【作者】 李斐；

【导师】 杨育；

【作者基本信息】 重庆大学 ， 工业工程， 2014， 博士

【摘要】 随着市场竞争的激烈化及客户需求的多样化与个性化，企业面临着各种新挑战，传统创新设计模式已经不能适应当前激烈的全球化市场竞争环境。网络信息技术与计算机技术的快速发展为企业创新模式的变革提供了技术支撑。客户协同产品创新模式是一种以创新客户为主导的多创新主体协同的开放式创新模式，极大地提高了企业的创新与市场竞争能力。知识学习与创造是客户协同创新的最主要工作，系统、有效的知识管理成为客户协同产品创新目标实现的重要保障。然而，由于客户协同创新过程的多主体性、非线性等复杂系统特性，客户协同产品创新中的知识管理变得异常复杂，所需考虑的因素众多，要进行知识的有效管理及优化已相当困难。如何从复杂系统的角度对客户协同创新组织的知识进行系统、有效的管理成为当前的研究热点。为此，本文在系统研究国内外有关客户协同产品创新、知识系统、复杂网络、超网络等相关理论与方法的基础上，对客户协同产品创新知识系统中的知识系统建模、创新客户知识主体识别、知识主体个体重要度评价、知识主体的创意知识产品组合决策、知识主体子系统的稳定性等关键问题进行了系统、深入的研究。全文的主要研究内容包括以下5个部分：第一，对客户协同产品创新知识系统进行总体研究。首先，对客户协同产品创新知识系统的内涵进行了分析，包括了其基本概念、构成及复杂性；然后，对系统建模的常用方法进行对比分析，对客户协同产品创新知识系统进行基础建模；最后，对客户协同产品创新知识系统的关键问题进行了分析，提出以客户协同创新过程为主线的客户协同产品创新知识系统研究思路及以知识主体为中心的创新客户知识主体识别、知识主体个体重要度评价等若干关键问题，并提出上述问题的研究框架。第二，对知识系统中创新客户知识主体的识别方法进行研究。首先，对创新客户知识主体识别的特点进行了分析，并在支持向量机等相关基础理论研究的基础上，构建了创新客户知识主体识别的研究框架；其次，对创新客户知识主体识别的要素进行了分析及混合量化，并基于粗糙集理论对识别要素进行了约简；然后，在上述研究的基础上，构建了基于改进代价敏感支持向量机的创新客户知识主体识别模型；最后，通过应用实例验证了所提识别模型的有效性。第三，对知识系统中知识主体的个体重要度评价方法进行研究。首先，对知识主体个体重要度评价的常用方法及评价指标进行了分析，并结合超网络、复杂网络等相关理论，构建了知识主体个体重要度评价的研究框架；其次，构建了加权知识主体-知识点超网络模型，为知识主体的个体重要度评价奠定了基础；然后，提出了知识主体网络的复杂网络特性验证方法，在此基础上，构建了基于加权超网络的知识主体个体重要度评价方法及过程模型；最后，通过实例与传统的仅通过知识水平进行知识主体个体重要度评价的方法与结果进行了对比分析，验证了所提模型的有效性和可行性。第四，对知识系统中创意知识产品的组合决策方法进行研究。首先，对创意知识产品组合决策问题的特点进行了分析，在此基础上，构建了创意知识产品组合决策问题的研究框架；其次，对创意知识产品组合决策的目标进行了分析与定量化表述，共包括创意价值、收益满意度、风险及交叉相似度等4个目标，在此基础上，构建了创意知识产品组合决策的多目标决策优化模型；然后，针对传统遗传算法容易局部收敛的问题，提出自适应遗传算法对上述多目标决策模型进行求解；最后，通过应用实例验证了该方法的可行性与有效性。第五，对知识系统中知识主体子系统的稳定性进行分析。首先，对客户协同创新知识主体子系统的稳定性进行了定义，并构建了知识主体子系统稳定性分析的研究框架；其次，基于无向加权图理论构建了客户协同创新知识主体子系统的无向加权网络，并对知识主体间的协同关联强度进行了分析；然后，针对复杂系统常用稳定性分析方法中的节点失效模式不能充分反映现实知识主体人才流失特点的问题，提出了不完全信息条件下的稳定性分析节点失效模式，并确定了知识主体子系统的稳定性测度，在此基础上，构建了客户协同产品创新知识主体子系统的稳定性分析过程模型；最后，通过应用实例验证了该方法的有效性与可行性。更多还原

【Abstract】 With drastic market competition and diversification and individuation ofcustomers’ demands, enterprises are facing various new challenges, so that thetraditional innovative design model has been unable to meet the current fierce globalmarket competition environment. The rapid development of network informationtechnology and computer technology provides technical support for changing theenterprise innovation model. The Customer Collaborative Product Innovation model,which asks innovative subjects must design collaboratively with innovation customersas the leading factor, is a more open approach to innovation. The model greatlyimproves innovation ability of the company and competitiveness in the market. Becauselearning and creation of knowledge are the most important factors in the customercollaborative innovation process, effective knowledge management has become animportant safeguard for customer collaborative product innovation to achieve theobjective. However, due to the complexity of the system features such asmulti-subjectivity, nonlinear in the customer collaborative innovation process, theknowledge management of customer collaborative product innovation has become verycomplex, which needs considerate many factors, so effective management andoptimization of knowledge have been quite difficult. How to manage effectivelyknowledge of customer collaborative innovation organization from the complex systemperspective has become a research hotspot. Therefore, based on researchingsystematically the relevant theories and methods in customer collaborative productinnovation, knowledge systems, complex networks, super-networks and so on, thecritical issues in the Customer Collaborative Innovation Knowledge System areresearched systematically and thorough, such as modeling knowledge system,identifying innovation customers knowledge bodies, evaluating the individualimportance degree of knowledge subjects, portfolio decision-making for creativeknowledge products, the stability of the knowledge agent system. The main contents ofthis paper include the following five parts:First, study overall the knowledge system of customer collaborative innovation.The first job is introducing basic concepts and structures of CCPIKS; Then, compareseveral common methods for building system models, construct the basal model of theCCPIKS; Finally, analyze key issues of CCPIKS in the customer collaborative product innovation process and propose research ideas for customer collaborative innovationknowledge systems which take knowledge bodies as the center and the customercollaborative innovation process as the main line, put forward some critical issues asidentifying innovation customers knowledge bodies, evaluating the individualimportance degree of knowledge subjects etc., and raise the overall framework of theabove-mentioned research questions.Second, research identifying approach of innovative customers knowledge agentsfor customer collaborative innovation knowledge systems. First of all, analyzecharacteristics of identifying, based on relevant basic theory researches as the supportvector machine propose the research framework to identify innovative customersknowledge agents; Then, analyze and quantize blendedly main factors on innovationcustomers knowledge subjects identification, and reduce these factors based onRough-Set theory; Next, on the basis of these studies, build a recognition model forknowledge subjects of innovative customers based on improved cost-sensitive supportvector machine; Finally, verify the validity of the identification model by an applicationexample.Third, study evaluation method for the individual importance degree of knowledgeagents. At first, analyze common evaluation methods and evaluation indicators for theindividual importance degree of knowledge agents, with combineing withsuper-network theory and complex networks, construct the research framework toassess the individual importance degree of knowledge subjects; secondly, build aweighted knowledge bodies and knowledge points super-network model, which sets thefoundation for the evaluation of the individual importance degree of knowledge subjects;Then present the authentication method to vertify complex network characteristics ofknowledge bodies networks, on this basis, constructe the evaluation method and theprocess model for the individual importance degree of knowledge subjects based onweighted super-networks; Finally, verify the effectiveness and feasibility of theproposed model by a practical example comparing the above method with thetraditional way which evaluates the individual importance degree of knowledge agentsonly according to knowledge level.Fourth, research portfolio decision-making method of innovation knowledgeproducts in the customer collaborative innovation system. Firstly, analyze thecharacteristics of portfolio decision problems of creative knowledge products, based onthis, build a research framework for creative knowledge products portfolio decisions; Secondly, analyze and quantified express targets of creative knowledge productsportfolio decision, including value of creations, income satisfaction, risks and crosssimilarity four goals, on this basis, build a multi-objective optimization model ofcreation knowledge products portfolio decisions; Then, due to traditional geneticalgorithm is easy to local convergence, this paper proposes adaptive genetic algorithmto solve the multi-objective decision-making model; Finally, use an application exampleto verify the feasibility and effectiveness of this method.Fifth, study the stability of knowledge agents subsystem in the customercollaborative innovation knowledge system. Firstly, define the stability of theknowledge agents subsystem, and build the research frame for analyzing the systemstability; Secondly, based on the undirected weighted graph theory build an undirectedweighted network of customer collaborative product innovation knowledge agentssubsystem, and analyze the collaborative strength among knowledge bodies; Then, forthe problem that in the complex system the common stability analysis methods as thenodes failure mode does not adequately reflect the characteristics of brain drainproblems of reality knowledge bobies, present the stability analysis node failure modeunder incomplete information conditions, and determine the stability measurement ofthe knowledge agents subsystem, and on this basis, construct the processs model toanalyze the stability of the customer collaborative product innovation knowledge agentssubsystem; Finally, verify the validity and feasibility of the method via an applicationexample.更多还原