important separators and parameterized algorithms
play

Important separators and parameterized algorithms Dniel Marx 1 1 - PowerPoint PPT Presentation

Important separators and parameterized algorithms Dniel Marx 1 1 Institute for Computer Science and Control, Hungarian Academy of Sciences (MTA SZTAKI) Budapest, Hungary School on Parameterized Algorithms and Complexity Bdlewo, Poland


  1. Important separators and parameterized algorithms Dániel Marx 1 1 Institute for Computer Science and Control, Hungarian Academy of Sciences (MTA SZTAKI) Budapest, Hungary School on Parameterized Algorithms and Complexity Będlewo, Poland August 19, 2014 1

  2. Main message Small separators in graphs have interesting extremal properties that can be exploited in combinatorial and algorithmic results. Bounding the number of “important” cuts. Edge/vertex versions, directed/undirected versions. Algorithmic applications: FPT algorithm for Multiway cut , Directed Feedback Vertex Set , and ( p , q ) -Clustering . Random selection of important separators: a new tool with many applications. Overview 2

  3. Definition: δ ( R ) is the set of edges with exactly one endpoint in R . Definition: A set S of edges is a minimal ( X , Y ) -cut if there is no X − Y path in G \ S and no proper subset of S breaks every X − Y path. Observation: Every minimal ( X , Y ) -cut S can be expressed as S = δ ( R ) for some X ⊆ R and R ∩ Y = ∅ . δ ( R ) Y X R Minimum cuts 3

  4. Theorem A minimum ( X , Y ) -cut can be found in polynomial time. Theorem The size of a minimum ( X , Y ) -cut equals the maximum size of a pairwise edge-disjoint collection of X − Y paths. δ ( R ) Y X R Minimum cuts 3

  5. There is a long list of algorithms for finding disjoint paths and minimum cuts. Edmonds-Karp: O ( | V ( G ) | · | E ( G ) | 2 ) Dinitz: O ( | V ( G ) | 2 · | E ( G ) | ) Push-relabel: O ( | V ( G ) | 3 ) Orlin-King-Rao-Tarjan: O ( | V ( G ) | · | E ( G ) | ) . . . But we need only the following result: Theorem An ( X , Y ) -cut of size at most k (if exists) can be found in time O ( k · ( | V ( G ) | + | E ( G ) | )) . Finding minimum cuts 4

  6. Theorem An ( X , Y ) -cut of size at most k (if exists) can be found in time O ( k · ( | V ( G ) | + | E ( G ) | )) . We try to grow a collection P of edge-disjoint X − Y paths. Residual graph: not used by P : bidirected, used by P : directed in the opposite direction. original graph residual graph X Y X Y Finding minimum cuts 5

  7. Theorem An ( X , Y ) -cut of size at most k (if exists) can be found in time O ( k · ( | V ( G ) | + | E ( G ) | )) . We try to grow a collection P of edge-disjoint X − Y paths. Residual graph: not used by P : bidirected, used by P : directed in the opposite direction. original graph residual graph X Y X Y Finding minimum cuts 5

  8. Theorem An ( X , Y ) -cut of size at most k (if exists) can be found in time O ( k · ( | V ( G ) | + | E ( G ) | )) . We try to grow a collection P of edge-disjoint X − Y paths. Residual graph: not used by P : bidirected, used by P : directed in the opposite direction. original graph residual graph X Y X Y Finding minimum cuts 5

  9. Theorem An ( X , Y ) -cut of size at most k (if exists) can be found in time O ( k · ( | V ( G ) | + | E ( G ) | )) . We try to grow a collection P of edge-disjoint X − Y paths. Residual graph: not used by P : bidirected, used by P : directed in the opposite direction. original graph residual graph X Y X Y If we cannot find an augmenting path, we can find a (minimum) cut of size |P| . Finding minimum cuts 5

  10. Fact: The function δ is submodular: for arbitrary sets A , B , | δ ( A ) | + | δ ( B ) | ≥ | δ ( A ∩ B ) | + | δ ( A ∪ B ) | Submodularity 6

  11. Fact: The function δ is submodular: for arbitrary sets A , B , | δ ( A ) | + | δ ( B ) | ≥ | δ ( A ∩ B ) | + | δ ( A ∪ B ) | Proof: Determine separately the contribution of the different types of edges. A B Submodularity 6

  12. Fact: The function δ is submodular: for arbitrary sets A , B , | δ ( A ) | + | δ ( B ) | ≥ | δ ( A ∩ B ) | + | δ ( A ∪ B ) | 0 1 1 0 Proof: Determine separately the contribution of the different types of edges. A B Submodularity 6

  13. Fact: The function δ is submodular: for arbitrary sets A , B , | δ ( A ) | + | δ ( B ) | ≥ | δ ( A ∩ B ) | + | δ ( A ∪ B ) | 1 0 1 0 Proof: Determine separately the contribution of the different types of edges. A B Submodularity 6

  14. Fact: The function δ is submodular: for arbitrary sets A , B , | δ ( A ) | + | δ ( B ) | ≥ | δ ( A ∩ B ) | + | δ ( A ∪ B ) | 0 1 0 1 Proof: Determine separately the contribution of the different types of edges. A B Submodularity 6

  15. Fact: The function δ is submodular: for arbitrary sets A , B , | δ ( A ) | + | δ ( B ) | ≥ | δ ( A ∩ B ) | + | δ ( A ∪ B ) | 1 0 0 1 Proof: Determine separately the contribution of the different types of edges. A B Submodularity 6

  16. Fact: The function δ is submodular: for arbitrary sets A , B , | δ ( A ) | + | δ ( B ) | ≥ | δ ( A ∩ B ) | + | δ ( A ∪ B ) | 1 1 1 1 Proof: Determine separately the contribution of the different types of edges. A B Submodularity 6

  17. Fact: The function δ is submodular: for arbitrary sets A , B , | δ ( A ) | + | δ ( B ) | ≥ | δ ( A ∩ B ) | + | δ ( A ∪ B ) | 1 1 0 0 Proof: Determine separately the contribution of the different types of edges. A B Submodularity 6

  18. Lemma Let λ be the minimum ( X , Y ) -cut size. There is a unique maximal R max ⊇ X such that δ ( R max ) is an ( X , Y ) -cut of size λ . Submodularity 7

  19. Lemma Let λ be the minimum ( X , Y ) -cut size. There is a unique maximal R max ⊇ X such that δ ( R max ) is an ( X , Y ) -cut of size λ . Proof: Let R 1 , R 2 ⊇ X be two sets such that δ ( R 1 ) , δ ( R 2 ) are ( X , Y ) -cuts of size λ . Y | δ ( R 1 ) | + | δ ( R 2 ) | ≥ | δ ( R 1 ∩ R 2 ) | + | δ ( R 1 ∪ R 2 ) | ≥ λ λ λ R 1 R 2 ⇒ | δ ( R 1 ∪ R 2 ) | ≤ λ X Note: Analogous result holds for a unique minimal R min . Submodularity 7

  20. Lemma Given a graph G and sets X , Y ⊆ V ( G ) , the sets R min and R max can be found in polynomial time. Proof: Iteratively add vertices to X if they do not increase the minimum X − Y cut size. When the process stops, X = R max . Similar for R min . But we can do better! Finding R min and R max 8

  21. Lemma Given a graph G and sets X , Y ⊆ V ( G ) , the sets R min and R max can be found in O ( λ · ( | V ( G ) | + | E ( G ) | )) time, where λ is the minimum X − Y cut size. Proof: Look at the residual graph. original graph residual graph X Y X Y R min R max R min R max R min : vertices reachable from X . R max : vertices from which Y is not reachable. Finding R min and R max 9

  22. Definition: δ ( R ) is the set of edges with exactly one endpoint in R . Definition: A set S of edges is a minimal ( X , Y ) -cut if there is no X − Y path in G \ S and no proper subset of S breaks every X − Y path. Observation: Every minimal ( X , Y ) -cut S can be expressed as S = δ ( R ) for some X ⊆ R and R ∩ Y = ∅ . δ ( R ) Y X R Important cuts 10

  23. Definition A minimal ( X , Y ) -cut δ ( R ) is important if there is no ( X , Y ) -cut δ ( R ′ ) with R ⊂ R ′ and | δ ( R ′ ) | ≤ | δ ( R ) | . Note: Can be checked in polynomial time if a cut is important ( δ ( R ) is important if R = R max ). δ ( R ) Y X R Important cuts 10

  24. Definition A minimal ( X , Y ) -cut δ ( R ) is important if there is no ( X , Y ) -cut δ ( R ′ ) with R ⊂ R ′ and | δ ( R ′ ) | ≤ | δ ( R ) | . Note: Can be checked in polynomial time if a cut is important ( δ ( R ) is important if R = R max ). δ ( R ) Y X δ ( R ′ ) R R ′ Important cuts 10

  25. Definition A minimal ( X , Y ) -cut δ ( R ) is important if there is no ( X , Y ) -cut δ ( R ′ ) with R ⊂ R ′ and | δ ( R ′ ) | ≤ | δ ( R ) | . Note: Can be checked in polynomial time if a cut is important ( δ ( R ) is important if R = R max ). δ ( R ) Y X R Important cuts 10

  26. The number of important cuts can be exponentially large. Example: Y 1 2 k / 2 X This graph has 2 k / 2 important ( X , Y ) -cuts of size at most k . Important cuts 11

  27. The number of important cuts can be exponentially large. Example: Y 1 2 k / 2 X This graph has 2 k / 2 important ( X , Y ) -cuts of size at most k . Theorem There are at most 4 k important ( X , Y ) -cuts of size at most k . Important cuts 11

  28. Theorem There are at most 4 k important ( X , Y ) -cuts of size at most k . Proof: Let λ be the minimum ( X , Y ) -cut size and let δ ( R max ) be the unique important cut of size λ such that R max is maximal. (1) We show that R max ⊆ R for every important cut δ ( R ) . Important cuts 12

  29. Theorem There are at most 4 k important ( X , Y ) -cuts of size at most k . Proof: Let λ be the minimum ( X , Y ) -cut size and let δ ( R max ) be the unique important cut of size λ such that R max is maximal. (1) We show that R max ⊆ R for every important cut δ ( R ) . By the submodularity of δ : | δ ( R max ) | + | δ ( R ) | ≥ | δ ( R max ∩ R ) | + | δ ( R max ∪ R ) | λ ≥ λ ⇓ | δ ( R max ∪ R ) | ≤ | δ ( R ) | ⇓ If R � = R max ∪ R , then δ ( R ) is not important. Important cuts 12

Recommend


More recommend