Computer Science > Computer Science and Game Theory
[Submitted on 18 Jun 2023 (this version), latest version 27 Apr 2024 (v2)]
Title:Bidder Selection Problem in Position Auctions via Poisson Approximation
View PDFAbstract:We consider Bidder Selection Problem (BSP) in position auctions motivated by practical concerns of online advertising platforms. In this problem, the platform sells ad slots via an auction to a large pool of $n$ potential buyers with independent values drawn from known prior distributions. The seller can only invite a fraction of $k<n$ advertisers to the auction due to communication and computation restrictions. She wishes to maximize either the social welfare or her revenue by selecting the set of invited bidders.
We study BSP in a classic multi-winner model of position auctions for welfare and revenue objectives using the optimal (respectively, VCG mechanism, or Myerson's auction) format for the selected set of bidders. We propose a novel Poisson-Chernoff relaxation of the problem that immediately implies that 1) BSP is polynomial time solvable up to a vanishingly small error as the problem size $k$ grows; 2) PTAS for position auctions after combining our relaxation with the trivial brute force algorithm; the algorithm is in fact an Efficient PTAS (EPTAS) under a mild assumption $k\ge\log n$ with much better running time than previous PTASes for single-item auction. Our approach yields simple and practically relevant algorithms unlike all previous complex PTASes, which had at least doubly exponential dependency of their running time on $\varepsilon$. In contrast, our algorithms are even faster than popular algorithms such as greedy for submodular maximization. Furthermore, we did extensive numerical experiments, which demonstrate high efficiency and practical applicability of our solution. Our experiments corroborate the experimental findings of [Mehta, Nadav, Psomas, Rubinstein 2020] that many simple heuristics perform surprisingly well, which indicates importance of using small $\varepsilon$ for the BSP and practical irrelevance of all previous PTAS approaches.
Submission history
From: Nick Gravin [view email][v1] Sun, 18 Jun 2023 23:29:56 UTC (57 KB)
[v2] Sat, 27 Apr 2024 10:50:13 UTC (73 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.