Stategic behaviour in multiagent systems

我们目前正在见证以互联网为代表的开放计算系统新范式的发展。这些系统(我们称之为代理)中的不同参与者通常具有不同的信息和利益。具体地说，在许多情况下，这些代理是自私自利的，并且会不顾他人的愿望而采取行动来实现自己的目标，也不合理地假设系统设计者能够迫使这些代理以某种社会最佳方式行事。这些特性的有趣应用范围从新的电子市场和拍卖，到互联网服务供应的差异化调度。

We are currently witnessing the growth of a new paradigm of open computing systems, as embodied by the Internet. The heterogeneous participants in these systems (which we call agents) often have diverging information and interests. In particular, in many situations these agents are self-interested and will act to further their own goals, irrespective of the desires of others, nor is it reasonable to assume that a system designer is able to force these agents to behave in some socially optimal way. Interesting applications with these properties range from new electronic marketplaces and auctions, to differentiated scheduling for Internet service provisioning.