A Hybrid Wavelet GP Model for Enhancing Forecasting Accuracy of Time Series Significant Wave Heights
A Hybrid Wavelet GP Model for Enhancing Forecasting Accuracy of Time Series Significant Wave Heights
INTRODUCTION Wind waves are very complex in nature. The representation of a wave field is normally done by significant wave height and significant time period which retains much of the insight gained from theoretical studies. Wave prediction is the prediction of wave parameters based on the meteorological and oceanographic data. Wave forecasting is extremely useful in the planning and maintenance of the marine activities. The representation of a wave field by significant height and period has the advantages of retaining much of the insight gained from theoretical studies. Its value has been demonstrated in the solution of many engineering problems. A significant wave height is defined as the average height of the one-third highest waves and it is about equal to the average height of the waves as estimated by an experienced observer. During recent decades, some black-box models have been applied to simulate the wave and the wave heights.
Although the ANNs are useful tools in the time series wave modeling, the obvious disadvantage of the ANNs is that they represent their knowledge in term of a weight matrix that is not accessible to human understanding at present (Savic et al, 1999); in other words, these types of models are so implicit that they cannot be simply used by other investigator.
Therefore, it is still necessary to develop an explicit model for overcoming this problem (Aytek and Kisi, 2008). From this point of view, genetic programming (GP), which is an evolutionary computing method that provides transparent and structured system identification, has been developed (Savic et al, 1999).
Genetic programming has been successfully applied to problems that are complex and nonlinear and where the size, shape, and overall form of the solutions are not explicitly known in advance (Whigham and Crapper, 2001). It also partially alleviates the problem necessary for conceptual model calibration. The state-of-the art applications of the GP in civil engineering have been listed by Shaw et al, (2004).