Global Hopf bifurcation analysis of a BAM neural network with delays (Q2471605): Difference between revisions

The paper studies the following three coupled delay differential equations (BAM neural network) \[ \begin{aligned} \dot{x}_{1}(t) & = -\mu_1 x_{1}(t) +c_{21}f_{1}(y_{2}(t-\tau_2)) +c_{31}f_{1}(y_{3}(t-\tau_2)) ,\\ \dot{y}_{2}(t) & = -\nu_2 y_{2}(t) +c_{12}f_{2}(x_{1}(t-\tau_1)),\\ \dot{y}_{3}(t) & = -\nu_3 y_{3}(t) +c_{13}f_{3}(x_{1}(t-\tau_1)), \end{aligned} \] where \(\mu_1>0\), \(\nu_j>0\), \(c_{j1}\) and \(c_{1j}\) \((j=2,3)\) are real constants, and \(\tau_1>0\), \(\tau_2>0\) delays. Sufficient conditions for the system to have multiple periodic solutions are obtained when the sum of delays is sufficiently large. The authors used the result of \textit{J. Wu} [Trans. Am. Math. Soc. 350, 4799--4838 (1998; Zbl 0905.34034)] about the global continuation of the Hopf branch.