Modeling and Planning of Step-Stress Accelerated Life Tests With Independent Competing Risks

This paper presents a method for planning multiple-step step-stress ALT (SSALT) with statistically independent competing risks. A statistical model is constructed and asymptotically c-, D-, and Ds-optimal ALT plans are developed. Because the asymptotical optimality might sometimes lead to impractical ALT plans with finite sample size, additional constraints are introduced, and the planning of ALT on the constrained design space is also studied. Finally, we provide a numerical example to illustrate the application of the proposed method.