An AEALSCE-based optimization method of sequential diagnostic strategy generation

As an important part of testability design and fault diagnosis, sequential fault diagnostic strategy generation plays a crucial role in improving the efficiency of system fault diagnosis and reducing the cost of system maintenance support. With the system complexity increasing, it is extremely difficult to generate an optimal diagnostic strategy. The existing methods are easy to fall into the local optimal solution and the time complexity is high. Therefore, this paper proposes a method based on a covariance matrix adaptation evolution strategy (CMA-ES) variant, named AEALSCE, to obtain the optimal strategy. Firstly, the fault diagnosis strategy problem model based on AEALSCE is established, and the fault diagnostic strategy problem is transformed into a continuous optimization problem to adapt the algorithm solving. Then, based on the diagnostic objective of the minimum diagnostic steps and test cost, the fitness function of the algorithm is constructed. At last, the optimal diagnosis strategy is obtained by solving the transformed continuous optimization problem with AEALSCE algorithm. The experimental results proves that the proposed AEALSCE-based approach is reasonable and feasible in obtaining the optimal diagnostic strategy, and outperforms other traditional methods and evolutionary algorithms in terms of accuracy and time efficiency.