Research On Trust Evaluation And Service Selection In The Mult Autonomous Region Internet Of Things Environment

In the Internet of things(IOT)environment,each company and organization has its own shared resources for internal entities to access and use,thus forming a large-scale,heterogeneous and dynamic autonomous domain of IOT.In essence,the Internet of things is an Internet that realizes the interconnection between people,people and things,things and things.There are a large number of littleknown or completely unknown entities within the network.Security has become a key problem affecting the performance of autonomous domain.Hard security mechanisms such as encryption and access control can resist malicious attacks from outside the domain,but can not ensure the reliability of the system when there are malicious activities in the system.Trust is a soft security mechanism that helps entities overcome perceptions of uncertainty and risk by analyzing the behavior of entities over time to maintain legitimacy.Previous studies have proposed various trust management mechanisms to gain the trust of Io T entities.However,due to the differences in trust requirements and evaluation indexes of each autonomous domain,each autonomous domain cannot use a unified trust model to evaluate the trust value of the entity.As a result,each autonomous domain is an independent autonomous system,and trust cannot be propagated across domains.However,entities may have the following requirements:(1)Cross-domain movement: Entities may leave an autonomous domain to join a new one at any time.(2)Cross domain service selection: When there is no service required by the entity in the autonomous domain,it needs to request service from the entity in other autonomous domains.Due to the neglect of the cross-domain trust evaluation in previous work,it has caused a lot of inconvenience :(1)The cross-domain mobile entity cannot migrate the accumulated trust value in the original autonomous domain to the new domain,so the entity needs to re-accumulate the trust value,resulting in unnecessary waste of time and resources;(2)When selecting services,entities cannot evaluate the reliability of entities in other autonomous regions,so they cannot choose reliable services from other autonomous regions,which hinders the establishment of service interaction between service providers and service requesters across domains.Therefore,this paper studies the trust evaluation and service selection of cross domain entities,and the specific work is as follows:1)Cross domain trust evaluation model based on social network relationship(CDTM)is proposed.With the increasing mobility of entities,entities may frequently cross domains.However,the previous cross domain model is centralized and needs the help of domain server,which is easy to cause a single point of failure.CDTM obtains the social path from the trustor to the trustee through breadth first search and loop removal algorithm,and calculates the social trust value between users from intimacy,content influence and attribute similarity.The model is a distributed cross domain trust evaluation model.Users can calculate the trust value of cross domain entities by themselves,without using domain servers.With the help of social relationship,the cross domain trust value can be calculated even when there is no direct or indirect trust between users and no trust information between domains.2)Based on the cross domain trust evaluation,an improved multi-objective ant colony algorithm is proposed for cross domain service selection.Traditional methods only consider the quality of service(Qo S)provided by service providers,but ignore the credibility and unreliability of services.Untrusted service providers may implement malicious services or spread false information.This method integrates multiple optimization objectives:(1)Qo S attribute(It is divided into qualitative QoS attributes and quantitative Qo S attributes)(2)trust value(3)distrust value.This work creatively regards distrust as a completely different concept from trust,and takes it as an optimization goal.By introducing four objective functions and Pareto optimal solution into the ant colony algorithm,the classical single objective ant colony optimization method is improved.In this paper,the CDTM model is simulated and evaluated based on opinions data set and twitter data set in Matlab environment.The experimental results show that the model has good convergence,effectiveness and accuracy.Based on the QWS2.0 dataset simulation,the performance of the proposed method for distrust evaluation,the accuracy of service selection method and the computational complexity are verified.The results show that the proposed method is superior to the traditional QoS-only method.