Research In Diffusion Models Of Innovation Based On Complex Social Networks

Spread of new product, techonology, thought, method, behavior, strategy, culture, fashion and other innovations in social economic system, which is significant to the control and development of social economic system, always draws much attention. The process of spread occurs in social group, which could be regarded as dynamic processes of mass individuals who make decision by contact with and learn from each others. Therefore, it is difficult to comprehend and predict. Researchers try to abstract the key factors of the complex dynamic process by means of different diffusion models of innovation, in order to provide theritical instruction to understand and control different diffusion phenominons of innovation.Complex network theory regards social economic system as relation network between individuals, provides research perspective on dynamic interation process between individuals, and shows that there are some new features of the topic under perspective of network, including: (1) individual interation could change result of innovation diffusion, (2) network effect influences diffusion process greatly, (3) different network structure could lead to different diffusion process. These features are not fully explained by current findings. Therefore, this dissertation focuses on four aspects as sally port of analysis, namely (1) pay attention to the micro-base of innovation diffusion and develop diffusion models on individual level which describes the social interaction and bounded nationality of social individual; (2) mass behavior on individual level lead to diversity of population effect on group level which should be considered in diffusion models; (3) diffusion models should consider feature of social networks in order to analyze influence of network structure; (4) real networks evolves with interaction process of individuals which should be considered into diffusion models.This dissertation analyzes the influence of network structure, relative dominace of innovation, individual interaction, population effect and dynamic evolving networks on innovation diffusion based on perspective of network. The research extends traditional theory of diffusion, promotes the application of complex network theory, and provides brand new perspective to topic of innovation diffusion in social system.The main works and results are concluded as follows:1. First, this dissertation comprehensively introduces the main research background and significance, points out that complex social network theory is important for understanding social economic system, emphasizes that perspective of network could transform traditional research of innovation diffusion, consequently denmonstrates the meaning of topic. Then the main contents and structure are summarized as well as the academic innovation of this dissertation.2. This dissertation summarizes the research findings up to date: Firstly, clarifies the topic scope and theoretical basis of innovation diffusion, concludes existing diffusion models and states that these models are inclined to prefer to individual level than population level, which proves the micro basis of diffusion deserves more attention in this dissertation. Secondly, conludes the theorictical finding of complex network theory by statistical feature, network generation models and diffusion models, which offers research perspective and analysis tools. Thirdly, elaborates the promotion of complex social networks on research of social economic system and diffusion, which extends research field by combination of different branch of knowledge. Finally, points out main research problems based on several limitations of existing research.3. Provides the framework for next chapters based on perspective of complex social networks, constructs the natural selection process of innovation based on topology of individual networks and relative dominance of innovation: an innovation of fitness r diffuses in networks and occupies the whole network with probability f ; while the higher the fitness r is, the bigger the fixation probability f is. The results show that the diffusion process of innovation is influenced by topology of network and behavioral model of individual. Heterogeneity and the lower connectivity of network favor selection of innovation while positive degree correlation and small world effect go against it. Moreover, behavioral model of firm shows distinct influence. Conscious behavioral model favors innovation of higher fitness while unconscious behavioral model even suppresses selection of innovation. These conclusions state that the structural difference of network will remarkably influences the dynamics on networks, which differ from results of traditional research of innovation diffusion.4. Constructs the network diffusion model with"neighbour effect function"which describes the influence of neighbors on individuals'choice, analyses the equilibrium and influencing factors under the collective diffusion model of external and local factors. The result shows that the property of neighbor effect function determines the equilibrium of diffusion; existence of multi-equilibrium increases the uncertainty of diffusion; neighbor effect and structural characteristics of complex social networks has great influence on adopters'equilibrium. Thus, network structure and interaction of neighbors should be considered in management decision.5. Model the characteristics of differential choice by anti-coordination game and analyse the process and influencing factors of innovation diffusion in complex social networks. The results show that this kind of diffusion could always occur. With the increase of risk dominant of adopted strategy, adopters increases in non-heterogeneous networks but may decrease in heterogeneous networks. Moreover, property of network structure could influence number of adopters significantly. Effect of small world depends on whether strategy is risk dominant while effect of connectivity and heterogeneity depends on degree of risk dominance in addition. Therefore, it is effective to adjust both payoffs of individuals and relationships between individuals to manage the diffusion with property of differential choice such as holiday resort, luxury product,"free-riding"behavior and fashion.6. Given condition of dynamic evolving networks, the dissertation models the social individuals'features of bounded nationality and strategic choice by coordination game, thus the co-evolving model of innovation diffusion and network structure of individuals is developed. The results show that influence of individual behavior and network structure on diffusion is determined by relative dominance of innovation. Moreover, structural feature of complex networks emerged from individuals'accumulative behavior. There exists interactive relationship between diffusion and network structure.The primary innovations include:1. This dissertation puts forward the diffusion model of innovation based on complex social networks and explores the influences of network structure on diffusion of innovation with relative dominance. Exsting findings believe that innovation with higher relative dominace diffuses more easily while the results of this dissertation show that this kind of diffusion could be promoted as well as prevented by variety of network structure. Therefore, the structural diversity of individuals'network would influence innovation diffusion greatly.2. This dissertation constructs the network diffusion model based on neighbour effect. Different from existing diffusion models, the"neighbour effect function"describes the influencing mechanism of neighbors on individuals'choice. The results show that the equilibria of diffusion depend on the property of neighbor effect function. Structural characteristics of complex social networks have great influence on adopters'equilibrium.3. This dissertation develops the diffusion model with the feature of differential choice of individuals, which explores, in perspective of strategic interaction of individual and property of network structure, the overall characters of diffusions with property of differential choice such as holiday resort, luxury product,"free-riding"behavior and fashion. The results show that this kind of diffusion could always occur. Risk dominant of adopted strategy and property of network structure could influence number of adopters significantly. These results are different with existing findings which suppose individuals prefer to choose same strategy with others.4. This dissertation try to extend perspective of static networks to dynamic evolving networks, provide co-evolving model of innovation diffusion and network structure of individuals and explore the frontier question of complex network theory: relationship between network structure and duffusion dynamics. The results show that a network payoff is determined by individual's payoff as well as network structure. Bilateral influence between diffusion process of innovation and network structure of individuals is significant which explains structural feature of complex network by means of individual-level factor.