A New Robust Stability Result for Delayed Neural Networks
This work proposes a further improved global robust stability condition for neural networks involving intervalized network parameters and including single time delay. For the sake of obtaining a new robust stability condition, a new upper bound for the norm of the intervalized interconnection matrices is established. The homeomorphism mapping and Lyapunov stability theorems are employed to derive the proposed stability condition by making use of this upper bound norm. The obtained result is applicable to all nondecreasing slope-bounded activation functions and imposes constraints on parameters of neural network without involving time delay.
KeywordsDelayed systems Neural networks Robust stability Lyapunov theorems
- 12.Wang, Z., Wang, J., Wu, Y.: State estimation for recurrent neural networks with unknown delays: a robust analysis approach. Neurocomputing 227, 20–36 (2017)Google Scholar