[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Help-glpk] Newbie - Having trouble with superfluous slack variables
Re: [Help-glpk] Newbie - Having trouble with superfluous slack variables?
Fri, 9 Aug 2002 08:55:42 +0200
On Thursday 08 August 2002 16:32, you wrote:
> >In my problem (which is an MIP problem), I encounter a situation like
> > this: for all i in some Set, a linear inequality must hold, where the
> > left hand side is a sum over "some" structural variables.
> >The actual variables that appear in this sum vary greatly depending on
> >several input constants to the problem, and the index variable used in the
> >sum ranges not between two constants but between the max of several values
> >and the minimum of several other values.
> Hmm... As I understand your problem includes inequalities like
> sum x[j] <= b[k],
> j in J[k]
> where J[k] are some index sets. Could you give more detailed description
> of your problem?
Thank you for your reply.
I have already identified sevral things were I was wrong:
Due to the inequalities like above the number of nonzero coefficients was not
readily visible and I had to keep track of that number when generating the
inequalities - and since array indexing (in rn, cn and a - for load_mat3)
starts at one (instead of 0), I was working with wrong indices (often off by
one) throughout my program.
I have found true what I started to suspect (after rereading 2.1) : one needs
different routines to obtain the solution to the relaxation and the solution
to the MIP (I was using the former)
The most simple MIP problem now works OK for me.
I'll see if I can fix my larger problem as well.