Modified semiinvariant methods of analysis of stochastic systems (Q1081574)

Methods of statistical analysis of nonlinear stochastic systems are described that require much less calculation than the existing approximate methods. In contrast to the well-known moment-semiinvariant method and the method of semiinvariants, not only the leading semiinvariants are set equal to zero, but also the lower-order mixed semiinvariants of order higher than the second. By such an approach it is possible to reduce considerably the number of equations for the moments (semiinvariants). Examples are also presented.