【作者】 李楠；

【导师】 陈燕；

【作者基本信息】 大连海事大学 ， 管理科学与工程， 2006， 硕士

【摘要】 随着现代物流信息化发展进程的加快,物流企业对信息管理的高层需求,即对决策支持的需求越来越多,而企业相关数据的不完全和离散性限制了一些数据挖掘模型的使用。灰色系统理论对统计数据少、信息不完全系统的建模与分析具有较好的效果。本文提出灰色系统的理论与方法在物流行业的应用,针对物流企业管理决策的实际问题建立基于灰色系统理论的数据挖掘模型。 本文首先对灰色系统理论的基本理论进行研究,阐述了灰色系统建模理论、灰色关联分析与灰色聚类方法和灰色预测模型。本文深入研究了GM(1,1)模型的数据生成、建立过程及检验方法,通过分析GM(1,1)模型误差产生的机理,根据影响模型精度的两个根本原因提出改进参数估计的GM(1,1)模型(GOM(1,1))和改进边界条件的GM(1,1)模型,并从GM(1,1)参数包为起始进行严格推导。 在理论研究的基础上,本文将灰色预测模型和灰色聚类模型应用到物流企业管理决策的实际问题中,实现了库存管理中不确定需求的灰色预测和供应链合作伙伴选择的聚类分析。在需求量预测模型中,分别用GM(1,1)模型、GOM(1,1)模型和改进GM(1,1)模型,根据物流信息系统生成的统计报表中的月出库量数据序列来预测未来库存的月需求量,并对结果进行检验。将经典GM(1,1)模型与改进模型的输出结果进行比较分析,并用时间序列法进行对比验证。在合作伙伴选择问题中,运用灰色聚类方法进行对供应链合作伙伴的因素指标分析,得到的聚类结果优化了选择并用来辅助决策。灰色数据挖掘模型在物流企业的管理决策问题中的应用证明了基于灰色系统理论的灰色预测和聚类模型是有效的、具有实用价值的数据挖掘模型。更多还原

【Abstract】 With the rapid development of modern logistics informationization, higher demand of information management of logistics enterprises, namely demand of decision support, is becoming more and more. The relevant data of enterprises is incomplete and scattered, which limits the application of some data mining model. Grey system model is especially efficient in modeling of system that is short of data and has incomplete information. This article introduces grey theory into logistics trade, and builds data mining model based on grey theory to solve the practical problem in logistics enterprises’ management and decision.Firstly this article studies the basic theories of grey system theory, concluding grey system model building theory, grey associating analysis, grey clustering and grey forecasting model. The model of GM(1,1) is carefully studied on data series forming, building of model and checking-up method. By analysis of mechanism of error, thisarticle advances two improved GM(1,1) model------parameter advanced GM(1,1)(GOM(1,1)) and border condition advanced GM(1,1), and strict prove is made from the basic parameter package.On the base of theory study, this article applies grey forecasting model and grey clustering model into the practical management and decision of logistics enterprises, realizing the grey forecasting of uncertain demand and the grey clustering analysis of supply chain partner selection. In forecasting model of demand, GM(1,1), GOM(1,1) and advanced GM(1,1) is used respectively. The monthly quantity shipping out of the stock which comes from the statistics report of logistics information system, is used to forecasting the demand in the future, and the results are checked up. This article made a contrast analysis of classic GM(1,1) model and improved model and used time series method to confirm the result. In the partner selection problem grey clustering method is used to analyze the factor index of supply chain partner, and the results optimize the choice and support the decision. The application of grey data mining model in the management and decision of logistics enterprises has proved that the grey forecasting model and clustering model is effective and of practical value.更多还原