I have to second simulated annealing if only because it gets too ignored normally. It's a wonderful technique for attacking the optimization of intractable problems, and one that belongs in any good mental toolkit for optimization.