Accuracy analysis of dynamic systems whose structure changes randomly

A nonlinear dynamic system is described whose structure can be in either of two possible states. The process of system structure switching is a Markov chain with two states and discrete time. Approximation of conditionald probability densities as a sum of Gaussian probability densities leads to a computing algorithm for determining the statistical characteristics of the system phase coordinate vector. The advantages of the method over some others are shown. An example is given.