基于粒子群优化智能体的创新行为与产业集聚研究 ——以广深科技创新走廊电子产业为例

On the Innovation Behavior and Industrial Agglomeration Based on PSO Agents: A Case Study of Electronic Industry in Guangzhou-Shenzhen Science and Technology Innovation Corridor

摘要:

基于企业间复杂的交互网络和知识迁移对企业创新收益的影响，采用粒子群优化（PSO）算法构建出理论模型并予以解释，将个体企业视为产业集聚环境下的微粒，并能感知、处理和影响外部环境的行为主体。模拟生物行为，通过对个体企业微观层面的创新行为进行分析，形成宏观层面的群体行为数据集，利用仿真实验方法探究企业创新行为与产业集聚之间的关系。仿真结果表明：首先，产业集聚过程中会形成以创新企业为中心的虹吸效应，这种分布由于企业创新行为选择差异造成存在多个中心的趋势；其次，创新企业是一个多智能体，其创新行为选择受到创新规则集和知识库的影响；再次，企业创新形成的知识溢出效应是产业集聚的动力系统，该动力系统的运作后果导致中心节点与其他节点最短路径之和最小；最后，产业集聚还会受到创新制度环境以及竞合行为的影响。

Abstract:

Based on the influence of complex interactive network and knowledge transfer among the enterprises innovation benefits, in this paper, the PSO algorithm is used to construct and explain the theoretical model, and the individual enterprises are regarded as the particles which can perceive, process and influence the external environment. Through simulating the behavior of biology, as well as analysis of innovation behavior of individual enterprises, the macro level data that relating to the group behavior are formed, the relationship between innovation behavior of enterprises and industrial agglomeration is explored by means of simulation experiment. The simulation results show that: First of all, the siphon effect is formed in the process of industrial agglomeration, which is caused by the trend of multiple centers due to the difference of innovation behavior. Secondly, the innovation enterprise is a multi-agent whose innovation behavior choices are affected by the innovation rule set and repository. Thirdly, the knowledge spillover effects are formed by enterprise innovation, which will change the dynamic system of industrial agglomeration, and the operation consequence of this dynamic system leads to the minimum sum of the shortest paths between the central node and other nodes. Finally, industrial agglomeration will also be affected by institutional environment and competing conduct.