BP神经网络在配方模拟评估系统中的应用  

Application on BP Neural Network in Formulating Simulated Evaluation System

作　　者：王涛[1] 

机构地区：[1]中国海洋大学

出　　处：《微计算机信息》2009年第18期166-167,184,共3页Control & Automation

摘　　要：人工神经网络是一种并行和分布式的信息处理网络,具有对外部环境自适应学习的能力,有很强的容错性和鲁棒性,善于联想、综合和推广,近十几年来在工程技术领域得到广泛应用,相关理论进而得到迅猛发展。卷烟配方涉及的理化、感官、烟气指标及加工工艺非常复杂,是一个典型的不确定的、非线性问题。本文结合卷烟配方设计的业务需求,阐述了BP神经网络的基本理论及其在配方辅助设计和维护过程中的应用。The artificial neural network is one kind of concurrence and the distributed information processing network, it has the ability of adaptive learning from outside environment. The artificial neural network is so fault-tolerant and stability that is broad used in many application fields. The theories about it have been fast developed more than 10 years.Cigarette Formulating is typical uncertain and nonlinear problem because it involves complex data and technique. This thesis expatiates computational intelligence＇s basic principle and application in cigarette formulating.

关 键 词：BP神经网络 感官质量 烟气指标 模拟评估 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]