NONGENTROPIC APPROACH TO ESTIMATING THE LEVEL OF INTELLIGENCE OF CONTROL SYSTEMS IN DECISION-MAKING TASKS

Abstract

A new multi-model approach to assessing the influence of the level of intellectualization of target selection procedures on a set of alternatives implemented through the use of information management systems is proposed. The methodological basis of decision-making procedures in these systems is human intuition, which divided into two types: Inside and Intueri. The first type related to the innate natural ability of people to make choices, based on their acquired life experience and is the most common. The second type constructed in the information environment as a composition/superposition of Inside options in order to increase the level of intellectualization of target choosing procedures. It is suggested that in developing rules of output stages of decision making they should be evaluated negentropically, the unit of measurement that should be "bit". After generalizing all the output procedures, this will allow a comparative analysis of the levels of intellectualization of ranking procedures, which are a preliminary stage in the implementation of the choosing process. Multivariance of solutions to the set tasks requires solving the problem of listing effective choices, their formalizing, modeling, and developing effectiveness criteria. To solve these problems, proposed a new multi-model approach. These approaches include linguistic contextual representations of ranking procedures based on the set of meanings (semantics) and "triangles" by N. Chomsky; formal systems for listing effective ranking procedures based on generating context-free grammar in the form of metalinguistic formulas and variables; mathematical algorithms for evaluating the levels of intellectualization of choice procedures using the negentropic principle. Solution of both simple and complex tasks selection is represented as a formal output system. The focus is on three classes of intelligent selection procedures: intuition-based selection of the Inside type; procedures for building artificial intelligence with the connection of intuition like Intueri; intelligent procedures for choosing alternatives using artificial intelligence. The classes of these intellectual procedures are presented as the resultant conclusions of the same "theorem", but differ in the efficiency of performing ranking / selection tasks and the level of their intellectualization. The solution of a simple choice problem characterized by low speed, low level of non-manipulability and documentability. Low effectiveness of ranking procedures leads to the need to introduce a large number of restrictions on the complexity parameters of the choice problems to solve. A high level of intellectualization of this procedure can be reveal in a variety of classes of problems to solved, which is due to the erudition of the subject of control. Realization of the second class of intelligent procedures can become the basis for the development of various options for using artificial intelligence. When once obtained algorithmic constructions can be available for repeated use by organizing standard access to these models. There will be an effect of "amortization" as an expenditure of intellectual resource when delegating authority for a new or repeated assignment. Intelligent alternative selection procedures using artificial intelligence are decision-making algorithms designed for multiple use, within its "authority" as a limited subject area.