Snihur O. Fuzzy models of groundwater deposits and methods of evaluating their state under conditions of uncertainty

The purpose of the thesis research is improving validity of groundwater deposits state evaluations by developing an aggregated groundwater deposit model and a method for deposit state evaluation under conditions of uncertainty based on interval type-2 fuzzy sets. An aggregated fuzzy groundwater deposit model based on interval type-2 fuzzy sets was proposed. The model was adapted for solving the task of artesian well prospect evaluation. The developed model can also operate under conditions of the decision maker being not sufficiently informed about the values of input variables. This allows to utilize the model on any stage of hydrogeological exploration. An automated informative feature selection method was developed, which utilizes expert knowledge of the subject area, as well as the accumulated experimental data. An aggregated criterion was proposed, which gives a generalized interval estimation of the output variable value based on a set of models. The PCM clustering method was improved by introducing interval membership grades. This allows to account for and model uncertainties related to the lack of knowledge about the object. The interval form of the method provides a generalized estimation by different validity indices, which allows to avoid erroneous results related to intrinsic properties of any one given criterion. A clustering based model for processing hydrogeological data as part of the well state evaluation task was developed. It was shown how the proposed model is used for analyzing ongoing groundwater deposit monitoring data for early detection of potentially negative tendencies in water quality changes. Specialized software for groundwater deposits evaluation based on the proposed models and methods was created.