Selecting optimal follow-up split-plot designs

Thumbnail Image
Date
2019
Authors
Yang, Bowei
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Split-plot designs and follow-up designs have been considered in literature for many years. However, no research has been done on selecting optimal follow-up split-plot designs for response surface methodology. In this thesis, we will find optimal follow-up split-plot designs. We modify Pareto-based coordinate exchange algorithm and apply it to follow-up split-plot designs. The algorithm allows for selecting optimal follow-up designs based on multiple criteria under user-specified models. We use DP- and I-criteria to select optimal follow-up split-plot designs. D$_{s}$ criterion and potential bias caused by model misspecification are also considered to further select optimal designs. Examples are given to illustrate the methods of selecting follow-up designs. Different follow-up run sizes are considered. Comparisons are made among the combined designs with the same follow-up run size.
Description
Keywords
Optimal Designs, Split-Plot Designs, Follow-Up Designs.
Citation