Bifurcation formula of transition paths in stochastic dynamical systems by spectral flow

Abstract: This paper investigates the bifurcation and stability of the most probable transition paths (MPTPs) in stochastic dynamical systems using variational methods and spectral flow. Drawing from Onsager-Machlup theory, we model MPTPs as minimizers of an autonomous Lagrangian functional dependent on both energy level and noise intensity. We establish existence criteria for minimizers using variational techniques and critical values. The spectral flow formula is obtained to detect bifurcation points and characterize stability under perturbations of noise intensity. Two cases of stability are analyzed in detail, revealing how slight changes in noise intensity can lead to loss of minimum property of MPTPs. The results link variational bifurcations to stochastic phase transitions, providing insight into the detection of transitions in noisy systems.