Generating correlated random variates based on an analogy between correlation and force

An approach for generating dependent random variates with specified marginal distributions and measure of correlation is developed based upon an analogy between correlation and force. We first examine the mathematical transformation implied by this analogy and attempt to construct a general algorithm for generating variates that have the desired properties. We then examine the limitations of this approach, noting some potentially surprising distributional consequences associated with linear combinations of random variables. We conclude with the development of an algorithm for generating correlated Uniform (0,1) random variates with (approximately) a specified product-moment correlation.