Study On Trust Model And Evolution Mechanism For Social Networks

In recent years,the study on trust mechanism of social networks has been a research hotspot.Social network is a kind of typical complex network and it evolves with time.New friends join in social networks,and old friends fade out them constantly.What's more,a variety of complex social activities occur in them.In social networks,trust is defined as the expectations and beliefs that a person to another person.Therefore,the trust of social networks is a kind of interpersonal trust between people.The interpersonal trust will change according to the changes of environment.However,the individuals can make trust stable through social learning.Trust degree can be obtained through the rational calculation and the cognition of emotion.Interpersonal trust exists in interpersonal relationships,and refers to the direct relationship in the process of communication between people.Therefore,this interpersonal trust relationship is the core of social networks.It can not only steady the social relationships,but also solve the complex problems of social relationships.Social networks have a typical evolution characteristic,and the environment that social networks exist in has a typical ecological feature.The trust evolution of social networks presents the self-organizing process from "disorder" to "order".In recent years,some scholars have done research work in the evolution,complexity and fuzziness of entities' trust behaviors in social networks.These research results have effectively promoted the development of the theory of dynamic trust evolution,and greatly enriched the further understanding for dynamic trust evolution.In order to study the trust mechanism of social networks,the research contents of this paper include the following four aspects.(1)In order to improve the trust level and the convergence rate of social networks,this paper introduces complex networks theory into the trust model of social networks.In addition,this paper studies the trust evaluation and search problem based on small-world and scale-free features of complex networks.In the process of search,considering the search efficiency and the credibility of nodes in networks,this paper proposes a new idea to solve the problem of trust searching strategy for the entities in social networks.(2)According to the problems that most of the trust measurement models fail to consider the timeliness of historical behavior data,and cannot deal with strategic dynamic changing behaviors of malicious nodes,this paper studies the dynamic measurement method that historical behavior data combines with trust attributes to build a multidimensional dynamic trust measurement model based on the forgetting curve.Through introducing the direct trust and recommendation trust,this paper establishes trust measurement model of social networks,and builds incentive mechanism to encourage nodes to select the strategies with high trust level.(3)According to the evolution of social networks,and the environment that social networks exist in has a typical ecological feature,this paper establishes trust evolution model of social networks.Because Logistic equation can reflect the self-organizing characteristics of complex networks,this paper introduces it into the trust evolution model of social networks based on Logistic model.In addition,this paper introduces multi-strategy idea into trust evolution model to accurately simulate the evolutionary trend of trust relationships between entities.Through adding incentive and punishment factor,this model can make a social network converge to an evolutionary stable state more quickly,and inspire entities to select the strategies with high trust degree effectively.(4)This paper proposes a trust penalty function of social networks based on redemption mechanism according to the On-off malicious attacks in credit systems of social networks.Once a malicious node attacks other nodes,the trust value of the malicious node will be punished.With the increasing of attacks,the punishment strength for malicious node will also be increased.However,if the behaviors of node recover to the normal state,its trust value will be redeemed(the weight of redemption is less than the weight of punishment).Therefore,this method can effectively improve the ability and the sensitivity of system to resist the On-off attacks.This paper studies the trust evaluation mechanism of social networks based on the characteristics of complex networks.In addition,the trust measurement model and the trust evolution mechanism of social networks are researched based on the psychology and biological evolutionism theories.Furthermore,the evolution of trust relationships between entities is studied.This paper establishes multi-strategy trust evolution model of social networks on the basis of the relationships between social individual and environment.What's more,the incentive and punishment mechanism is built to optimize the performance and to promote the trust level of social networks.