Research on Failure Prognostics Method of Electronic System Based on Improved Fruit Fly Algorithm and Grey Fast Relevance Vector Machine

Abstract

In order to solve the failure prognostics problem of electronic system, a method of fast relevance vector machine (FRVM) based on improved fruit fly optimization algorithm (FOA) is proposed. Grey data generation operation is introduced to process the original data and the output data for enhancing the regularity and reducing the randomness. Furthermore, the kernel function parameter of FRVM model is optimized based on the improved FOA which adds the annealing parameter to establish the prediction model. In addition, the performance of the proposed model is studied and evaluated by a radar transmitter fault prediction experiment in this paper. The results demonstrate that the presented method performs significantly better than the traditional methods in terms of global optimization, convergence speed, training time and prediction accuracy.