The purpose of this paper is to enhance trust in e-commerce multi-agent systems by presenting a model, called RUU, to select the most trustworthy provider agent based on learning from previous interactions and computing reliability, unreliability and uncertainty.
The methodology comprises analyzing the most representative existing trust models, while a new concept was proposed and measured as unreliability. To make decision about the agents, RUU integrated reliability, unreliability and uncertainty components and used the TOPSIS multi-criteria decision method to select the most trustworthy provider agent. To evaluate the RUU model, the experimentation was carried out in two stages. First, the average accuracy of the model was investigated by simulating RUU in a multi-agent environment. Second, the performance of the model was compared with other related trust models.
The experimental results revealed that RUU model outperforms current models in providing accurate credibility measurements and computing an accurate trust mechanism for agents, also presenting a decision-making process to choose the most trustworthy provider agent.
The model presented based on different mathematical computations that take time to be calculated, which is a big limitation of computational models.
RUU enables an agent to make effective and sound decisions in light of the uncertainty that exists in e-commerce multi-agent environments.
This paper is beneficial to enhance the fulfilment of purchasing between provider and requester agents. In fact, the proposed model can ensure critical transactions performed securely in e-commerce multi-agent environments.
