Core Concepts
Developing a versatile greedy-based technique for Single-Sample Prophet Inequalities directly, improving competitive guarantees.
Abstract
The paper introduces a novel approach to Single-Sample Prophet Inequalities (SSPIs) through a greedy-based technique, providing improved competitive guarantees. The algorithm is analyzed in two scenarios: general graphs with edge arrivals and bipartite graphs with vertex arrivals. The study focuses on maximizing expected rewards under limited information constraints.
General Graphs:
- Introduction to prophet inequalities from optimal stopping theory.
- SSPIs with only one sample per distribution yield strong results.
- Comparison of SSPIs to Order-Oblivious Secretary algorithms.
Bipartite Graphs:
- Algorithm design for SSPI matching in bipartite graphs with vertex arrivals.
- Equivalence between online and offline algorithms for analysis.
- Proof of correctness and competitiveness of the proposed algorithm.
Stats
一つのサンプルから強力な結果を得るために、SSPIが重要である。
単一のサンプルからのSSPIは強力な結果をもたらす。
提案されたアルゴリズムの正確性と競争力の証明。