基于思维进化算法优化的BP神经网络对蒸发波导高度的预测

Prediction of evaporation duct height based on BP neural network optimized by mind evolutionary algorithm

预先掌握海上的波导高度参数对海上作战有重要意义,然而准确探测或预测波导参数十分困难。本文使用南海铁塔波导探测平台的实测数据与ERA-Interim公开数据集,提取所需相关的水文气象要素的点数据,利用NPS模型计算波导高度值,组建数据集;之后使用MEA优化的BP神经网络模型预测波导高度值。优化后的BP神经网络可避免陷入局部极小值点,预测结果有较高的准确率,且MEA在求最优个体时有比较快的收敛速度,为水平均匀的蒸发波导高度预测提供了一种方法。

It is very important to predict the evaporation duct height parameters at sea,but it is difficult to detect or predict duct parameters accurately.In this paper,measured data from evaporation duct detection platforms in the South China Sea and the ERA-Interim data set were used to extract point data of relevant hydrological and meteorological elements.The NPS model was used to calculate the height of evaporation duct to form the data set.The model of MEA-BP was then used to predict the height of the evaporation duct.The optimized BP neural network can avoid falling into the local minimum point,and the prediction results have a high accuracy rate.Moreover,MEA has a relatively fast convergence rate when calculating the optimal individual.Finally,this paper provides a method for predicting the height of horizontally uniform evaporative duct.