Hierarchical optimization of complex systems with changing criterion functions

In the paper a hierarchical optimization method for complex systems, including control systems , characterized with replacement of the nonlinear global criterion function F(x,y) with a simpler and faster for calculation one - Psi (x,y) (unidirectional with F(x,y) ) for a subset of variable parameters in the second hierarchical level is suggested. A definition of unidirectional functions is proposed. The described method is convergent if the criterion functions Psi (x,y) , F(x,y) are unidirectional, which is proven with theorem. Two examples are given, one of them on the control of induction motors.