Semester

Fall

Date of Graduation

2008

Document Type

Thesis

Degree Type

MS

College

Statler College of Engineering and Mineral Resources

Department

Industrial and Managements Systems Engineering

Committee Chair

Feng Yang.

Abstract

Computer simulation is an appealing approach for the reliability analysis of structure-based software systems as it can accommodate complexities present in realistic systems. When the system is complex, a screening experiment to quickly identify important factors (components) can significantly improve efficiency of the analysis. The challenge is to guarantee the correctness of the screening results with stochastic simulation responses. Control Sequential Bifurcation (CSB) is a new method for factors screening using simulation experiments, when only main effects models are considered. By grouping factors, CSB can identify the importance of factors while reducing the simulation effort. With appropriate hypothesis testing procedures embedded, CSB procedure can simultaneously control the Type I error probability and the power. The existing work has focused on normally distributed output responses. This thesis extends the existing CSB procedure by embedding Meeker's conditional sequential test to deal with binary responses and guarantee the desired error control for factor screening results. The effectiveness of the extended factor screening procedure is demonstrated with the application on a software system.

Share

COinS