Computing the distribution of the sum of dependent random variables via overlapping hypercubes (Q894208)

For the computation of VaR (value-at-risk), Arbenz, Embrechts and Puccetti [\textit{P. Arbenz} et al., Bernoulli 17, No. 2, 562--591 (2011; Zbl 1248.60018)] have proposed a new algorithm, called AEP after the names of the authors. In this paper, the author proves that the AEP algorithm converges for any \(d \geq 2\) and any absolutely continuous distribution \(H\) (with bounded density in a neighborhood of the simplex diagonal). The proof of the convergence is very detailed and precisely given.