计算与应用数学拔尖博士生系列论坛——Generalized high-index saddle dynamics method for non-gradient systems
2019-12-06 12:00-13:00 Room 1560, Sciences Building No. 1
Abstract: The saddle points of a dynamical system are helpful to understand the evolution of space structure. Yin et al. proposed a high-index optimization-based shrinking dimer (HiOSD) method for gradient systems to compute the index-k saddle points and construct the pathway map to help understand the structure of the energy landscape. In this talk, we will present a generalized high-index saddle dynamics (GHiSD) method for non-gradient systems. We have proved the linear stability of the index-k saddle point for the dynamics we proposed under some assumptions and demonstrated the validation of the algorithm by constructing the
solution landscape of a toy model and the phase field model in the presence of shear flow.