摘要
编制预制构件生产计划时,生产过程中一些不可控因素会导致理论工时与实际工时存在偏差,进而使生产计划在执行时不能有效指导实际生产。为解决该问题,提出一种基于动态集成学习的预制构件加工工时预测方法。首先,采用长短期记忆深度学习模型作为基学习器,引入密度峰值聚类方法将训练集样本划分为多个子集合,并训练基学习器;然后,依据K近邻算法找出待测样本的近邻样本,根据其占子集合的比例,动态赋予各基学习器相应权值,实现多个基学习器联合预测;最后,通过预制构件的实际生产数据对所提方法进行仿真验证。结果表明,所提方法可以有效提升预制构件加工工时的预测准确率。
When preparing the production plan of prefabricated components,some uncontrollable factors in the production process will lead to the deviations between the theoretical and actual machining hours,which will make the production plan cannot effectively guide the actual production.In order to solve this problem,a prediction method for machining hours of prefabricated components based on dynamic ensemble learning is proposed.Firstly,the long short-term memory depth learning model is used as the base learner,and the density peak clustering method is introduced to divide the training set samples into multiple subsets,and the base learner is trained.Then,based on the K-nearest neighbor algorithm,the nearest neighbor samples of the test sample are found out.According to their proportions in the subsets,corresponding weights are dynamically assigned to each base learner to achieve joint prediction of multiple base learners.Finally,the proposed method is simulated and validated by using actual production data of prefabricated components.The results show that the proposed method can effectively improve the prediction accuracy of the machining hours of prefabricated components.
作者
韩忠华
张文缤
李曼
孙亮亮
HAN Zhonghua;ZHANG Wenbin;LI Man;SUN Liangliang(School of Electrical and Control Engineering,Shenyang Jianzhu University,Shenyang 110168,China;Department of Digital Factory,Shenyang Institute of Automation,Chinese Academy of Sciences,Shenyang 110016,China;Yatai Group Shenyang Modern Building Industry Co.,Ltd.,Shenyang 110136,China)
出处
《控制工程》
CSCD
北大核心
2023年第7期1338-1345,共8页
Control Engineering of China
基金
国家自然科学基金资助项目(61873174)
辽宁省科技厅应用基础研究计划资助项目(2022JH2/101300253)
辽宁省重点研发计划资助项目(2020JH2/10100039)
辽宁省教育厅高等学校基本科研重点项目(LJKZ0583)
关键词
预制构件
生产计划
动态集成学习
K近邻算法
Prefabricated component
production plan
dynamic ensemble learning
K-nearest neighbor algorithm
