Spiking and Resetting
Cédric Bernardin, Vsevolod Vladimirovich Tarsamaev
TL;DR
This work rigorously derives the singular spiking limit for a one-dimensional PDMP with resetting, showing that the time-space pre-spike points converge to a decorated Poisson point process with intensity f(0)^2 dt ⊗ x^{-2} dx. The authors develop a robust renewal- and Laplace-transform framework, proving conditional no-jump rectangle convergence and then full Poisson convergence, while providing explicit asymptotics for key generating-function components E^ε, D^ε, and C^ε. These results vindicate conjectures from prior physics literature and clarify the mechanism by which spiking events emerge in the small-ε limit. The work highlights a deep connection between resetting dynamics, boundary-layer behavior, and Poissonian spiking structures that may inform broader resetting and PDMP analyses.
Abstract
We consider a one-dimensional piecewise deterministic Markov process (PDMP) on $[0,1]$ with resetting at $0$ and depending on a small parameter $\varepsilon>0$. In the singular vanishing limit $\varepsilon \to 0$ we prove that the `` resetting '' simple point process associated to the PDMP converges to a point process described by a jump Markov process decorated by ``spikes'' distributed as a time-space Poisson point process with intensity proportional to $dt \otimes x^{-2} dx$. This proves rigorously results appeared previously in \cite{SBDKC25} and also justifies partially a conjecture formulated there.
