[Symphony] [SYMPHONY] error in lp_rowfunc.c, line 444

S R sr.pointe at gmail.com
Fri Aug 13 11:15:10 EDT 2010


Hi.  I have developed a custom application using symphony version
5.2.4 with MSVC (windows) to solve a specific type of integer program.
 As part of my application, I generate cuts in user_find_cuts and add
them using cg_add_explicit_cut.  For some problems, I get an error
"Unhandled exception at 0x008b80b4 in Test.exe: 0xC0000005: Access
violation reading location 0xf0bdff60" in file lp_rowfunc.c at line
444.  The line is

	 lhs += matval[j] * x[matind[j]];

If I inspect the values of each variable, they seem reasonable except
matind[j] = -33686019.  This then causes the exception when
x[matind[j]] is evaluated.  I have not been able to figure out much
else.  The call stack is:

	compute_violations(LP_PROB * p=0x00d8b670, int new_row_num=1,
WAITING_ROW * * new_rows=0x00d17f28)  Line 444 + 0x15 bytes	C++
 	add_new_rows_to_waiting_rows(LP_PROB * p=0x00d8b670, WAITING_ROW * *
new_rows=0x00d17f28, int new_row_num=1)  Line 295 + 0x11 bytes	C++
 	generate_cuts_in_lp_u(LP_PROB * p=0x00d8b670)  Line 2289 + 0x18 bytes	C++
 	receive_cuts(LP_PROB * p=0x00d8b670, int first_lp=0, int
no_more_cuts_count=0)  Line 371 + 0x9 bytes	C++
 	fathom_branch(LP_PROB * p=0x00d8b670)  Line 445 + 0x12 bytes	C++
 	process_chain(LP_PROB * p=0x00d8b670)  Line 184 + 0x9 bytes	C++
 	solve(TM_PROB * tm=0x00dabe98)  Line 397 + 0x18 bytes	C++
 	sym_solve(SYM_ENVIRONMENT * env=0x00336010)  Line 1150 + 0xc bytes	C++
 	OsiSymSolverInterface::branchAndBound()  Line 89 + 0xc bytes	C++
 	main(int argc=5, char * * argv=0x00333b48)  Line 6102	C++

I have not been able to make much progress in figuring out what is
going on.  Looking at the output, I see that the one iteration
completes successfully.  The error occurs after "Starting iteration 2"
is displayed.  This seems to indicate that my general
approach/syntax/etc. are correct.  Of course, if I remove my user
generated cuts, I do not receive an error.  If I examine the cuts I am
trying to add, I don't see anything incorrect about them.  But it
seems likely that I am doing something wrong.

Does anyone have any suggestions/hint/wisdom?  Any guidance on how to
debug this?

Thanks,
Sam




More information about the Symphony mailing list