[Symphony] SYMPHONY 5.5.7 ..

[Mikko Alinikula - SUBNIC OY]

Hello Ted,

I calculate the my truss-problem in 2 phases.  First a 
fixed-geometry-solution is optimized with SYMPHONY and it takes around 
100 ms to get a solution.  In phase 2 I input that solution as a initial 
solution using sym_set_col_solution-function to calculate the real problem.

I got this working with small models but larger models with 1000 
variables or so crashed when I called sym_solve.  Sometimes a memory 
error was issued but more often the system just crashed.

But there seems to be a workaround to this .. if I set 
sym_set_int_param(env, "prep_level", 0) I got the system working with 
models having more than 3000 variables and 10000 restricting equations.

Maybe there is a bug in the preprosessor in the Win32-environment?

BR

Mikko