【作者】 阴艳超；

【导师】 孙林夫；

【作者基本信息】 西南交通大学 ， 计算机应用技术， 2009， 博士

【摘要】 产业链是21世纪企业应对市场竞争,获得竞争优势的主要组织形式。基于ASP的公共服务平台建设已成为制造业产业链发展的大趋势。ASP模式下的产业链协作,需要克服原有信息技术对产品资源管理业务支持的局限性,在更高层次上和更宽范围内扩展产品资源及相关业务。产品应用服务系统采用基于公共服务平台的产品应用服务模式,为产业链上、下游供应商、制造商、经销商、售后服务商之间的协同采购、协同销售、协同售后服务业务提供产品资源信息服务,对于消除产业链上协作企业之间鸿沟,形成产业链中中小企业群体的产品应用服务协同环境,实现产品资源及其支持业务的产业链业务层扩展具有重要意义。产品应用服务系统的构建是ASP模式下产品资源信息服务实现的前提,其构建过程需要一系列方法的支持。本文结合国家“863”高技术发展项目“基于国产三维设计平台或核心的汽车及零部件产品应用服务系统”(项目编号：2007AA040404),研究了支持产业链协作的产品应用服务建模与系统构建的相关理论、方法和关键技术。具体包括以下几个方面：(1)针对ASP模式下产业链协作特点建立了业务驱动的产业链协作模型,提出了基于公共服务平台的产品应用服务模式,构造了该模式下产品应用服务系统的构建模型,抽象出了产品应用服务系统的构建进程,提出了产品应用服务系统的推进式构建方法,研究了推进式构建方法的构建机理、过程模型、进化模型及实现过程。(2)结合产品应用服务业务与MDA框架,构建了面向产品应用服务的模型驱动架构。该架构包括可拓统一建模语言EUML元模型、产品应用服务需求模型、产品资源全性能模型、产品应用服务模型四个建模层次。各层模型之间通过细化求精、选择变换和映射组合完成了从业务模型到实现模型的转换,实现了从服务需求空间到产品应用服务功能空间再到产业链价值空间的映射；(3)针对产品应用服务构件的全局组合优化问题,提出了基于模糊物元分析和改进微粒群(ERTPSO)算法的双层组合优化策略。首先将模糊物元分析理论作为算法的第一层,建立多构件组合优化问题的模糊物元形式化模型,并应用模糊物元优化方法将多目标优化问题转化为单目标优化问题；然后通过对PSO算法参数的动态优选,较好解决了PSO算法的早熟问题,并将改进PSO算法作为第二层,依据组合目标函数被优化过程中的反馈信息,以较快的收敛速度搜索全局最优组合方案；(4)针对产品应用服务系统的多层次特点,提出了面向全局需求适应度的产品应用服务构件组合策略。在建立多构件自适应组合优化过程框架的基础上,给出了基于稳定服务链的产品应用服务构件选择方法,建立了以产品应用服务系统整体性能、构建成本及效率为目标函数的多构件组合优化数学模型,并采用双层组合优化算法对其进行多目标折衷优化,确定最优组合方案。(5)将研究成果应用于某汽车集团公司的汽车及零部件产品应用服务系统,验证了所提方法的有效性和实用性。更多还原

【Abstract】 Industry chain as the important organization mode for enterprises to create and capture value, is a critical supporting to deal with market competition and obtain competitive dominance in 21 century. Public service platform based on Application Service Provider has become the effective solution to better facilicatate the development of industry chain. Such new collaborative mode of industry chain requires complete product resources in product lifecycle for supporting, which should overcome limitations of product data management at present. Product application service system (PASS) is considered as a product resource service system composed of multiple types of service elements, in which public service platform as service provider that combines data, information and knowledge in product lifecycle to support collaborative business between manufacturing, supplier, distributor and facilitator. This dissertation attempts to study the product application service modeling and system constructing method, using which to eliminate gap between collaborative enterprises of industry chain and enable them to operate effectively together in a product application service collaborative environment. Besides, the convergence of product resouces and industry business is an additional promotion of product resource management capacity.In order to support execution a product resources service for industry chain, there should be a well-constructed PASS for supporting. The complex construction process is composed of series of methods and techniques cross service system lifecycle. We define the concept "Product Application Service" studying domain-specific theory, methods and key techniques integrately, using which to effectively support the pusher convergent method for construction of product application service system. Research work in this dissertation is partially supported by the National High-Tech Research and Development Plan of China (No.2007AA040404), which mainly concentrates on the follwing several aspects, i.e.(1) According to the characteristics of industry chain cooperation in application service provider mode, the business driven cooperative model of industry chain is established, and the product application service mode based on public service platform is presented, based on which, the construction model of product application service system supporting industry chain cooperation is constructed, and the construction process of product application service system is studied. In the end, the pusher convergent method for construction of product application service system is presented, and the construction mechanics, process model, evolutionary model and implementation procedure of the method are studied.(2) Model driven architecture supporting system contruction of product application service is established combined with service business and MDA, in which the extension unified modeling language (EUML) is employed as the tool to describe the four layers driven models. Consider EUML meta-model, Product Application Service Demand Model (PASDM), Product Resource Whole Performance Model (PRWPM) and Product Application Service Model (PASM) four layers driven model as a core of system convergence process, and the transformation from business model to implementation model is completed by refining precision, selecting transformation, mapping composition. In the end, it validates that the extestion meta architecture could realized the mapping from service requirement space to service function space, and to value space of industry chain.(3) In order to support global optimization for a service system, a two-layer combinatorial heuristic algorithm is proposed based Embedded Rhombus Thought particle swarm optimization (ERTPSO) combined with fuzzy matter element analysis (FMEA). The FMEA is used in the first layer to establish the fuzzy matter-element model of problem and transform the multi-objective optimization problems to the single-objective optimization problems; The ERTPSO is used in the second layer, on the basis of standard PSO algorithm, the performance and efficiency of algorithm are enhanced, which is intend to deal with the complex multi-objective optimization problem in engineering.(4) Component selection and composition will impact product application service system quality and efficiency to a large extent. According to the multi-layer characteristics of product application service system, the optimization strategy of product application service component composition for global demand sufficiency is put forward. In successtion, the stable service chain is established to select heterogeneous product application service component, in which business system of industry chain is decomposed into fuctional module according to stability and variability of business model, and the corresponding component sequences are selected with the the characteristic mapping strategy. Finally, a composition optimization mathematical model of components is given to satisfy the maximum global demand sufficiency of system, i.e., performance, cost and efficiency, based on which, an improved particle swarm optimization method was employed to optimize the multi-component composition scheme.(5) through the construction practice of product application service system the automobile and parts product application service system in Sichuan Truck Manufacture Plant Co., Ltd, using the modeling and constructing methods, the implementation procedure is given to verify the practicability and effectiveness of methods framework.更多还原