分析了已有的多属性拍卖方法中存在的问题,给出了一种广义的多属性拍卖模型.进而提出了一种暗标叫价多属性拍卖方法——MAV,并证明了MAV的一些重要性质,如卖方激励相容性等,还给出了卖方的占优策略和买方的Bayes最优策略.论证了MAV中的买卖方总效用在Bayes均衡点上达到最大.最后与已有的多属性拍卖方法进行了比较,结果表明MAV改进了Esther等人的工作.
A general multi-attribute auction model and a new auction method MAV under this model are presented. MAV is an extension for Vickrey auction from single attribute to multiattribute. Authors discuss strategies and profits of buyer and sellers in MAV. Some main properties of MAV are proved. Seller agents have dominant strategy and buyer agent has Bayes strategy in MAV. Buyer agent and seller agents can achieve maximal utility in the Bayes equilibrium point, etc. Compare to Esther David's works, MAV has a more generalized model and better performance.