In-service evaluation method of naval degaussing system optimized by BP-SA hybrid learning strategy
Equipment in-service evaluation is an important means to verify the operational and support effectiveness after the equipment is commissioned,and to promote the iterative upgrading of the equipment.Considering the characteristics of naval degaussing system in service evaluation and the shortcomings of conventional evaluation methods,the in-service evaluation index system of naval degaussing system is established from five aspects:combat effectiveness,applicability,reliability,maintainability,and testability.Based on the conventional back-propagation neural network,the simulated annealing learning strategy is introduced to improve the convergence and stability of the neural network,which is used to randomly find the optimal solution.After training,testing and validating 70 data samples,the evaluation model finally obtained the experimental result of 9.309 3 residual prediction residual,which indicates that the model not only overcomes the problems of the traditional BP neural network algorithm such as local minimization and poor fitting effect,but also has a good ability to predict and evaluate the in-service evaluation results of shipboard degaussing system.
万方数据
维普
