[Coin-discuss] OsiOsl bug ?

[Tim Hultberg]

Hi,
   The attached code demonstrates what in my eyes seems to be a bug in OsiOsl.

Basically, the same problem is loaded into OsiOsl twice. The problem has the 
trivial optimal solution: all variables zero.

After the first solve, an optimal solution is reported.

After the second solve, the problem is declared primal infesible.

PS1 (The MPS files of the two problems are identical)
PS2 The same behaviour is observed even if the problem is only loaded once, 
but solved twice.
PS3 This may seem an academic problem, but in fact it is a distilled version 
of a "real problem" occurring in the context of subgradient optimization.

Am I missing something or is this a bug in OsiOsl?


Best regards,
                            Tim Hultberg

-------------- next part --------------
A non-text attachment was scrubbed...
Name: osiOslBug.cpp
Type: text/x-c++src
Size: 1491 bytes
Desc: not available
URL: <http://list.coin-or.org/pipermail/coin-discuss/attachments/20030530/dd213011/attachment.bin>