Most research in automated negotiation focuses on strategy development in preset scenarios where decisions about what to negotiate about, whom to negotiate with, and on which issues are given to the agents. Moreover, in many cases, the agents' utility functions are predefined, static, and independent of other negotiations. NegMAS (Negotiations Managed by Agent Simulations/Negotiation Multia-gent System) was developed to facilitate the research and development of autonomous agents that operate in a rich multiagent system where negotiations are paramount, such as a supply chain. The richness of the setting creates what we call situated negotiations, where negotiations naturally interdepend and agents' utility functions arise endogenously from system dynamics. This paper introduces NegMAS-a platform for autonomous negotiation within a rich simulated multiagent system-and evaluates its use in a sample application.
展开▼