Stability and bifurcation of a two-dimension discrete neural network model with multi-delays (Q2425882): Difference between revisions

A forward Euler method is applied to the \(2\)-dimensional Hopfield-type neural network model \[ \begin{aligned} \dot u_1(t)&=-u_1(t)+a_1F_1(u_2(t-\tau_1))\\ \dot u_2(t)&=-u_2(t)+a_2F_2(u_1(t-\tau_1)) \end{aligned} \] where the \(C^2\)-functions \(F_1,F_2\) vanish at the origin. Provided the step-size satisfies \(h=1/m\), this yields a planar difference equation with delay \(m\). For this discretization, the authors derive step-size independent conditions for instability and absolute stability of the zero solution. Moreover, there occur Hopf bifurcations when the delay passes a sequence of critical values. Using normal form and center manifold techniques, explicit formulas to determine the direction of the Hopf bifurcations are derived, as well as results on the stability of the bifurcating objects. Finally, computer simulations for an explicit example illustrate the obtained results.