[Help-glpk] Problem Diagnostics

[Reginald Beardsley]

What input problem and internal diagnostics are available?  The "search this 
book" link in the wiki produced 91 million hits for "diagnostic".  I'm 
particularly interested in diagnostics applicable to the initial input problem.

For example, is there currently a way to obtain the determinant of the 
non-basic matrix in glpsol?  I found lots of instances of "determine"  with 
"egrep determin", but none of "determinant".  If there is not, would an 
optional dependency on the Gnu Scientific Library be acceptable?  That has code 
to compute the determinant from an LU decomposition that I might be able to 
adapt to work in GLPK to monitor the evolution of the solution.

I'd also like to be able to examine the basic variables as the solution 
progresses. The only internal diagnostics I've found so far are compile time 
controlled assertions guarded by GLP_DEBUG.  However, it's obvious that there 
are other internal diagnostics at key points in the code.  It would be helpful 
to know about them and what they are checking.

I'm trying to find out why a small number of instances in a large suite of 
large, dense problems don't work as I expect.  There doesn't appear to me to be 
a significant difference between the few that fail and the many that work as 
expected.  I'm clearly missing something and looking for a way to find it.  My 
knowledge of LP is limited to what I've learned from reading Vanderbei and the 
GLPK docs so I'm likely missing something obvious.  My past experience with 
problems like these has been based on using singular value decomposition.  And 
to make matters worse that was a long time ago.

Thanks,
Reg

NB I'm using a hardcopy of the reference manual marked 4.43 (it's from the 4.45 
distribution) so if there are changes to the manual that would help please let 
me know.  I much prefer to read physical books, but don't want to kill trees 
for minor changes.