Neural Network-Based State Modeling of Nonlinear Discrete Descriptor Systems

Mathematical Methods of Operations Research · Springer Nature · 2026-09-01 · DOI: 10.1007/s10559-026-00919-7

Аннотация

A method is proposed for modeling the states of nonlinear discrete-time descriptor dynamic systems based on the use of artificial neural networks. The system dynamics is described by a nonlinear vector difference equation that cannot be solved for the state at the next time step. Trial states dependent on adjustable parameters (weight matrices and bias vectors) are generated using a recurrent neural network. The results are illustrated using model examples, in particular, a Leontief-type input–output production model.

Читать у издателя →

Похожие статьи