Description: Implements the NonuniformDMD class for a new DMD algorithm with the flexibility to non-uniform time stepping. The class is inherited from DMD class, similar to AdaptiveDMD. Example runs and results are added to heat_conduction.cpp (with uniform time stepping) and dg_euler.cpp (with non-uniform time stepping).
Thoughts: This new algorithm may be more friendly to parametric case of varying time-stepping applications. We do not need to bother about the time step size and interpolation.
Discussion: The two algorithms are currently named AdpativeDMD and NonuniformDMD have similar goals to do DMD on non-uniform time stepping data, but the names may not be informative enough. In fact, I found literatures doing similar jobs as well, and they are termed NU-DMD, where NU stands for non-uniform. I am thinking if InterpolatedNUDMD and ExponentialNUDMD are suitable names.
Description: Implements the
NonuniformDMDclass for a new DMD algorithm with the flexibility to non-uniform time stepping. The class is inherited fromDMDclass, similar toAdaptiveDMD. Example runs and results are added toheat_conduction.cpp(with uniform time stepping) anddg_euler.cpp(with non-uniform time stepping).Thoughts: This new algorithm may be more friendly to parametric case of varying time-stepping applications. We do not need to bother about the time step size and interpolation.
Discussion: The two algorithms are currently named
AdpativeDMDandNonuniformDMDhave similar goals to do DMD on non-uniform time stepping data, but the names may not be informative enough. In fact, I found literatures doing similar jobs as well, and they are termed NU-DMD, where NU stands for non-uniform. I am thinking ifInterpolatedNUDMDandExponentialNUDMDare suitable names.