AVERAGE RUN LENGTH COMPUTATION OF ARMA CHARTS FOR STATIONARY PROCESSES
Author:
- W. JiangWJW. Jiang
- 31 August 2001
- journal article
- Published by Taylor & Francis in Communications in Statistics - Simulation and Computation
- Vol. 30 (3) , 699-716
Abstract
The ARMA chart is a unified family of statistical process control techniques proposed in [1] for monitoring the mean level of stationary processes. Based on simulation studies, the ARMA chart has been shown to be comparable to the optimal EWMA chart for monitoring IID processes and outperform other conventional charts proposed for monitoring autocorrelated processes. In this paper, a Markov chain model is developed for evaluating the run length performance of the ARMA chart applied to an ARMA(p, q) process. The algorithm is implemented using Sparse Matrix operations in MATLAB and the approximation results are consistent with the simulation results.Keywords
This publication has 14 references indexed in Scilit:
- A numerical study of large sparse matrix exponentials arising in Markov chainsComputational Statistics & Data Analysis, 1999
- A Statistical Control Chart for Stationary Process DataTechnometrics, 1998
- The Matlab® 5 Handbook1998
- On EWMA Charts for Time Series1997
- Run-Length Distributions of Special-Cause Control Charts for Correlated ProcessesTechnometrics, 1994
- Analysis of Some Krylov Subspace Approximations to the Matrix Exponential OperatorSIAM Journal on Numerical Analysis, 1992
- Effects of autocorrelation on control chart performanceCommunications in Statistics - Theory and Methods, 1992
- Statistical process control procedures for correlated observationsThe Canadian Journal of Chemical Engineering, 1991
- Exponentially Weighted Moving Average Control Schemes: Properties and EnhancementsTechnometrics, 1990
- An approach to the probability distribution of cusum run lengthBiometrika, 1972