基于VDM与APSO优化极限学习机的船舶运动姿态预报

为提高船舶在海上运动的耐波性与适航性,并为解决具有非线性、随机性和非平稳性特点的船舶运动姿态难以准确预测的问题,提出运用一种基于变分模态分解和自适应粒子群算法优化极限学习机的组合预测模型。该算法首先利用变分模态分解将船舶运动姿态序列分解为一系列限带内本征模态函数,并且变分模态分解可以避免经验模态分解技术所产生的模态混叠和端点效应,可以降低序列的非平稳性对预测精度的影响;然后对各模态分量分别建立极限学习机预测模型,并用改进的粒子群算法对极限学习机的初始权值和阈值进行优化;最后将各模态分量预测结果进行叠加,得到最终的船舶运动姿态预测值。通过模拟试验测试并与其他传统的预测方法进行比较,结果表明所建立的组合预测模型具有更高的预测精度。

Ship Motion Attitude Prediction Based on VDM and APSO Optimized Extreme Learning Machine

In order to improve the seakeeping and seaworthiness of ship motion at sea and to solve the problem of difficult to accurately predict ship motion attitude series with the characteristics of non-linearity, randomness and non-stationarity, a combined prediction model is proposed based on Variational Mode Decomposition (VMD) and Adaptive Particle Swarm Optimization (APSO) for extreme learning machine. Firstly, VMD is used to decompose the ship motion attitude sequence into a series of intrinsic mode functions in the limited band. Moreover, VMD can avoid the mode aliasing and endpoint effect caused by empirical mode decomposition technology, and reduce the influence of non-stationarity of the sequence on the prediction accuracy. Then, the prediction models of the extreme learning machine for each modal component are established separately, and the improved particle swarm optimization algorithm (APSO) is used to optimize the initial weights and thresholds of the extreme learning machine. Finally, superimposing the predicted results of each modal component to obtain the final predicted values of ship motion attitude. The results of simulation experiments show that the combined prediction model has higher prediction accuracy when compared with other traditional prediction methods.