Adaptive EWMA control charts for monitoring dispersed count data: a COM-Poisson-based self-adjusting framework and applications
Year: 2026
Type: article
Abstract: Control charts for count data are based on the Poisson (equi-dispersed), Negative Binomial (over-dispersed), or Binomial (under-dispersed) distributions. However, the Conway-Maxwell-Poisson (COM-Poisson) distribution can flexibly model data with varying dispersion levels (under-, over-, and equi-dispersion). Over the past few years, adaptive control charts have gained attention compared to traditional Exponentially Weighted Moving Average (EWMA) charts because of their ability to dynamically adjust the smoothing constant in response to process shifts. Existing adaptive EWMA charts are primarily restricted to Poisson-based models and therefore fail to provide a unified monitoring framework for under-, equi-, and over-dispersed count data. To overcome this gap, this study is the first to develop two Adaptive Exponentially Weighted Moving Average (AEWMA) control charts for the COM-Poisson distribution, allowing the smoothing constant to dynamically adjust estimated process shifts for a fixed rate parameter under-, equi-, and over-dispersed settings. The COMP-AEWMA-I chart is constructed using either of the two given smoothing strategies, whereas the COMP-AEWMA-II chart utilises dispersion-specific upper-tail quantiles for the COM-Poisson distribution to define intervals that govern the adaptive selection of the smoothing constant. The primary objectives of this study are to detect shifts promptly and to enhance process monitoring accuracy across various dispersion levels. Using the Monte Carlo simulation method, the performance of the proposed control charts is evaluated using key measures, including the average, standard deviation, and percentiles of run lengths for both zero- and steady-state cases. To assess the practical applicability, validity, and robustness of the proposed charts, real-life datasets from different fields are also used.
Source: Scientific Reports
Language: English
Subfield: Statistics, Probability and Uncertainty
Field: Decision Sciences
Domain: Social Sciences
Open Access status: gold
Estimated APC paid: $2,690
Funder: King Saud University
Award: G1365612741
OpenAlex ID: W7233718553
