Identification of Continuous-Time Dynamical Systems: Neural Network\n Based Algorithms and Parallel Implementation

1993/05/13 by Robert M. Farber, Farber, Robert M., Alan S. Lapedes +5
Computer Science · Engineering · #Adaptation and Self-Organizing Systems (nlin.AO) #Advanced Adaptive Filtering Techniques #Cellular Automata and Lattice Gases (nlin.CG) #Control Systems and Identification #FOS: Physical sciences #Neural Networks and Applications

paper · pdf · doi:10.48550/arxiv.comp-gas/9305001

openalex publication_date 1993/05/13 · openalex created_date 2019/06/27 · openalex updated_date 2026/07/28

Abstract

Time-delay mappings constructed using neural networks have proven successful\nin performing nonlinear system identification; however, because of their\ndiscrete nature, their use in bifurcation analysis of continuous-time systems\nis limited. This shortcoming can be avoided by embedding the neural networks in\na training algorithm that mimics a numerical integrator. Both explicit and\nimplicit integrators can be used. The former case is based on repeated\nevaluations of the network in a feedforward implementation; the latter relies\non a recurrent network implementation. Here the algorithms and their\nimplementation on parallel machines (SIMD and MIMD architectures) are\ndiscussed.\n

Related

• Systematic construction of continuous-time neural networks for linear dynamical systems
• Identification and control of dynamical systems using neural networks
• dynoNet: a neural network architecture for learning dynamical systems
• Neural Integration of Continuous Dynamics
• Learning Time Delay Systems with Neural Ordinary Differential Equations