Our research focuses on the question of how best to make use of the ability to evaluate the (potential) quality of a candidate solution cheaply, so that we can decide when to spend the time required for the expensive evaluation of the validity of the solution and when to instead search for more promising candidates.