Optimal inspection policy in a serial production system

Abstract

This thesis is concerned with the optimal quality control inspection policy in a serial production system with inspection errors. It is assumed that the fraction defective varies from lot to lot with some prior distribution. One of the following three decisions may be made at each inspection stage; no inspection, rectifying sampling inspection and 100% screening inspection. Defectives found are replaced by nondefective units. The goal is to seek the right combination of inspection levels and locations of inspection stations in order to minimize the total expected cost of inspection and replacement within the specified constraint on quality of the finished products. A mathematical model is formulated and solved using a dynamic programming approach. A numerical example is also given.