Study Of Individual Interaction And Cooperation On Complex Networks

Complex network is an interdisciplinary science for complex system, including maths, physics, biology, economics, social science, and so on. Almost all complex systems can be represented by graphs with individuals denoted by nodes and interactions by links, such as Internet, railway networks, air line network, power grids and all kinds of biomolecular network. Besides, human society itself is also a complex network. Therefore, the study of complex networks is meaningful for both science and practice.A system is called complex system not only because of its large scale, but also the heterogeneous and complex interaction between individuals. In general, the aim of the complex networks is to reveal the macroscopic behavior of the complex systems though analyzing the microscopic mechanism and process. In real life, many complex systems are concerned with human being, such as social networking services and financial systems. Thus, exploring human behaviors characteristics, and understanding the relation between microscopic human behaviors and society system’s movement is of great significance. This thesis will study the evolutionary game and human dynamics on complex networks from the perspective of individual interplaying.In the spatial public goods games, individuals interact with each other by playing games in more than one group, and evolve by learning strategy form their neighbors. This thesis investigates the impact of social diversity on the evolution of cooperative behavior in spatial public goods games where individuals are nodes of a BA scale-free network. Due to the existence of social diversity, we assume the allocation of payoff on individual is proportional to the power of its degree, i.e. kα. There the power exponent a is an adjustable parameter that controls the degree of diversity of individuals’profits. We found an optimal value of a which yields the highest level of cooperation, suggesting that a suitable degree of diversity among individuals can promote the emergence of cooperation.In this thesis, we also focus on non-Poisson property in temporal and spatial distributions of human behavior, of which power-law distribution is the most representative law. But we notice that many distributions deviate from the power-law. Based on the widespread individual interplay in our daily life, we propose a minimal human dynamics model, which is different with the traditional queuing theory model and dose not incorporates mechanisms such as interest and rhythm yet. This model explains the emergence of non-Poisson temporal and spatial properties in human behaviors from a novel perspective. This simple model well reproduces almost all empirical observations, including exponential distributions, power-law distributions, bimodal distributions and so on. This model bridges priority queues and punctuated equilibrium, both of which are mechanisms leading to scaling laws. These reveal that individual interplay drives system self-organize to critical state, giving rise to the emergence of power laws. Non-Poisson property in individual level may be the reflection in the micro level of the critical state. Moreover, this thesis investigates impact of topological properties on dynamics and statistical properties by simulating on different networks. We find that the strength and topological structure of individual interaction could provide an explanation for the deviation of power law, which indicates that this deviation could be an essential property of the social complex system. This model first obtains multi-modal power law on a real social network, subverting previous explanation that different power exponents due to different mechanisms. Our modeling and analysis is likely to shed light on non-Poisson phenomena in many complex systems.