On a multivariate strong renewal theorem (Q1800934)

For a probability distribution \(F\) on \(\mathbb R^{d}\), the renewal measure is given by \(U(\cdot)=\sum_{n=1}^{\infty}F^{\ast n}(\cdot)\), where \(F^{\ast n}(\cdot)\) is the \(n\)-fold convolution of \(F\). In this paper, the author assumes that \(F\) is in the domain of attraction -- without centering -- of a nondegenerate law, i.e., there exists a sequence \(\{ a_{n}\} \) of positive real numbers such that \(F^{\ast n}(a_{n}dx)\) weakly converges to an \(\alpha\)-stable law. Let \(A(x)\) denote a regularly varying function so that \( A(a_{n})/n\rightarrow 1\). For \(x\in \mathbb R^{d}\), let \(|x|\) denote its Euclidean norm and define \(S^{d-1}=\{ x\in \mathbb R^{d}:|x|=1\} \). In the main result of the paper, the author states conditions so that for \(x\in S^{d-1}\), we have \(\lim_{t\rightarrow \infty}t^{d}U(tx+E)/A(t)=g(x)u(E)\), where \(g(\cdot)\) is a function not identically equal to zero, and \(u(\cdot)\) is a nonzero measure on \(\mathbb R^{d}\). To prove the result, the author considers into detail the contribution of the two parts \(\sum_{n\geq A(\delta |x|)}F^{\ast n}(x+E)\) and \(\sum_{n<A(\delta |x|)}F^{\ast n}(x+E)\). In the paper several corollaries and examples are given.