Deterministic and probabilistic ship pitch prediction using a multi-predictor integration model based on hybrid data preprocessing, reinforcement learning and improved QRNN. (October 2022)

• Record Type: Journal Article
• Authors: Wei, Yunyu; Chen, Zezong; Zhao, Chen; Chen, Xi; Yang, Rui; He, Jiangheng; Zhang, Chunyang; Wu, Sitao
• Is Part Of: Advanced engineering informatics. Volume 54(2022)
• Publication Date: 2022-10
• Subjects: Multi-predictor integration -- Hybrid data preprocessing -- Reinforcement learning -- Quantile regression neural network -- Deterministic and probabilistic prediction -- Ship pitch motion; AR autoregressive -- GRU gated recurrent unit -- ORELM outlier-robust extreme learning machine -- BiLSTM bidirectional long short-term memory -- EMD-SVR empirical mode decomposition-support vector regression -- EMD empirical mode decomposition -- SVR support vector regression -- EWT empirical wavelet transform -- SSA singular spectrum analysis -- GWO grey wolf optimizer -- ENN Elman neural network -- BPNN back propagation neural network -- GRNN generalized regression neural network -- WNN wavelet neural network -- QR quantile regression -- QRNN quantile regression neural network -- PSO particle swarm optimization -- WHO wild horse optimizer -- SEGOBQWQ SSA-EWT-GRU-ORELM-BiLSTM-Q-learning-WHO-QRNN -- SVD singular value decomposition -- AM-FM amplitude modulation-frequency modulation -- ELM extreme learning machine -- LSTM long short-term memory -- MPICD Mean Prediction Interval Centre Deviation -- SampEn sample entropy -- MAE mean absolute error -- MAPE mean absolute percentage error -- RMSE root mean square error -- PMAE Promoting percentages of the MAE -- PMAPE Promoting percentages of the MAPE -- PRMSE Promoting percentages of the RMSE -- PICP Prediction Interval Coverage Probability -- PINAW Prediction Interval Normalized Average Width -- CWC Coverage Width-based Criterion -- GOBQ GRU-ORELM-BiLSTM-Q-learning -- SGOBQ SSA-GRU-ORELM-BiLSTM-Q-learning -- EGOBQ EWT-GRU-ORELM-BiLSTM-Q-learning -- SEGOBQ SSA-EWT-GRU-ORELM-BiLSTM-Q-learning -- WT wavelet transform -- SVM support vector machine -- RBFNN radial basis function neural network -- WSBGG WT-SVM-BP-GRU-GWO -- DRLOQ EMD-RBFNN-LSTM-ORELM-Q-learning -- SWBGBQ SSA-WT-BP-GRU-BiLSTM-Q-learning -- SEGLOG SSA-EWT-GRU-LSTM-ORELM-GWO -- SDGOBQ SSA-EMD-GRU-ORELM-BiLSTM-Q-learning -- CLs confidence levels; Computer-aided engineering -- Periodicals; Engineering -- Data processing -- Periodicals; 620.00285
• DOI: 10.1016/j.aei.2022.101806
• Languages: English
• ISSNs: 1474-0346

Abstract

Highlights: Propose a deterministic and probabilistic ship pitch forecasting system. Develop a denoising-decomposition data preprocessing strategy. Reinforcement learning are used to achieve multi-predictor integrated prediction. Use the improved QRNN model for probabilistic prediction. Abstract: The deterministic and probabilistic prediction of ship motion is important for safe navigation and stable real-time operational control of ships at sea. However, the volatility and randomness of ship motion, the non-adaptive nature of single predictors and the poor coverage of quantile regression pose serious challenges to uncertainty prediction, making research in this field limited. In this paper, a multi-predictor integration model based on hybrid data preprocessing, reinforcement learning and improved quantile regression neural network (QRNN) is proposed to explore the deterministic and probabilistic prediction of ship pitch motion. To validate the performance of the proposed multi-predictor integrated prediction model, an experimental study is conducted with three sets of actual ship longitudinal motions during sea trials in the South China Sea. The experimental results indicate that the root mean square errors (RMSEs) of the proposed model of deterministic prediction are 0.0254°, 0.0359°, and 0.0188°, respectively. Taking series #2 as an example, the prediction interval coverage probabilities (PICPs) of the proposed model of probability predictions at 90%, 95%, and 99% confidence levels (CLs) are 0.9400, 0.9800, and 1.0000, respectively. This study signifies that the proposed model can provide trusted deterministic predictions and can effectively quantify the uncertainty of ship pitch motion, which has the potential to provide practical support for ship early warning systems.