The concept of information is fundamental in the IT education. Understanding the meaning and significance of information is required by modern educational standards of higher school. However, in existing educational courses, as a rule, the problem of defining this term is ignored. Learning is based on intuitive ideas. There is a methodological contradiction: the development and application of information technologies requires accuracy and rigor, but all constructions are based on a vague, intuitive concept. The purpose of this work is to develop a new approach to studying the concept of information within the framework of IT-education. Here information is understood as the result of a subset selection from a set of elements. The significance of the proposed approach is revealed using the example of numerical methods of multimodal optimization. It is shown that the new approach makes it possible to formulate effective stopping conditions for a stochastic optimal search algorithm. This stopping criterion guarantees a given amount of information about the location of the global optimum. The developed approach was introduced into the educational process of training IT students within the academic disciplines “Information Theory”, “Bio-inspired optimization methods” and “Mathematical modeling of selection processes”.