Research On Quantization Model On Dynamic Trust-relationship Of Trusted Network Based-on Multi-attribute Strategy

Along with the popularization of network in real life and the increasing demandsof people on the network, the quality of the network is becoming more and morepopular. And so, trusted network study becomes a hot issue. Considering manyuncertain factors in the network, research on the dynamic trust-relationship ofnetwork has important significance. This paper studies the quantization model ondynamic trust-relationship based on the mult-attribute strategy.This paper first analyzes the quantization model on dynamic trust-relationshipbased on multible attribute which include the direct trust, the history trust, thefeedback trust and the risk function. Considering the problem that all informationfrom nodes feedback would be not true or reliable, this paper proposes a improvedrule to calculate the feedback trust. The new rule is taken into account in thecredibility of the feedback information to draw the new feedback trust. Using theresult, this paper get a new trust aggregation with other attributes. Via the simulationexperiment, the new model with the improved feedback trust can more accurately assessthe dynamic trust-relationship the network. Meanwhile, it improves the success rate ofinteractive behavior. Secondly, given to the affect to network performance from itscomplex structure characteristics, this paper put forward a new quantization model ontrust-relationship of trusted network, which is based-on nodes similarity. Combainedwith other four attributes, the node similarity based on resource allocation isintroduced into the multiple attribute model to get a new evaluation result. Thesimulation experiment show that the new quantization of trust-relationship model ondynamic network improve the success rate, making the result of evaluating thetrust-relationship of network more stable. Also, it proves that network structurecharacteristics affects to evaluate its credibility. At last, base on the above two improvements that improved feedback and the introduced nodes similarity, this paperpresents a new improved model. The simulation experiment further prove thesuperiority of the new model. It shows better Dynamic adaptability and get betteraccuracy and stability than the model of dynamic trust-relationship based onmulti-attribution.