Optimizing Life‐Testing Plans With Progressive Hybrid Censoring Through Compound Design Strategy

ABSTRACT In life‐testing experiments, while single objective and constraint‐based optimal designs have been well‐explored, research on multi‐objective designs is still emerging. This paper addresses this concept by applying compound optimal designs to life‐testing scenarios. We implement a novel approach that integrates compound design strategies with considerations of cost and variability, focusing on products with Weibull lifetimes under progressive hybrid censoring. Our approach utilizes a graphical method to develop and evaluate efficient test designs, balancing cost with factors such as trace, determinant, and variance of the inverse Fisher information matrix (FIM). We provide an illustrative example to demonstrate the advantages of compound optimal designs over traditional constraint‐based methods and use sensitivity analysis to assess design robustness and limitations. Additionally, we analyze real‐life data to validate the practical applicability of our approach, showcasing its effectiveness in optimizing life‐testing experiments.