帆船VMG预测遗传算法优化  

VMG Forecasting of Sailboat by Optimizing BP Neural Network Based on Genetic Algorithm

作　　者：潘明杰[1] 任久春[1] 汪建波[1] 朱谦[1] 

机构地区：[1]复旦大学信息科学与工程学院,上海200433

出　　处：《微型电脑应用》2017年第3期12-15,共4页Microcomputer Applications

基　　金：上海市科研计划项目(14231202102)

摘　　要：VMG(Velocity Made Good)泛指帆船在迎风阶段船速于迎风方向上的分量,反映了帆船运动员利用风的能力。在前期研究中通过设计BP神经网络模型对帆船VMG进行预测。为提高BP神经网络预测的准确性,提出了采用遗传算法对原网络模型的权值及阈值进行优化,并对网络重新进行学习训练。对比结果表明,使用遗传算法优化后的BP模型在多项指标上都有了明显提高。VMG（Velocity Made Good） means the projection of sailboat speed in the direction of true wind, and its value shows athletes＇ ability in sailing against wind as well as ability in making use of wind. We have designed a back propagation model to predict VMG to predict the VMG. In order to increase the accuracy of the prediction, we take genetic algorithm to optimize the weights and threshold of the BP network. The result shows that the genetic algorithm can make significant improvement in many aspects of the BP network.

关 键 词：VMG 时间序列预测 遗传算法 BP算法 神经网络 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]