An improved neural network method based on the differential evolution algorithm to predict the tidal level in the tidal power station

Numerous methods have been used to extract energy from the sea. The character of tidal level makes it possible to predict. In order to obtain the energy more effectively, a highly precious method is demanded to predict the tidal level. In this paper, a hybrid method which combines the traditional neural network method and the differential evolution algorithm is proposed. By using this method, the ability of the global search are improved and the optimal weight values and the threshold values can be calculated through the neural network algorithm. These approximate optimal values obtained from the differential evolution algorithm are trained in the neural network, thus a more accurate solution can be achieved. In the end of this issue, the tidal prediction model is verified by the tidal data of the Qingdao harbor. The experimental results prove the effectiveness of the proposed algorithm.