基于混合遗传神经网络的百米跑成绩预测方法  

Prediction Method of 100Sprint Performance Based on Hybrid Genetic Neural Network

作　　者：陈海英[1] 郭巧[1] 徐力[1] 

机构地区：[1]北京理工大学机器人研究中心,北京100081

出　　处：《计算机仿真》2004年第2期89-91,共3页Computer Simulation

基　　金：国家自然科学基金资助项目(60171018)

摘　　要：在遗传算法(GeneticAlgorithm)与BP(BackPropagation)网络结构模型相结合的基础上,设计了用遗传算法训练神经网络权重的新方法,并把这种方法用于运动员百米跑成绩预测。与BP算法和LM(LevenbergMarquardt)算法相比,基于混合遗传算法的神经网络不仅有较快的学习速度和较好的学习精度,而且网络的泛化能力(GeneralizationAbility)得到了很大提高。Based upon the combination of Genetic Algorithm and BP Neural Network, this paper puts forward a new approach to learn the weights of BP using GA, and applies this method firstly to predict 100m sprint performance of athletes. Compared with BP algorithm and LM algorithm, the neural network based on hybrid GA not only has quick convergence speed and better learning precision, but also the generalization ability of network has improved greatly.

关 键 词：百米跑 成绩预测 混合遗传神经网络 遗传算法 人体运动能力 权值优化 

分 类 号：G822.1[文化科学—体育训练] TP183[文化科学—体育学]