Plans for Multiple Agents
Abstract: "Research in Distributed Artificial Intelligence has historically focused on two distinct classes of problems. One paradigm, 'planning for multiple agents, ' considers issues inherent in centrally directed multiple-agent execution (Smith's Contract Net [15,14] falls into this category, as does other DAI work such as [11,10]). In the second paradigm, 'distributed planning, ' multiple agents participate more autonomously in deciding upon and coordinating their own actions [2,12,3].