A Probabilistic Reputation Algorithm for Decentralized Multi-Agent Environments

The importance of trust in electronic transactions is well understood. The majority of current trust models consist of a central entity that verifies compliance with the trust requirements, using standardized evaluation methods and criteria. In decentralized environments, the communication scenarios are more complex, and no universally accepted objective requirements or evaluation criteria exist. It should be noted that the situation would get even more complicated when agents are interacting with each other. The goal of this research is to model trust and reputation in decentralized multi-agent systems. To achieve this, we have chosen the Ntropi model, among several other models, as a starting point, The efficiency of the model in such scenarios has been significantly improved by introducing a new probabilistic reputation algorithm for the Ntropi model.