A Bayesian-based optimization approach for accelerated degradation test plan of a LED component with self-heating impact

Constant-stress Accelerated Degradation Testing (ADT) has become a common method to get degradation data of components/systems. The ADT design aims to build an optimal ADT plan under specific constraints, e.g., limited test resources. Bayesian optimal design is a method of decision theory under uncertainty, which uses historical data and expert information to find the optimal test plan. In this paper, we develop a Bayesian-based approach for optimizing ADT plan of a LED component considering the self-heating phenomenon. In that way, D-optimal is used as a main criterion to find the optimal ADT plan under limited resources. Experimental data from LED components demonstrate the proposed method. The obtained results highlight the robustness and performance of the proposed approach.