On two approaches to approximation of multidimensional stable laws (Q697476)

The most discouraging feature of multivariate stable distributions is that none of them is known in closed form, even in the simplest case of spherically invariant stable distributions and their densities. The authors obtain estimates on the accuracy of approximations of multivariate stable laws, in terms of certain probability metrics, using two different approaches: The first one involves discretisation of the spectral measure, whereas the second is based on the central limit theorem for multivariate stable laws using generalized Pareto-variables. A discussion on simulation of multivariate stable laws using either of these approaches is also included, as well as a comparison to the so called LePage series representation as a method of simulation.