Prior-free cost sharing design: group strategyproofness and the worst absolute loss
Prior-free cost sharing design: group strategyproofness and the worst absolute loss
复制标题
无先验成本分摊设计：群体策略证明和最坏绝对损失
DOI：
10.1007/978-1-4419-0056-2_16
复制
发表时间：
2009
期刊：
影响因子：
2.5
通讯作者：
R. Juarez
中科院分区：
文献类型：
--
作者：
R. Juarez
以一定成本为一组代理提供服务。服务是二元的，每个代理要么接收服务，要么不接收服务。一种机制激发代理商为获得服务而付费的意愿，为一组代理商提供服务，并通过向这些代理商收取费用来支付成本。我们考虑这样的问题：机制的设计者几乎没有关于参与者的信息，无法形成关于其可能类型分布的先验信念。本文研究了满足两个在此设置中有意义的稳健属性的机制。一方面，我们研究阻止任何代理组之间协调的机制（以下称为组策略证明）。另一方面，我们研究使用最差绝对盈余损失衡量标准的最佳机制。最优和群体策略证明的机制针对不同形状的成本函数进行了表征。
A service is produced for a set of agents at some cost. The service is bi- nary, each agent either receives service or not. A mechanism elicits the willingness to pay of the agents for getting service, serves a group of agents and covers the cost by charging those agents. We consider the problem in which the designer of the mechanism has little infor- mation about the participants to be able to form a prior-belief about the distribution of their possible types. This paper investigates mechanisms that meet two robust properties that are meaningful in this setting. On one hand, we look at mechanisms that prevent coordination between any group of agents (henceforth called group strategyproof). On the other hand, we look at mechanisms that are optimal using the worst absolute surplus loss measure. The mechanisms that are optimal and group strategyproof are characterized for different shapes of cost functions.