利用前馈神经网络计算Lyapunov指数

Computing Lyapunov exponents with feedforward neural networks

Ｌｙａｐｕｎｏｖ指数是混沌动力学系统的重要不变量．用基于前馈神经网络的自适应快速学习算法计算混沌动力学系统的全部Ｌｙａｐｕｎｏｖ指数．结果表明，本算法在有限的样本点及出现外部噪声的情况下，是非常可靠的．

Lyapunov exponents are the important invariants of a chaotic dynamical system.Based on the fast learning algorithm of the feedforward neural networks with the adaptive variable stepsize,all Lyapunov exponents of a chaotic dynamical system are calculated.The results indicate this algorithm is very reliable in the case of limited sample points with the presence of external noise.