Reproduce “Utility Maximizer or Value Maximizer: Mechanism Design for Mixed Bidders in Online Advertising” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hongtao Lv, Zhilin Zhang, Zhenzhe Zheng, Jinghan Liu, Chuan Yu, Lei Liu, Lizhen Cui, Fan Wu
Researchers from Shandong University, Shanghai Jiao Tong University, and Alibaba Group developed truthful and robust auction mechanisms for online advertising platforms accommodating both Utility Maximizer and Value Maximizer advertisers. Their Mechanism for Private Classes (MPR) achieves an approximation ratio of at most 2 on Liquid Social Welfare and is proven individually rational and incentive compatible, even when bidder types are private information.
Key insights
- A simple mechanism (MPU) combining VCG and GSP based on public bidder types achieves optimal Liquid Social Welfare, but it is untruthful if bidder types are private.
- The MPR mechanism ensures truthfulness for private bidder types by designing slot payments that effectively prevent manipulation of reported class or value.
- MPR is robust, meaning it naturally generalizes to VCG when all bidders are Utility Maximizers and to GSP when all bidders are Value Maximizers.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.