Comparison of a US Approach and a Russian Approachto the Historical Analysis of Inventory Differences

During the past few years, there have been several collaborations between US DOE laboratories and the Tomsk Chemical Combine in Russia. This paper describes a task that compares a US approach and a Tomsk approach to the historical analysis of inventory differences (IDS). We focus here on one open-literature 12-year ID sequence. We first describe the aspects that both approaches have in common: attention to serial correlation, and control limits for abrupt diversion. We then describe aspects that are unique to the Tomsk approach: attempts to separate systematic from random errors, and control limits for protracted diversion that do not widen over time. Next, we describe the aspects that are unique to the US approach: removal of serial correlation by transformation, the use of Page’s test to test for protracted diversion, and widening control limits over time.