Modified semiinvariant methods of analyzing stochastic systems

New methods for statistical analysis of nonlinear stochastic systems are proposed which require a much lower computer load that the conventional approximate methods. Unlike the moment-semiinvariant method and the semiinvariant method in the proposed method, junior mixed semiinvariants of order above the second as well as senior semiinvariants are assumed equal to zero in closing the equation set for moments. In this way the number of equations for moments (semiinvariants) is significantly reduced. Examples are provided.