摘要
为解决非线性对象的预测控制问题,提出基于混沌神经网络的控制方法。将海洋经济作为非线性对象,建立混沌神经网络模型的基本结构,对其时间序列进行相空间重构,利用信息熵及假近邻法确定时间延迟及最佳嵌入维数。进而确定非线性对象的混沌特性和混沌神经网络的拓扑结构。仿真中建立测试样本、实验样本,对比预测值及真实值,仿真结果证明了该方法的有效性,且混沌神经网络训练时不会陷入局部极小值的前提下,减少训练时间,保证预测精度。
Taking the nonlinear system, the basic structure of the chaotic neural network model is proposed. The time series for phase space reconstruction. Using the information entropy and the false neighbor method the time delay and the best embedding dimension are determined. The nonlinear chaotic characteristics of objects are determined. And a chaotic neural network topology structure is established, making the chaos neural network training under the premise of will not get into local minimum value. Reducing the training time, guarantee the accuracy. Comparing predicted values and real values. The simulation results prove the effectiveness of the proposed method.
引文
[1]李克平.混沌神经网络和非线性时间序列预测[D].天津:南开大学,2002
[2]AIHARA K,TAKABE T,TOYODA M.Chaotic Neural Networks[J].Physics Letter-A,1990,144(6-7):333-340
[3]EARL C.JOSEPH.Chao&Postmodernism Forecasting&Futuring Insights[J].Fururics,1994,18(6):3-4
[4]王耀南,余群明,袁小芳.混沌神经网络模型及其应用研究综述[J].控制与决策,2006,21(1):111-128
[5]刘华杰.论浑沌对决策论的影响[J].自然辩证法研究,1994,10(11):11-12
[6]蒋夕平,吴凤凰.混沌理论和神经网络的农业经济预测仿真研究[J].计算机仿真,2013,30(4):394-397
[7]黄登仕,李后墙.非线性经济学的理论和方法[M].成都:四川大学出版社,1993