Existence of periodic solutions for a system of delay differential equations (Q1044502)

The work is mainly devoted to the existence of periodic solutions of a system of delay differential equations representing a two-neuron network model of Hopfield type with discrete time delays. The dynamics of this model is governed by the following system: \[ x'(t)=-x(t)+p f(x(t))+s f(y(t-\tau_1)), \] \[ y'(t)=-y(t)+p f(y(t))+r f(x(t-\tau_2)), \] where \(x\) and \(y\) represents the voltages of the neurons, and the nonnegative time delay satisfy \(\tau_1+\tau_2>0\). The nonlinear function (fire rate) \(f: \mathbb R\rightarrow \mathbb R\) is assumed to be odd, \(f'(0)=\lambda>0\), \(f(\pm\infty)=\pm M\), \(f\) is concave on \((0,\infty)\). The Poincaré-Bendixson Theorem for monotone cyclic feedback systems with delay is applied to establish, roughly speaking, that if the unique equilibrium \((0,0)\) is unstable, then the considered system possesses a periodic solution. In contrast, sufficient conditions are proposed that guarantee a global asymptotic stability. The theoretical analysis is supported by numerical examples.