The study conducted in this paper started from a real-life problem encounted by a software test manager. Testing a software consist in passing a set of tests, many of them are interdependent, they have di erent weights and the software testing team have a limited time resource. It is clear that it takes di erent times to pass di erent tests. The goal of the team is to conduct tests in the allowed time interval following a path such that the overall weight of passed tests is maximal. In the present paper we present a mathematical model of the problem and propose a solution to it based on developed evolutionary algorithm.