The performance of the synthetic control chart for detecting increases in the fraction nonconforming is examined and compared with a broader range of control charts. The strong performance previously reported for the synthetic chart is shown to be due to the inclusion of a head-start feature when computing the values of average time to signal. Without this head-start feature, it is shown that the synthetic chart performed no better than alternative charts, and that the use of a CUSUM of the run lengths of conforming items gave better results. The RL(2) chart also performs quite well under most circumstances.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados