Improved Mechanisms and Prophet Inequalities for Graphical Dependencies

Vasilis Livanos,Kalen Patton,Sahil Singla

2024-06-08

Abstract:Over the past two decades, significant strides have been made in stochastic problems such as revenue-optimal auction design and prophet inequalities, traditionally modeled with $n$ independent random variables to represent the values of $n$ items. However, in many applications, this assumption of independence often diverges from reality. Given the strong impossibility results associated with arbitrary correlations, recent research has pivoted towards exploring these problems under models of mild dependency. In this work, we study the optimal auction and prophet inequalities problems within the framework of the popular graphical model of Markov Random Fields (MRFs), a choice motivated by its ability to capture complex dependency structures. Specifically, for the problem of selling $n$ items to a single buyer to maximize revenue, we show that the max of SRev and BRev is an $O(\Delta)$-approximation to the optimal revenue for subadditive buyers, where $\Delta$ is the maximum weighted degree of the underlying MRF. This is a generalization as well as an exponential improvement on the $\exp(O(\Delta))$-approximation results of Cai and Oikonomou (EC 2021) for additive and unit-demand buyers. We also obtain a similar exponential improvement for the prophet inequality problem, which is asymptotically optimal as we show a matching upper bound.

Computer Science and Game Theory

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to design the optimal auction mechanism and prophet inequality in an environment with mild dependencies. Specifically:

1. **Problem Background**:
- In the past two decades, stochastic problems (such as revenue - optimal auction design and prophet inequality) usually assume that item values are independent random variables. However, in many practical applications, this independence assumption often does not conform to reality.
- Due to the strong impossibility results brought about by arbitrary correlations, recent research has turned to exploring the performance of these problems under mild - dependency models.

2. **Research Objectives**:
- To study how to solve the optimal auction and prophet inequality problems under the framework of Markov Random Fields (MRFs).
- Pay special attention to how to maximize revenue when buyers value \(n\) items, and explore the applicability of these methods in the case of subadditive buyers.

3. **Main Contributions**:
- For subadditive buyers, it is proved that \(\max\{\text{SRev}, \text{BRev}\}\) is an \(O(\Delta)\)-approximate solution of the optimal revenue, where \(\Delta\) is the maximum weighted degree of the underlying MRF.
- The previous results on additive and single - demand buyers are improved, from \(\exp(O(\Delta))\)-approximate to \(O(\Delta)\)-approximate.
- A similar exponential - level improvement is also obtained for the prophet inequality problem, and the matching upper bound is shown.

4. **Technical Means**:
- Use the new "approximate marginal mechanism" lemma (Lemma 2.5) to divide the item set into a core set and a tail set, and analyze their revenues respectively.
- Introduce the exponential binning technique to handle the core and tail boundaries in MRF correlations.

Through these studies, the paper aims to make theoretical results closer to practical applications while avoiding the difficulties brought about by dealing with arbitrary distributions.