Title :
Study of exponential stability of equilibrium neural networks
Author_Institution :
Dept. of Comput. Sci., Univ. of Zabol, Zabol, Iran
Abstract :
In this article, we consider the generalized Hopfield neural networks and some sufficient criteria are derived for study exponential stability of the equilibrium of Hopfield neural networks of the form Cidui/dt=Σj=1n Tijvj-ui/Ri+Ii i=1,2,...,n. by means of Lyapunov method and some analysis techniques, in which vi=(λiui), and T=[Tij]n×n is a connection matrix, not requiring symmetric and all output responses gi(·) are the same. Ci > 0, Ri > 0 and Ii are called the capacitance, the resistance and the network external current input constants respectively of the ith neuron. The criteria are easy to check and apply in practice.0
Keywords :
Hopfield neural nets; Lyapunov methods; asymptotic stability; matrix algebra; Hopfield neural networks; Lyapunov method; equilibrium neural networks; exponential stability; matrix connection; Artificial neural networks; Asymptotic stability; Circuit stability; Hopfield neural networks; Lyapunov method; Optimization; Stability analysis; equilibrium neural networks; stability;
Conference_Titel :
Electronics and Information Engineering (ICEIE), 2010 International Conference On
Conference_Location :
Kyoto
Print_ISBN :
978-1-4244-7679-4
Electronic_ISBN :
978-1-4244-7681-7
DOI :
10.1109/ICEIE.2010.5560615