【作者】 王星汉；

【导师】 明新国；

【作者基本信息】 上海交通大学 ， 机械工程， 2010， 博士

【摘要】 多级供应商协同是以复杂产品、重大设备或大型工程的实现过程为目标,借助各种信息及管理技术和手段,将分布在不同空间及时间,隶属于不同合作方的企业资源迅速有效地组织为统一的有机体,对内实现合作方之间业务、知识、信息及数据的安全有效沟通,对外形成虚拟企业联盟,向社会推出高品质、高质量创新型产品的复杂过程。多级供应商协同项目则是指在这种工业应用场景下,多个企业依照一定的层级关系组成的动态联盟所进行的一系列企业间协同任务的总称。多级供应商协同项目管理就是为实现多级供应商协同项目目标所必需的一切活动的计划、安排与控制过程。在航空、航天、船舶、汽车、装备等行业,由于产品的零部件数目极其繁多,零部件设计、制造相关供应商数目十分庞大,同样的零部件可能存在多个潜在的供应商,而不同的供应商在实力、声誉、交货期、可靠性等方面不尽相同,且供应商之间存在合作、竞争等多种复杂关系。为提高业务协同效率,需要进行多级供应商协同项目管理。通过对多级供应商协同项目管理模式的建立,本论文重点研究在多级供应商、供应商资源有限、分布式企业协同环境下的协同项目管理技术,包括协同供应商选择、多级计划制定、动态进度偏差识别与调整等,并开发了原型演示系统。具体研究内容如下:多级供应商协同项目管理模式与协同流程分析方法。协同项目管理过程中各种需求的实现,需要改变主制造商和供应商的传统关系思维方式,实现跨企业业务流和知识共享,进而实现供应商协同项目管理模式的创新。本论文提出了基于协同工程思想和集成产品开发团队方法的新型项目管理模式,通过多维度、多层次分析,构建了相应的管理架构及各组成部分功能。此外,还建立了涵盖协同层次、协同要素、协同过程互动、协同信息流分析等内容的协同流程分析方法,并在后续研究中进行了广泛的应用。基于多种协同约束的供应商选择方法。通过供应商选择过程,能够逐步积累合格供应商的详细资料信息,形成较为稳定的供应商战略合作伙伴。本论文提出一种协同供应商选择方法,其中包括:协同供应商选择准则/属性的确定、选择流程、定性评定向定量化数据的转换方式、综合评优算法等。借助层次分析法与模糊理论(模糊数、模糊运算),建立了协同供应商选择模型以及基于梯形模糊数和层次分析法的TrFN-AHP协同供应商选择优化算法。多约束条件下协同项目计划制定技术。本论文对协同多级计划制定过程进行了细致的分析;通过建立项目相似度模型以及基于灰色系统理论的求解方法,为新项目计划的制定提供相似项目计划制定经验;建立了多种约束条件下的多级供应商协同项目最短工期计划制定模型,借助遗传算法,实现任务工期快速求解;提出了多级协同项目计划缓冲区设置的流程。协同项目进度实时跟踪方法。项目进度控制包括进度跟踪、偏差识别、动态纠偏以及计划重排几个环节。进度控制过程是保证项目在执行期能够尽可能按照原有计划执行的保障手段,是项目管理过程至关重要的一环,保证整个项目进度的闭环可控性。本论文详细分析了多级供应商协同项目管理进度控制过程;提出了改进进度前锋线法,用以实现进度偏差识别;提出了基于缓冲区的协同项目进度动态纠偏方法,并建立了面向工期最短的补充计划制定模型。在此研究基础之上,开发了相应的原型演示系统,通过实例对上述研究内容予以验证。本研究来源于实际工业需求(基于863课题),以工业原型系统开发与案例验证需求为出发点研究多级供应商协同项目管理过程中的若干技术,一定程度上理清了协同项目管理中存在的若干问题。本论文成果为多供应商参与的协同项目管理研究与应用提供了若干创新性框架、方法、模型与算法,有助于缩短协同项目管理理论研究与工业实际需求的差距。更多还原

【Abstract】 During the development of complex product, heavy equipment or big engineering project, multi-tier supplier collaboration combined enterprise resources located in different areas to a unity to coordinate business, knowledge, information and data in a safe, effective way across partners. Multi-tier supplier collaborative project is the general concept of all temporary tasks in finishing the design and development process across enterprises in a unique time, resource, performance and scale restriction situation. The planning, scheduling, controlling activities to achieve collaborative project goals are called multi-tier supplier collaborative project management (MSCPM).In aviation, aerospace, ship, automobile and equipment industries, there are many components and parts designed and developed by large amounts of engineers of different supplier companies. Even for each part, there are still several potential suppliers to select, which are different in strength, credit, security, etc. To enhancing the efficiency of business collaboration, suppliers should combine to a virtual enterprise alliance under the directions of the main manufacturer.This dissertation puts efforts in some key points of MSCPM in multi-tier suppliers, parallel projects, limited resources and distributed scenarios. These technologies include collaborative supplier selection, multi-tier plan scheduling, dynamic planning deviation identification and controlling, and associated system. Detailed research works includes:The multi-tier supplier collaborative project management framework proposed in this research employs a new analysis model to settle the requirements come from complex product development industry. By analyzing the collaboration level, collaboration element, collaborative interaction process, and collaborative information stream, collaborative project management ontology and application scenario have been draw out. This is the theoretical basis in future research.This study presents a new supplier selection criteria and attributes towards collaborative product development between suppliers and customers. Five main aspects are summarized as R&D capacity, information system, willing to collaborate, enterprise background and risk management. Another novel point in this study is the algorithm named TrFN-AHP, which helps to qualify the uncertain linguistic weights made by decision makers and give a more reasonable result to the enterprise. This new approach is based on AHP and fuzzy number.By detailed analysis of multi-tier project scheduling making, this study established a new method to solve the project similarity model based on grey system theory. With the help of reformed generic algorithm, we can get the global best solution for task start time in a multi-restrictions shortest project date making model. Based on critical chain theory, a general process is introduced to set buffers in current schedule.Project schedule controlling includes schedule tracing, deviation identification, dynamic correcting and task rescheduling. Good controlling methods can help project execution process in a close-loop with high robustness. In this study, we analyzed the collaborative process of MSCPM schedule controlling phase. We reformed traditional schedule front line method with resource utility ratio and buffer. By detailed analysis of effecting scopes in delay and rescheduling, we proposed a new deviation correction method based on critical chain buffers.A prototype software system is developed in this research, which aims to verify the research contents above. Research results in this study provided several novel methods and framework to shorten distances between theoretical research and requirements in industries.更多还原