As no closed-form expressions exist, simple approximations are being sought-after to solve the inventory models.
Ghalebsaz-Jeddi et al. [14] use UNLLIðkÞ, the unit normal linear loss integral, for U1. In order to solve the investigated multi-item
problem they make use of a piecewise quadratic approximation. Axsäter [15] develops a table and polynomial
approximations to easily find the reorder point r and the order quantity Q in case of a fill rate S limitation, where S ¼ 1 A.
Within inventory management U1inv computations are needed on a large scale, e.g. in inventory problems with a huge
amount of items. So one can appreciate the direct benefit of having a highly efficient and effective approximation within
the range and the precision of double precision floating point numbers.
Other statistical applications where these and other repeated integrals can be used are listed in Fisher [16]:
Calculation of moments of truncated normal distribution;
Expression of the non-central t density;
In the posterior distribution of a Poisson variate with chi-squared prior for the squared mean parameter of the Poisson
variate.