Control Systems: Intellectualization As a Response to Challenges of the Big Data Era

Abstract—

This article analyzes the problem of processing massive amounts of data under strict time constraints in control systems. One strategy for solving this problem involves the use of artificial intelligence (AI) technologies. The range of mathematical models traditionally used in control systems has been supplemented by AI-based solutions that involve computer-oriented formalizations of strategies used by human experts to solve problems of this type: so-called interpolation/extrapolation (I/E) models. This article discusses certain significant features of I/E-type solutions, in particular, their ability to generate effective solutions in open big data environments (where the behavior of the control object is not characterized by a single NORMAL state), the problem of identifying stable (inheritable) solutions in the set of permissible solutions when updating empirical data on behaviors of the control object, and finally the problem of identifying empirical dependencies of a causal nature in the current data to compile informal interpretations of alternatives (recommendations) generated by the digital control system to be presented to decision makers (DMs).