| As today's manufacturing firms are moving towards agile manufacturing, quick and economic on-line statistical process control solutions are in high demand. Timely detection of small shifts in the process prevents production of defective items therefore reduces the cost of production and improves the quality of products.; Because of the efficiency of the double sampling (DS) X¯ charts in detecting small shifts in process means, in this research, the basic principle of the DS X¯ chart procedure is further extended to develop multiple sampling charts for controlling process mean and variability. The objective of this dissertation is to advance the state-of-the-art of quality control charts. The contribution of this thesis is the development of univariate and multivariate multiple sampling charts that are sensitive and fast in detecting small shifts in process mean and variance.; First, for univariate case the improved DS s charts are developed. In the statistical design of the improved DS s charts the assumption in the previously designed DS s charts that the process standard deviation follows a normal distribution is relaxed. Second, the joint DS X¯ and s charts are developed because in statistical quality control, usually the mean and variance of a manufacturing process are monitored jointly.; The multivariate control procedures take advantage of the relationships among the variables and therefore are more sensitive to assignable causes that are poorly detected by univariate control charts on individual variables. Third, the multivariate multiple sampling (MMS) χ2 charts are developed for controlling the process mean vector. Finally, in order to control process covariance matrix the multivariate double sampling (MDS) | S| charts are developed.; The newly developed multiples sampling charts showed significant improvement over certain ranges in the efficiency measured in average run lengths (ARL) compared to the competing schemes. |
