PascalX - Supplementary information

with Fχ2 m+2k the CDF of a χ distribution with m + 2k degrees of freedom and β > 0 an arbitrary constant. For ak ≥ 0 and ∑ k ak = 1 the above sum constitutes a so-called (infinite) mixture representation. In particular, every β ≤ λmin yields such a mixture representation, with λmin denoting the smallest coefficient of the linear combination of χ distributed random variables. Ruben showed that in the mixture case, the truncation error, and so the precision ε, can be bounded by ∣∣∣∣1− K−1 ∑