Bifurcation analysis for a two-dimensional discrete-time Hopfield neural network with delays (Q925380)

The authors consider discrete-time Hopfield neural network of two neurons with four different delays \(\tau_i\geq 0\), \(i=1,\dots,4\), \[ \begin{aligned} x_1(n+1)&=\beta x_1(n)+a_{11}f_1(x_1(n-\tau_1))+ a_{12}f_2(x_2(n-\tau_2)),\\ x_2(n+1)&=\beta x_2(n)+a_{21}f_3(x_1(n-\tau_3))+a_{22}f_4(x_2(n-\tau_4)),\end{aligned} \] where \(x_1\), \(x_2\) are the activities of the neurons, \(\beta\in (0,1)\) is the internal decay of the neurons, the constants \(a_{ik}\) denote the connection weights, \(f_i:\mathbb{R}\to\mathbb{R}\), \(f_i(0)=0\) are continuous transfer functions, with the aim to study the stability of the null solution and the Neimark-Sacker bifurcation in this system at the usage of \textit{Yu. A. Kuznetsov}'s technique [Elements of applied bifurcation theory. 2nd updated ed., New York, NY: Springer (1998; Zbl 0914.58025)]. The direction and the stability of the Neimark-Sacker bifurcation are investigated by applying the normal form and center manifold theory.