基于遗传算法学习的一类多层神经网络

Learning a Class of Multi-Layer Neural Network Based on Genetic Algorithm

针对非线性、不确定性问题，采用了遗传算法和神经网络相结合的办法，成功地实现 了非线性模型的辨识问题及倒立摆的控制．仿真结果证明：遗传算法和神经网络的结合，可兼有 神经网络广泛映射能力和遗传算法快速全局收敛等性能．

To counter the problems of nonlinearity and uncertainties, the combining method of genetic algorithm and neural network is adopted, by which a nonlinear model can be identified and an inverted pendulum controlled. The simulation results show that extensive mapping ability of neaural network and rapid global convergence of genetic algorithm can be obtained by combining genetic algorithm and neural network.