摘要
将模糊逻辑系统、ＧＡ算法与ＢＰ算法相结合，形成一种有效网络学习方法。提出了基于神经网络的模糊逻辑系统辨识器的结构形式，利用自适应交叉率、变异率ＧＡ算法对其进行参数训练。通过计算机仿真和加氢裂化装置航煤干点模型预测，表明该方法的可行性和实用性，可望为石油化工模型预测、优化操作提供一种辅助性有效手段。
An effective network learning method is formed by combining GA,BP algorithms with fuzzy logic system. The structure of fuzzy logic system predicator based on neural network is proposed. Its parameter training is carried out by using mutation GA algorithm and self-tunning adaptivity aross. Computer simulation and practical model predication for hydrocracking have shown the feasibility and practability of this method. This method is expected to provide an effective and helpful means for optimizing operation and model predication in petrochemical industry.
出处
《石油化工高等学校学报》
EI
CAS
1998年第3期41-44,共4页
Journal of Petrochemical Universities
基金
国家攀登计划(神经网络模型)项目
广东省科学基金
关键词
神经网络
GA算法
模型辨识
学习系统
柴油
生产
Neural networks
Fuzzy logic system
Genetic algorithms
Model prediction
