266 – Contributed Oral Poster Presentations: Quality and Productivity Section
Bootstrap-Based Confidence Intervals in Partially Accelerated Life Testing Under the Generalized Exponential Distribution
Ahmed Eshebli
Missouri University of Science
V.A. Samaranayake
University of Missouri-Rolla
Partially accelerated life testing (PALT) is preferable over accelerated life testing (ALT) in situations where a model linking the stress to the distribution parameters is unavailable. Under the assumption of a generalized exponential life distribution and Type I censoring, a bootstrap-based method of obtaining confidence intervals for the distribution parameters, the acceleration factor, and the mean life, is introduced. Its performance is studied against that of intervals obtained using the traditional delta method, using Monte Carlo simulation. Results show that the bootstrap-based method performs better than the traditional approach.