Stochastic simulation and Monte-Carlo methods
A system with randomly changed structure is characterised by the state vector ${bf Y}(t)$ and the structure number $l(t)=1,...,N_0$; $N_0$ is the number of determinate structures. Vector equation for $l$--structure is the Stochastic Differential Equation (SDE) in the sense of Stratonovich. Number $l(t)$ is a random discrete scalar integer-valued process. The paper considers dynamic systems of random structure, when the structure number $l(t)$ is the conditional Markovian process depending on the vector ${bf Y}(t)$. Approximate methods of probabilistic analysis of these systems usually are based on the two-moment parametrization approximation of probability density functions and statistical linearization. The algorithm for statistic simulation of dynamic non-linear stochastic systems with the conditional Markovian change of structure was construted in this paper . It is based on numerical methods for solving the SDEs and "the method of maximum cross-section" for simulation of the moments of the structure change. The mathematical model of follow-up system with possible interchangeable tracking loss and the reconstruction was solved by this algorithm. The constructing algorithm is distinguished by universality and also by the possibility to calculate the estimates of different probabilistic characteristics of a solution including the estimates for probability density functions.
Note. Abstracts are published in author's edition
Mail to Webmaster |
|Home Page| |English Part| |
Go to Home |