A One-Sided EWMA Control Chart for Monitoring Process Means

Abstract

Great efforts have been devoted to two-sided exponentially weighted moving average (EWMA) control charts, while less attention has been paid to one-sided EWMA charts. This article proposes an improved one-sided EWMA chart (IEWMA) for rapid detection of upward or downward changes in the process mean. The basic idea is to accumulate positive (or negative) observations only and reset negative (or positive) observations to zero in the iterative computation of the EWMA estimate for the purpose of quickly catching the upward (or downward) change in the process mean. The comparison result of the IEWMA chart with other control charts favors the former.

Keywords:

• Average run length
• Markov chain
• Statistical process control

Mathematics Subject Classification:

• Primary 93C80
• Secondary 60J10

Acknowledgments

The authors would like to thank the department editor and three referees for their valuable suggestions. L. J. Shu's work was supported in part by the Research Committee under the grant RG013/05-06S. W. Jiang's work was partially supported by NSF-DMI #0200224 and NSF-IIS #0542881.

Notes

Note. ∗indicates out-of-control signals.