Optimal Design of Accelerated Life Test Plan for Test Standard of a Manufacturer Making Multi-Series Products

Abstract

To win fierce competition, a manufacturer would produce multi-series products to meet the diverse needs of customers. These products shall comply with the same test standard to estimate reliability before being launched into market. Extensive research has been conducted on the optimal design of accelerated life test (ALT) for a single type of product. However, these methods only utilize the information of a certain type of product and hence might not obtain an optimal plan for other products. This motivates us to study the planning of ALT for a manufacturer making multi-series products. Two criteria, i.e., the minimum sum and the minimum maximum of the asymptotic variance (Avar) of the life at a given percentile, are proposed to design the ALT plan optimally. The analytical solutions are derived for some special cases and numerical methods are provided for more general cases. We further explore the difference between test plans obtained by the two criteria in the cases of small and large number of product categories. Two numerical examples are presented to illustrate the application of the proposed method. The methodology proposed in this paper would be vital for a manufacturer to design its test standard.

Keywords

• Multi-series products
• accelerated life test
• asymptotic variance
• maximum likelihood estimation
• Fisher information matrix