Research on One-Dimensional Chaos Maps for Fuzzy Optimal Selection Neural Network

For solving the optimization problems with slow convergence speed and local optimum in fuzzy optimal selection neural network, this paper applies the chaos optimization algorithm by using a chaos variable from one-dimensional iterative map to optimize the network weight For selecting the reasonable chaos variable, multi one-dimensional chaos maps, such as Logistic Map, Sine Map, Cosine Map and Cubic Map, are researched and compared To verify feasibility of one-dimensional chaos map for fuzzy optimal selection neural network in the practical application, the case of Yamadu Hydrological Station located in Yili River for annual runoff forecast is analyzed and discussed The results show that the chaos optimization algorithm is an efficient learning algorithm which has the advantage of speed convergence and high precision for fuzzy optimal selection neural network.