This is a report on Simurgh team's participation in the 2011 multi-agent contest. The design and development process, architecture details, and team strategies for the multi-agent system have been discussed, along with experiences of the developers. Gaia methodology was used for the design and analysis of the Simurgh multi-agent system. The main strategy was obtaining a higher score through the support of agents with a better perceived strategic placement. Decision correction strategy was used to change the agent behavior, by taking the other conflicting team members' decisions, into account. Simurgh was implemented using Java language. Agents have the same degree of autonomy and the team is implemented in a completely decentralized fashion.
展开▼
