Aerospace Instrument-Making Annotation << Back METHODOLOGY FOR IDENTIFYING THE BEST SAMPLES IN THE DESIGN OF PROMISING UNMANNED AERIAL VEHICLES S.V. Mikoni For the design of new products of technology, the analysis of the best world-class samples is in demand. The properties of complex objects are represented by a large number of indicators. Averaging the values of indicators when evaluating the object as a whole hides the characteristics of its individual parts. Since it can be difficult to achieve better results in all indicators within a complex object, the task of evaluating the object in parts is relevant. Using the example of unmanned aerial vehicles, a method is proposed for finding the best sample in terms of flight and tactical characteristics. The best sample is considered to be the one that received the highest rating for the relevant indicators. The use of the Pareto dominance relation does not guarantee the identification of the best sample, and other methods of dominant analysis have limitations on the number of indicator values. For these reasons, sample ordering is done using multiobjective optimization techniques and value-utility theory. The construction of a multidimensional evaluation model is divided into the construction of a domain model and a preference model for decision makers. The construction of a multilevel domain model is based on the use of system analysis patterns, and the construction of a decision maker’s preference model is based on the use of axiom logical and numerical methods for comparing objects. It is shown on a specific example what errors occur when these axioms are not observed, in particular, when ignoring the statement about class boundaries according to the classification indicator. Eliminating the mistakes made allows you to get reasonable UAV ratings. The use of a common scale by computational methods allows not only to calculate overall scores for groups of indicators, but also to conduct a comparative analysis of scores at any level of the UAV evaluation model. Keywords: complex object, unmanned aerial vehicle, domain model, decision maker’s preference model, comparison axioms, indicator scale, rating of objects, dominant analysis, multipurpose optimization. DOI: 10.25791/aviakosmos.10.2022.1304 Pp. 43-56.