Genetically optimized prediction of remaining useful life

The application of remaining useful life (RUL) prediction is very important in terms energy optimization, cost-effectiveness, and risk mitigation. existing RUL algorithms mostly constitute deep learning frameworks. In this paper, we implement LSTM GRU models compare the obtained results with a proposed genetically trained neural network. current solely depend on ADAM SGD for optimization learning. Although have worked well these optimizers, even little uncertainties prognostics can result huge losses. We hope to improve consistency predictions by adding another layer using Genetic Algorithms. hyper-parameters – rate batch size are optimized beyond manual capacity. These architecture tested NASA Turbofan Jet Engine dataset. predict given autonomously provide superior results.