提出改进的自动竞价机制，显著提升广告拍卖效率。
Efficiency of Proportional Mechanisms in Online Auto-Bidding Advertising
- 设计新支付方案，优化比例机制的竞价均衡。
- 理论证明效率损失小于2，且随代理数量增加趋近最优。
- 方法简洁可推广，适合研究机制设计与竞价系统者参考。
在线广告中自动化竞价策略的兴起带来了新的机制设计挑战。本文聚焦于自动竞价场景下的比例机制，研究纯纳什均衡在液态福利目标下的效率，即价格无政府性（PoA）。我们首先证明标准比例机制的紧PoA界为2。随后引入一种新支付方案的改进版本，其PoA界为$1 + \frac{O(1)}{n-1}$，其中$n \geq 2$表示竞价代理数。该结果突破了现有PoA的2的上限，并随代理数量增加趋近完全效率。方法基于线性与凸规划的对偶性及KKT条件。由于概念简洁，该方法可能广泛适用于建立PoA边界。
原文摘要 · Abstract (English)
The rise of automated bidding strategies in online advertising presents new challenges in designing and analyzing efficient auction mechanisms. In this paper, we focus on proportional mechanisms within the context of auto-bidding and study the efficiency of pure Nash equilibria, specifically the price of anarchy (PoA), under the liquid welfare objective. We first establish a tight PoA bound of 2 for the standard proportional mechanism. Next, we introduce a modified version with an alternative payment scheme that achieves a PoA bound of $1 + \frac{O(1)}{n-1}$ where $n \geq 2$ denotes the number of bidding agents. This improvement surpasses the existing PoA barrier of 2 and approaches full efficiency as the number of agents increases. Our methodology leverages duality and the Karush-Kuhn-Tucker (KKT) conditions from linear and convex programming. Due to its conceptual simplicity, our approach may offer broader applications for establishing PoA bounds.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。