Current fluctuations for independent random walks in multiple dimensions (Q662872): Difference between revisions

Let \(\eta_0(x)\) be the number of particles at site \(x\in\mathbb{Z}^d\). Assume that these particles evolve like i.i.d. continuous random walks. Denote by \(X_{m,j}(t)\), where \(m\in\mathbb{Z}^d\) and \(j=1,\ldots, \eta_0(m)\), the position at time \(t>0\) of the \(j\)-th random walk starting at site \(m\). The common jump rates of the random walks are given by a probability kernel \(\{p(x), x\in\mathbb{Z}^d\}\). Set \(\vec{v}=\sum_x xp(x)\). Scaling space by \(\sqrt{n}\) and time by \(n\), the author studies (under specified conditions) a sequence of current processes \[ \xi_n(t,\phi):= n^{-\frac{d}{4}}\sum_{m\in\mathbb{Z}^d} \sum_{j=1}^{\eta_0(m)}\left[\phi\left(\frac{X_{m,j} (nt)-[n\vec{v}t]}{\sqrt{n}}\right) -\phi\left(\frac{m}{\sqrt{n}}\right)\right], \] where \(\phi\) belongs to the Schwartz space \(S(\mathbb{R}^d)\). It is proved that \(\xi_n(\cdot,\cdot)\to \xi(\cdot,\cdot)\) in law in the Skorokhod space \(D([0,T],S'(\mathbb{R}^d))\) as \(n\to \infty\); here, \(S'(\mathbb{R}^d)\) is the dual space to \(S(\mathbb{R}^d)\). The limiting process \(\xi\) is a Gaussian one with mean zero and explicit covariance function. Also, the corresponding result is established for the scaled ``box-current process'' (describing the difference between the number of particles inside the box at time \(t\) and the number of particles initially inside the box). The limiting current process is equal in law to the solution of a given stochastic partial differential equation which is related to the generalized Ornstein-Uhlenbeck process.