Modeling of Complex Systems over Bayesian Belief Network

The article analyzes the factors that have an impact on the activation of complex natural processes and their catastrophic consequences. Solar activity, precipitation and seismic activity have chosen as the most significant factors, a sharp increase in which leads to activation of natural processes such as landslides, mudflows, natural and forest fires, snow avalanches, hurricanes, tornados, squalls, volcanic eruptions, etc. Intermediate and resulting indicators have chosen to describe and evaluate the course of the processes under consideration: Natural Hazards, Time Frame and Cash. The construction of the Bayesian Belief Network for modeling of complex natural processes on the Northern Mountainous Coast of Black Sea, its structure and training by filling in probability tables are considered. Experts fill conditional probabilities for dependent or intermediate vertices. The values of unconditional probabilities have filled according to observation data for top-level vertices. The managing or controlling factor is determined as Funds Invested and its importance has discussed. Simulation of various scenarios of the course of natural processes has carried out, for which the probability of occurrence of certain states of the resulting outcome indicators is calculated.