[Help-glpk] Linear Integer Arithmetic

[Paulo Jorge Matos]

Hi all,

First time using glpk I have a question which was already made here
some times (some results came up on archives search).
I'm working on Linear Integer Arithmetic and I've seen that in glpk to
solve such a problem you first relax it with simplex and then use
lpx_integer. Can I use interior point instead of simplex in this case
(guess not but need oppinions)? (Would it be faster? In general
problems, the manual asserts this is the case).

As you might imagine Linear Integer Arithmetic is 'not' Linear
Programming so I don't have an objective function. What's the best way
to say I don't need an objective function?
In cplex I do: 'maximize 0'

Now, the central issue has to do with strict inequalities. Are they
supported in glpk? I've seen on some posts (some of them years ago,
2002) that they are not but I've also read about a function able to
read cplex files and cplex files accept strict inequalities so... how
does glpk handle this? Andrew Mahhorin on Aug 9th 2002 mentioned a
solution to solve this issue which was solved by adding an epsilon to
the constraint bound. Is this 'still' the number one solution for us
all with glpk?

Thanks a lot,
--
Paulo Jorge Matos - pocm at sat inesc-id pt
Web: http://sat.inesc-id.pt/~pocm
Computer and Software Engineering
INESC-ID - SAT Group