After reading through the thread, I think its timely to call your attention that Genetic Algorithm (GA) is not the one and only choice when solving the classroom scheduling problem. While I am not proficient in the field, in my understanding of combinatorial optimization, your problem is a special form of the broader range of assignment problems, and as such, it quite naturally lends itself to be formulated as an integer linear program (ILP). Here you can read a pretty good introduction to the topic with the most important backgrounds.
in reply to Looking for help with AI::Genetic and classroom scheduling
Now, the only thing that remained to be done is to actually solve the resultant ILP. For this I coded and maintain (unfortunately, out of CPAN) a nice Perl module, Math::GLPK, which wraps the GNU Linear Programming Toolkit (GLPK). Based on my thorough experience with GLPK and ILPs, you have pretty good chance to obtain 100 percent optimal solution in all cases as long as the size of the problem remains of similar range as yours.
It is clear that both approaches have its pros and cons. However, for me at least, the ILP approach proved itself much more natural and concise than any other formulations of hard combinatorial optimization problems...