An embedding for the Kesten-Spitzer random walk in random scenery

From MaRDI portal
(Redirected from Publication:1613617)



Abstract: For one-dimensional simple random walk in a general i.i.d. scenery and its limiting process we construct a coupling with explicit rate of approximation extending a recent result for Gaussian sceneries due to Khoshnevisan and Lewis. Furthermore we explicity identify the constant in the law of iterated logarithm.


The Kesten-Spitzer random walk in random scenery is the process \(K = (K(n))_{n\in N_0}\) defined by \(K(n) = \sum _{k=0}^n \sigma (S_k)\), where \(\sigma =(\sigma (x))_{x\in Z}\) denotes a sequence of i.i.d. random variables (quoted as random scenery) with \(E \sigma (0)=0\), \(E \sigma (0)^2=1\) and \(E|\sigma (0)|^p < \infty \) for all \(p>0\), and \(S = (S_k)_{k\in N_0}\) is a simple symmetric random walk on \(Z\) starting at \(S_0=0\) and independent of \(\sigma \). As the limiting continuous analogue can serve the process \(G\), which is called Brownian motion in Brownian scenery. The main result of the paper says that there is a coupling for \(K\) and \(G\), such that, with probability one, for every \(\varepsilon >0\) \[ \max _{0\leq m\leq n} \left |K(m)-G(m)\right |= o(n^{5/8}+\varepsilon),\quad n\to \infty . \] The result is a bit weaker to compare with the (earlier proved) special case of Gaussian scenery, but strong enough to yield the law of iterated logarithm for \(K\), namely \[ \limsup _{n\to \infty } {K(n)\over (n\log \log n)^{3/4}} = c_0 \quad \text{a.s.} \] with the exact value for the constant \(c_0 = 2^{5/4}/3\) which is the second main result of the paper.











This page was built for publication: An embedding for the Kesten-Spitzer random walk in random scenery

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1613617)