changing the cache without telling GLPK, similar to the
> implementations in OsiDylp and OsiXpr. I'm still not convinced this
> is the best long-term solution.
>
> Because of this and other recent corrections, I modified
> OsiSolverInterfaceTest.cpp by removing almost all of the explicit checks
> for GLPK. That is, OsiGlpk now passes almost all of the tests in
> OsiSolverInterfaceTest.cpp (previously there were about 5 that had to
> be explicitly skipped).
>
> The one exception is the test for setDblParam. This test sends in a
> variety of values. Since GLPK does some of its own sanity checks, it
> aborts when unrealistic values are given. The function setDblParam
> works fine (as far as I know); it just won't pass this test. I don't
> think this is a problem.
>
> Brady
>
_______________________________________________
Coin-discuss mailing list
Coin-discuss at www-124.ibm.com
http://www-124.ibm.com/developerworks/oss/mailman/listinfo/coin-discuss
More information about the Coin-discuss
mailing list