Open AccessA homogenously weighted moving average scheme for observations under the effect of serial dependence and measurement inaccuracy
Author(s) -
Maonatlala Thanwane,
Sandile Charles Shongwe,
Muhammad Aslam,
JeanClaude MalelaMajika,
Mohammed Albassam
Publication year - 2021
Publication title -
international journal of industrial engineering computations
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.564
H-Index - 26
eISSN - 1923-2926
pISSN - 1923-2934
DOI - 10.5267/j.ijiec.2021.5.003
Subject(s) - independent and identically distributed random variables , scheme (mathematics) , variance (accounting) , observational error , statistics , monte carlo method , sampling scheme , algorithm , dependency (uml) , computer science , mathematics , process (computing) , autocorrelation , constant (computer programming) , random variable , artificial intelligence , mathematical analysis , accounting , estimator , business , programming language , operating system
The combined effect of serial dependency and measurement errors is known to negatively affect the statistical efficiency of any monitoring scheme. However, for the recently proposed homogenously weighted moving average (HWMA) scheme, the research that exists concerns independent and identically distributed observations and measurement errors only. Thus, in this paper, the HWMA scheme for monitoring the process mean under the effect of within-sample serial dependence with measurement errors is proposed for both constant and linearly increasing measurement system variance. Monte Carlo simulation is used to evaluate the run-length distribution of the proposed HWMA scheme. A mixed-s&m sampling strategy is incorporated to the HWMA scheme to reduce the negative effect of serial dependence and measurement errors and its performance is compared to the existing Shewhart scheme. An example is given to illustrate how to implement the proposed HWMA scheme for use in real-life applications.