Abstract
We propose a simple and efficient way to approximate multivariate normal probabilities using univariate and bivariate probabilities. The approximation is computationally tested for the trivariate and quadrivariate normal probabilities. A few problems of higher dimensions were also tested.