基于神经网络的非线性离散动态系统DISOPE算法

提出了基于神经网络的非线性离散动态系统优化与参数估计集成算法。先用动态神经网络辩识系统得出一个近似模型，在存在模型实际差异的情况下，从神经网络模型出发通过迭代运算得到实际非线性离散动态系统的真实最优解，系统优化采用Ｈｏｐｆｉｅｌｄ 神经网络算法，求解速度快。仿真结果表明了算法的有效性和实用性。

Dynamic Integrated System Optimization and Parameter Estimation Based on Neural Network for Nonlinear Discrete-Time Dynamic System

In thispaper,A dynam icintegrated system optim ization and param eter estim ation algorithm based on neuralnetwork fornonlineardiscrete tim e dynam ic system isintroduced.AtFirst,an approach m odelcan be obtained by using dynam ic neuralnetw ork.Then, realexactoptim alsolution fornonlinear discrete tim e dynam ic system w ith m odel reality differences can beobtained from neuralnetw ork m odelby iterativesolutions.System optim ization isim plem ented by using Hopfield network which can solve rapidly.The efficiency and applicability of this algorithm are dem onstrated by the sim ulation results.