Oscillation Series Gray Model-based Ship Motion Prediction

This paper focus on prediction of ship motion, studied the characteristics of ship motion, vibration sequence for the forecast result is not an ideal situation, for processing, so that transformation is monotonically increasing sequence, then the model predicts reach to the desired results. The specific applicable theory and transform as follows:1. The oscillation sequence is not regular, we must change the sequence into a monotone increasing sequence, so that it can be applied to GM model, we apply on the accelerated translation transformation and weighting mean value generating transformation dead down oscillation sequence randomicity theory, making the sequence suit to establish GM model prediction; And has made the proof with the theory of probability knowledge to weighting mean value generating's nature, enhance the feasibility and correctness of the modeling2. In modeling the ship motion, curve fitting and linear fitting are used for the transformation data of ship motion. Carries on the curve and the linear fitting using the least squares method principle, we proved the fact that the oscillation sequence can be transformed into a linear sequence, not only reduced the difficulty of forecasting, but also reduced the workload;3. GM(1,1) model and the GM(0,1) model are used for the monotone transformation sequences to establish the simulation model, and then predict the data. The results show that this prediction method is more suitable to apply GM (0,1) model. Meanwhile, as a result of the traditional thought using GM (1,1) the model firstly, forecast implies that the result is not ideal. After study on the results, we find the fact that the transformation sequence have the obvious property of linear feature. Therefore, we adopt the GM (0,1) model to predict the ship motion. Finally, better results are obtained.