Rates of convergence to Brownian local time (Q689459): Difference between revisions

Let \(X_ 1,X_ 2,\ldots\) be i.i.d. r.v.'s with \(\mathbb{E} X_ i=0\), \(\mathbb{E} X^ 2_ i=1\) and let \(S_ n=X_ 1+X_ 2+\cdots+X_ n\). Define the ``local time'' of \(S_ n\) by \[ \eta(k,n)=\#\bigl\{j:j\leq n,| S_ j- k|\leq 1/2\bigr\}. \] The authors' goal is to prove a strong invariance principle claiming that \(\eta(k,n)\) can be approximated by the local time \(L(x,n)\) of a Wiener process. The main result tells us that for a suitable construction of the sequence \(\{X_ i\}\) and the Wiener process we have \[ \sup_{x\in\mathbb{Z}}|\eta(x,n)-L(x,n)|=O(n^{1/4}(\log n)^{1/2}(\log\log n)^{1/4}) \] if \(\mathbb{E}| X_ i|^{5+\varepsilon}<\infty\).