摘要
为了解决人工工时估算不准确且费时等问题,准确测算船舶铁舾件舾装作业的工时定额,提出了基于CBR和BP神经网络的铁舾件装配工时定额制定方法。针对铁舾件细分工种安装工艺相似但型号各异的特点,提出基于非数值型和数值型特征参数相结合的相似性规律,在已知部分型号舾装工时情况下,利用相似性检索方法找出与新铁舾件最相似的铁舾件集,根据铁舾件特征参数和工时信息,利用BP神经网络算法找出工时测算模型,从而推导出新铁舾件的舾装工时定额。经实际测得的船舶铁舾件工时数据验证,该工时定额估算方法可以高效、准确地估算铁舾件安装工时。
In order to solve the time-consuming problem and improve the estimation accuracy,the work quota estimation method of Ship Outfitting pieces based on CBR and BP neural network is proposed.According to the feature that the installation technology of the subdividing work-types of the Iron Outfitting pieces is similar but different in models,based on the similarity rules of the combination of non-numeric and numeric characteristic parameters and under the condition that the outfitting work-hours of some models are known,the most similar iron outfitting set to the new iron outfitting is searched out according to the similarity principle.And the work-hour regularity is found out according to the parameters and work-hour information and using the neural networks algorithm,thus to derive the work quota of the new iron outfitting pieces.Verified by the actual data of ship iron outfitting pieces measured in company,the method can efficiently and accurately estimate the installation hours of iron outfitting pieces.
出处
《工业工程与管理》
北大核心
2011年第4期96-102,共7页
Industrial Engineering and Management
基金
国家自然科学基金资助项目(70872076)
关键词
工时定额
铁舾件
CBR
BP神经网络
相似性检索
work quota
iron outfitting pieces
CBR
BP neural network
similarity retrieval
