Much contemporary decision-making requires choosing from a huge set of possible combinations. We want the best solution regarding certain objectives, for example maximizing productivity while minimizing costs. If nice mathematical equations accurately describe how each objective depends on our decisions, then standard mathematical approaches (such as calculus) will lead us to the optimal solution. But what if we have no idea how our decisions affect the objectives, and only have a few experimental results? How can we make the best decisions with limited information?