Error distributions for random grid approximations of multidimensional stochastic integrals (Q1948705)

The authors study discrete approximations of stochastic integrals with respect to multidimensional Brownian semimartingales. Specifically, a functional central limit theorem (in the sense of stable convergence) is proved for the approximation error, extending the result of \textit{H. Rootzen} [Ann. Probab. 8, 241--251 (1980; Zbl 0428.60068)] to a multidimensional setting, going beyond equidistant/non-random approximation grids and Brownian motions as driven processes. The limiting process is a continuous semimartingale with mixed Gaussian finite-dimensional laws. Additionally, the authors discuss how to choose the (random) approximation grids optimally, in order to minimize the asymptotic approximation error. Finally, the methods are applied to discrete hedging of derivatives in the Black--Scholes model.