| Artificial Neural Network which have strong local search ability can impresscomplex nonlinear relationship and shows great advantage in solving complexPractical problems like Process Control, fault diagnosisa and System Discrimination.But the convergence rate of the Algorithm is slow and the learning efficiency is low.Genetic Algorithm is a global optimization algorithm. HGA(Hierarchic GeneticAlgorithm) is one of many improved algorithm. We combine BP Algorithm and HGAtogether to train the weight of neural network. The application of optimization ofneural network based on Genetic Algorithm have made great progress, it enhance theglobal search ability of neural network.Fencing is a hold equipment directly to fight class project. We use theinternational tournament data from the information platform of national fencing teamand figured out the nonlinear relationship between fencing winning rate and the maindata of technical and tactical. We build a Diagnostic Model of fencing competition byusing genetic algorithm optimize BP neural network. According to this model, we cananalyse the influence to the athlete and express the result through graphs, then drawup a programme for the single athlete. It shows great significance in sports area, andcan give other teams a example.we first introduce the basic theory,application and shortage of Simple GeneticAlgorithm, and then study the principle and process of Hierarchical Neural Network.Then introduce the basic theory of ANN, and emphasize BP Neural network Which isrelative perfact in Neural network area. Neural network can not only dig out the causeand effect implied in samples, but impress complex nonlinear relationship. But theconvergence rate is slow and the learning efficiency is low. According to its shortcoming,we introduce a model of BP neural network prediction based onHGA(Hierarchic Genetic Algorithm), we combine these two theories together,it has agreat application in this paper. |
PDF Full Text Request