abstract
- We develop optimal sublot policies to minimise total expected costs in a failure-prone two-stage, high-volume repetitive production process which manufactures batches of the same item over a long period of time. The first stage starts in control but may fail at an exponentially distributed random time, producing only good units before failure and only defective units afterward. Failure is detected upon encountering defective units in the second stage, at which time the first stage process may be restored to good condition. To enhance quality, a production lot in Stage 1 is subdivided into sublots which are transferred as soon as possible to Stage 2 for inspection and processing. We determine exact closed-form continuous solutions as well as an efficient algorithm for finding optimal integer sublots. We provide managerial insights as to what control mechanism would be adopted depending on the state of the production process parameters. [Received 14 May 2015; Revised 16 February 2016; Accepted 1 March 2016]