Imitation models in decomposition method of nested processes

The problem is considered to enhance adequacy of the nested process decomposition method used to resolve the queuing networks containing blocking. Imitation procedure is proposed to be used for analysis of closed systems on the nested level in the framework of decomposition method. The stochastic approximation method by Robbins-Monroe is also proposed to be applied due to stochastic nature of the output imitation processes. The performed analysis of Robbins-Monroe procedure convergence allows to determine its parameters, mathematic expectation value, and dispersion of the solved equation root estimation. It is shown that the value of control interval should be diminished. The proposed procedure allows to reduce by several fold the computing time spent to search a solution for the balance equation for the models of nested level.