On fire at 4:30 am, after a night-out :-)

-
Did you know that excel solver (+ all its add ons downloadable from the net) give only localised optimal values for non-linear concave formulations?

I did not...till yesterday night. So I packed my humungously huge model into that blasted solver and it sweetly told me 'No feasible solutions'. And I spent hours trying to figure out which variable or constraint I got wrong. This local optima thing is crazy, if you have a feasible initial solution, the solver condescends to optimise certain variables (atleast that's what I gathered), so you never know on which side of the fence you're standing.

So what did I do? I coded a mini-solver (solver for the specific kind of problems I'm working on) in VB....that's taken me almost the night. And now the VB compiler is sweetly telling me 'Sub not found'.

I'm on fire, literally. The names of the files are the only mute transitory testimony to this (they are called blasted.xls, death.xls....so on, don't worry it is transitory, I'll try to change those names when I reach saner senses).

For now, to hell with sanity, I'm back to work... at 4:45 am :-)