A note on the weak invariance principle for local times (Q1359778)

Let \(\{X_k\}\) be a sequence of i.i.d. integer-valued random variables with common distribution \(F\) which belongs to a domain of attraction of a strictly \(\alpha\)-stable distribution with \(1< \alpha\leq 2\). Let \(S_n= X_1+ \cdots+ X_n\) and put \(Y^{(n)} (t)= S_{[nt]}/ a_n\). Let \(D[0,\infty)\) be as usual. Then, it is known that under some assumptions \(Y^{(n)}\) converges weakly to a strictly \(\alpha\)-stable process \(Y\) in \(D[0,\infty)\) and for the process \(Y\) there exists a version of local time \(\{L(t,x)\}\) which is jointly continuous in \((t,x)\) and satisfies the so-called time density formula. Let \(N(n,x)\) be the local time of random walk \(\{S_k\}\) and put \(L^{(n)} (t,x)= {a_n \over n} N([nt], [xa_n])\). The authors prove that it is possible to construct processes \(\widetilde Y^{(n)}\) and \(\widetilde Y\) in \(D[0,\infty)\) such that (i) \(\widetilde Y^{(n)} {\buildrel {\mathcal D} \over =} Y^{(n)}\), \(\widetilde Y{\buildrel {\mathcal D} \over =} Y\), (ii) \(\widetilde Y^{(n)}\) converges to \(\widetilde Y\) a.s. in \(D[0,\infty)\), and (iii) for the processes \(\widetilde L^{(n)} (t,x)\) and \(\widetilde L(t,x)\), defined with respect to \(\widetilde Y^{(n)}\) and \(\widetilde Y\), and for any \(T>0\), \[ \sup_{(t,x)\in [0,T] \times \mathbb{R}^1} \bigl|\widetilde L^{(n)} (t,x)- \widetilde L(t,x) \bigr|\to 0\text{ in } L^2, \] where ``\(\buildrel {\mathcal D} \over =\)'' means that two processes have the same finite-dimensional distributions.