Characterisation and optimal design of a new double sampling c chart

Abstract

[EN] This paper proposes a new double sampling scheme for c control chart (DS-c), which was designed to improve the performance of c chart or to reduce the inspection cost. The mathematical expression required to do an exact evaluation of ARL and ASN is deduced. Further, a bi-objective genetic algorithm is implemented to obtain the optimal design of the DS-c scheme. This optimisation is aimed to simultaneously minimising the error probability type II and the ASN, guaranteeing a desired level for the error probability type I. A performance comparison between the double sampling (DS), fixed parameters (FP), variable simple size (VSS) and exponential weighted moving average (EWMA) schemes for the c chart is carried out. The comparison shows that with the implementation of DS-c scheme is obtained a significant reduction of the out of control ARL with a lower ASN respect to FP and a better ARL profile than VSS and EWMA.