Resilience Analysis And Application In Complex Networks Based On Individual Optimal Decisions

Networks with some or all characteristics of self-organized,self-similarity,attractor,small world,scale-free are called complex networks.Due to the complexity of network structure,com-plex networks and relevant studies could help people understand the complexity of things in nature and in real life,and could also provide new perspectives and methods for researches on complex systems.Therefore,research on complex networks and its characteristics is a hot topic for schol-ars,which includes study on resilience of complex networks.Resilience describes the ability of systems to resist external influence and the ability to recover quickly from the influence.So far,the concept of resilience has been applied to more than ten fieldsDue to dynamic characteristics of systems are involved,there are big challenges in the re-search of resilience.And resilience studies in complex networks are still in its early days.Ex-isting works generally have several drawbacks:1)they cannot quantify and analyze the network resilience accurately;2)they could only apply to complex systems with special dynamic character-istic;3)they only consider resilience pattern of the whole network,but could not analyze the role of individuals in network resilience;4)due to lack of analysis for individuals,existing works have weak interpretability for network resilience.Given that,this paper proposes a research method for complex networks resilience based on individual optimal decisions,the main contributions are listed as follows:1.The role of individuals(nodes)in resilience of complex networks is studied.Consider the huge difference of contributions for nodes in network resilience,a centrality index named resilience centrality is proposed based on dynamic characteristics of networks,with which to quantify the difference in importance of nodes in network resilience effectively.Through an-alyzing the equation of resilience centrality,the paper finds that the role of nodes in network resilience is highly dependent on degree and weighted nearest neighbor degree of nodes.Re-silience centrality helps people understand network resilience from the angle of individual,and establishes the relation between role of nodes in network resilience and network struc-ture,and provides theoretical basis for the research on network resilience.Finally,simulation results based on real data demonstrate the performance of resilience centrality index.2.The problem of customer churn which affects the resilience of telecommunication network is studies.Consider the impact of customer churn on resilience of telecommunication net-work,a theoretical framework is proposed,which mitigates the problem of customer churn and strengthens the resilience of telecommunication network.The theoretical framework analyzes customer churn from two perspectives:nodes of customer and nodes of service packages.Based on the difference of the impact of customers on telecommunication net-work,the optimal decision of classification for customer churn is provided,with which to identify those customers who are prone to churn and also have relatively great influence on operators.Based on impact of service packages on customers,resilience patterns of telecom-munication network is analyzed,and the conclusion that some features of service packages could affect the failure of nodes for customers is provided.The framework analyzes re-silience in telecommunication networks well,and provides theoretical guidance for several practical problems such as the reasonable design of service packages and the performance improvements of resilience of telecommunication network.3.The problem of voltage collapse which affects the resilience of power network is studied.Consider the impact of voltage collapse on resilience of power network,an analysis method based on network reduction and equivalent is proposed.In order to quantify the ability of different load buses to affect the resilience of power network,a stability index for voltage of load buses in power network is proposed.In order to figure out the mechanism of the impact of voltage collapse on resilience of power network,a network reduction method and the concept of effective distance is proposed based on matrix calculation,which predicts the critical point of voltage collapse,and discovers that the impact of voltage collapse on resilience of power network is mainly determined by initial power intake and local connec-tion of the perturbed load bus.A network reduction based on path searching is proposed,which identifies the role of nodes and links of power networks in network resilience,builds the long-sought relations between resilience of power network and network structure,and provides theoretical guidance for a series of practical problems such as effective design of power grids and resilience enhancement of power grids.Finally,the simulation results based on IEEE standard cases demonstrate the good performance of the method.In the end,the thesis is concluded and some future research works are discussed.