Flocking of a Kuramoto type model with local interaction functions (Q6110924)

The authors consider the model \[ \frac{d\mathbf{x}_i}{dt} = (\cos{\theta_i},\sin{\theta_i}) \] \[ \frac{d\theta_i}{dt} =\frac{\lambda}{N_i(t)}\sum_{k=1}^N\psi(|\mathbf{x}_k-\mathbf{x}_i|)\sin{(\theta_k-\theta_i)} \] where \(\mathbf{x}_i\) and \(\theta_i\) are the position and phase, respectively, of the \(i\)th agent in the plane. \(\lambda\) is the coupling strength between agents, \(N_i(t)\) is the number of other agents within a distance \(r\) of agent \(i\) at time \(t\), and \(\psi\) is positive, non-increasing, and equal to zero for an argument greater than \(r\), i.e.~agents more than a distance \(r\) apart do not interact. The connectivity of the system can be described by a time-dependent symmetric matrix whose \((i,j)\)th entry is 1 if agents \(i\) and \(j\) are less than \(r\) apart, and zero otherwise. This matrix then describes a graph, with agents corresponding to vertices. The main result is: assuming that the initial graph is connected, if certain inequalities on the initial conditions hold then the system will achieve flocking. The flocking state is defined by having all phases approach constant values, and the distances between all pairs of agents being less than \(r\). The phases approach their final values exponentially.