In real life, we often need to make a decision, i.e., we need to select one of the possible alternatives. In many practical situations, our objective is financial: we need to select an alternative that will bring us the largest financial gain. The problem is that usually, we only know the gain corresponding to each alternative with some uncertainty: instead of the exact numerical value of this gain, there is a whole set of possible values of this gain. How can we make decisions under such interval-valued uncertainty? An answer to this question is known for the case when these sets are closed. In this paper, we extend this result of the general case, when sets are not necessarily closed.