ABSTRACT

The paper concerns the modeling problem for controlled dynamic plants. With the aim of resolving the contradiction between the completeness of formal description of the plant properties and the algorithmic simplicity of description, it is suggested to build up a hierarchy of models over a sequence of nested subsets of a set of feasible motions. The way this sequence is generated depends on a goal of undertaken investigations. For modeling of optimally controlled plants, a procedure of sequential reduction of a feasible controls set to a family of subsets of feasible controls satisfying the necessary optimality conditions is applied. The approach is illustrated by mathematical modeling of liquid and nuclear rocket engines.