In a social system or production line,the restrictions of the cost and the due-time exist in each period.Generally,whether these restrictions are satisfied is dependent not only on the risks of this period,but also on...In a social system or production line,the restrictions of the cost and the due-time exist in each period.Generally,whether these restrictions are satisfied is dependent not only on the risks of this period,but also on the risks generated beforehand.We consider controlling the production process by switching the processing rate to a faster one at a given period.This paper deals with the optimal switching period to minimize the total expected cost of the production process.We first propose the optimal switching period model,and then the mathematic formulation of the total expectation is presented.Finally,the policy of optimal switching period is investigated in detail by numerical experiments.展开更多
为了推动大数据技术在制造车间的应用,针对复杂产品晶圆制造过程中海量制造数据时序性、强噪音影响加工周期预测精度的问题,提出考虑特征学习的改进粒子群优化长短期记忆网络(improved particle swarm optimization-long short term mem...为了推动大数据技术在制造车间的应用,针对复杂产品晶圆制造过程中海量制造数据时序性、强噪音影响加工周期预测精度的问题,提出考虑特征学习的改进粒子群优化长短期记忆网络(improved particle swarm optimization-long short term memory,IPSO-LSTM)的加工周期预测方法。采用降噪自编码器和稀疏自编码器联合构建深度自编码器,增强特征学习能力和抗噪能力;运用IPSO优化LSTM参数,克服时间依赖性,提升预测模型性能。实例验证了所提方法的预测精度优于传统机器学习方法,其平均绝对误差低于3%;并分析特征学习方法的有效性,将支持向量回归和多层感知器等传统方法加入特征学习方法,R^(2)分别提高了1.46%、1.05%,为晶圆加工周期的有效预测提供新的解决方法。展开更多
基金Project partially supported by the Ministry of Education,Science,Sports and Culture,and a Grant-in-Aid for Scientific Research (C),20510130 in 2008,Japan
文摘In a social system or production line,the restrictions of the cost and the due-time exist in each period.Generally,whether these restrictions are satisfied is dependent not only on the risks of this period,but also on the risks generated beforehand.We consider controlling the production process by switching the processing rate to a faster one at a given period.This paper deals with the optimal switching period to minimize the total expected cost of the production process.We first propose the optimal switching period model,and then the mathematic formulation of the total expectation is presented.Finally,the policy of optimal switching period is investigated in detail by numerical experiments.
文摘为了推动大数据技术在制造车间的应用,针对复杂产品晶圆制造过程中海量制造数据时序性、强噪音影响加工周期预测精度的问题,提出考虑特征学习的改进粒子群优化长短期记忆网络(improved particle swarm optimization-long short term memory,IPSO-LSTM)的加工周期预测方法。采用降噪自编码器和稀疏自编码器联合构建深度自编码器,增强特征学习能力和抗噪能力;运用IPSO优化LSTM参数,克服时间依赖性,提升预测模型性能。实例验证了所提方法的预测精度优于传统机器学习方法,其平均绝对误差低于3%;并分析特征学习方法的有效性,将支持向量回归和多层感知器等传统方法加入特征学习方法,R^(2)分别提高了1.46%、1.05%,为晶圆加工周期的有效预测提供新的解决方法。
