On an Adaptive Acceptance Control Chart for Autocorrelated Processes
Standard control chart techniques for process surveillance are usually assessed under the assumption that sampled values from a process in control may be represented by independent, identically distributed random variables. In some applications, however, the process exhibits an inherent dynamics with slow variations of the process level, even in a state of statistical control. In order to describe these variations we suggest a simple time—series model for the process level. The model implies that an explicit estimate of the current process level may be obtained by the Kalman Filter technique. In the paper we show how this estimate may be utilized in an acceptance control chart to vary the sample size in accordance with the estimated actual process level, without sacrificing a specified consumer protection.
KeywordsControl Chart Acceptance Criterion Process Level Precision Parameter Variable Sampling Interval
Unable to display preview. Download preview PDF.
- FREUND, R.A. (1957): Acceptance Control Charts, Industrial Quality Control, XIV, No. 4. October 1957Google Scholar
- HUNTER, J.S. (1986): The Exponentially Weighted Moving Average, Journal of Quality Technology, 18: 203–210Google Scholar
- JASWINSKI, A.H. (1970): Stochastic Processes and Filtering Theory.. New York: Academic Press.Google Scholar
- MEINHOLD, M.R. and SINGPURVALLA, N.D. (1983): Understanding the Kalman Filter, The American Statistician, 37: 123–127Google Scholar