[Help-glpk] dual simplex

[H . Georg Büsching]

Hey,

 

I have written a problem which really produces the dual problem of normal LP.

For my sake, the solutions values are always the same, so I suppose everything is okay.

 

But the normal simplex method on my dual problem is sometimes more faster than the dual simplex on the normal problem. This was not expected.

 

Is there an explanation for this?

 

Naturally I can give examples, some members of the MIPLIB of 2003 have this property.

 

Best regards

 

H. Georg Büsching.