Research On Multi-layer Network Innovation Propagation Model Based On Agent

This thesis aims to study the agent-based multilayer network innovation propagation model and explore the propagation mechanism of innovation in complex network structures.Innovation includes new products,new technologies,new ideas and new behaviors,and innovation communication has important innovation value and development significance in modern society and economy.However,the traditional single-layer network model is difficult to accurately describe the extensive complex system propagation process of innovation in the real world,and the current behavioral decision-making model for studying innovation dissemination often ignores the influencing factors at the level of individual concepts of decision-making,which can easily lead to the deviation of the overall innovation communication decision due to the slight fluctuation of group concepts.Therefore,this thesis takes the adoption process of enterprises releasing innovative products as an application scenario,analyzes the innovation propagation process into a three-layer coupling multi-layer network model such as physical layer,concept layer and consensus layer through the concept of concept dynamics,and puts forward the concept of preference entropy to establish a behavioral decision-making model,and finally proposes an agent-based modeling method to study the process of innovation propagation in multi-layer networks from a microscopic perspective.This thesis completes the following work:(1)Firstly,the existing typical network model is developed,the different topological models of multi-layer networks are studied,and finally according to the concept of conceptual dynamics,a multi-layer network model based on the coupling of three layers such as physical layer,conceptual layer and consensus layer is proposed,and the coupling relationship between the three layers is explained through cognitive strategies and consensus mechanisms.(2)The concept of preference entropy is proposed,and the characteristics such as inconsistency and stability of attitude elements of Boltzmann entropy and Gibbs entropy are studied,in which the main theoretical basis is based on the causal attitude network model,and the Ising model based on Glauber dynamics is studied.Then,based on the five stages of Rogers’ innovation adoption,an entropy-based behavioral decision-making framework is constructed.Finally,the influence of different dependence parameters β,node threshold ι and connection strength ω on the Ising model is analyzed experimentally.(3)Using the agent-based modeling and simulation experimental method,the Netlogo simulation platform was selected to complete the experimental analysis of the actual interaction network(physical layer)of the individual,and then the causal attitude network in the concept layer was modeled through R language,and different cognitive strategies were compared and analyzed.Finally,the influence of different intensities of external environmental stimuli on innovation dissemination is studied.This thesis shows that the multi-layer network model based on conceptual dynamics has better explanatory properties for the innovation propagation process than the single-layer network.The innovative propagation of agent-based scale-free network structure is more in line with the actual data,and the subsequent experiments adopt the scale-free network structure,and it is concluded that the agent-based model is better than the linear threshold model and can better describe the process of innovation propagation in practical application scenarios.In addition,this thesis conducts experimental analysis according to consumers’ conceptual attitudes towards new energy vehicles,and concludes that the average strategy rule is more in line with the actual trend.Finally,the impact of external event stimuli on innovation communication under different degrees is studied,and it is concluded that under a certain degree of external policy stimulation,it is conducive to the rapid development of innovation communication in the early stage,and when it exceeds a certain limit,it will play a negative role.