An approach is proposed to generate random vectors using transformation with multiple roots. This approach generalizes the one-dimensional inverse transformation with multiple roots method to higher dimensions, i.e., to random vectors with or without densities. In this approach, multiple roots of the transformation and probabilities of selecting each of the roots are derived. The strategies for constructing such a transformation are discussed and several examples are presented to motivate this simulation approach.
Peng, Qidi; Schellhorn, Henry; and Zhu, Lu
Generating Random Vectors Using Transformation with Multiple Roots and its Applications,
Applications and Applied Mathematics: An International Journal (AAM), Vol. 10,
1, Article 4.
Available at: https://digitalcommons.pvamu.edu/aam/vol10/iss1/4