Reassessing Acceptance Sampling Plan Utilizing EWMA Statistic with Application

This paper redesigns an acceptance sampling plan using the exponentially weighted moving average (EWMA) statistic, specifically for time-series autocorrelated data. The proposed plan balances producer and consumer risks by ensuring that sample sizes and acceptance criteria meet the required Lot Quality Levels (LQL) and Acceptable Quality Levels (AQL). By incorporating the autocorrelation present in the data, the plan parameters are determined through Monte Carlo simulation, optimizing the sampling process for better quality control. The study compares the new plan with existing sampling plans based on normal distribution approximations, highlighting the advantages of the EWMA-based approach in terms of improved accuracy, responsiveness, and efficiency in monitoring quality over time. To facilitate real-world application, the paper includes tables for the plan parameters, obtained by fixing LQL and AQL while varying the smoothing constant. A practical example from the system engineering process, specifically the inside diameter measurement of forged automobile engine piston rings, demonstrates the plan's effectiveness. The results show that the redesigned acceptance sampling plan, incorporating the EWMA statistic, offers superior performance compared to traditional methods. This real-world case study underscores the dependability and usefulness of the proposed plan in actual quality control scenarios.