Abstract: This seminar presents investigations of the lumen degradation of light emitting diodes (LEDs) subject to stress loadings. Cumulative damage measurements are collected from a two-variable constant-stress accelerated degradation test (ADT). The underlying process for the data is a Wiener diffusion process with a drift which depends on the stress loadings. General statistical inferences on the parameters and percentiles of the LED lifetime distribution are presented. Approximate lower confidence bounds of the LED percentile lifetime are given using the Fisher information of the maximum-likelihood estimates and Bonferroni's inequality. This work establishes optimal strategies on the constant-stress ADT plan for a compromised decision between experiment budget and estimation precision. The study provides an algorithm to search the optimal strategy for the ADT. Finally, an example of LED tests is used to illustrate applications of the proposed methods.

Bio: Dr. Yuhlong Lio received Ph.D. in Statistics from University of South Carolina in 1987. Since then he has served as assistant professor, associate professor and professor in the Department of Mathematical Sciences, University of South Dakota.  Dr. Lio is an Associate Editor for the Journal of Statistical Computation and Simulation.  His research interest includes Kernel smooth quantile estimation, reliability, survival analysis and statistical quality control.