Network Trustworthy Assessment Based on Advanced DyTrust Model

Abstract

According to the algorithm of direct trust value in DyTrust Model, a time attenuation factor was presented, and the influence of time factor was considered. At the same time, a risk function was put forward, and a new model named A-DyTrust was proposed. Based on the proportional relationship to the quality and risk of request services by node, a formula of the risk function was raised to improve the DyTrust Model. The simulation results show that A-DyTrust Model is more effective than DyTrust Model, which has better dynamic adaptive capacity, higher ability to detect malicious nodes and more effectively aggregate capacity of the information than DyTrust Model. Aim at the question of network trust rating, a method of assessment for network trustworthy based on the associative memory capability of the Hopfield Neural Network was given. The simulation results show that this method can determine the level of trust of current network accurately, then achieve the quantification of network trust.