Trust and Reputation in Multi-Agent Systems

摘要: Multi-Agent systems (MAS) are artificial societies populated with distributed autonomous agents that intelligent and rational. These self-independent capable of independent decision making towards their predefined goals. These goals might be common between or unique for an agent. Agents may cooperate with one another to facilitate progresses. One the fundamental challenges in such settings is do not have a full knowledge over environment regarding decision making processes, they need request other piece of information service. The crucial issues then how rely on the information provided by agents, consider the collected data, select appropriate ask for the required information. There some proposals addressing how an agent can on compute the overall opinion about particular In this context, the trust value reflects extent which other agents reputation represents public a particular Existing approaches reliable information propagation fail capture dynamic relationships between agents influence further making process. Therefore, these models adapt frequent environment changes. general, well-founded trust and reputation system prevents malicious acts emerged by selfish multi-agent systems. We propose a trust mechanism measures analyzes reliability of agents cooperating with one another. This concentrates on key attributes related their relationships. also measure analyze reputation of large-scale environments utilizing sound reputation mechanism. mechanism, we aim at maintaining public reputation assessment actions are accurately under analysis. On top theoretical analysis, we experimentally validate our approaches through different simulations. Our preliminary results show that our approach outperforms current frameworks providing accurate credibility measurements accurate mechanisms.