Persistent search in confined domains: a velocity-jump process model

TL;DR

This paper models persistent search behavior in confined domains using a velocity-jump process, analyzing how domain boundaries and target placement affect search efficiency through mathematical and numerical methods.

Contribution

It introduces a Markov chain approximation for velocity-jump search models in bounded domains and derives optimal transition rates for partitioned domains.

Findings

Search time decreases for targets near the boundary.

Small turning probabilities slightly reduce search time.

Optimal transition rate scales as the inverse square root of mean transition time.

Abstract

We analyze velocity-jump process models of persistent search for a single target on a bounded domain. The searcher proceeds along ballistic trajectories and is absorbed upon collision with the target boundary. When reaching the domain boundary, the searcher chooses a random direction for its new trajectory. For circular domains and targets, we can approximate the mean first passage time (MFPT) using a Markov chain approximation of the search process. Our analysis and numerical simulations reveal that the time to find the target decreases for targets closer to the domain boundary. When there is a small probability of direction-switching within the domain, we find the time to find the target decreases slightly with the turning probability. We also extend our exit time analysis to the case of partitioned domains, where there is a single target within one of multiple disjoint subdomains. Given an average time of transition between domains $\langle T \rangle$, we find that the optimal rate of transition that minimizes the time to find the target obeys $\beta_{{\rm min}} \propto 1/ \sqrt{\langle T \rangle}$.