Unbiased statistics and simulation of continuous mail processing

Abstract

Simulation is used to determine the number of sorting devices needed in an automated postal system facility. The model deals with mail types with great variability in processing characteristics. Interactions between continuously and sequentially processed batches at the sorting stations introduces a bias in conventionally collected statistics. This type of interaction in effect necessitates computation of statistics based on moving averages. Unfortu nately, none of the simulation languages provide this capability as a standard output. Although the logic of the basic simulation is straightforward, the model must be significantly modified to produce unbiased or undistorted statistics. SIMNET and SLAM II are used for independently developed final models.