A Comparison Study of Second-Order Screening Designs and Their Extension
Abstract
Recent literature has proposed employing a single experimental design capable of preforming both factor screening and response surface estimation when conducting sequential experiments is unrealistic due to time, budget, or other constraints. Military systems, particularly aerodynamic systems, are complex. It is not unusual for these systems to exhibit nonlinear response behavior. Developmental testing may be tasked to characterize the nonlinear behavior of such systems while being restricted in how much testing can be accomplished. Second-order screening designs provide a means in a single design experiment to effectively focus test resources onto those factors driving system performance. Sponsored by the Office of the Secretary of Defense (ODS) in support of the Science of Test initiative, this research characterizes and adds to the area of second-order screening designs, particularly as applied to defense testing. Existing design methods are empirically tested and examined for robustness. The leading design method, a method that is very run efficient, is extended to overcome limitations when screening for non-linear effects. A case study and screening design guidance for defense testers is also provided.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2013
- Accession Number
- ADA590760
Entities
People
- Shane A. Dougherty
Organizations
- Air Force Institute of Technology