A Special Criteria to Globally Exponentially Stability for Discrete-Time Recurrent Neural Networks
On average, each of the 1011 neurons has 1000 synaptic connections with other neurons in reality. In order to simulate a biological genuine model, the stability of a special discrete-time recurrent neural networks model that every neuron only has one input neuron is considered. And a main result is obtained. It provides some theoretical basis for the application.
Qi Luo and Yuanzhi Wang
J. M. Yuan et al., "A Special Criteria to Globally Exponentially Stability for Discrete-Time Recurrent Neural Networks", Advanced Materials Research, Vols. 181-182, pp. 293-298, 2011