一种模糊神经网络自适应预测控制方案的研究

针对现有模糊推理方法的缺点，结合神经网络的学习能力和插值能力，提出了一种新型的模糊神经网络结构。在此基础上设计了一种模糊神经网络自适应预测控制方案，并导出了相应的学习算法。同时引入了预测误差的智能补偿，以提高预测及控制精度。仿真实验表明，该算法能实现模糊控制和神经网络控制的优势互补，在非线性复杂系统的控制方面具备较高的性能。

Investigation on a Self-Adaptive Predictive Control Program Based on a Fuzzy Neural Network

In this paper, aself adptive controlalgorithm based on fuzzy neuralnetwork hasbeen developed, and a new learning algorithm is derived theoretically. Based on the error com pensation m ethod, the accuracy ofthe algo rithm is im proved. Experim entresultsshow thatthe algorithm can obtain the advantagesofthefuzzy system and the neuralnetw ork and can have better perform ance in controlling the nonlinear and com plex system .