ABSTRACT
The process continues until the adjustment point turns out to be the “hot point.” Searching the “hot point” is an iterative process. At the beginning, one should choose the point by approximation and linearization. Local normalization of the initial values should be made at the adjustment point. The modified Monte-Carlo method cures this defect. Instead of two steps, in the modified method, a stratified sample base is formed at once at necessary sub-intervals with the given volume of sub-samples. If the probability is estimated by the frequency of the event, then a sufficiently large number of statistical tests are performed according to the Bernoulli scheme, that is, random realizations of all the initial values are generated at each test. The method is extremely simple and universal, but it requires a mandatory analysis of the proximity of the estimate ν to the desired probability Pf, which depends on the number of tests m.
