[Bug-glpk] [Fwd: glpk terminating abnormally]

[Andrew Makhorin]

-------- Forwarded Message --------
From: address@hidden
To: address@hidden
Subject: glpk terminating abnormally
Date: Wed, 19 Mar 2014 15:36:09 -0400

Hello everyone,
 
We have been using the GLPK software in our statistical system for over
10 years now and it is working very well for us, thank you for this
excellent product!  We are using GLPK 4.07 for the LP component in one
of our systems written in C and we have come up with a strange result. 
 
In this particular problem, it appears we have numerical instability ,
which calls the spx_prim_feas() function, then it is triggering a call
to the insist function as shown below, which causes the system to
terminate abnormally.  It appears that at one point, in our LP problem,
the spx->p has a zero value.
 
From source program glpxpx2.c
 
/* the sum of infeasibilities can’t be negative, therefore the modified
problem can’t have unbounded solution */
Insist(spx->p != 0);
 
Is anyone aware if this part has been modified in one of the more recent
versions of glpk ?  Do you have any recommendations on how to deal with
this situation ?
 
Best regards,
Marc Thomson