Optimal fluctuation pathways to desynchronization in coupled oscillator networks

There is great interest in understanding how topology, dynamics, and uncertainty conspire to produce rare and extreme events in networks. This is particularly the case for coupled oscillator networks since they appear at the core of many biological and physical systems where noise and uncertainty play a significant role. A primary example is desynchronization in power grids from input-power fluctuations. In this talk, we develop theory for the most-likely, or optimal, pathway of noise-induced desynchronization in phase-oscillator networks (with and without inertia) and in Stuart-Landau oscillator networks. We quantitatively characterize the scalings and patterns for the optimal path and the probability of desynchronization as a function of network topology and local dynamics, and compare the behavior for the various models. Lastly, we discuss the effects of non-Gaussian, “pulse” noise, and controls on the input power, on desynchronization. Such effects are especially relevant for power grids with renewable energy sources.