A partition experimental design for a sequential process with a large number of variables

Statistically designed experiments provide a systematic approach to study and analyze the effects of multiple factors on process performance. In order to improve the performance across an organization, one must investigate multiple processes from supplier to customer. Partition experimental designs provide a design to model several sequential processes simultaneously. The partition design provides the investigator with an efficient design where potential cost and time constraints exist. That is, partition designs are applicable when runs are costly and a large number of variables exist in the serial process. In this paper, we present the application of partition designs using 10 factors in a simulated case study. The design and analysis procedures are discussed, as well as the benefits.