【作者】 吴小虎；

【导师】 徐琪；

【作者基本信息】 东华大学 ， 管理科学与工程， 2011， 硕士

【摘要】 自2009年以来,我国跃居成为世界上汽车产销第一大国。在感叹经济增长,享受现代汽车工业丰硕成果的同时,应该考虑到汽车产销的增加必定会导致将来报废量上升这一事实。汽车报废问题倘若处理得不好,会带来严重的资源短缺、环境污染、经济甚至社会问题。相反,如果能够科学规划报废汽车逆向物流的发展,将报废汽车的问题解决好,不仅可以节约资源,降低对环境的负面影响,而且可以提升经济效益,有利于社会的稳定和可持续发展。作为逆向物流的基础和核心,网络设计的合理与否决定了逆向物流的管理绩效。文献研究发现关于逆向物流的研究比较充分,尤其是网络设计领域,但是汽车行业的逆向物流研究就显得较为罕见,这和汽车产业的发展是不平衡的。鉴于此,本文针对报废汽车的逆向物流网设计问题进行研究,旨在科学化网络规划,具体研究工作如下：首先,考虑报废汽车多级逆向物流网络,包括客户区、回收中心、拆解中心、再制造中心以及填埋点。以逆向物流运营成本最低作为目标函数,建立混合整数规划模型(Mixed Integer Linear Program, MILP)网络设计优化模型。据此可以确定网络中的设施选址以及各个节点的流量分配。其次,汽车报废量是决定逆向物流网络规模的重要指标,是MILP模型的重要参数。鉴于此,第五章引入了汽车报废量的预测模型,采用基于主成分分析(Principal Component Analysis, PCA)的BP(Back Propagation)神经网络针对上海市汽车报废量进行预测,为第六章的实证分析提供参数。第三,考虑上海市报废汽车现有的逆向物流网络,将预测以及通过查阅文献得到的参数代入MILP模型,对网络进行重新设计、规划,确定网络选址以及流量分配。实证分析验证了MILP模型和PCA-BP模型的有效性。更多还原

【Abstract】 Since 2009, China has been the first place of automobile in both production and marketing. While enjoying economic growth and achievement of automobile industry development, a fact that growth of production and marketing will certainly raise the quantity of end of life automotive (EOLA) should be considered. If the EOLA is misdealt, the problems of resource, environment, economics and society will appear. On the contrary, if reverse logistics (RL) network is designed reasonably and this problem of EOLA is handled well, resources will be saved, pollution of the environment will be reduced and economic returns will be improved. The most important it will be beneficial to social harmony and sustainable development.As the foundation and core of the RL, whether it is reasonable determine the performance of RL. After literature research, a phenomenon is found, RL especially network design has been discussed adequately, but research on automobile industry RL is not enough. This is out of all relation to the development of automobile industry. In view of this, in order to make network design more scientific, EOLA RL network design is been researched. More details as follows:First of all, a multilevel RL network including client areas, recycling points, dismantling points, reproducing points and landfill is considered. A MILP model whose objective function is minimum cost is proposed to optimize network design. After doing this, a plan including site location and flow distribution will be obtained.Secondly, the quantity of EOLA is an important index that determines the scale of the RL network and it is also a crucial parameter of the MILP model. In consideration of this, a model of forecasting the quantity of EOLA is introduced in Chapter 5. This model that based on principal component analysis (PCA) and back propagation (BP) neural networks provides parameter for the demonstration of Shanghai in Chapter 6.Thirdly, on the basis of Shanghai existing RL network, forecasted and collected parameters are applied to the MILP model to replan and redesign. The demonstration of Shanghai indicates the effectiveness of MILP model and PCA-BP model.更多还原