【作者】 王增强；

【导师】 蒲云；

【作者基本信息】 西南交通大学 ， 物流工程， 2013， 博士

【摘要】 经济全球化的到来加剧了企业的竞争,只有生产满足顾客需求的产品,才能在激烈的竞争中占据有利地位,在这种情形下,企业的大规模生产方式必须向大规模定制(Mass Customization, MC)转变。MC是以大批量的生产效率满足不同客户个性化需求的生产模式,同时又不增加成交的总成本。毋庸置疑,MC将成为21世纪生产企业追求的理想模式。合理的物流系统可以降低成本和增加利润,同时保证产品快速到达最终客户,因此,实现MC所面临的一个重要决策问题就是选择物流系统。在实际的MC环境下物流系统选择过程中,能够获取的信息通常是非常有限的,尤其在行业中优先实施MC时,此外,各种信息的来源不同、表达信息的方式多种多样以及包含较多人为判断的输入信息,不可避免地导致信息环境充满了主观性和模糊性,对系统的合理选择提出了挑战。传统的评价方法都尝试摆脱单纯依靠定性分析和逻辑判断的方法,试图将决策中的所有指标定量化,实际上,许多指标的量化是非常困难的,如何充分利用有限的信息和决策者的经验是本研究的核心内容,具体体现在以下几个方面：(1)针对MC环境下物流系统关键指标权重的不确定多属性群决策问题,提出了几种基于主观和模糊信息的权重分析方法,分别适用于不同规模的企业。首次将语义决策理论引入到MC环境下物流系统指标权重确定的过程中,充分反映信息环境的模糊性和主观性,同时能较大程度上避免评估信息的丢失,中小型企业决策者的工作环境和经历差异较小,使用相同语义短语集给出偏好信息,而大中型企业决策者的文化水平和工作经历差异较大,选用不同粒度的语义标度,分别提出了基于群体语义信息的处理方法和基于复杂语义信息的处理方法以确定物流系统关键指标的权重。鉴于跨国公司或超大型企业的评估成员在文化背景、语言和价值体系等方面存在明显差异,多个评估成员通常会选择自己熟悉或者偏好的信息形式给出偏好信息,显然,这会增加问题的处理难度,但这也更符合真实情况,据此提出了基于多种偏好信息的关键指标权重确定方法。(2)为有效反映MC环境下物流系统评价的复杂性,提出了基于群体语义信息的物流系统评价方法。在该方法中,首先,评估人员利用语义短语表征物流系统备选方案的偏好信息,并使用转换函数将语义短语转化为三角模糊数；其次,两次使用广义的导出有序加权平均算子将评估人员给出的偏好信息与指标权重和评估人员权重集结,得到各个方案的综合评估值；最后,对方案综合评估值排序,得出最优方案。(3)通过对MC环境下物流系统复杂性和主观性的分析,MC实施团队通常会给出物流系统一些指标的目标期望,在考虑指标的目标期望情形下,提出了基于公理设计的物流系统评价方法。在该方法中,首先,通过计算每个指标的评估值符合指标的目标期望程度,构建符合期望程度的评估矩阵；其次,根据公理设计对专家团队核定备选方案的期望符合程度进行处理,得到各方案的综合信息量,最后,对综合信息量排序,确定最合理的方案。(4)为了有效地评估MC环境下中物流系统的模糊性,每个评估成员在评估过程中针对同一组物流方案可能给出多种不同形式的信息,提出了基于多种偏好信息的物流系统选择方法。首先,决策者采用偏好的信息形式给出系统备选方案的偏好信息；其次,利用转换函数将多种形式的偏好信息一致化为模糊互补判断矩阵,应用行和归一化方法计算排序向量,进而集结决策成员的权重获得备选方案的群体评估矩阵,再次,集结关键指标的权重信息,计算各个备选方案的综合评估值；最后,依据综合评估值的大小进行排序,获得最佳方案。最后,研究使用一些物流系统选择的实例,证明了所提方法的合理性和有效性。从理论方面看,基于不确定信息的Mc环境物流系统选择理论,在很大程度上增强了系统的针对性,并且对系统选择理论是一种完善。从实践方面分析,使用不确定信息的系统选择方法可以增强物流系统的适用性,有助于提高企业的顾客满意度,优化资源配置。更多还原

【Abstract】 The arrival of the economic globalization induces an ever more competitive environment in which manufacturers operate. In order to occupy favorable position in fierce competitive battle, organizations must provide high-quality products to meet customer’s requirements, therefore, it is imperative to transform from mass production to mass customization(MC). Mass customization refers to the production mode that produces varied and often individually customized products and services with the efficiency of mass production, without any increase in the total cost. No doubt MC is an ideal mode that the manufacturing industry will pursue far in21century. The using of reasonable logistic system can help manufacturers to reduce costs and increase profits, and at the same time ensure that speed products to customers, therefore, it is an important group decision making problem to choose the suitable logistics system for the companies which are ready for implementing MC.In reality, the inherent vagueness or uncertainty in the process of evaluating logistics system under the environment of MC presents a special challenge to the effective selection of the suitable system. The uncertainty and vagueness are due to a number of reasons:the data available for evaluating is often limited, especially when a company first implements MC in the industry; the obtained information was from a different source; the information was presented by various ways and the evaluating process involves various inputs in the form of human judgment. In the traditional evaluation method, most of the input variables were assumed to be precise and were treated as crisp numerical data, as a matter of fact, it is too difficult to quantify all the indices, and how to make full use of limited information and the decision makers’experience is the core of this study, the major contents of the dissertation are summarized as follows:(1)In order to deal with the uncertain multiple attribute group decision making problems of logistics system key indices analysis in MC, several analysis methods based on uncertain and vague information were put forward, which can be applied in different scale manufacturers. For reflecting the vagueness and uncertainty of information environment and avoiding the loss of the judgment information to a greater degree, the linguistic decision-making theory was introduced to determine the attributes weights of logistics system in MC. In small and medium-sized businesses, the differences of experts’cultural level and work experience were lower, they can use the same linguistic variable set to express their preference values, while the experts in large and medium enterprises have large differences, they usually express their judgments using different linguistic term sets. Then, based on the group linguistic information and complex linguistic information, two approaches were presented to determine the weights of key attributes. Experts in multinational companies or conglomerates express their preferences in their preferred or familiar formats, depending on their cultural background, language and value systems. Although the free expressions of preferences make the problem complicated, it is obviously that they make it more realistic, therefore, a multi-format information-based approach of determining the weights of key attributes was proposed.(2)In order to reflect the complexity of logistics systems under the environment of MC in an effective way, a group linguistic information-based approach of evaluating logistics systems was proposed. In this approach, firstly, decision-makers used linguistic variables to express their preference values of the logistics system alternatives, and a transformation function was used to change the variables into triangular fuzzy numbers; secondly, a generalized induced ordered weighted operator was introduced to integrate the given preference values with the weights of key attributes and decision-makers, and then the comprehensive values of each alternative were obtained; finally, the suitable alternative could be chosen based on the priority of comprehensive values.(3)According to studying the uncertainty and complexity of logistics systems in MC, the team in charge of implement MC mode often gave orient expectation of some logistics system indices, a methodology based on axiomatic design was proposed considering index expectation. In this methodology, firstly, based on the degree which each logistics system alternative’s preference values with respect to the orient expectation, the evaluation matrix of expectation accordance degree was obtained; secondly, estimators checked the degree to meet expectations of each alternative, and then calculated the comprehensive information content by using axiomatic design; finally, on the basis of the order of the comprehensive information, managers can judge which logistics system has more advantage than the other systems.(4)For reflecting vagueness of logistics systems under the environment of MC in an effective way, estimators tend to give evaluation values of the same logistics system alternative about their personal preferences in many different ways, a multi-format information-based analysis approach of system selection was proposed. Firstly, decision-makers express their preferences of the relative importance weights of logistics system alternatives in their preferred formats; secondly, some transformation functions were used to unify the multi-format information into complementary judgment matrix, and priorities of complementary judgment matrix were resolved by using normalizing rank aggregation method, then the group evaluation matrix was determined by aggregating the weights of decision-makers; thirdly, the comprehensive values of each alternative were obtained by integrated with the weights of key attributes; Finally, on the basis of ranking the comprehensive values, the suitable scheme was acquired.Finally, some real-world logistics system selection cases were presented to demonstrate the feasibility and validity of the proposed approach. In theory, the uncertain information-based methods of evaluating logistics system under the environment of MC, to a great extent, improves the flexibility of system, and enriches the system evaluation theory. In practical, when manufacturers use the logistics system selection methods based on uncertain information, the applicability of the logistics system can be enhanced, more customer satisfaction can be achieved, and resource allocation can be optimized.更多还原