[Cbc] Regression in CBC Stable

Christoph Cullmann cullmann at absint.com
Thu Apr 9 06:29:06 EDT 2015


Hi,

since this change in the CBC repo:

--- Cgl/src/CglGomory/CglGomory.cpp
+++ Cgl/src/CglGomory/CglGomory.cpp
@@ -72,10 +72,10 @@ void CglGomory::generateCuts(const OsiSolverInterface & si, OsiCuts & cs,
          intVar[i] = 2; // other
        } else {
          // negative bounds - I am not sure works
-         intVar[i] = 0;
+         intVar[i] = 3;
        }
       } else {
-       intVar[i] = 0;
+       intVar[i] = 4;
       }
     } else {
       intVar[i]=0;
@@ -1239,7 +1239,7 @@ CglGomory::generateCuts(
            CoinFillN(cutElement,number+1,0.0);
          } else {
            // relax rhs a tiny bit
-           //#define CGL_GOMORY_OLD_RELAX
+#define CGL_GOMORY_OLD_RELAX
 #ifndef CGL_GOMORY_OLD_RELAX
 #if 0
            double rhs2=rhs;


I have a regression that some LPs now solve different for e.g. Win vs. Linux compiles.

Without that change, current stable branch is nice.

The

+#define CGL_GOMORY_OLD_RELAX

seems to be the issue, I get now slightly too small optimal solutions.

If I deactivate that define again, all is fine.

Greetings
Christoph

-- 
----------------------------- Dr.-Ing. Christoph Cullmann ---------
AbsInt Angewandte Informatik GmbH      Email: cullmann at AbsInt.com
Science Park 1                         Tel:   +49-681-38360-22
66123 Saarbrücken                      Fax:   +49-681-38360-20
GERMANY                                WWW:   http://www.AbsInt.com
--------------------------------------------------------------------
Geschäftsführung: Dr.-Ing. Christian Ferdinand
Eingetragen im Handelsregister des Amtsgerichts Saarbrücken, HRB 11234



More information about the Cbc mailing list