Images and level sets of additive random walks

From MaRDI portal





Let \(X_1,\dots,X_n\) be \(N\) independent random walks on \(\mathbb{Z}^d\) starting at the origin, and define \(X({\mathbf n})=X_1(n_1)+ \cdots+X_N(n_N)\) for \({\mathbf n}=(n_1, \dots,n_N) \in\mathbb{N}_0^N\), where \(\mathbb{N}_0= \{0,1,\dots\}\). The gauge function \(\Phi\) is defined by \(\Phi({\mathbf m}, {\mathbf n})=\mathbb{P} (X({\mathbf m)}= X({\mathbf n}))\), \({\mathbf m},{\mathbf n}\in \mathbb{N}_0^N\). For any probability measure \(\mu\) on \(\mathbb{N}^N_0\), the \(\Phi\)-energy of \(\mu\) is defined as \({\mathcal E}_\Phi(\mu) =\sum_{{\mathbf m},{\mathbf n}\in\mathbb{N}_0^N}\Phi ({\mathbf m},{\mathbf n}) \mu(\{{\mathbf m}\}) \mu(\{{\mathbf n}\})\). Further, the \(\Phi\)-capacity \({\mathcal C}_\Phi (E)\) of \(E\subset\mathbb{N}_0^N\) is defined by \({\mathcal C}_\Phi(E)^{-1}=\inf \{{\mathcal E}_\Phi (\mu):\mu\) is a probability on \(\mathbb{N}_0^N\) such that \(\mu(E)=1\) and \({\mathcal E}_\Phi (\mu)>0\}\). The main result of this article provides the estimate \({\mathcal C}_\Phi(E) \leq\mathbb{E} (\#X(E))\leq 16^N{\mathcal C}_\Phi(E)\), \(E\subset \mathbb{N}_0^N\), where \(\#X(E)\) denotes the cardinality of \(X(E)\). If \(N=1\), it is shown that the constant 16 can be improved to 2. This general estimate is then used to study the level sets of \(X\) under the assumption of symmetry (i.e. \(X({ \mathbf n})\) and \(-X({\mathbf n})\) have the same distribution). For the entire collection see [Zbl 0948.00040].











This page was built for publication: Images and level sets of additive random walks

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4524751)