Predictability of solitary decoherence events in a symmetrically equidistant chain of coupled Kuramoto oscillators
Phys. Rev. E 114, 044215 – Published 19 October, 2026
DOI: https://doi.org/10.1103/fltc-hfvt
Abstract
In the modern world, many prediction algorithms are based on data collected over a long time. Could we still use such an algorithm after a major change takes place? In the present study we explore predictability of events in order to see how it changes with system parameters if the prediction algorithm remains the same. We keep both the model and the algorithm simple enough to avoid unnecessary complexity and maximize the clearness of the presentation. We explore the chain model of coupled Kuramoto oscillators, whose typical state is synchronization. We consider a slowly oscillating forcing applied to the edges of the chain. The imposed symmetries allow the model to be represented by two key parameters: fragility and forcing amplitude. A special type of desynchronization event, namely, solitary decoherence, is the subject of prediction. We consider four precursors based on the order parameter and quantify the predictability by the sum of type I and type II errors. The study reveals four parametric zones of predictability. We suggest that the predictability of desynchronization events may be considered as a measure of complexity and is similar to the order parameter.
