Probabilistic Models of Knowledge Representation to Support Decision-Making in Conditions of Risk and Uncertainty in Atmospheric Air Protection Example

Abstract

In order to systematize and integrate the acquired experience necessary for decision-making in conditions of war and man-made danger, as well as for the purpose of controlling emissions of greenhouse gases or other harmful substances, knowledge presentation models have been de-veloped that take into account both the results of the analysis of available data and probabilistic assessments of the state safety of man-made enterprises and adjacent territories. In order to im-prove the decision-making process, a number of probabilistic models are considered, which are based on the calculation of subjective probability estimates regarding the occurrence of danger-ous events and forecasting the corresponding risks. Factors of various nature are considered during modeling: external influences, concentrations of harmful substances, greenhouse gas emissions, indicators of the state of safety of man-made productions, efficiency of equipment, accounting of violations, and other indicators. Also, the knowledge system provides for calcu-lating the risks of dangerous events, the probability of which increases under the interaction of two or a number of hazardous factors. On the basis of the conducted research, an algorithm for building and the structure of a probabilistic model of knowledge focused on software implementation in the decision-making support system for managing the safety of man-made enterprises that pose threats to the popula-tion and the natural environment has been developed.