Dominance relations approach to design and control configuration of robotic groups

摘要

The article describes a general interdisciplinary approach to the design of robotic groups, using ideas from the multi-criteria analysis of individual decision-making problems. Instead of analysing the preferences of a person (a decision-maker), it is proposed to automatically determine preferences for adapting robots to the operating conditions and achieving goals. The statement of the problem assumes that the operating conditions, and hence the preferences, may not be known in advance, and may also change dynamically. A method for increasing the efficiency of the procedure for automatic determination of preferences based on the Criteria Importance Theory is proposed. In addition, a general concept of a goal has been introduced, including both enemy or competing objects (targets), and specific tasks. It is proposed to use the concepts of potential optimality and potential non-dominance to find the composition of a robotic group covering the maximum number of goals.