A General Mixture-Truncation Method for Generating Bivariate Random Variables from Univariate Generators

摘要

This thesis deals with the generation of bivariate random variables with identical marginal distributions and specified correlation coefficients. Many schemes have been put forward for different cases; they almost always involve computing the inverse probability distribution of the marginal distributions or exploiting special properties of the random variables in question. A simple mixture-truncation method is put forward here to generate bivariate distributions with any marginal distribution from univariate generators.