Date Added: May 2009
Decentralised multi-agent systems have empowered individual agents to actively select their course of action to achieve their goals. However, this leaves agents with a limited view of their environment, limiting awareness mainly to their interaction partners. The authors believe that through interactions and recommendation exchange, agents can build a network of their environment and expand their local view. Agents can extract emergent information from the relationships and interactions they observe. Analyzing and understanding such information can improve an agent's efforts in minimising risk associated with the uncertainty of agent interactions, including collusion detection and the discovery of new providers.