基于DEA方法的多产区生产配送一体化问题研究
发布时间：2018-09-11 13:16
【摘要】：在当今激烈的竞争环境下,越来越多的优化模型和算法、决策支持系统、程序化分析工具等等方法被用来改善企业整体运营绩效,以此帮助企业在当今激烈的产品竞争中获得长期的竞争优势。 本文提出的生产配送一体化模型主要考虑的是如何最优的决定该问题中生产环节和运输环节的各个相关变量,其中生产环节主要涉及投入变量和产出变量,配送环节主要涉及各个分销中心之间的是否存在运输及对应的运输量。 本文相对于之前的研究主要有两点区别,其一是我们在考虑生产计划时不假设任何先验信息。特别的,本文将数据包络分析(data envelopment analysis,DEA)引入到统一的生产运输问题当中。数据包络分析作为一个非参数的方法,不同于之前该领域的用来描述生产关系的已有方法。在当下,已经有很多涉及到统一的生产和配送规划问题,其中有些考虑跨期操作的甚至包括了对于库存管理的运筹。但是这些文章对于生产关系的刻画都是添加了很多外生假设的。数据包络分析是一个提供满意解的主要工具,也就意味着本文提出的模型是基于中央决策者所能获得有限消息下做出的最满意的决策。 另外一个不同之处在于本文在处理生产关系中将生产关系从确定性情形中进一步推广到随机性情形。事实上,不确定性是处处存在的,而不是生活中的一个特例。在本质上生产过程都是具有不确定性的。例如,在劳动力密集型产业中,产品将会受到工人情绪的影响。随着机械化的不断普及和提高,生产过程将受到机器故障、停电等其他不可控制的因素的影响。这种随机性情形拓展了本文的应用范围,缩小了理论和实际的距离,能更好的反应生产过程中的实际情况。 此外,本文分别针对确定型和随机型情景下的模型进行了最优解存在性的讨论,以及二者的最优解的生产效率问题和两个情景下最优解的比较。本文使用了一个统一的例子来具体展示文中的一系列模型,在确定型情景中还使用了灵敏度分析,即不断变换单位投入费用所在的数量级与单位运输费用所在的数量级之比的比值,而在随机型情景中分别给出了在三个置信水平下的解,用以加深对于模型理解。
[Abstract]:In today's competitive environment, more and more optimization models and algorithms, decision support systems, program analysis tools and other methods are used to improve the overall operating performance of enterprises. In order to help enterprises in today's fierce product competition to obtain a long-term competitive advantage. The integrated model of production and distribution proposed in this paper is mainly concerned with how to determine the optimal variables of the production and transportation links, in which the production links mainly involve input variables and output variables. Distribution is mainly related to the existence of transportation between distribution centers and the corresponding traffic volume. There are two main differences between this paper and previous studies. One is that we do not assume any prior information when considering production planning. In particular, this paper introduces the data Envelopment Analysis (data envelopment analysis,DEA) into the unified production and transportation problems. As a nonparametric method, data Envelopment Analysis (DEA) is different from the existing methods used to describe the relations of production in this field. At present, there are many problems related to unified production and distribution planning, some of which consider intertemporal operations, including even inventory management. However, these articles add a lot of exogenous assumptions to the depiction of the relations of production. Data envelopment analysis is a main tool to provide satisfactory solutions, which means that the model proposed in this paper is based on the most satisfactory decision made by central decision makers under limited information. The other difference lies in the fact that the relations of production are further extended from deterministic to random in dealing with the relations of production in this paper. In fact, uncertainty exists everywhere, not as a special case in life. In essence, the production process is uncertain. In labor-intensive industries, for example, products will be affected by worker sentiment. With the popularization and improvement of mechanization, the production process will be affected by other uncontrollable factors such as machine failure, power failure and so on. This random situation expands the scope of application of this paper, reduces the distance between theory and practice, and can better reflect the actual situation in the process of production. In addition, this paper discusses the existence of optimal solutions for the models under deterministic and stochastic scenarios, and compares the production efficiency of the optimal solutions between the two scenarios and the optimal solutions under the two scenarios. In this paper, a unified example is used to show a series of models in the text, and sensitivity analysis is also used in deterministic scenarios. That is to say, the ratio of the order of magnitude of the unit input cost to the order of magnitude of the unit transportation cost is constantly changed, and the solution under the three confidence levels is given in the random scenario to deepen the understanding of the model.
【学位授予单位】：中国科学技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：O221.4;F274
本文编号：2236787
[Abstract]:In today's competitive environment, more and more optimization models and algorithms, decision support systems, program analysis tools and other methods are used to improve the overall operating performance of enterprises. In order to help enterprises in today's fierce product competition to obtain a long-term competitive advantage. The integrated model of production and distribution proposed in this paper is mainly concerned with how to determine the optimal variables of the production and transportation links, in which the production links mainly involve input variables and output variables. Distribution is mainly related to the existence of transportation between distribution centers and the corresponding traffic volume. There are two main differences between this paper and previous studies. One is that we do not assume any prior information when considering production planning. In particular, this paper introduces the data Envelopment Analysis (data envelopment analysis,DEA) into the unified production and transportation problems. As a nonparametric method, data Envelopment Analysis (DEA) is different from the existing methods used to describe the relations of production in this field. At present, there are many problems related to unified production and distribution planning, some of which consider intertemporal operations, including even inventory management. However, these articles add a lot of exogenous assumptions to the depiction of the relations of production. Data envelopment analysis is a main tool to provide satisfactory solutions, which means that the model proposed in this paper is based on the most satisfactory decision made by central decision makers under limited information. The other difference lies in the fact that the relations of production are further extended from deterministic to random in dealing with the relations of production in this paper. In fact, uncertainty exists everywhere, not as a special case in life. In essence, the production process is uncertain. In labor-intensive industries, for example, products will be affected by worker sentiment. With the popularization and improvement of mechanization, the production process will be affected by other uncontrollable factors such as machine failure, power failure and so on. This random situation expands the scope of application of this paper, reduces the distance between theory and practice, and can better reflect the actual situation in the process of production. In addition, this paper discusses the existence of optimal solutions for the models under deterministic and stochastic scenarios, and compares the production efficiency of the optimal solutions between the two scenarios and the optimal solutions under the two scenarios. In this paper, a unified example is used to show a series of models in the text, and sensitivity analysis is also used in deterministic scenarios. That is to say, the ratio of the order of magnitude of the unit input cost to the order of magnitude of the unit transportation cost is constantly changed, and the solution under the three confidence levels is given in the random scenario to deepen the understanding of the model.
【学位授予单位】：中国科学技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：O221.4;F274
【参考文献】
相关期刊论文 前8条
1 陈建华;马士华;;基于集配中心的供应链物流整合方式[J];当代经济管理;2006年04期
2 赵秋红;谢稳;;生产-分销物流系统的Robust优化模型与算法[J];系统工程;2006年04期
3 冯立杰;供应链生产分销协同博弈研究[J];机械传动;2004年06期
4 王竹泉;马广林;;分销渠道控制:跨区分销企业营运资金管理的重心[J];会计研究;2005年06期
5 赵晓煜,汪定伟;供应链中二级分销网络优化设计的模糊机会约束规划模型[J];控制理论与应用;2002年02期
6 周金宏,汪定伟;分布式多工厂、多分销商的供应链生产计划模型[J];信息与控制;2001年02期
7 田俊峰,杨梅;供应链生产—分销运作一体化研究[J];信息与控制;2004年06期
8 李应;杨善林;;供应链分布式多层协同生产-分销计划模型与求解[J];中国机械工程;2008年21期
，本文编号：2236787
本文链接：https://www.wllwen.com/guanlilunwen/gongyinglianguanli/2236787.html
