Exponential periodicity and stability of delayed neural networks (Q1877713)

The paper is concerned with the system of functional-differential equations \[ \dot x_i(t) = - f_i(x_i(t)) + \sum_{j=1}^n a_{ij} g_j(x_j(t)) + \sum_{j=1}^n b_{ij} g_j(x_j(t-\tau_j)) + I_i(t), \quad i=1,\dots,n, \tag{1} \] where \(t\geq 0\), \(x_i(t)=\phi_i(t)\) for \(-\tau\leq t \leq 0\), \(0\leq \tau_j\leq \tau\), \(a_{ij}\) and \(b_{ij}\) are constants. \(I(t)=(I_1(t),\dots,I_n(t))^T\) is an input periodic vector function. The authors obtain sufficient conditions for the existence and uniqueness of periodic solutions of system (1) without assuming differentiability and boundedness of the activation functions \(g_j\).