An LMI approach to asymptotical stability of multi-delayed neural networks

From MaRDI portal
Publication:707178





The authors consider the following neural network with multiple delays \[ \frac{du(t)}{dt}=-Au(t)+W^{(0)}g(u(t))+\sum_{k=1}^{r}W^{(k)}g(u(t-\tau _{k}))+I, \] where \(u(t)=[u_{1}(t),\dots ,u_{n}(t)]^{T}\) is the neuron state vector, \(A=\text{diag}(a_{1},\dots ,a_{n})\) is a positive diagonal matrix, \(W^{(k)}=(w_{ij}^{(k)}) _{n\times n},\) \(k=0,\dots ,r\), are the interconnection matrices, \(g(u)=[g_{1}(u_{1}), \dots ,g_{n}(u_{n})]^{T}\) denotes the neuron activation with \(g(0)=0\) and \(I=[I_{1},\dots,I_{n}]^{T}\) is a constant input vector, while \(\tau _{k}>0,\) \(k=1,\dots ,r,\) being the delay parameters. The authors present sufficient conditions for the origin to be asymptotically stable by constructing a Lyapunov-Krasovskii functional for the cases where the time delays are constants or are time-varying, respectively. The main results are generalizations of some results reported in the literature. Three examples are given in this paper.




Cited in
(34)








This page was built for publication: An LMI approach to asymptotical stability of multi-delayed neural networks

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q707178)