| The advance of science and technology has pushed human beings into a new information era.Information and information technology have changed our lifestyle in many ways,which have made them the most dynamic factor of production and the most important strategic resources in the modern world.There are a lot of different information in various fields of the society nowadays.On the one hand,information brings great convenience to people’s life and work;on the other hand,the appearance of information security has gradually become a major issue that people have to pay attention to.Theories and technologies have been researched and developed during the conflict when human beings tried to maintain information security.However,the traditional security technology based on Cryptography are increasingly unsuited to solve the large-scale security problems in this open and distributed environment.With the development of various emerging information security technologies,trust technology emerges as the times require.The emergence of trust technology provides not only a new way of thinking to solve the problem of security,but also a security technology guarantee for the safe operation of the dynamic cooperation mode among different components.Trust relationship modeling and management has become a hot issue in the research of information security field.However,due to the complexity of the trust concept and the difference between trust and security,it becomes difficult to grasp and understand the essence of trust relationship,and that makes it complex and difficult to effectively model and manage trust relationship.Despite the painstaking researches many scholars have made,the phenomenon,“the practice of technology develops rapidly,while its theory lags behind”,has not changed in the realm of trust research.The research project of this dissertation is proposed under this kind of background,which is expected to contribute to the research of modeling and management of trust relationship.Based on the essence of trust relationship,this dissertation applies the research results of fuzzy theory to the modeling and management of the trust relationship so that we can model the measurement,spread,decision-making,and such things in the modeling andmanagement of this difficult-to-describe trust relationship.Main research works were carried out as follows:1.The relationship between trust and security is analysed from the view of trust,security,and menace.By doing so,the trust relationship among entities can simplify the complex problem of information security in the complex and open network interactive environment,which ensures the safety of the interaction in a certain extent.This article suggests that security can not be replaced by trust completely.Trust mechanism can not work without the coordination of traditional security mechanisms.The two complement and cooperate with each other.2.Starting from the natural symbiotic relationship between trust and risk,this paper reiterates the concept of trust,and puts forward the theoretical model and decision-making model of the trust relationship in the perspective of risk.The model defines the trust relationship from the perspective of risk,and introduces the decision preference and transaction importance of trust subject.By analysing the security decision-making process of trust,risk and credible decision,security decisions would be more rational,comprehensive and objective.3.Constrained by the fuzzy integration operator orness measure level,this part proposes a direct trust aggregation algorithm based on the minimum variance time sequence weight.According to the reducing changes of trust relationship with time,the algorithm guarantees that time series weight factor would be a part of calculation,which avoids the influence of human factors.The algorithm shows that the trust degree of the object is stable to the majority of subjects,and the credibility of the object is finally convergent to the object.Without complex iterative process,the algorithm has simple calculation,good scalability and dynamic adaptability.4.Under the influence of fuzzy weight,the forecast model of dynamic trust relationship is proposed,which overcomes the limitation of the existing model for the high demand of the sample size and the data distribution.By setting the relevant trust threshold,the model effectively filter out malicious recommendations of the malicious entity.By means of non statistical estimation and ranking linear estimation method,this dissertation abandons the false recommendations of effective entities,which makes the prediction of feedback trustdegree and overall trust degree more accurate and objective.Without complicated calculation,the model is simple,unified,and better in convergence and expansibility.5.A decision-making algorithm of multi-attribution dynamic trust relationship group which is based on the maximum deviation model is proposed.The model converts the obscure trust value into the trust relationship described in natural language.It is clear in concept,specific in meaning,standard in process,precise in decision,convenient in program realization,and adapted in the requirements of open and dynamic network environment.The algorithm can be adapted to many applications by making simple modifications,which indicates its expansibility and practical value. |
