kth.sePublications KTH
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Maximum Flow by Augmenting Paths in n2+o(1) Time
Rutgers University.
KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science, Theoretical Computer Science, TCS.ORCID iD: 0009-0004-0874-2356
University of Michigan.
Stanford University.
2024 (English)In: 2024 IEEE 65rd Annual Symposium on Foundations of Computer Science (FOCS), Chicago, USA: IEEE, 2024Conference paper, Published paper (Refereed)
Abstract [en]

We present a combinatorial algorithm for computing exact maximum flows in directed graphs with n vertices and edge capacities from {1,…,U} in n2+o(1)logU time, which is almost optimal in dense graphs. Our algorithm is a novel implementation of the classical augmenting-path framework; we list augmenting paths more efficiently using a new variant of the push-relabel algorithm that uses additional edge weights to guide the algorithm, and we derive the edge weights by constructing a directed expander hierarchy.

Even in unit-capacity graphs, this breaks the long-standing O(m⋅min{√m,n2/3}) time bound of the previous combinatorial algorithms by Karzanov (1973) and Even and Tarjan (1975) when the graph has m=ω(n4/3) edges. Notably, our approach does not rely on continuous optimization nor heavy dynamic graph data structures, both of which are crucial in the recent developments that led to the almost-linear time algorithm by Chen et al. (FOCS 2022). Our running time also matches the n2+o(1) time bound of the independent combinatorial algorithm by Chuzhoy and Khanna (STOC 2024) for computing the maximum bipartite matching, a special case of maximum flow.

Place, publisher, year, edition, pages
Chicago, USA: IEEE, 2024.
Keywords [en]
Maximum Flow, Combinatorial Algorithms
National Category
Computer Sciences
Research subject
Computer Science
Identifiers
URN: urn:nbn:se:kth:diva-355868OAI: oai:DiVA.org:kth-355868DiVA, id: diva2:1910542
Conference
Foundations of Computer Science (FOCS)
Note

QC 20241105

Available from: 2024-11-05 Created: 2024-11-05 Last updated: 2024-11-05Bibliographically approved
In thesis
1. Matchings, Maxflows, Matroids: The Power of Augmenting Paths and Computational Models
Open this publication in new window or tab >>Matchings, Maxflows, Matroids: The Power of Augmenting Paths and Computational Models
2024 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Matchings, Maximum Flow, and Matroid Intersections are fundamental combinatorial optimization problems that have been studied extensively since the inception of computer science. A series of breakthroughs in graph algorithms and continuous optimization in the past decade has led to exciting almost-optimal algorithms for maximum flow and bipartite matching. However, we are still far from fully understanding these problems. First, it remains open how to solve these problems in modern models of computation, such as parallel, dynamic, online, and communication models. Second, as algorithms become more sophisticated in pursuit of efficiency, they often sacrifice simplicity, potentially obscuring valuable combinatorial insights. This raises a fundamental question: can we develop efficient algorithms that maintain the combinatorial nature of these problems, rather than relying on linear algebra and continuous methods?

This thesis returns to the classic augmenting paths framework---the original approach to matchings, maximum flow, and matroid intersection---with the goal of developing new efficient combinatorial algorithms. Our key contributions include the first combinatorial algorithm achieving almost-linear time for maximum flow on dense graphs, and the first subquadratic independence-query algorithm for matroid intersection. For modern computational models, our contributions include an improved online rounding scheme for fractional matching (leading to an optimal online edge coloring algorithm), a resolution of the query and communication complexity for bipartite matching, and the first sublinear-round parallel algorithms for matroid intersection.

Abstract [sv]

Matchningar, Maximala Flöden och Matroidsnitt är grundläggande kombinatoriska optimeringsproblem som har studerats ingående sedan datorvetenskapens början.  En serie genombrott inom grafalgoritmik och kontinuerlig optimering under det senaste årentiondet har lett till imponerande nästan optimala algoritmer för maximalt flöde och bipartit matchning. Vi är dock fortfarande långt ifrån att fullt förstå dessa problem. För det första återstår frågan om hur man kan lösa dessa problem i andra beräkningsmodeller, såsom parallell-, dynamisk-, online- och kommunikationsmodeller. För det andra, när algoritmer blir allt mer sofistikerade i deras effektivitetsträvan, offrar de ofta enkelhet, vilket potentiellt kan dölja värdefulla kombinatoriska insikter. Detta motiverar en grundläggande fråga: kan vi utveckla effektiva algoritmer som bevarar den kombinatoriska karaktären hos dessa problem, istället för att förlita sig på linjär algebra och kontinuerliga metoder?

Denna avhandling återgår till de klassiska augmenting-pathsalgoritmerna---det ursprungliga angreppssättet för matchningar, maximalt flöde och matroid-snitt---med målet att utveckla nya effektiva kombinatoriska algoritmer. Våra viktigaste bidrag inkluderar den första kombinatoriska algoritmen som uppnår nästan linjär tid för maximalt flöde på täta grafer, och den första subkvadratiska independence-query-algoritmen för matroidsnitt. För moderna beräkningsmodeller inkluderar våra bidrag förbättrade  onlineavrundningsalgorithmer för fraktionell matchning (vilket leder till en optimal onlinealgoritm för kantfärgning), en lösning av query- och kommunikationskomplexiteten för bipartit matchning, och de första sublinjära parallella algoritmerna för matroidsnitt.

Place, publisher, year, edition, pages
Stockholm: KTH Royal Institute of Technology, 2024. p. x, 67
Series
TRITA-EECS-AVL ; 2024:84
National Category
Computer Sciences
Research subject
Computer Science
Identifiers
urn:nbn:se:kth:diva-355377 (URN)978-91-8106-091-1 (ISBN)
Public defence
2024-11-27, F3, Lindstedtsvägen 26 & 28, KTH, Stockholm, 14:00 (English)
Opponent
Supervisors
Note

QC 20241105

Available from: 2024-11-05 Created: 2024-11-03 Last updated: 2025-12-03Bibliographically approved

Open Access in DiVA

No full text in DiVA

Authority records

Blikstad, Joakim

Search in DiVA

By author/editor
Blikstad, Joakim
By organisation
Theoretical Computer Science, TCS
Computer Sciences

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 141 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf