A deep reinforcement learning method for structural dominant failure modes searching based on self-play strategy.
Xiaoshu GuanHuabin SunRongrong HouYang XuYuequan BaoHui LiPublished in: Reliab. Eng. Syst. Saf. (2023)
Keyphrases
- high precision
- detection method
- dynamic programming
- computational cost
- search strategy
- synthetic data
- clustering method
- significant improvement
- optimization algorithm
- high accuracy
- failure modes
- optimization method
- segmentation method
- classification accuracy
- experimental evaluation
- prior knowledge
- artificial neural networks
- multi agent
- detection algorithm
- evolutionary algorithm
- preprocessing
- classification method
- reinforcement learning
- convergence rate
- machine learning
- selection strategy
- neural network