A functional LIL for stochastic integrals and the Lévy area process (Q1780924): Difference between revisions

Consider the class of stochastic integrals of the form \(X(t) = \int_0^t \langle A W(s), \,dW(s) \rangle\), \(t\geq 0\), where \(A\) is a real, non-zero, skew-symmetric \((d \times d)\)-matrix, and \(\{W(t)\}_{t \geq 0}\) a standard Brownian motion. For \(d=2\), this class in particular contains Lévy's stochastic area process. The authors establish a functional law of the iterated logarithm for the process \(\{X(t)\}_{t \geq 0}\) as well as an almost-sure convergence result for the weighted occupation measures. This extends earlier results by \textit{Z. Shi} [in: Séminaire de probabilités XXVIII. Lect. Notes Math. 1583, 122--137 (1994; Zbl 0810.60076] and \textit{B. Rémillard} [Ann. Probab. 22, No.~4, 1794--1802 (1994; Zbl 0840.60030]. The proofs draw on small-ball estimates for Brownian motion and provide small-ball estimates for the weighted sup-norm of \(\{X(t)\}_{t \geq 0}\).