Encounter-based model of a run-and-tumble particle II: absorption at sticky boundaries

TL;DR

This paper develops a mathematical model for run-and-tumble particles confined in a finite interval with sticky, partially absorbing boundaries, incorporating occupation times and non-Markovian absorption processes to analyze first passage times and absorption probabilities.

Contribution

It introduces an encounter-based framework for modeling absorption at sticky boundaries, extending previous models to include occupation time dependence and non-exponential absorption distributions.

Findings

Derived a boundary value problem for the joint probability density of position, velocity, and occupation time.

Calculated mean first passage times and splitting probabilities for absorption.

Unified and extended previous non-sticky boundary results within the new encounter-based model.

Abstract

In this paper we develop an encounter-based model of a run-and-tumble particle (RTP) confined to a finite interval $[0,L]$ with partially absorbing, sticky boundaries at both ends. We assume that the particle switches between two constant velocity states $\pm v$ at a rate $\alpha$. Whenever the particle hits a boundary, it becomes stuck by pushing on the boundary until either a tumble event reverses the swimming direction or it is permanently absorbed. We formulate the absorption process by identifying the first passage time (FPT) for absorption with the event that the time $A(t)$ spent attached to either wall up to time $t$ (the occupation time) crosses some random threshold $\hat{A}$. Taking $\Psi({a})\equiv\mathbb{P}[\hat{A}\gt{a}]$ to be an exponential distribution, $\Psi[{a}]=e^{-\kappa_0a}$, we show that the joint probability density for particle position $X(t)$ and velocity state $\sigma(t)=\pm $ satisfies a well-defined boundary value problem (BVP) with $\kappa_0$ representing a constant absorption rate. The solution of this BVP determines the so-called occupation time propagator, which is the joint probability density for the triplet $(X(t),A(t),\sigma(t))$. The propagator is then used to incorporate more general models of absorption based on non-exponential (non-Markovian) distributions $\Psi(a)$. We illustrate the theory by calculating the mean FPT (MFPT) and splitting probabilities for absorption. We also show how our previous results for partially absorbing, non-sticky boundaries can be recovered in an appropriate limit. Absorption now depends on the number of collisions $\ell(t)$ of the RTP with the boundary. Finally, we extend the theory by taking absorption to depend on the individual occupation times at the two ends.