scientific article; zbMATH DE number 1330538

From MaRDI portal
Publication:4261111





Consider the Gaussian process defined as follows: \[ \xi(t)= B(t)- \int_0^1 B(s) ds,\qquad t\in [0,1], \] where \(B\) is the standard Brownian bridge on \([0, 1]\). The rough asymptotics of the tail probability for \(L^1\)-norm of the centered Brownian bridge \(\xi\), the norm of \(\xi\) in the Lebesgue space \(L^1\), is defined by \(\|\xi\|_1= \int_0^1|\xi(s)|ds\). The main theorem states that \[ \lim_{r\to\infty} \tfrac {1} {r^2} \ln P(\|\xi \|_1> r)= -24. \] To prove this result the general theorem of Donsker and Varadhan is used. As a remark the extremal function \(f_0\) \((|f_0(s)|\leq 0\) a.e.), i.e. \(\|Kf_0 \|_1= \frac{1}{48}\), will remain extremal under the shift transformation \(f_0(x)\to f_0 (s-\tau)\), where \(K(s,t)\) is the covariance function of \(\xi\): \(K(t,s)= E(\xi(t), \xi(s))\).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4261111)