Residual Based Control Charts for Zero-inflated Poisson Processes

Revolution in manufacturing and service industries brings a considerable change in the quality of the products and services. Most systems produce near-zero defects; therefore, data related to defects has many zeros. For estimation, the traditional Poisson model cannot deal with the excess number of zeros. Hence, a possible alternative solution is to use the zero-inflated Poisson model. From a quality control perspective, many control charts monitor zero-inflated Poisson processes. However, very few have considered covariates along with the zero-inflated Poisson variable in monitoring and termed model-based monitoring. This study is designed to propose the model-based Homogenous Weighted Moving Average (HWMA) and Double Homogenous Weighted Moving Average (DHWMA) control charts based on the Pearson residuals of ZIP models. In addition, a simulation-based comparative study is designed where findings are reported using run-length metrics. The findings revealed that the PR-DHWMA chart performs relatively better than the PR-HWMA chart.