Random walks in varying dimensions

From MaRDI portal



Abstract: We establish recurrence criteria for sums of independent random variables which take values in Euclidean lattices of varying dimension. In particular, we describe transient inhomogenous random walks in the plane which interlace two symmetric step distributions of bounded support.


Let \(X_1, X_2,\dots\) be a sequence of independent and identically distributed random variables with values in the three-dimensional lattice \(\mathbb{Z}^3\), with zero mean and finite variance. Let \(X_i'\) be the vector that results if the third component of \(X_i\) is replaced by 0. It is well-known that the random walk with steps \(X_i\) is transient, but that recurrence obtains if the steps are \(X_i'\) instead. An inhomogeneous random walk with dimension between 2 and 3 results if we take the steps to be \(X_i'\) or \(X_i\), depending on whether \(i = a_n\) for some \(n\) or not, with \((a_n)_{n \in \mathbb{N}}\) a fixed sequence of positive integers. The main result of the paper shows that, in somewhat informal language, the watershed between transience and recurrence is provided by the sequence \(a_n = \text{exp}(\text{exp}(n^{1/2}))\).











This page was built for publication: Random walks in varying dimensions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1908213)