Features of the choice of factor space in the estimation and prediction of the state of a complex object

Abstract

Modeling of complex objects is always associated with the solution of problems not only of a methodo- logical, but also of a methodical nature. Such problems include, for example, the problem of the formation and use of the factor space in solving the problem of estimating and predicting the state of the complex object. As part of the problem statement, the main rules are formulated that an expert should use when choosing a factor space for the synthesis of various classes of models, including the example of constructing a fuzzy-possibility model of complex object based on explicit and implicit expert knowledge. An example of constructing a factor space and the corresponding model for a technological process associated with solid-phase roasting of a sulfide concentrate is considered using two fundamental- ly different approaches to describing this process. These approaches are the classical balance approach and fuzzy- possible approach. On a practical example, it is shown that a reasonable choice of the factor space composition and structure, especially dependent variables (output parameters), plays an important role in achieving the desired result - building a model for estimating and predicting the state of the complex object, which describes the process under study with the required degree of adequacy.