Abstract
New model-reduction numerical procedure for a class of stable nonlinear systems is proposed. The proposed design is devoted to a spacial class of nonlinear systems whose nonlinearities are not necessarily Lipschitz with respect to its arguments. Additionally, the systems under consideration may contain uncertain parameters, having known lower and upper bounds. The computation of the reduced-model matrices is achieved by solving a set of linear matrix inequalities in iterative manner. An illustrative example is studied to approve the proposed theoretical results.