A limit theorem for isotropic random walks on \(R^d\) as \(d\to \infty\) (Q1264871)

Let \(X_i\) be a sequence of i.i.d. random vectors in \(R^d\) having spherically symmetric distribution with radial distribution \(\mu\). Then the variables \(\| \sum_{i=1}^{k}X_i\| \) are known to satisfy the central limit theorem [see \textit{J. F. C. Kingman}, Acta Math. 109, 11-53 (1963; Zbl 0121.12803)]. In the present paper it is shown that for isotropic random walks in Euclidean spaces of growing dimensions the CLT is no longer valid. Namely, the main result of the paper states that if \(X^d_i\) is a sequence of spherically invariant i.i.d. random vectors in \(R^d\) with the radial distribution \(\mu\) (the same for all \(d\)), then for any sequence of dimensions \(d_k\to\infty\) the variable \(\| \sum_{i=1}^{k}X_{i}^{d_{k}}\| /\sqrt k\) tends in probability to some constant as \(k\to\infty\).