Shved A. Models and Methods of the Normative Theory of the Synthesis of Information Technologies under Different Forms of Ignorance

The dissertation research is devoted to solving the actual scientific and practical problem of creation and development of bases of the normative theory of synthesis of the information technologies focused on intellectual maintenance of decision-making processes under various types of ignorance. The basic concepts and theoretical provisions of modeling different types of ignorance are studied; the approaches to their classification are analyzed. The main features of data extraction and processing of data and expert knowledge under the influence of incompleteness, uncertainty, inhomogeneity, vagueness, inaccuracy and their possible combinations has been considered, modern mathematical theories of their modeling has been defined. The basic concepts of normative theory and methodology of synthesis of information technologies focused on intellectual support of decision-making processes under non-factors of different nature is offered; a set of rules and formalized mathematical models of the process of synthesis of information technologies under non-factors of different nature has been developed. The procedure of non-factors modeling methods choosing, which allows to generate a system of decisive rules of reasonable choice of mathematical formalism of modeling of different types of ignorance, which have a negative effect on the processes of obtaining and processing a set of source data, is offered. A system of criteria for identification of various analyzed types of ignorance that affect the processes associated with the extraction and analysis of data and expert knowledge, in order to formalize them and a reasonable choice of mathematical apparatus for their modeling, has been proposed. A set of methods and information technologies for structuring and managing data and expert knowledge, which are formed and processed in conditions of uncertainty, heterogeneity, inaccuracy, and combined types of ignorance has been proposed: the method of clustering the initial set of group expert assessments, in order to identify agreed (with a given maximum level of conflict) subgroups; the method for aggregating group expert, in order to synthesize a group decision taking into account different forms of presenting expert judgments (interval, fuzzy, crisp); the method for selection the optimal combination rule, in order to obtain a combined probability mass with the lowest achievable level of uncertainty. The method of non-criteria selection of Pareto-optimal solution has been improved due to the ranking of solutions included in the Pareto set; the case-based reasoning method has been improved due to the use of specific types of approximations in order to classify precedents and their further search in the base of precedents; the method of constructing probability trees has been improved by determining a priori estimates of the probability of realization of events on the basis of the results of group examination in the conditions of complex forms of ignorance and conflict. Further development received models and methods of synthesis the individual and group decisions in the conditions of complex forms of ignorance (including their combination), multi-alternativeness and multi-criteria. The concept and architecture of the decision support system that implements the mechanisms that underlie the proposed methodology for the synthesis of information technologies for decision support. A set of information technologies for the implementation of a number of practical tasks has been developed.