Automating large-scale LEMUF calculations

To better understand material unaccounted for (MUFs) and, in some cases, to comply with formal regulatory requirements, many facilities are paying increasing attention to software for MUF evaluation. Activities related to improving understanding of MUFs are generic (including the identification, by name, of individual measured values and individual special nuclear material (SNM) items in a data base, and the handling of a wide variety of accounting problems) as well as facility-specific (including interfacing a facility's data base to a computational 'engine' and subsequent uses of that engine). Los Alamos efforts to develop a practical engine are reviewed and some of the lessons learned during that development are described. Major obstacles to implementation do not involve lack of software or statistical theory, but the lack of resources required for implementation and (in some cases) realistic measurement uncertainties.