Multivariate process-oriented statistical process control: A delivery chain application

Multivariate quality control is a demand requiring the development of new tools that have potential for successes similar to those seen in univariate quality methods. This research introduces a strategy for delivery chain SPC and two multivariate methods for statistical process control (SPC) from a process-oriented perspective. Delivery chains are a new application for the process-oriented approach. Therefore, a new strategy for delivery chain SPC is presented that focuses on the special considerations needed when plotting multivariate time-based variables.; This research proposes an alternate SPC charting strategy that involves accurately representing multiple SPC charts with a single chart. The problem addressed is the need to reduce time spent investigating false alarms in a delivery chain, and the need to create new process-oriented variables that improve root-cause analyses.; Another method this research addresses is a new process-oriented multivariate capability index. Barton and Gonzalez-Barreto (1996) developed a process-oriented basis representations (POBREP) method as a means to diagnose pre-identified manufacturing irregularities. This method, however, did not provide a multivariate capability measure. Other multivariate capability indices exist, but are not created from a process-oriented perspective. The new capability index is applicable for both manufacturing and delivery chain systems.