Re: [Help-glpk] xmalloc: no memory available

[Matthieu Bollot]

Le Mercredi 19 decembre 2007 a 10:22 +0300, Andrew Makhorin a ecrit :
> > On solving it. The file is about 30 lines long. glpsol works for about
> > 3-4 minutes until 400Mo of RAM (but not sure about the real value I can
> > watch them exactly if it is useful)
> 
> > And it seems that it is able to run all the "subject to" but after that
> > I've got the malloc error.
> 
> > I mean that on log I've got :
> > subject to const1
> > subject to const2
> > ...
> > subject to constx
> > subject to consty
> > malloc error
> 
> > and withtout consty I've got
> > subject to const1
> > subject to const2
> > ...
> > subject to constx
> > malloc error
> 
> Looks like the error happens on generating the model. Try to run
> glpsol with the option '--check', and if there will be the same
> picture, check constraints 'constx' and 'consty' as well as sets,
> parameters, and/or variables they refer to more carefully.
> 
> If you wish, you can post your model to me. Then I could say
> something more definite.
> 
> 
> Andrew Makhorin
> 
> 

Hi all,
'--check' doesn't change anything. 
Model has been successfully generated
xmalloc: no memory available

I'll not give you all my model. Actually, I could give it to you but I
don't want you to optimize it because it is a part of my school work and
I don't want to cheat ;) (until I could find alone)

by the way, this constraint is the problem :
subject to cont7 {i in 1..nbjob, j in 1..nbopparjob[i], k in
1..nbmachines, t in 1..BS-1} :
tpsopparjob[i,j]*(x[i,j,k,t]-x[i,j,k,(t+1)])+sum{l in 1..nbmachines, r
in (t+2)..BS}x[i,j,l,r]<=tpsopparjob[i,j];

x is a var and everything else are parameters.

But, actually, what I really want to do is :
subject to cont7 {i in 1..nbjob, j in 1..nbopparjob[i], k in
1..nbmachines, t in 1..BS-1} :
if (x[i,j,k,t]-x[i,j,k,(t+1)] == 1) then tpsopparjob[i,j]+sum{l in
1..nbmachines, r in (t+2)..BS}x[i,j,l,r]<=tpsopparjob[i,j];


thanks, and merry christmas,
Matthieu Bollot.