多物品拍卖机制设计

The purpose of this project is to reexamine the auction problems for multiple objects which are substitutes to each other from the view point of partial availability of information. It is known that there is a shortcoming in the existing auction mechanisms for substitutes: on one hand, the exact auction mechanism, which requires the full information about buyer's preferences, is not practical because it is very demanding for the buyers to describe preferences among bundles of objects completely, let alone a potential incentive problem for revealing correct information. On the other hand, the approximate auction mechanism, which only requires partial information about preferences, is not efficient in the sense that the auction does not necessarily find an efficient allocation of the objects, leading to an economic loss. Thus a challenging task is to find an auction mechanism which improves upon these deficiencies. .We intend to accomplish it in this project, by proposing a practical and efficient auction mechanism for substitutes. Moreover, we intend to design it in such a way that it can be readily performed on line, where the anonymity of buyers holds. Hence we contend that this project is valuable not only as a contribution to economic theory, but also as a practical tool which directly benefits commercial activities. .This project can be divided into three steps: first, we consider the simplest case of unit demand, i.e., each buyer is interested in at most one of the objects. In this set up, we develop two new concepts of "tgtightly demanded set" and "revealed demand set" which constitutes the key elements of the mathematical analysis. We then design a new general English auction for multiple objects with partial information. .Secondly, we study the structure of demand sets for general preferences, with the assumption that the objects are substitutes. Elaborating on the two new concepts above, we propose a new method for decomposing the demand set, which helps us to translate this general version of the auction problem into the case of unit demand. This idea is particularly important in the anonymous internet auction application, where it is difficult to follow the identity of the bidders. Our method takes each anonymous bidder independently, and works well even if one economic agent uses multiple identities for bidding purpose. .Thirdly, we shall "fine tune" the mechanism in such a way that the mechanism is robust against the bidders' manipulable behavior and the seller's shill bidding. This is an important step to make sure that the auction works as intended in the real world. At least for the case of unit demand, we expect that the mechanism can be made strtegy-proof as a dynamic game, which offers a theoretical check. Meanwhile we shall use simulation and experiment to see if our auction mechanism performs as intended.

随着互联网的出现多个物品的同步拍卖成为可能，关于多物品拍的研究也越来越受到人们的重视。本申请课题将研究替代物品的拍卖问题。我们将针对现有的替代物品拍卖机制中存在的"近似拍卖机制配置效率低，而精确拍卖机制又不实用"这一缺陷，设计介于现有的精确拍卖机制与近似拍卖机制之间的适用于拍卖替代物品的实用型精确拍卖机制，开发相应的实用型网络拍卖平台。为此，我们首先研究最简单的替代物品拍卖，即单位需求多物品拍卖的问题，给出"紧需求集"、 "显示需求集合"这两个新的经济学概念，并用其设计全新的单位需求多物品拍卖机制。其次，研究替代性条件下需求集合的结构，给出"替代性需求集合的分解"这一新的概念和方法，利用该分解法可以巧妙地将替代物品的拍卖问题转变成单位需求多物品拍卖的问题。另外，我们还计划运用计算机模拟、经济学实验的方法来比较和评价我的拍卖机制。

随着互联网的出现多个物品的同步拍卖成为可能，关于多物品拍的研究也越来越受到人们的重视，多物品拍卖在现实中的应用也越来越广泛。本课题紧紧围绕着“多物品拍卖”这个中心问题以及与其紧密相关的双向匹配问题展开了系列性的理论和应用研究，取得了多项具有较高学术价值和应用前景的研究成果，发表了几篇高水平学术论文。课题立项以来，我们在本课题的资助下主要针对以下六项子课题展开了研究、取得了研究成果：. （1）研究互补物品的拍卖问题，运用非歧视性非线性价格体系设计出一个适合于互补物品拍卖的、有效且具有激励相容性的动态拍卖机制。. （2）研究我国汽车牌照配置问题，构建了分析汽车牌照配置问题的理论框架，给出了公共资源配置机制公平性的合理定义，设计出了一类更科学的混合配置机制，并且在统一框架下比较了目前国内采用的上海、北京、广州、天津和杭州机制。. （3）从拍卖中间人的角度设计具有让买卖双方诚实报价、期望收益非负、防止卖家找托等性质的双向拍卖（double auction）机制。. （4）在招生（或招聘）过程中如何确定发多少offer以及向谁发offer，是一个众所周知的难题。针对这一难题，我们提出了返还报名费的招生规则，从而激励学生披露学校所需信息，从而降低过多录取学生的风险、提高招生效率。. （5）设计介于现有的精确拍卖机制与近似拍卖机制之间的、且具有一定纠错功能的实用型精确拍卖机制。. （6）研究婚姻再匹配问题，建立了一种全新的匹配模型，从而发展了匹配理论（matching theory）。. 过去4年间，课题组在经济学国际顶级期刊《政治经济学期刊》（Journal of Political Economy ）,.中文顶级期刊《经济研究》，以及中文一流期刊《财经研究》发表了3篇学术论文。另外，课题组还完成了多篇工作论文，并投稿或即将投稿到经济学国际顶级和一流期刊。. 在队伍建设与人才培养方面取得了很好的成果。课题立项以来，围绕本课题研究共培养4名硕士毕业生，4名博士毕业生。孙宁教授本人2014年入选上海市领军人才（地方队），2016年入选国家“万人计划”哲学社会科学领军人才。