| Summary: | 碩士 === 朝陽科技大學 === 工業工程與管理系碩士班 === 94 === With the awareness of the consumer moves ahead, the pattern of manufacturing has shifted from mass-production to customer-oriented build-to-order production. The major characteristic of this production pattern is the short-run process. A short-run process is a process that can manufacture various products with small batch size of each product type. Traditional control charts are not adequate for short-run processes in two ways. One is that it is difficult to calculate accurate control parameters with such small batch size of each product type. The other is that it takes more samples (or more time) to detect small process shift than a short-run process can offer.
Most commonly used control charts for short-run processes are CUSUM and EWMA. The other option is the Zone Control Chart (ZCC), which is more straightforward and easy-to-use. In this study we have developed a set of modified ZCC for the short-run processes. We thereafter utilize Monte Carlo Simulation for ARL comparisons of modified ZCC, CUSUM, and EWMA while taking into account the distribution types of the process data (Normal, Uniform, and Lognormal) and possible process shifts (shift in mean, standard deviation, and both).
This study have found that the two proposed modified ZCC have better ARLs than CUSUM in all possible process shifts when the process data are normally distributed. Modified ZCC also outperform EWMA when process shift only occurred in standard deviation. On the other hand, when process data are not normally distributed, modified ZCC have better ARL1 than CUSUM and EWMA in all shifts except shift in mean. In addition, modified ZCC have larger ARL0 than others in all shifts and when the process data are uniformly distributed, that is, the likelihood of false alarm is lower.
|
|---|
