Subtopic Deep Dive

Stochastic Matching in Online Bipartite Graphs
Research Guide

What is Stochastic Matching in Online Bipartite Graphs?

Stochastic matching in online bipartite graphs develops approximation algorithms for matching vertices where weights are drawn from known distributions in an online arrival model.

Algorithms achieve competitive ratios better than adversarial online matching's 1-1/e bound by exploiting stochastic weight distributions (Mahdian and Yan, 2011; 231 citations). Research extends to prophet inequalities and applications like ad allocation and ride-sharing. Over 10 key papers span from foundational work in 1992 to recent advances.

15
Curated Papers
3
Key Challenges

Why It Matters

Algorithms enable efficient matching in ride-sharing platforms with reusable resources, improving allocation under uncertainty (Dickerson et al., 2018; 67 citations). Prophet-inequality approaches optimize ad allocation by balancing online decisions against known distributions (Alaei et al., 2012; 100 citations). Competitive ratios guide resource allocation in dynamic marketplaces like caching and prefetching (Vitter and Krishnan, 1996; 139 citations).

Key Research Challenges

Achieving Optimal Competitive Ratios

Stochastic models improve over 1-1/e but tight bounds remain elusive for general distributions. Mahdian and Yan (2011) show random arrivals yield better ratios than adversarial. Prophet inequalities require balancing threshold policies against full information optima (Alaei et al., 2012).

Handling Reusable Resources

Ride-sharing introduces offline reusable supply complicating online matching. Dickerson et al. (2018) develop algorithms for bipartite markets with reusable resources. Competitive analysis must account for multiple matching opportunities per resource.

Prophet Inequality Extensions

Extending prophet inequalities to weighted bipartite graphs with capacities challenges threshold design. Alaei et al. (2012) address bidder capacities and online item streams. Geometry of feasible regions impacts packing constraints (Molinaro and Ravi, 2013).

Essential Papers

1.

An optimal on-line algorithm for metrical task system

Allan Borodin, Nathan Linial, Michael Saks · 1992 · Journal of the ACM · 328 citations

In practice, almost all dynamic systems require decisions to be made on-line, without full knowledge of their future impact on the system. A general model for the processing of sequences of tasks i...

2.

Online bipartite matching with random arrivals

Mohammad Mahdian, Qiqi Yan · 2011 · 231 citations

In a seminal paper, Karp, Vazirani, and Vazirani show that a simple ranking algorithm achieves a competitive ratio of 1-1/e for the online bipartite matching problem in the standard adversarial mod...

3.

Optimal prefetching via data compression

Jeffrey Scott Vitter, P. Krishnan · 1996 · Journal of the ACM · 139 citations

Caching and prefetching are important mechanisms for speeding up access time to data on secondary storage. Recent work in competitive online algorithms has uncovered several promising new algorithm...

4.

Online prophet-inequality matching with applications to ad allocation

Saeed Alaei, MohammadTaghi Hajiaghayi, Vahid Liaghat · 2012 · 100 citations

We study the problem of online prophet-inequality matching in bipartite graphs. There is a static set of bidders and an online stream of items. We represent the interest of bidders in items by a we...

5.

Allocation Problems in Ride-Sharing Platforms: Online Matching With Offline Reusable Resources

John P. Dickerson, Karthik Abinav Sankararaman, Aravind Srinivasan et al. · 2018 · Proceedings of the AAAI Conference on Artificial Intelligence · 67 citations

Bipartite matching markets pair agents on one side of a market with agents, items, or contracts on the opposing side. Prior work addresses online bipartite matching markets, where agents arrive ove...

6.

Online matching: haste makes waste!

Yuval Emek, Shay Kutten, Roger Wattenhofer · 2016 · 54 citations

This paper studies a new online problem, referred to as min-cost perfect matching with delays (MPMD), defined over a finite metric space (i.e., a complete graph with positive edge weights obeying t...

7.

Improved Achievability and Converse Bounds for Erdos-Renyi Graph Matching

Daniel Cullina, Negar Kiyavash · 2016 · ACM SIGMETRICS Performance Evaluation Review · 52 citations

We consider the problem of perfectly recovering the vertex correspondence between two correlated Erdos-Renyi (ER) graphs. For a pair of correlated graphs on the same vertex set, the correspondence ...

Reading Guide

Foundational Papers

Start with Borodin et al. (1992) for online competitive analysis foundations (328 citations), then Mahdian and Yan (2011) for stochastic bipartite matching (231 citations), followed by Alaei et al. (2012) for prophet extensions (100 citations).

Recent Advances

Study Dickerson et al. (2018) for ride-sharing applications (67 citations), Emek et al. (2016) for delay-tolerant matching (54 citations), and Lykouris and Vassilvitskii (2021) for ML-augmented algorithms (46 citations).

Core Methods

Competitive analysis via randomized ranking (Mahdian and Yan, 2011), dynamic threshold prophet inequalities (Alaei et al., 2012), online packing LPs (Molinaro and Ravi, 2013), and reusable resource LP relaxations (Dickerson et al., 2018).

How PapersFlow Helps You Research Stochastic Matching in Online Bipartite Graphs

Discover & Search

Research Agent uses searchPapers to find 'stochastic online bipartite matching' yielding Mahdian and Yan (2011; 231 citations), then citationGraph reveals 50+ descendants like Dickerson et al. (2018), and findSimilarPapers expands to prophet inequality variants. exaSearch queries 'ride-sharing stochastic matching' surfaces application papers.

Analyze & Verify

Analysis Agent applies readPaperContent to extract competitive ratio proofs from Alaei et al. (2012), verifies claims via verifyResponse (CoVe) against Borodin et al. (1992) metrical task bounds, and runPythonAnalysis simulates matching algorithms with NumPy for ratio computation. GRADE grading scores algorithm novelty on 1-5 evidence scale.

Synthesize & Write

Synthesis Agent detects gaps in prophet inequalities for reusable resources via gap detection, flags contradictions between adversarial and stochastic ratios, and uses exportMermaid for competitive ratio comparison diagrams. Writing Agent employs latexEditText to draft proofs, latexSyncCitations integrates 20+ references, and latexCompile generates camera-ready survey sections.

Use Cases

"Simulate competitive ratios for stochastic matching algorithms"

Research Agent → searchPapers → Analysis Agent → runPythonAnalysis (NumPy Monte Carlo simulation of Mahdian-Yan algorithm vs 1-1/e baseline) → matplotlib ratio plots.

"What are current bounds for ride-sharing matching?"

Research Agent → citationGraph (Dickerson 2018) → Synthesis Agent → gap detection → Writing Agent → latexEditText + latexSyncCitations + latexCompile → LaTeX survey with bounds table.

"Find code for online prophet matching implementations"

Research Agent → paperExtractUrls (Alaei 2012) → Code Discovery → paperFindGithubRepo → githubRepoInspect → verified algorithm code + runPythonAnalysis sandbox test.

Automated Workflows

Deep Research workflow conducts systematic review: searchPapers (250+ hits) → citationGraph pruning → DeepScan 7-step analysis with GRADE checkpoints on 15 core papers. Theorizer generates prophet inequality conjectures from Borodin (1992) + recent ride-sharing papers. DeepScan verifies ratio claims across Mahdian-Yan (2011) lineage via CoVe chain.

Frequently Asked Questions

What defines stochastic matching in online bipartite graphs?

Vertices arrive online with weights drawn from known distributions; algorithms exploit this for competitive ratios exceeding adversarial 1-1/e (Mahdian and Yan, 2011).

What are main methods used?

Ranking algorithms for random arrivals (Mahdian and Yan, 2011), prophet threshold policies (Alaei et al., 2012), and reusable resource matching (Dickerson et al., 2018).

What are key papers?

Foundational: Borodin et al. (1992; 328 citations), Mahdian and Yan (2011; 231 citations). Recent: Dickerson et al. (2018; 67 citations), Lykouris and Vassilvitskii (2021; 46 citations).

What open problems exist?

Optimal ratios for reusable resources in general distributions; extending prophet inequalities to metric spaces (Emek et al., 2016); machine-learned advice integration (Lykouris and Vassilvitskii, 2021).

Research Optimization and Search Problems with AI

PapersFlow provides specialized AI tools for Computer Science researchers. Here are the most relevant for this topic:

See how researchers in Computer Science & AI use PapersFlow

Field-specific workflows, example queries, and use cases.

Computer Science & AI Guide

Start Researching Stochastic Matching in Online Bipartite Graphs with AI

Search 474M+ papers, run AI-powered literature reviews, and write with integrated citations — all in one workspace.

See how PapersFlow works for Computer Science researchers