AccSamplingDesign: An R Package for Optimizing Acceptance Sampling Plans

Abstract

This paper introduces an R package for the optimization and visualization of acceptance sampling plans. It supports classical distributions (Binomial, Poisson, Normal) and provides the first R implementation of Beta-based plans tailored for bounded, non-normal data such as proportions. The package leverages nonlinear programming to compute optimal designs significantly faster and employs less memory than traditional grid search methods, achieving over 1000× speedup for Beta-based plans and substantial improvements for Normal-based plans. The package also delivers stable performance across a wide range of sampling plan settings and constraints. Users can visualize operating characteristic curves, compare sampling plans, and apply the tools in real-world quality control settings. We outline the statistical foundation, software architecture, and key applications.