大脑神经网络认识的开发性研究:基于混合神经网络与遗传算法方法的注塑参数优化系统(英文)  

Development of cognitive function of cerebral neural network: an optimization system for injection molding parameters based on hybrid neural network and genet ic algorithm

作　　者：郑生荣[1] 陈敏 

机构地区：[1]南昌工程学院机电工程系,江西省南昌市330099 [2]南昌铁路中心医院干部病房,江西省南昌市330003

出　　处：《中国临床康复》2005年第1期209-213,共5页Chinese Journal of Clinical Rehabilitation

摘　　要：背景:塑料注塑成型技术可用于康复工程的多种实用矫形器中,遗传算法所表现出的易于实现及健壮性等特点,使它在许多领域,特别是近年来在机器学习、模式识别、智能控制和最优化等领域得到了广泛应用。目的:建立基于混合神经网络与遗传算法方法的注塑工艺参数优化系统。设计:开发性研究。单位:南昌工程学院机电工程系及南昌铁路中心医院干部病房。方法:用Matlab语言编制了应用程序,对神经网络的参数预测与遗传算法的优化过程进行求解。将网络预测结果与CAE(computeraidedengi-neering)模拟结果进行比较和误差分析,显示出反向传播网络的稳定性和可靠性。结果:优化结果经CAE模拟和实验验证,证明是正确的,结论:基于混合神经网络与遗传算法方法的注塑工艺参数优化方法是可行的。该优化系统具有工程实用价值。BACKGROUND:Injection molding can be used in producing a variety of orthotic de vices of practical use in rehabilitation medicine. With its merits of convenient realization and stability,genetic algorism(GA) has found wide application in ma ny fields,especially in machine learning,pattern recognition, intelligent contro l and optimization. OBJECTIVE:To establish an optimization system for injection molding parameters based on hybrid neural network and genetic algorithm. DESIGN:An open trial. SETTING:Department of Mechanical and Electrical Engineering, Nanchang Institut e of Technology;Ward for Cadres,Nanchang Railway Central Hospital. METHODS:The a pplication program was developed in Matlab language and used in the parameter pr ediction with the neural network and genetic algorithm optimization.The comparis on and error analysis were made between the neural network predicted results an d those generated by simulation with computer aided engineering (CAE),which show ed that the back propagation network was stable and reliable. RESULTS:The results of optimization,after verified by CAE simulation and teste d by experiment,are proved to be correct. CONCLUSION:It is feasible to optimize the injection molding parameters based o n hybrid neural network and genetic algorithm approach.The optimization system i

关 键 词：神经网络 模型 遗传学 脑 

分 类 号：R74[医药卫生—神经病学与精神病学]