The authors consider the problem of determining the values of collective expert evaluation of alternative decisions on the basis of models whose structure is described by fragments of the Kolmogorov-Gabor polynomial. The approach is proposed that allows us to formalize the uncertainly of the definition of model parameters of multifactor evaluation based on fuzzy intervals; to define collective fuzzy estimates of alternatives and use them to range the alternatives in accordance with the decomposition of fuzzy intervals at the a-levels.Keywords: collective expert estimate, utility function, multifactor estimate of an alternative, Kolmogorov-Gabor polynomial, ranking of fuzzy intervals.

Retroactive analysis of scientific and technical progress reveals that all previous stages were linked with attempts to enhance the physical power of man and develop new construction materials. In this sense, the present stage of scientific and technical progress is unique. The advent of computers has provided us, for the first time in human history, with a tool for enhancing our intellectual powers. This is attributable to the implementation of all the relatively complex intellectual functions by computers. The process began with the implementation of algorithmic computational procedures, and is rapidly developing toward computerization of progressively more complex intellectual processes: situation recognition, semantic analysis of information, acquisition of new knowledge, decision making, etc. On the whole, this is the artificial intelligence problem. The central topic of this problem is the construction of a system of formal models of human intelligent activity. PHENOMENOLOGICAL MODEL OF HUMAN INTELLIGENT ACTIVITYSubstantively, a general model of intelligent activity can be described as follows. An individual endowed with a sensory system creates an individual information image of a certain situation in the context of the environment and then analyzes the information to construct his behavior in the context of broad understanding. This may lead to manifestation of activity in any form, logical deduction, acquisition of new knowledge, etc. We denote by @ the set of external stimuli on the human sensory system in a specific situation, and by Ji its information image for individual i. Then J~(t) = ~'i [o (t -r), t].(1)Here t is the current time, ,t, i is the individual information mapping operator, r is the delay associated with the inertial propertie ~ of the sensory system. The individual behavior is described by the equationwhere 1-[ i is the individual behavior operator. It is useful to distinguish between two models: the model of reflex (unconscious) behavior, which involves execution of rigid algorithms similar to automatic control, such as stabilization of the body center of mass when walking, and the model of active (conscious) behavior, based on intelligent processing of information. In what follows, we only consider models of the second type. The transition to formal models requires identification of the operators q/i and H i. IDENTIFICATION PROBLEMThe synthesis of formal models involves solution of two problems: structural identification, i.e., determination of the form of the operator that establishes the relationship between input and output, and parametric, or qualitative, identification.

A method is proposed for comparative structural-parametric identification of a model of individual multifactor estimation. An experimental solution to this problem is obtained using evolutionary methods.

System identification has become a highly relevant problem with the increasing demand for high-quality control and prediction in various branches of science and engineering. To be able to make inferences about system operation, we need a model that runs in real time. If the dynamic model does not fit adequately the actual behavior of the s~,stem, the estimation errors build up leading to divergence of the recursive estimation process.The recursive least squares (RLS) methoql has a central role in recursive identification of processes. This rote is attributable to its simple implementation and availability of numerically stable algorithms. Adaptive control relies on construction of new algorithms that exploit the good properties of the least squares method and include an additional qualitatively new property. Thus, we construct algorithms that are capable of tracking time-dependent parameters in the presence of unobservable noise and inexact model specification. Numerical aspects guaranteeing convergence are also important [ 1, 2].Stability, convergence, and dynamic properties of the algorithm essentially depend on the method of determination of the updated covariance matrix, which provides important information about the operation of the algorithm. The RLS method cannot be used to track time-dependent parameters, because the algorithm gains tend to zero. We accordingly use modified algorithms for estimation of time-dependent parameters, which preserve global time convergence (the invariant case) while ensuring a covariance matrix with nonzero elements.The tracking of time-dependent properties in the exponential estimation paradigm still requires attention to the specification of the forgetting factor, which should be correlated with the variation of the unknown parameters. Moreover, controls must be introduced to suppress exponential growth of the covariance matrix P.We consider standard modifications of the lease squares method for tracking time-dependent parameters.In the algorithm with constant forgetting, the eigenvalues of P vary from zero to infinity. In the algorithm with correction, the trace of the matrix P remains constant, the eigenvalues are upper bounded, but the lower bound for the eigenvalues is still zero. In the algorithm with resetting of the covariance matrix to the given matrix or to a matrix that depends on the latest values, the main problem is to choose the resetting point. The algorithm with modification of the covariance matrix adds a positive definite matrix to the updated covariance matrix. This ensures a lower bound on the minimum eigenvalue of P, but the maximum eigenvalue may go to infinity.Thus, standard modifications of the least squares method achieve some desirable properties, while other properties are lost [3].Let us consider an algorithm which keeps the trace constant and uses variable forgetting. Here

No abstract