Choice problems in modeling the behavior of complex control systems are often characterized by multiple criteria that the decision maker (DM) must optimize. The importance of these criteria is assessed subjectively, which can lead to different decisions. Numerous methods have been developed for solving multicriteria problems, but their use is hampered by the need for preliminary assessment of criterion importance. This article examines an approach in which the DM determines criterion importance using intervals, given incomplete information. A modification of the EDAS method with interval weights defined on a standard simplex is proposed, and two new decision concepts are formalized: a strong decision and a soft EDAS decision. The practical application of these conceptual concepts is illustrated using a steam boiler selection problem as an example.