科研合作网络的知识扩散机理研究

Knowledge Diffusion Mechanism of Scientific Cooperation Network

知识扩散是知识生产过程中的核心环节,对知识继承和知识创新具有重要作用。本文结合知识在科研合作网络中的流动特征,引入复杂网络理论,构建知识扩散模型,模拟知识在科研合作网络中的扩散过程。考虑不同个体的知识自我增长以及知识吸收能力,采用网络平均知识水平、知识扩散速率、知识均衡程度等作为衡量知识扩散效果的评价指标,探究不同合作网络结构、知识遗传继承和知识变异重组与知识扩散的动态关系。研究显示:知识在合作网络中的知识水平、扩散速率、分布均衡程度很大程度上取决于网络拓扑结构的动态变化,网络的随机化程度越大,知识扩散的速度越快,知识的分布越均匀;合作网络的规模越小,专家高知识溢出效应越显著,越能促进知识的有效扩散;知识继承吸收和知识自我创新对知识扩散的影响在某一时刻可达到最佳均衡状态。研究合作网络中各影响因素对知识扩散的震荡作用,有利于形成更稳健的合作模式,发挥科研合作的最大效能。

The study on the mechanism of knowledge diffusion is playing an increasingly important role in communication and cooperation. Many efforts contribute noticeably to improving the understanding of knowledge diffusion in all sorts of networks. Researchers have concluded some key elements that affect knowledge diffusion, such as the absorptive capacity of receivers, the stickiness of knowledge and network topology structure for knowledge diffusion. Successful knowledge diffusion is difficult to achieve and some important features of real scientific collaboration network are unsuccessfully explained by simple model networks. Complex networks, which can well mimic the interactions between individuals in real systems, provide a substrate for the researchers to study many interesting dynamical processes. Taking scale-free small-world network as basic model, this study presents a knowledge diffusion model by considering the differences of self-knowledge growth and knowledge absorption capacity βi, where βi is correlated with the knowledge stock of receivers, the knowledge spillover effects and the correlation intensity of both. Scientific research collaboration is a process of absorbing each other＇s knowledge and co-creating new knowledge, when knowledge overflow from the high knowledge level to the low, existing genetic inheritance of its original knowledge manifested as restorative combination of constitution elements and fixed structure, and abrupt variation of node represented as a critical behavior of thinking subjective initiative so as to realize the innovative elements reorganization or restructuring. Accordingly, applying the average knowledge levelμ（t）, the knowledge diffusion rater （t）, the knowledge balance degree σ （t） and knowledge transient increment ρ （ t ） to measure the growth and diffusion of knowledge and combining with the sticky phenomenon of knowledge, we study numerically the dynamic relation of knowledge diffusion with network topology structures, knowledge genetic inheritance and abrupt variation, knowledge spillover effects of experts. Results indicate the following conclusions. 1） The average knowledge level, diffusion rate and balance degree of knowledge in collaboration network depend on the dynamics network topology, the higher stochastic diffusion probability of the network is, the higher the diffusion rate is and the more the uniform distribution of knowledge stock has. Knowledge level has a rapid increase up to a maximum value in the earlier time based on random network and the diffusion advantage is no longer obvious following repeat exchange of knowledge in the network, yet the diffusion characteristics in the small-world network are more in line with the reality of scientific cooperation network. 2） The smaller the scale of the cooperation network is, the more significant the spillover effects of experts are, and the more effectively the knowledge diffusion is promoted. With the spreading of knowledge, v （ t ） increase first and then remain the same, ultimately which has nothing to do with the network scale. The standard deviationtr （ t ） of knowledge stock in different network scales obtain a single-peaked curve and the smallest network generates the uniform distribution of knowledge level. 3） Genetic inheritance and abrupt variation work on the performance of knowledge diffusion in some different degree and achieve a state of equilibrium at a certain moment, and when purely rely on one element to facilitate the knowledge diffusion, the growth of knowledge level is far less than that when the two exist together. Meanwhile, the genetic inheritance mainly affects the early evolution stage of knowledge diffusion and abrupt variation has a bigger stock in the later period. Research on the influence factors of the cooperative network on knowledge diffusion, which is conducive to the formation of a more robust cooperation model, maximizes the effectiveness of scientific research cooperation. Through identifying the innovation tension and knowledge growth advantage in the process of scientific research cooperation to amplify the benign interaction between team members and separate the differences that hinder cooperation and exchange, more effective knowledge exchange strategy and innovation can be achieved. 9 figs. 1 tab. 22 refs.