Zemlianskyi O. Information technology of forecasting of hazardous chemicals concentration after accidental release under uncertainty

The object of research is the process of forecasting of hazardous chemicals concentration in chemical accidents. The aim of the dissertation research is to improve the decision-making processes in chemical accidents by developing models and methods for predicting the concentration of a hazardous chemical under uncertainty. To solve the problems elements used systems approach and systems analysis, methods of artificial intelligence, neural network technologies, evolutionary modeling and elements of the theory of fuzzy sets. Scientific novelty and practical significance of the results: first developed neural network post-forecasting method concentration of the dangerous substance under conditions of uncertainty, which, unlike existing methods, allows for refined prediction in post-accident period; first developed a method of parametric optimization model aimed concentration of the dangerous substance as a neuro-fuzzy network based on the use of evolutionary strategies and elements of the analytic hierarchy, thus improving the accuracy of the results using the model; first constructed the model to adjust the initial values of the parameters of an accident on the basis of production rules, which allowed to reduce the uncertainty of decision-making, as well as to carry out the operational forecasting of chazardous chemicals oncentrations; was developed further technology of dependencies identification based on neuro-fuzzy network with fuzzy inference in the form of Tsukamoto, where the consequent membership function is not monotonic, as a result of which it became possible to carry out forecasting of hazardous chemicals concentrations in the period of growth and reduction; improved forecasting technology of hazardous chemicals concentrations in the post-accident period through the use of expert advice in the construction of appropriate models, which allowed to objectify the decision-making process to minimize the consequences of the accident. Scientific and technical novelty of research publications and confirmed by acts of implementation. Key provisions obtained in the thesis focused on the further development of technology forecasting of hazardous chemicals concentrations in the post-accident period and construction the fields of concentration. The developed models and methods constitute a methodological framework to optimize decision-making processes for the prevention of chemical accidents and mitigation of their consequences.