The Box-Jenkins Manual Adjustment Chart

摘要

This paper introduces the Box-Jenkins Manual Adjustment Chart and Bounded Adjustment Chart with an example employing a series of 100 sequentially recorded observations. The first 50 observations are reviewed in a Shewhart individual observation chart. the time dependence of these observations is investigated by computing their autocorrelations and variogram. The EWMA (Exponentially Weighted Moving Average) is then chose to model the data, its parameter #lambda# estimated, and the fitted model used to control process performance for the remaining 50 observations. The Box-Jenkins Manual Adjustment Chart is demonstrated. An obvious dimunition of process variability about target is obtained by taking action after every observation. The example continues by demonstrating a nearly equal dimunition in variance using many fewer process changes using a Bounded Box-Jenkins chart. Only simple graphical methods are employed although detailed computations are included for completeness. The Box-Jenkins Manual Adjustment Chart procedures offer the professional quality control professional a valuable tool for enhancing process performance using simple graphics.