An offline inspection and disposition model incorporating discrete weibull distribution and manufacturing variation

Abstract

To solve the quality control problem for a batch produced from an unreliable production system, an economic offline inspection and disposition (ID) model was previously proposed where the process is assumed to have a constant failure rate, i.e., a geometric shift distribution. Through setting the proper value of parameters in the ID model, three commonly used policies: cost minimizing, perfect information and zero-defects are readily obtained. To facilitate the adoption of the developed ID model in practical applications, this study investigated the ID model in which two additional features are considered: (i) the process has a non-constant failure rate (NCFR). (ii) A manufacturing variation (MV) in the process exists. Two sets of recursive cost equations are obtained to establish an algorithm for the optimal ID policy of a given production lot. Numerical examples are used to show that MV and/or NCFR have significant effects on the minimum cost. The results also show that it is infeasible to reach the perfect information policy once MV exists, but that the zero-defects policy remains feasible. Finally, a concluding remark is made.