企业知识网络复杂系统的结构与演化：产业集群情境下的实证研究

【作者】 李文博；

【导师】 郝云宏；

【作者基本信息】 浙江工商大学 ， 企业管理， 2009， 博士

【摘要】 随着知识经济的到来,知识已经取代土地、劳动力、资本等传统的生产要素,跃升为最主要的经济资源和占支配地位的竞争优势之源泉,所以,知识管理在理论界和管理实践中蓬勃地发展起来。通过对国内外知识管理理论研究脉络的梳理和前沿动态的追踪,知识网络日益成为研究的热点问题,对知识网络的研究也从质点化、简单化日益转为整体化和复杂化。知识网络是产业集群内在的一种优势效应,产业集群中的知识溢出和组织学习为知识网络存在提供了现实基础。本文所研究的知识网络区别于企业内部的员工知识网络等,是指存在于特定产业集群情境变量下的以企业、研究机构等为节点,节点间具有知识交换关系的复杂网络。从空间范围看,由于知识的可传播性,特定产业集群情境中的知识网络也是本地知识网络和全球知识网络的嵌套。以此为背景,本论文以复杂系统为理论基点,结合产业集群情境下的实证研究,对知识网络的结构与演化问题进行了探索性研究。论文的主要研究结论如下:(1)产业集群中的知识网络是一个复杂适应性系统。以CAS为理论基点,阐述了知识网络所蕴含的系统思想及层面结构,建构了知识网络的系统概念模型,提出了若干具有执行力的知识网络演进机制。研究表明,知识网络具备明确的复杂系统特征:①知识网络中节点与节点之间联结的多样性。②知识网络系统内部各个知识节点之间存在正负反馈机制,通过这种机制,系统实现信息、知识的流动。③各个知识节点之间的关系具备重复博弈的特征,表现出长期性。(2)基于层面间的深层耦合机理,知识网络具有要素、联系及演化三个层次的复杂性。指出知识网络涌现出的令人关注的复杂性似乎会挫败研究者对知识网络的建模兴趣,但同时也带来了极大的挑战。知识网络系统的结构复杂多样,具有层次性且相互嵌套,各子系统之间彼此影响,相互耦合,十分复杂,任一微小输入变量发生变化,都可能会导致网络系统突变。知识网络中的局部互动关联性,涌现为网络整体上的动态演化行为模式,而这种行为模式又导致网络结构的不断变化与更替。(3)知识网络结构的测量可由4个维度刻画。产业集群中的知识网络问题受到来自各领域学者的和企业实践者的关注,其中,知识网络结构的测量问题日显突出。在现有研究基础上,通过理论分析建立了理论框架,构建了结构方程,选取中国浙江省117家企业为样本进行实证研究。结果表明:该测量模型具有较好的建构效度,理论贡献在于确定的知识网络结构测量工具包括四个维度:节点中心性、联结强度、关系质量和网络密度。(4)知识网络演化过程具备开放性、远离平衡态、非线性及涨落等耗散系统特征。囿于对知识管理的狭义理解,以往人们关注的多是知识网络系统中的特定子集或知识管理过程中的某个片断。随着知识管理活动本身与外界环境中其它主体的关系日益复杂化,这种狭义理解的局限性更加显露出来,因此需要启用一种有机的方法对知识网络演化进行系统思考。本部分的创新性体现在从耗散结构的理论视角剖析知识网络演化过程的动力学特征,将注意力聚焦于开放性、远离平衡态、非线性及涨落四个方面。(5)知识网络演化的涌现行为可由神经网络动力学模型模拟。针对知识网络演化建模其本质属于系统仿真方法这一特点,提出了基于人工神经网络的系统演化建模方法,并归纳了建模步骤。神经网络是由许多人工神经元按一定的方式联结形成的网络,可被用来作为信息处理的一个整体。精密的仿真方法是对现有研究的改进,仿真试验表明:以神经网络逼近知识网络结构开展动力学分析是恰当的。该方法具有自适应、自学习功能,实现了对知识网络演化结果的智能化测量,采用该方法获得的结果是令人满意的。与已有研究相比,论文的创新点主要有如下几点:(1)构建了知识网络复杂系统诠释的新的理论视角。基于众多学者的相关研究,本文将知识网络视为一个CAS,并以此为理论基点对知识网络的内在性质、机制、环境和协同演化等进行了诠释,形成了本研究全新的理论视角。本文业已构建了一个新的分析范式剖析知识网络系统,其技术路线是“蕴含的系统思想—→系统层面—→概念模型→复杂性涌现”,研究表明该架构可以更好地分析与描述产业集群中知识网络这一复杂系统发生、创新、学习和适应等行为的本质。这种研究视角和得出的研究结论对我国学者进一步开展类似研究具有参考价值。(2)基于结构和演化两个维度勾勒出比较清晰的知识网络研究的一般框架。本文提出的研究框架对产业集群中的知识网络给出了较为系统的理论解释,可以对现实中广泛存在于知识网络中的知识溢出、知识转移、知识整合等现象给出较为有力地解释,同时也拓宽了知识网络研究的思路和视角,探索性的研究发现为后续相关研究提供了参考和借鉴。该框架的有效性表现为企业嵌入知识网络,被视为生存于产业集群网络和环境中的能动的行为主体,对于其与大学、研发机构、政府、中介机构等关系的研究,审视之间动态关系和相互作用,有助于认识其规律的内在性。(3)采用严谨的量表开发与精炼技术,基于特定产业集群情境的知识网络结构的实证研究。产业集群中知识网络结构测量问题是知识网络理论研究的基本问题,国内研究知识网络结构测量问题的文献相对较少,本文的研究结果将对知识网络理论体系形成产生积极作用,尤其为未来的实证研究提供有价值的参考。在现有研究的基础上,基于个体层面、企业间层面、网络层面3个层面构建了知识网络结构的测量模型,经预测试、先导测试以及问卷调查等方式对该工具进行精练,并实证检验了该量表的可靠性和有效性。(4)将抽象的动态演化细化为具体的演化分析框架,为知识网络的管理实践提供了具体的策略指导。知识网络研究领域,很多术语和概念尚未统一,研究基础、框架和理论模型正处在发展阶段,缺少一致性意见。本部分提出的分析框架为:“动态演化机制、过程—→动力学模型及仿真→知识网络结构优化”,这一研究框架实质上是基于两个得到普遍认可的基本假设:一是知识网络系统只能在保持物质流、信息流和能量流的开放系统中维持,是一种非均衡有序的耗散结构;二是知识网络系统演化的终极动因在于各子系统(耦合元素)之间的相互作用,关键是非线性相互作用。从这个意义上,本文所构建的逻辑模型对任何企业都具有重要的指导价值。上述研究相当部分内容属于探索性研究,所得结论有助于丰富、完善知识网络理论,并对企业管理实践具有一定的启示和指导意义。但受困于数据采集的限制,加上笔者时间和能力的限制,本文尚存在一些不足之处,希望在今后的研究中能够得到完善和补充。可能的研究拓展是基于知识网络结构量表验证产业集群中网络结构与知识共享、企业创新绩效等变量间的关系。同时,随着数据采集技术的完善,采用更先进有效的检验方法研究知识网络演化结果的测度也值得进一步的探索研究。更多还原

【Abstract】 With the coming of knowledge-based economy,instead of the traditional production factors such as earth,labor and capital,knowledge has exceeded to be the uppermost economic resource and the dominant source of competitive advantage. Therefore,knowledge management has been blooming in the fields of theory and practise.A comprehensive literature study was conducted to find that knowledge network has become a hot issue.Much research on knowledge network is converted from quality and point,simplicity to integration and complexity.Knowledge network is a dominate effect existed in industry clusters,and knowledge spillovers and organization learning provide a foundation for knowledge network.The concept of knowledge network of this paper is different from staff knowledge network.It is a kind of complex network existed in special industry clusters,which has knowledge exchange relation among nodes.Because of knowledge diffusion,cluster knowledge network is also a part of local and global knowledge network.This paper,from the view of complex system,gives a thorough study on network structure and its evolution,by means of an empirical study of industry clusters.The paper research obtains main conclusions as follows:(1)Knowledge network in industry clusters is a complexity adaptive system.The author first use complexity adaptive system theory to interpret the system thought and multi-level facet structure.Then,propose a conceptual model and some evolutionary strategies with high practical ability.Research shows knowledge network has explicit characters of complex system.The first,knowledge network has different kinds of coupling among actors.The second,knowledge network is provided with multi-feedback mechanism,positive or negative,through which system can make information and knowledge flow.The third,the relationship among actors possesses characters of repeated game during a long time.(2)Based on the deep-rooted coupling mechanism,we elaborate the complexity including three levels:elements,ties and evolution.Realizing the attracting complexity emerging from knowledge network seems to frustrate scholars’ interest of modeling,but it also brings us huge challenges.With multi-hierarchy and nesting each other,the structure of knowledge network is complex and different.There are much influence and coupling among the subsystems.Any small perturbation would cause network system catastrophe.The whole behavior pattern emerging from the local and mutual relationship leads network structure to durative change and alternation.(3)The measurement of knowledge network structure can be depicted by four dimensions.An increasing number of attentions to knowledge network in industry clusters are paid by practitioners and scholars,and it is becoming important how to measure knowledge network structure.Through theory analysis,the paper builds a research framework and constructs SEM,and then goes on an empirical research on 117 firms which come from Zhejiang province.The results show that the measurement model has good construct validity and the instrument includes four dimensions:centrality of an actor,strength of a tie,quality of relationship and density.(4)The evolutionary process is provided with some characters of dissipative system,such as an open system far from the balanced state,the non-linear and fluctuation.Limited to narrow understanding of knowledge management,in the past, people focus on special subset or some part in knowledge network system.With the relationship between knowledge management and its environment complex,the shortcomings of narrow understanding are emerging.Therefore,we need an organic method consider knowledge network evolution.The innovation of this part is the author analyses dynamic characters of evolutionary process,from the perspective of dissipative structure theory,and focuses on four aspects,that is,an open system far from the balanced state,the non-linear and fluctuation.(5)We can simulate the evolutionary behavior of knowledge network with the help of artificial neural networks(ANN).With the opinion evolutionary modeling of knowledge network is a kind of system simulation,we have established ANN-based method of evolutionary model and concluded its steps.A neural network has a massively parallel and distributed structure,which is composed of many artificial neurons.And the neurons can be used in applications for information processing.An exact simulation method improves the present research and the simulation shows it is right to carry through dynamics analysis.An intelligent measurement was achieved by neural networks with self-adaptation and self-study,and the results from the simulation are satisfied.Compared with previous research findings,the following aspects are unique points of this research:(1)The paper constructed a new perspective to interpret complex knowledge network system.Based on many relative studies,we use complexity adaptive system to interpret the internal attribute,mechanism,environment and synergy evolution. This paper constructed a new analytical paradigm to interpret knowledge network system,and the technology path is "embedded system thought-system layers-concept model- complexity emerging from network system".The result shows the analytical paradigm can better describe the essence of the appearance,innovation,learn and adaptation behavior of complex knowledge network system in industry clusters.The perspective and the conclusions have referenced value for future research.(2)This paper outlined a general framework for studying knowledge network, based on two dimensions including structure and evolution.The framework can give a systemic interpret for knowledge network in industry clusters,such as knowledge spillover,knowledge transfer and knowledge integration.Meanwhile,the study broadens research path and perspective,and the explorative findings will provide reference and experience for future research.Being an active actor in cluster network and environment,firms are embedded in knowledge network.It is useful to study the dynamic relationships and reactions among firms,universities,R&D institutions, government and agencies,and to understand the internal mechanism.Through the former study,the validity of the framework is tested.(3)Using the purified technique,an instrument of knowledge network structure is developed through an empirical study based on Zhejiang industry clusters.The measurement of knowledge network structure is a radical issue of knowledge network theory and unwonted in relative research in China.The results will contribute great value to forming system info of knowledge network,especially for the future empirical study.The measure model of knowledge network structure is developed based on three levels,including actor level,inter-actor level and network level.Pretest, pilot test,and survey are implemented to purify the items.And,reliability ad validity of the instrument is tested empirically.(4)The evolutionary framework is purified from abstract evolution,which provides some strategies for firms’ management practice of knowledge network.In the research fields on knowledge network,many terms and concepts aren’t unified, and the foundation,framework and theory model belongs to the seedtime.The analytical framework of this part is "evolutionary mechanism/process-dynamic model and simulation-optimizing network structure",and this framework,in fact,is based on two hypothesis accepted by us.One is that knowledge network system is a dissipative structure with non-equilibrium and order;and can maintain only in the opening system with matter flow,information flow and energy flow.The other is that the ultimate evolutionary cause is the non-linear reactions of subsystems(coupling elements).From this significance,the concept model constructed by the paper will contribute great value to any firms.Based on the explorative study,the results enrich the theory of knowledge network,and it is instructive to direct management practice of enterprise.Restricted to time,ability and data collection,this paper has some shortcomings waiting for improvement in future.Possible development is to use the instrument to test the relationship of knowledge network structure,knowledge sharing and innovation performance.In addition,with the improvement of data collection,based on more effective test methods,the measurement of the evolving results is worthy to be further explored.更多还原