基于多Agent系统的科研合作网络知识扩散建模与仿真

Modeling and Simulation of Knowledge Diffusion in Scientific Collaboration Network Based on a Multi-agent System

科研合作网络的知识扩散主要受知识溢出和知识创新这两个过程的影响,据此提出知识扩散的机制。采用多Agent系统建模方法构建科研合作网络知识扩散仿真演化模型,分析了网络结构、知识溢出效应和个体知识创新能力对知识扩散的影响。通过知识扩散效果评价指标的对比分析,得出如下结论:科研合作网络的拓扑结构对知识扩散效果造成影响,BA无标度网络结构优于其他网络结构(规则网络、小世界网络、随机网络);知识溢出效应主要影响知识扩散前期,随着知识溢出效率因子的递增,网络平均知识存量总体震荡上升,知识扩散速率递增,同时网络知识存量分布均衡度递减;个体知识创新能力主要影响知识扩散后期,个体知识创新能力因子大的网络表现出较强的平均知识存量增长,也加剧了网络知识存量的不均衡分布。

It has been proposed that knowledge diffusion in scientific research cooperation networks is affected mainly by knowledge spillover and knowledge innovation. On the basis of this idea, this paper proposes a knowledge diffusion mech‐anism. This multi-agent system modeling method is used to build a simulation evolution model of knowledge diffusion in scientific research cooperative networks. Thereafter, the influence of network structures, knowledge overflow effects, and individual knowledge innovation ability on knowledge diffusion is analyzed. Through a comparative analysis of evaluation indexes of the knowledge diffusion effect, the following conclusions are drawn. The topology of scientific research cooper‐ation networks has an impact on the knowledge diffusion effect, and a BA scale-free network structure is superior to other network structures(regular network, small world network, random network). The knowledge spillover effect affects mainly the early stage of knowledge diffusion. With the increase in the knowledge spillover efficiency factor, the average network knowledge stock increases in oscillations, the knowledge diffusion rate increases, and the balance degree of network knowledge stock distribution decreases. Individual knowledge innovation ability affects mainly the later stage of knowl‐edge diffusion. Networks with large individual knowledge innovation ability factors show strong average knowledge stock growth, which also aggravates the unbalanced distribution of knowledge stock among individuals.