【作者】 汪奇；

【导师】 黄洪；

【作者基本信息】 浙江工业大学 ， 管理科学与工程， 2008， 硕士

【摘要】 随着移动计算技术的发展,笔记本电脑应用日趋广泛,市场规模也不断扩大。影响笔记本电脑品牌竞争力的两个重要因素:面向客户个性化与需求快速响应成为笔记本电脑制造企业关注的焦点,由此笔记本电脑BTO制造模式应运而生。在实际情况中,由于受制于行业特有的不利状况,笔记本电脑BTO生产是通过一种“推拉”结合的生产模式来实现的,因此生产计划的准确性对该模式的正常运行至关重要。本文对某笔记本电脑制造企业的实际状况进行了调查,分析了该企业生产计划制定中的难点以及实现方式上的特殊性,提出了一种新思路,即用关联规则数据挖掘方法从企业ERP订单模块与PDM数据中提取有关电脑组件配置选择关联度的信息,辅助生产管理部门制定生产计划的思路,帮助企业改变过分依赖个人经验的状况,制定出更为科学、合理的生产计划。具体的工作有:1、根据笔记本电脑BTO生产计划的特点,对经典关联规则挖掘算法——Apriori算法进行了分组启发式改进并结合多最小支持度、增量更新改进,最终组合形成了一种综合算法,提高了实用性。2、设计实现了笔记本电脑组件关联规则挖掘系统,并对数据仓库的建立,数据ECTL以及数据挖掘结果的展示做了详细的阐述。3、结合实例解释了关联规则挖掘结果对生产计划的作用和意义,并介绍了一种基于马尔柯夫链的预测方法,用于电脑配置关联度变化趋势的预测。更多还原

【Abstract】 Due to the development of mobile computing technology, notebook computers have been gradually used in diverse areas with an increasing market. What major notebook computer manufacturers concern mostly are two key factors—customized computer-design and immediate response to customer demands, which influence the competitiveness of a notebook computer brand. As a result, the build-to-order model is created. In fact, because of several restraints of the industry, BTO production of notebook computers is carried out with the usage of a "push-pull" mixed model. Therefore, the precision of the production plan is crucial to the operation of this model.Based on the research of a particular notebook computer manufacturer, this paper analyses the difficulties of formulating a proper production plan and several special considerations required to carry out the plan. The thesis also proposes to find out the relations between computer components from ERP orders module and PDM data of the company, by employing association rules data mining, in order to help relevant departments formulate the production plan. The main research work includes:1. Regarding the characteristics of the BTO production of notebook computers, this paper suggests improving the classic algorithm of association rules data mining—Apriori algorithm, with multi minimum supports, incremental renewal and heuristic grouping, to enhance the practicality in its implementation.2. This paper explains how to build the data warehouse, the ECTL processing and the user interface in designing and implementation of association rules data mining system.3. The effects of association rules data mining results to production planning are explained with examples. A method based on the Markov chain is proposed to predict the changes of computer configuration relations.更多还原