Stability Analysis of Delayed Recurrent Neural Networks Based on a Flexible Terminal Inequality
The stability issue of recurrent neural networks (RNNs) with time-varying delay is studied in this brief. By adding a set of flexible terminals, a flexible terminal-based reciprocally convex inequality (FTRCI) relying on one adjustable parameter <inline-formula> <tex-math notation="LaTeX">$\varrho $ </tex-math></inline-formula> is suggested. Different from the existing estimation techniques, FTRCI employs more slack matrices and utilize more delay information. A novel stability criterion based on linear matrix inequalities (LMIs) is developed using FTRCI. The recently developed stability criterion is demonstrated through the use of a numerical example.
Paper
Full text
Stability Analysis of Delayed Recurrent Neural Networks Based on a Flexible Terminal Inequality
Semantic Scholar · Computer Science · 2024
Abstract
The stability issue of recurrent neural networks (RNNs) with time-varying delay is studied in this brief. By adding a set of flexible terminals, a flexible terminal-based reciprocally convex inequality (FTRCI) relying on one adjustable parameter <inline-formula> <tex-math notation="LaTeX">$\varrho $ </tex-math></inline-formula> is suggested. Different from the existing estimation techniques, FTRCI employs more slack matrices and utilize more delay information. A novel stability criterion based on linear matrix inequalities (LMIs) is developed using FTRCI. The recently developed stability criterion is demonstrated through the use of a numerical example.