基于复杂网络的产业创新扩散与演化仿真研究

发布时间：2018-11-08 19:34

【摘要】：复杂网络理论是当前科学理论的重要研究热点,其广泛应用的根本原因之一在于跨学科的良好交叉性。在复杂网络研究中,通过建立合适的网络模型来分析网络统计性质的产生机理并预测网络的行为,可以使人们能够更加合理地控制整个网络的性能和设计更加合理的网络结构。而产业网络本质上是一个以资本、知识及技术为基础进行交流合作,在各种因素的共同促进下形成一个动态的、开放的复杂关系系统,具有复杂网络的结构特征。本文在复杂网络理论以及产业演化研究的基础上,首先定性分析了产业演化机制,然后分别从横向的创新扩散和纵向的产业结构演化两个方面进行了建模与仿真,最后以生物科技产业为实例,说明模型的适用性及对政府策略的启发性。本文主要研究工作如下:(1)在创新扩散层面,综合节点自身异质性、网络效应以及政策市场环境三个方面的因素影响,在EWA学习模型的基础上,对参数设置和效用函数都进行了改进,使其针对产业网络中企业主体的创新策略采纳决策更具适用性;然后,运用理论推导的方式计算出初始扩散阈值,当满足条件时产业中创新将会出现;最后,合理的设置参数值,运用matlab,对影响扩散的三个因素:网络结构、创新及学习失败率以及政策扶植力度进行了仿真与分析,并且就结果对政府如何根据产业情况采取相关政策给出建议。(2)在产业拓扑结构层面,首先,设置了原产业网络中节点重连及断边机制以及新节点加入和连接机制;然后综合创新扩散过程,建立了一个混合的产业复杂网络演化模型并设置了仿真规则;最后,运用matlab仿真分别研究了重连及断边速度、新节点加入速度、择优和随机连接概率等因素对创新扩散均衡和网络拓扑结构的影响,并就结论给出政策建议。(3)以美国生物科技产业为例进行了案例分析,从理论上验证了模型的适用性,并且针对产业发展情况进行总结,对政府引扶持国内高新技术产业健康发展具有较大借鉴意义。
[Abstract]:The theory of complex network is an important research hotspot in current scientific theory. One of the fundamental reasons for its wide application lies in the good intersectionality of disciplines. In the study of complex networks, by establishing appropriate network models to analyze the generation mechanism of network statistical properties and predict network behavior, people can more reasonably control the performance of the entire network and design a more reasonable network structure. In essence, industrial network is a dynamic and open complex relationship system based on capital, knowledge and technology. It has the structural characteristics of complex network. Based on the theory of complex network and the research of industrial evolution, this paper first analyzes the mechanism of industrial evolution qualitatively, and then models and simulates the horizontal innovation diffusion and vertical industrial structure evolution. Finally, take the biotechnology industry as an example to illustrate the applicability of the model and enlighten the government strategy. The main research work of this paper is as follows: (1) on the basis of EWA learning model, the main research work is as follows: (1) in the aspect of innovation diffusion, considering the heterogeneity of node itself, network effect and policy market environment, The parameter setting and utility function are improved to make it more applicable to the innovation strategy of the enterprise in the industrial network. Then, the initial diffusion threshold is calculated by theoretical derivation. When the conditions are satisfied, innovation in the industry will appear. Finally, by setting the parameter value reasonably and using matlab, the paper simulates and analyzes the three factors that influence diffusion: network structure, innovation and learning failure rate, and policy support. And on the basis of the results, some suggestions are given on how the government should adopt relevant policies according to the industrial situation. (2) at the level of industrial topology, firstly, the mechanism of node reconnection and disconnection and the mechanism of joining and connecting new nodes in the original industrial network are set up. Then a hybrid industrial complex network evolution model is established and simulation rules are set up by synthesizing innovation diffusion process. Finally, the effects of reconnection and breaking speed, new node adding speed, optimal selection and random connection probability on innovation diffusion equalization and network topology are studied by matlab simulation, respectively. Finally, some policy suggestions are given. (3) A case study of American biotech industry is given, which verifies the applicability of the model in theory, and summarizes the development of the industry. It is of great significance for the government to support the healthy development of domestic high-tech industry.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：F124.3;F276.44

【参考文献】