Abstract
In literature, the exact non-null distributions of one and two-sample rank statistics are known for small sample sizes (say N ≤ 12) only. To fill the gap between these small sample sizes and sample sizes for which approximations are accurate, new algorithms are developed. These algorithms are based on probability generating functions and we show that they give us the opportunity to compute the exact non-null distribution for much larger samples than the existing algorithms. Finally, we present some applications for nonparametric control charts and power curves.