Where to Sell: Simulating Auctions From Learning Algorithms
Where to Sell: Simulating Auctions From Learning Algorithms
复制标题
在哪里销售：通过学习算法模拟拍卖
DOI：
10.2139/ssrn.2783938
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
Umar Syed
中科院分区：
文献类型：
--
作者：
Hamid Nazerzadeh;R. Leme;Afshin Rostamizadeh;Umar Syed
广告交换平台连接在线发布商和广告商，每天促进数十亿展示的销售。我们从出版商的角度研究这些环境，出版商希望找到利润最大化的交易所来出售他的库存。理想情况下，出版商将在交易所之间进行拍卖。然而，由于实际的业务考虑，这通常是不可能的。相反，出版商必须将每一次展示只发送给其中一个交易所，沿着附上要价。我们模型的问题作为一个变化的多武装土匪问题，其中交易所（武器）可以采取战略行动，以最大限度地提高自己的利润。我们提出了电子机制，找到最好的交流与次线性遗憾，并具有理想的激励性能。
Ad exchange platforms connect online publishers and advertisers and facilitate the sale of billions of impressions every day. We study these environments from the perspective of a publisher who wants to find the profit-maximizing exchange in which to sell his inventory. Ideally, the publisher would run an auction among exchanges. However, this is not usually possible due to practical business considerations. Instead, the publisher must send each impression to only one of the exchanges, along with an asking price. We model the problem as a variation of the multi-armed bandits problem in which exchanges (arms) can behave strategically in order to maximizes their own profit. We propose e mechanisms that find the best exchange with sub-linear regret and have desirable incentive properties.