Not every optimisation problem needs a neural network. We break down when classic algorithms outperform ML and when they don't.