基于神经网络的储粮测控专家系统的研究与应用  

RESEARCH AND APPLICATION OF MEASUREMENT AND CONTROL EXPERT SYSTEM FOR GRAIN STORAGE BASED ON NEURAL NETWORK

作　　者：张兴红[1] 甄彤[1] 包晖[1] 丁伟[1] 张俊兰[2] 

机构地区：[1]河南工业大学信息科学与工程学院,河南郑州450001 [2]中央储备粮曹县直属库,山东曹县274400

出　　处：《河南工业大学学报（自然科学版）》2010年第3期76-79,共4页Journal of Henan University of Technology：Natural Science Edition

基　　金："十一五"国家科技支撑计划(2006BAD08B01);河南工业大学校科研基金资助(08XPT003)

摘　　要：将神经网络应用于专家系统中,利用神经网络系统的学习功能解决传统专家系统在知识获取、推理能力、自学习能力上的缺陷,与粮情测控系统相结合,提出了一种改进BP算法.该算法加入动量项,然后对自适应算法进行了一些适当改进,进而给出了该算法的理论推导和具体实现步骤.与未改进的BP算法相比,该方法协调了训练次数多而引起的学习效率和收敛速度之间的矛盾,提高了BP算法的熟练速度和收敛速度.实验结果也说明了该方法的快速性、有效性、稳定性.The paper applied neural network into an expert system to overcome the disadvantages of the conventional expert system on knowledge acquisition,reasoning ability and self-learning ability based on the learning function of the neural network.By combining the neural network with the grain measurement and control system,the paper put forward an improved BP algorithm.By introducing a momentum term into the algorithm and improving the self-adaptive algorithm,the paper provided the theory derivation and the specific steps of the algorithm.Compared with the unimproved BP algorithm,the method solved the contradiction between the learning efficiency and the convergence speed caused by the frequent training so as to improve the skilled speed and the convergence speed of the BP algorithm.The results show that the method has high speed,high validity and high stability.

关 键 词：神经网络 BP算法 专家系统 储粮测控 

分 类 号：TS210[轻工技术与工程—粮食、油脂及植物蛋白工程] TP18[轻工技术与工程—食品科学与工程]