Research On Subject Behavior Of Industrial Cluster Innovation Based On Complex Adaptive System Theory

In recent years, with the development of IT technology and t he trend of globalization, industrial cluster has become the core driving force of regional economic development. The force mainly grows out of the innovative ability of industrial cluster. The industrial cluster innovation which develops into networ k is playing a significant role during the formation of regional innovation system, and presents significant characteristics of complex system. In order to improve overall innovation ability of industrial clusters, the laws of evolution of innovation system and the association relationship and laws of relevant innovation subjects in the interacting process should be clarified. While the industrial cluster innovation system is a complex nonlinear system, it is composed of many retroactions and emergence problems which can’t be explained by reductionism theory. Otherwise, traditional research are maily based on fixed structure of economic system, and the relationships between industrial cluster innovation subjects in reality are changing contuniously. It’s hard to precisely depict the reality by fixed structure. In the view of traditional economics inherent limitations in the research of industrial cluster innovation problems, the charateresitics of subjects when they are participating in innovation process are fully discussed in this research. Through the analysis on the characteristics of industrial cluster multi-subject while participating in innovation process, and following the “from bottom to top” paradigm of complex adapative system, the characteristics of “complexity out of adaptability” in the process of multi-subject innovation are depicted. From the perspective of multi-suject the paper reveals the source of innovation advantage and provides reference for related economic decision-making.Based on the collating and summarizing relevant research, the meanings about behaviors from industrial cluster innovation subjects are defined. From three aspects, the characteristics of complex adaptive system are described including basic attributes, evolution process and subject behavior. Furthermore, the idea of application of complex adaptive system theory is extracted. Four basic theories introduced in this paper are knowledge evolution, cooperation innovation, and group behavior and innovation network. Meanwhile, the four tools are the theoretical tools for the study of complex adaptability.On the basis of summarizing the theoretical tools, this paper further abstracts theory into the process that can be studied using quantitative methods by means of mathematical analysis. Based on complex adaptive system feature, the paper illustrates the complex adaptive features including the basic attributes of industry cluster innovation system, its evolution process and subject behavior. The paper points out that the essence of the industrial cluster innovation system is a complex adaptive system. Through the introduction of complex adaptation system theory, respectively, from aspects of single subject, dual subjects and multi subjects, the paper describes the adaptive learning behavior, interaction and cooperation behavior, and group behavior. With stimulus response model, logistic growth model and constrained generating procedure model, the principle of the above core behaviors is summarized. According to the different levels of the main behavior characteristics, the paper selects evolutionary computation, stochastic differential equations, multi-agent simulation, complex network, and case study as analysis methods. After all, the theoretical concepts become visual, operable specific behavior models, building a bridge between theoretical assumption and quantitative model.First, the paper analyzes the adaptive learning behavior of industrial cluster innovation subject from the perspective of single subject. Based on the characteristics of the subject, the quantitative model is constructed by using the NK model which is widely used in biological genetics science. Regarding the adaptation degree of the landscape as the main characterization, the paper considers the impact factors of learning process as the input parameters of the model in order to get the results of fitness landscape under different conditions. By adjusting the parameters, the paper analyzes the degree of variation quantitatively, and then describes the complex adaptability of the behavior of single subject in innovation process. The results show that the innovation of the industrial cluster is not static, but dynamic, which means the subject learn and grow through self-organization and retroaction to adapt to changes in the environment and seek their own survival and development. The results of quantitative model also show that the overall fitness value is not only related to the elements of the state level, but also is related to the elements of the nonlinear interactions(the k value of model).Secondly, considering the interaction between dual subjects, the process of interaction and cooperation behaviors between the industrial cluster innovation subjects and the balance between the subjects is analyzed in this paper. Taking cooperation innovation cost, cooperation innovation benefits and government support into consideration, this paper also uses two key tools, evolutionary stable strategy and replicated dynamic equation, to construct the classic evolutionary game model. Meanwhile, the adaptability of industrial cluster innovation subject is discussed and regarded as the system noise in setting up stochastic differential equations. The results show that without consideration of the adaptability of subject, the collaboration between industrial cluster innovation subjects depends on factors including cooperation innovation cost, cooperation innovation benefits, and government support. Certain conditions can promote cooperation between subjects, namely the steady state of game. Considering the industrial clusters innovation subject adaptability, it could influence the decision-making process and the time taken by both sides in the game strategy. Meanwhile, the adaptability could increase the complexity of the game strategy and display the nonlinear character more obvious. The subjects involved in the game would make their own decisions of their own benefits, but the final steady-state strategy would remain unchanged.Finally, adaptive learning behavior of single subject and i nteraction and cooperation behavior of dual subjects are taken as the inherent motive forces. From the macro layer of industrial cluster innovation system, this paper analyzes the group behavior emergence effects of multi-subject. By defining the behaviors of enterprise, university, research institution, government and intermediary, the Repast simulation platform is used to simulate the scale effect of the group behavior emergence. By describing the characteristics of complex network topology structure and the modified BA network model, the structure effect of group behavior emergence are simulated. The results show that when the industrial clusters innovation resource endowments and environmental uncertainty are relatively high, the majority of subjects would take cooperation innovation strategy to promote the innovation degree of system, and it would also enlarge the scale effect of the system emergence. In the procession of the subject’s continuous evolution that presents scale-free characteristics, the subject would keep adjusting its behavior regulation to adapt to the environment in order to optimize the innovative advantages of the network structure, while the structure effect of emergence would come out as a result.The theoretical significance of the study on innovation subjects lies in the application of current complex adaptive system theory and traditional industrial cluster innovation theory. By using “complexity out of adaptability” idea and taking the perspective of multi-subject, the complex adaptive research framework of industrial cluster innovation is established, and it is not only beneficial to expanding the industrial cluster theory and innovation theory, but also extends the scope of application of complex adaptive system theory. The practical meaning is that from the angle of multi-subject, industrial cluster innovation could be understood profoundly, which is also more close to reality. Besides, Zhongguancun industrial cluster is taken in the case study part, so as to enhance the underst anding of industrial cluster innovation mechanism. A better way is suggested for the industrial cluster subjects, including enterprises, government, research institutions, universities and other intermediaries to participate in innovation activities. Some policy suggestions are provided for fostering industrial cluster to possess competitive advantage.