用改进的遗传神经网络预报冷连轧轧机的轧制压力

为了充分利用ＢＰ神经网络的函数逼近能力并克服其易陷入局部极值点的主要缺陷，提出了利用遗传算法（ＧＡ）和ＢＰ算法联合进行网络权值优化调整的思想，系统具有遗传算法（ＭＧＡ）和ＢＰ网络有机结合建立了ＭＧＡ－ＢＰ网络系统。该网络系统具有遗传算法性能参数优选、网络结构参数优选、网络性能参数优选以及ＧＡ－ＢＰ自满联合进行网络权值修改几种功能，利用ＭＧＡ－ＢＰ网络进行了带钢冷连轧机轧制压力预报，结果良好。

PREDICTION OF ROLLING PRESSURE OF CONTINUOUS COLD ROLLING MILL BY MODIFIED GENETIC NEURAL NETWORK

Synopsis To allow full play to the approximation capability of the function of BP neural network and overcome the main shortcomings of its liability to fall into the local extreme value the present paper proposed a concept of applying the genetic algorithm and BP algorithm jointly to optimize the network value and incorporating the modified genetic algorithm (MGA) into BP network to establish a MGA-BP network system. The new network system functions the optimization seeking for parameters of the genetic algorithm performance, network structure and network performance and revision of the network value with GA-BP algorithm. Satisfactory results have been yielded in application of MGA-BP network in prediction of the rolling pressure of the continuous cold rolling mill.