Publication Date

Spring 5-14-2021

Abstract

When developing a system, considering system performance from a user perspective can be done through operational testing|assessing the ability of representative users to accomplish tasks with the system in operationally representative environments. This critical process can be expensive and time-consuming. We show how to leverage an existing design of experiments (DOE) process to construct a Bayesian adaptive design. This method allows for interim analyses using predictive probabilities to stop testing early for success or futility. Furthermore, operational environments are directly used in product evaluation. Representative simulations demonstrate reductions in necessary test events. Next, priors are built using developmental testing data. The novel proposal for creating priors using developmental testing data allows for more flexibility than the current process and demonstrates it is possible to get more precise parameter estimates. The methods presented will allow future testing to be conducted in less time and at less expense, on average.

Degree Name

Statistics

Level of Degree

Doctoral

Department Name

Mathematics & Statistics

First Committee Member (Chair)

Fletcher G. W. Christensen

Second Committee Member

Gabriel Huerta

Third Committee Member

Raymond Hill

Fourth Committee Member

Laura Freeman

Language

English

Keywords

Bayesian adaptive design, conditional normalized partial borrowing power prior, defense acquisition, mission set analysis, operational testing, power priors

Document Type

Dissertation

Comments

Document has obtained USAF Public Affairs release; page including this information has been added to the dissertation. For questions please contact me at: vcarrillo314@gmail.com, 509-768-4853, or vcarrillo314@unm.edu.

Download

Share

COinS