Examples of interacting particle systems on \(\mathbb {Z}\) as Pfaffian point processes: annihilating and coalescing random walks (Q1626829)

This paper is concerned with a class of interacting particle systems consisting of annihilating and coalescing random walkers. These are Markovian processes \((\eta_t)_{t \geq 0}\) on \(\{0,1\}^{\mathbb{Z}}\), where \(\eta_t(x)=1\) means that a random walker is present at site \(x\in \mathbb{Z}\) at time \(t\). Given the initial distribution of all walkers at time \(t=0\), these walkers jump independently of each other either left or right. The rates at which they do so are allowed to depend on their precise location on \(\mathbb{Z}\). At the occurrence of a collision, that is, a walker jumping to a site where another walker is already present, one of the following two things happens: either the walkers instantaneously annihilate each other; this happens with probability \(\theta \in [0,1]\), or they coalesce into one walker; happening with probability \(1-\theta\). In contrast to the jump rates, the parameter \(\theta\) does not depend on the location of the walkers. \par Theorem \(1\), the main result of the paper, says that, for any initial configuration \(\eta \in \{0,1\}^{\mathbb{Z}}\), and at any fixed time \(t>0\), the variable \(\eta_t\) is a Pfaffian point process on \(\mathbb{Z}\) with a given kernel \(\mathbf{K}\) constructed from a single scalar function. The key tool in the proof of this is a Markov duality property. Moreover, this scalar function can be characterised as the solution to a system of differential equations. This again enables the derivation of explicit Pfaffian kernels for a variety of coalescing and annihilating Brownian motions by taking continuum limits (after diffusive scaling) of the discrete-time processes, as outlined in Theorem \(2\). All these results generalize the work in [the last two authors, Electron. J. Probab. 16, Paper No. 76, 2080--2103 (2011; Zbl 1244.60097)] to several very interesting cases.