基于神经网络的机械传动方案决策系统的建模与仿真  

Modeling and Simulation of Mechanical Drive Type Decision System Using Neural Network

作　　者：孙华丽[1] 谢剑英[1] 温正忠[2] 陈建军[2] 

机构地区：[1]上海交通大学自动化系,上海200030 [2]西安电子科技大学机电工程学院,陕西西安710071

出　　处：《计算机仿真》2005年第9期146-149,共4页Computer Simulation

基　　金：国家863计划(2001AA413110);国家自然科学基金项目(60174024)

摘　　要：针对机械传动方案决策系统难以建立准确数学模型的特点,提出了利用BP及BP改进神经网络的非线性映射能力,进行机械传动方案决策系统模型辨识建模的新方法。介于网络的输入输出数据难以精确描述的特点,运用模糊理论对其进行了处理。通过仿真实验验证了利用BP及BP改进网络建立其决策模型的可行性,表明了BP改进算法收敛速度更快,模型更准确,从而为机械系统传动方案的决策提供了一种更为有效的推理模型。The paper analyzes the difficulties of the mathematical modeling of mechanical drive type decision system. A novel method is proposed that uses BP and improved BP network to identify the mathematical model of mechanical drive type decision system. The input and output data are dealt with by using fuzzy theory. The possibility of modeling of mechanical drive type decision system using BP and improved BP network is verified by the sirrmlation experiment. The results show that the converge speed of the improved algorithm is faster. A better reasoning model is presented for the decision of mechanical drive type.

关 键 词：机械传动系统 系统辨识 特征因素集 模糊逻辑 建模 

分 类 号：TP271.2[自动化与计算机技术—检测技术与自动化装置] TP29[自动化与计算机技术—控制科学与工程]