📝 SummarXiv

Abstract

PDEs with periodic boundary conditions are frequently used to model processes in large spatial environments, assuming solutions to extend periodically beyond some bounded interval. However, solutions to these PDEs often do not converge to a unique equilibrium, but instead converge to non-stationary trajectories existing in the nullspace of the spatial differential operator (e.g. $\frac{\partial^2}{\partial x^2}$). To analyse this convergence behaviour, in this paper, it is shown how such trajectories can be modeled for a broad class of linear, 2nd order, 1D PDEs with periodic as well as more general boundary conditions, using the Partial Integral Equation (PIE) representation. In particular, it is first shown how any PDE state satisfying these boundary conditions can be uniquely expressed in terms of two components, existing in the image and the nullspace of the differential operator $\frac{\partial^2}{\partial x^2}$, respectively. An equivalent representation of linear PDEs is then derived as a PIE, explicitly defining the dynamics of both state components. Finally, a notion of exponential stability is defined that requires only one of the state components to converge to zero, and it is shown how this stability notion can be tested by solving a linear operator inequality. The proposed methodology is applied to examples of heat and wave equations, demonstrating that exponential stability can be verified with tight bounds on the rate of decay.