.png){kind=icon}
.png){kind=icon}
Return
A self-optimised informer-based model for nonlinear ship roll motion forecasting
.png){kind=icon}
PRE
AI
DOI：10.1080/23249935.2025.2606152.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Ship motion exhibits strong nonlinearity due to the combined effects of wind, waves, and currents, which brings difficulties to ship motion forecasting. To overcome the above challenges, this paper attempts to propose a novel nonlinear forecasting approach for ship motion. Firstly, focusing on the nonlinear characteristics of ship motion, an instantaneous input data encoder is designed, and an Instantaneous-Informer (I-Informer) is established based on the Informer network. Then, targeting the hyper-parameter optimisation problem of the I-Informer and focusing on the inherent defects of the Pelican Optimisation Algorithm (POA), a new quantum POA (QPOA) is proposed to construct a hyper-parameter optimisation method for the I-Informer. Finally, based on the established I-Informer and the proposed QPOA, a novel ship motion forecasting approach is proposed, namely I-Informer_QPOA. Subsequently, rolling data of a C11 container ship in irregular waves is used to test the performance of the new approach. Through example analysis, this paper discusses the recommended value ranges of output length, input and output ratio, and hyper-parameter for the I-Informer, and conducts comparative tests within the recommended value ranges. The experimental results demonstrate that the established I-Informer can obtain more accurate forecasting results than the model selected in this paper. In addition, the proposed QPOA can provide a better and more stable hyper-parameter combination for the I-Informer compared with the algorithm selected in this paper.
Keywords:
Ship motion forecasting
Nonlinear dynamical system
Deep learning model
Informer
Pelican optimisation algorithm
Journal
IF:
3.1
Papers:
927
Citations:
2.2K