Comprehensive Optimization Training Framework for TSK Fuzzy Systems: Applications of MBGD, AMPO, and RDA

In this study, a method combining advanced machine learning techniques to enhance the prediction accuracy of equipment remaining life is proposed. Given the task of predicting the remaining life of turbofan engines involves complex nonlinearities and high-dimensional data issues, Principal Component Analysis (PCA) is applied to reduce the dimensions of the data, extract key features, and decrease computational complexity. Subsequently, a training and prediction model is developed using an Adaptive Multi-population Optimization (AMPO) and Mini-batch Gradient Descent (MBGD) strategy combined with a regularization technique, DropRule, and AdaBound (termed RDA) in a type-1 Takagi-Sugeno-Kang (TSK) fuzzy system. Through this integrated approach, a precise TSK fuzzy neural network prediction model is established, capable of effectively processing and predicting the remaining life of equipment. Finally, the model is validated using a dataset on the remaining life of turbofan engines, and the results demonstrate that the model achieves high accuracy and reliability in predicting equipment remaining life. This research not only enhances prediction accuracy but also provides robust decision support for the maintenance and management of complex equipment like turbofan engines.