in reply to
Re: Looking for help with AI::Genetic and classroom scheduling
in thread Looking for help with AI::Genetic and classroom scheduling
The problem is that the result is not simply a "valid" solution. I'm looking for a "best" solution. I know enough Prolog to code the search space for a valid solution, even using AI::Prolog (thanks Ovid!). What I need is for a way to move from a valid solution to a "better" solution, based on soft things like attendee preference, natural ordering of materials, and room size, after first matching hard things like a single-purpose room and a single-location instructor.