创新网络空间格局形成机理及演化分析——基于多元邻近的Agent仿真

Formation Mechanism and Evolution Analysis of Innovation Network Space Pattern——Based on Multivariate Proximity Agent Simulation

针对创新网络空间格局如何形成及演化的问题,基于多元邻近视角剖析创新网络空间格局形成机理,运用Agent理论设计仿真算法,并基于Netlogo构建创新网络空间格局演化仿真平台。通过模拟实验发现:(1)多元邻近更有利于创新网络区域内、外部合作共生,且(跨区、内部)合作边数与区域面积、毗邻距离呈正、负相关;(2)社会邻近或技术邻近更有利于创新网络产生(绝对)空间集聚,且集聚态势与区域位置中心性呈正相关;(3)创新网络空间格局呈现随机型至斑块型(或均匀型)的动态演化趋势,且社会邻近或技术邻近更有利于创新网络产生均匀型空间格局,地理邻近或多元邻近更有利于创新网络产生斑块型空间格局;(4)地理因素作用下的创新网络(空间格局)仍保持幂律分布特性;(5)斑块型空间格局的创新绩效高于均匀型,且多元邻近下斑块型空间格局的创新绩效最高。

In order to reveal the question that formation and evolution of innovation network spatial pattern, we analyze the forma tion mechanism of innovation network spatial pattern based on multi adjacent. And agent theory is used to design the simulation algorithm and innovation network spatial pattern evolution simulation platform is built based on NetLogo. Through simulation experiments, we find： ①Multiple proximity is more conducive to cooperation of the internal region and external re gion of innovation network, and cooperation （cross, internal） number of edges is positive correlation to the regional area while negative correlation to the adjacent distance; ②The social proximity or technology proximity is more conducive to in novation network （absolute） spatial agglomeration, meanwhile the agglomeration and the center of regional position are positive; ③ Innovation network spatial patterns presented the dynamic evolution trend that random type to plaque type （or uniform type）, the social proximity or technology proximity is more benefit to innovation network （absolute） engender u niform spatial patterns, while geographical proximity or multiple is more advantage to innovation network produce the plaque type spatial pattern; ④Innovation network（ spatial pattern） still maintain a power law distribution features under the effect of the geographical factors; ⑤The plaque type spatial pattern of innovation performance is higher than the uni form type, and under the multiple proximity innovative performance of the patch type spatial pattern is highest.