Robust Design of Loss-Based Ideal Repetitive Group Sampling Plan under Uncertainty of Input Parameters

Integrating both loss and minimum angle method (MAM) as objective functions in the economic-statistical modeling of variable acceptance sampling plans (VASPs) can yield the most cost-effective plan with the ideal operating characteristic (OC) curve. Nevertheless, occurring crises can disrupt organizational input parameters, causing inefficiencies in providing solutions. This study develops the first robust designs of VASPs, accounting for the uncertainty of input parameters. Unlike previous studies that assume fixed inputs, this research considers deviations from nominal values to address parameter uncertainty. In this way, the challenge of parameter uncertainty's impact on the effectiveness of designs is investigated. We propose a solution procedure based on Particle swarm optimization (PSO). Findings from case studies reveal that (1) a marginal cost increase in the Cost-MAM model significantly reduces overall risks, (2) the repetitive group sampling plan yields lower costs and risks, and (3) tolerating increased costs is imperative to manage potential uncertainty.