Optimum target levels for a network of processes

These days, it is common to have the same item produced with varying features and levels of quality and offered to the consumer as different products. This research considers such manufacturing systems and determines their optimum machine settings. Each process in the system generates a random quality characteristic that has specification limits. Depending on the value of the quality characteristic, a product can be reworked, scrapped or forwarded to the next process. The processes are continuously running; hence, the "long-term" probabilities of meeting the specifications are developed. These are used to construct the profit function. The case of a network of processes is considered; that is where each product passes through varying and possibly unique routes with different feeding and exit points. A novel technique to readily model such systems is presented, and its objective function is developed.