Assessing the economic-statistical performance of variable acceptance sampling plans based on loss function

Acceptance sampling plans (ASPs) for attributes are sometimes misapplied to normal quality characteristics. When inspection costs and quality levels are high, using variable ASPs (VASPs) can be preferable. Among developed approaches to design ASPs, few studies have incorporated losses into the cost objective function. Their limited attention, such as focusing on limited random scenarios, considering only the activation of one specification limit, failing to compare VASPs with military standards still in use, and relying on time-consuming solution procedures, motivated us to utilize the advantages of loss-based economic-statistical design, evaluate four VASPs and two military standards, and presenting detailed results. Additionally, we develop the first Particle swarm optimization (PSO)-based solution procedure for designing VASPs. Numerical and real case studies, which consider the activation of lower and upper specification limits, demonstrate the superior performance of (1) the repetitive group sampling plan, (2) MIL-STD-414 over MIL-STD-105E, and (3) PSO compared to other approaches.