Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf ·...
Transcript of Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf ·...
![Page 1: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/1.jpg)
Algorithms and complexity of graph convexity problems
Vinıcius F. dos Santos
Centro Federal de Educacao Tecnologica de Minas Gerais (CEFET-MG)Belo Horizonte, Brazil
Joint Work with:Igor da Fonseca Ramos
Jayme L. Szwarcfiter
December 2014
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 1 / 36
![Page 2: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/2.jpg)
Introduction Graph Convexity
Convexity
Definition: convexity
Given a finite ground set V and a family C of subsets of V , C is aconvexity if the following conditions are satisfied:
(a) The sets ∅ and V belong to C; and
(b) C is closed for intersections.
Definition: convex setA set is convex if it belongs to C.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 2 / 36
![Page 3: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/3.jpg)
Introduction Graph Convexity
Convexity
Definition: convexity
Given a finite ground set V and a family C of subsets of V , C is aconvexity if the following conditions are satisfied:
(a) The sets ∅ and V belong to C; and
(b) C is closed for intersections.
Definition: convex setA set is convex if it belongs to C.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 2 / 36
![Page 4: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/4.jpg)
Introduction Graph Convexity
Graph convexity
For a graph G = (V ,E ), the ground set is V .
The family C contains subsets of V .
Path convexity
We can use a family of paths FG to define C:
S ⊆ V (G ) is convex if, for every pair u, v ∈ S , every w ∈ V (G ) in auv -path in FG is also in S .
Interval function: I (S) is the set of vertices that belong to at leastone uv -path of FG , u, v ∈ S
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 3 / 36
![Page 5: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/5.jpg)
Introduction Graph Convexity
Graph convexity
For a graph G = (V ,E ), the ground set is V .
The family C contains subsets of V .
Path convexity
We can use a family of paths FG to define C:
S ⊆ V (G ) is convex if, for every pair u, v ∈ S , every w ∈ V (G ) in auv -path in FG is also in S .
Interval function: I (S) is the set of vertices that belong to at leastone uv -path of FG , u, v ∈ S
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 3 / 36
![Page 6: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/6.jpg)
Introduction Graph Convexity
Convexity
Some well studied graph convexities
Geodetic Convexity
Monophonic Convexity
P3 Convexity
P∗3 Convexity
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 4 / 36
![Page 7: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/7.jpg)
Introduction Main path convexities
Geodetic convexity
Family of paths FG consists of all the shortest paths of G .
Convex setsS ⊆ V (G ) is convex if and only if for every pair u, v ∈ S , every vertex in auv -shortest path in G is also in S .
Figure: Convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 5 / 36
![Page 8: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/8.jpg)
Introduction Main path convexities
Geodetic convexity
Family of paths FG consists of all the shortest paths of G .
Convex setsS ⊆ V (G ) is convex if and only if for every pair u, v ∈ S , every vertex in auv -shortest path in G is also in S .
Figure: Convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 5 / 36
![Page 9: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/9.jpg)
Introduction Main path convexities
Geodetic convexity
Family of paths FG consists of all the shortest paths of G .
Convex setsS ⊆ V (G ) is convex if and only if for every pair u, v ∈ S , every vertex in auv -shortest path in G is also in S .
Figure: Non-convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 5 / 36
![Page 10: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/10.jpg)
Introduction Main path convexities
Monophonic convexity
Family of paths FG consists of the induced paths of G .
Convex setsS ⊆ V (G ) is convex if and only if for every pair u, v ∈ S , every vertex in auv -induced path of G also belongs to S .
Figure: Convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 6 / 36
![Page 11: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/11.jpg)
Introduction Main path convexities
Monophonic convexity
Family of paths FG consists of the induced paths of G .
Convex setsS ⊆ V (G ) is convex if and only if for every pair u, v ∈ S , every vertex in auv -induced path of G also belongs to S .
Figure: Convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 6 / 36
![Page 12: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/12.jpg)
Introduction Main path convexities
Monophonic convexity
Family of paths FG consists of the induced paths of G .
Convex setsS ⊆ V (G ) is convex if and only if for every pair u, v ∈ S , every vertex in auv -induced path of G also belongs to S .
Figure: Non-convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 6 / 36
![Page 13: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/13.jpg)
Introduction Main path convexities
P3 convexity
Family of paths FG contain the paths of order 3 of G .
Convex setsS ⊆ V (G ) is convex if and only if for all u, v ∈ S , every path in a uv -pathof order 3 is also in S .
Figure: Convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 7 / 36
![Page 14: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/14.jpg)
Introduction Main path convexities
P3 convexity
Family of paths FG contain the paths of order 3 of G .
Convex setsS ⊆ V (G ) is convex if there is no v ∈ V (G ) \ S with two or moreneighbors in S .
Figure: Convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 7 / 36
![Page 15: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/15.jpg)
Introduction Main path convexities
P3 convexity
Family of paths FG contain the paths of order 3 of G .
Convex setsS ⊆ V (G ) is convex if there is no v ∈ V (G ) \ S with two or moreneighbors in S .
Figure: Convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 7 / 36
![Page 16: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/16.jpg)
Introduction Main path convexities
P3 convexity
Family of paths FG contain the paths of order 3 of G .
Convex setsS ⊆ V (G ) is convex if there is no v ∈ V (G ) \ S with two or moreneighbors in S .
Figure: Non-convex set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 7 / 36
![Page 17: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/17.jpg)
Introduction Convex Hull
Convex Hull in graphs
Convex hull of S : smallest convex superset containing S .
Denoted by H(S).
If v ∈ H(S) \ S , we say that v is generated by S .
Example on the P3 convexity:
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 8 / 36
![Page 18: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/18.jpg)
Introduction Convex Hull
Convex Hull in graphs
Convex hull of S : smallest convex superset containing S .
Denoted by H(S).
If v ∈ H(S) \ S , we say that v is generated by S .
Example on the P3 convexity:
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 8 / 36
![Page 19: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/19.jpg)
Introduction Convex Hull
Convex Hull in graphs
Convex hull of S : smallest convex superset containing S .
Denoted by H(S).
If v ∈ H(S) \ S , we say that v is generated by S .
Example on the P3 convexity:
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 8 / 36
![Page 20: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/20.jpg)
Introduction Convex Hull
Algorithm to determine the convex hull
Greedy algorithm.
Start with H(S) = S .
Repeat:
Let R = I (H(S));H(S)← H(S) ∪ R.
Example on the P3 convexity:
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 9 / 36
![Page 21: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/21.jpg)
Introduction Convex Hull
Algorithm to determine the convex hull
Greedy algorithm.
Start with H(S) = S .
Repeat:
Let R = I (H(S));H(S)← H(S) ∪ R.
Example on the P3 convexity:
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 9 / 36
![Page 22: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/22.jpg)
Introduction Convex Hull
Algorithm to determine the convex hull
Greedy algorithm.
Start with H(S) = S .
Repeat:
Let R = I (H(S));H(S)← H(S) ∪ R.
Example on the P3 convexity:
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 9 / 36
![Page 23: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/23.jpg)
Introduction Convex Hull
Algorithm to determine the convex hull
Greedy algorithm.
Start with H(S) = S .
Repeat:
Let R = I (H(S));H(S)← H(S) ∪ R.
Example on the P3 convexity:
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 9 / 36
![Page 24: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/24.jpg)
Introduction Convex Hull
Algorithm to determine the convex hull
Greedy algorithm.
Start with H(S) = S .
Repeat:
Let R = I (H(S));H(S)← H(S) ∪ R.
Example on the P3 convexity:
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 9 / 36
![Page 25: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/25.jpg)
Introduction Convex Hull
Algorithm to determine the convex hull
Greedy algorithm.
Start with H(S) = S .
Repeat:
Let R = I (H(S));H(S)← H(S) ∪ R.
Example on the P3 convexity:
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 9 / 36
![Page 26: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/26.jpg)
Introduction Convex Hull
Hull set
Definition: hull setIf H(S) = V (G ), S is a hull set of G .
Figure: A hull set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 10 / 36
![Page 27: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/27.jpg)
Introduction Convex Hull
Hull set
Definition: hull setIf H(S) = V (G ), S is a hull set of G .
Figure: A hull set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 10 / 36
![Page 28: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/28.jpg)
Introduction Some parameters of a graph convexity
Parameters of a graph convexity
Definition: hull numberThe hull number of a graph is the cardinality of its smallest hull set.
Definition: interval numberThe minimum size of a set S such that I (S) = V (G ).
Definition: convexity number
The maximum size of a convex set S such that S 6= V (G ).
Definition: partition number
The maximum integer k such that V (G ) can be partitioned into p convexsets for any p ∈ {1, . . . , k}.
Definition: percolation time
The maximum number of iteractions of the greed algorithm to generatethe whole graph.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 11 / 36
![Page 29: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/29.jpg)
Introduction Some parameters of a graph convexity
Parameters of a graph convexity
Definition: hull numberThe hull number of a graph is the cardinality of its smallest hull set.
Definition: interval numberThe minimum size of a set S such that I (S) = V (G ).
Definition: convexity number
The maximum size of a convex set S such that S 6= V (G ).
Definition: partition number
The maximum integer k such that V (G ) can be partitioned into p convexsets for any p ∈ {1, . . . , k}.
Definition: percolation time
The maximum number of iteractions of the greed algorithm to generatethe whole graph.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 11 / 36
![Page 30: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/30.jpg)
Introduction Some parameters of a graph convexity
Parameters of a graph convexity
Definition: hull numberThe hull number of a graph is the cardinality of its smallest hull set.
Definition: interval numberThe minimum size of a set S such that I (S) = V (G ).
Definition: convexity number
The maximum size of a convex set S such that S 6= V (G ).
Definition: partition number
The maximum integer k such that V (G ) can be partitioned into p convexsets for any p ∈ {1, . . . , k}.
Definition: percolation time
The maximum number of iteractions of the greed algorithm to generatethe whole graph.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 11 / 36
![Page 31: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/31.jpg)
Introduction Some parameters of a graph convexity
Parameters of a graph convexity
Definition: hull numberThe hull number of a graph is the cardinality of its smallest hull set.
Definition: interval numberThe minimum size of a set S such that I (S) = V (G ).
Definition: convexity number
The maximum size of a convex set S such that S 6= V (G ).
Definition: partition number
The maximum integer k such that V (G ) can be partitioned into p convexsets for any p ∈ {1, . . . , k}.
Definition: percolation time
The maximum number of iteractions of the greed algorithm to generatethe whole graph.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 11 / 36
![Page 32: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/32.jpg)
Introduction Some parameters of a graph convexity
Parameters of a graph convexity
Definition: hull numberThe hull number of a graph is the cardinality of its smallest hull set.
Definition: interval numberThe minimum size of a set S such that I (S) = V (G ).
Definition: convexity number
The maximum size of a convex set S such that S 6= V (G ).
Definition: partition number
The maximum integer k such that V (G ) can be partitioned into p convexsets for any p ∈ {1, . . . , k}.
Definition: percolation time
The maximum number of iteractions of the greed algorithm to generatethe whole graph.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 11 / 36
![Page 33: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/33.jpg)
Introduction Some parameters of a graph convexity
Radon number
Definition: Radon partition
Given a set R, a partition R = R1 ∪ R2 such that H(R1) ∩ H(R2) 6= ∅ is aRadon partition.
Definition: Radon-independent set
If a set does not admit a Radon partition we say it is a Radon-independentset.
Definition: Radon numberThe Radon number r(G ) is given bymax{|R|,R is a Radon-independent set}+ 1.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 12 / 36
![Page 34: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/34.jpg)
Introduction Some parameters of a graph convexity
Radon number
Definition: Radon partition
Given a set R, a partition R = R1 ∪ R2 such that H(R1) ∩ H(R2) 6= ∅ is aRadon partition.
Definition: Radon-independent set
If a set does not admit a Radon partition we say it is a Radon-independentset.
Definition: Radon numberThe Radon number r(G ) is given bymax{|R|,R is a Radon-independent set}+ 1.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 12 / 36
![Page 35: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/35.jpg)
Introduction Some parameters of a graph convexity
Radon number
Definition: Radon partition
Given a set R, a partition R = R1 ∪ R2 such that H(R1) ∩ H(R2) 6= ∅ is aRadon partition.
Definition: Radon-independent set
If a set does not admit a Radon partition we say it is a Radon-independentset.
Definition: Radon numberThe Radon number r(G ) is given bymax{|R|,R is a Radon-independent set}+ 1.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 12 / 36
![Page 36: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/36.jpg)
Introduction Some parameters of a graph convexity
Caratheodory number
Definition: Caratheodory-independent set
A set S such thatH(S) \
⋃v∈S
H(S \ {v}) 6= ∅.
Definition: Caratheodory number
Largest cardinality of a Caratheodory-independent set.
Smallest integer k such that for every U ⊆ V and every u ∈ H(U), thereis a set F ⊆ U with |F | ≤ k and u ∈ H(F ).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 13 / 36
![Page 37: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/37.jpg)
Introduction Some parameters of a graph convexity
Caratheodory number
Definition: Caratheodory-independent set
A set S such thatH(S) \
⋃v∈S
H(S \ {v}) 6= ∅.
Definition: Caratheodory number
Largest cardinality of a Caratheodory-independent set.
Smallest integer k such that for every U ⊆ V and every u ∈ H(U), thereis a set F ⊆ U with |F | ≤ k and u ∈ H(F ).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 13 / 36
![Page 38: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/38.jpg)
Introduction Some parameters of a graph convexity
Helly number
Definition: Helly-independent set
A set S such that ⋂v∈S
H(S \ {v}) = ∅.
Definition: Helly number
The cardinality of a maximum Helly-independent set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 14 / 36
![Page 39: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/39.jpg)
Introduction Some parameters of a graph convexity
Helly number
Definition: Helly-independent set
A set S such that ⋂v∈S
H(S \ {v}) = ∅.
Definition: Helly number
The cardinality of a maximum Helly-independent set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 14 / 36
![Page 40: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/40.jpg)
The rank and the Convexly Independent Setproblem
Convexly independent sets
Definition: convexly independent set
S ⊆ V (G ) is a convexly independent set if, for all v ∈ S , v /∈ H(S − v).Equivalently, no vertex of S is generated by the others.
If a vertex of S ⊆ V (G ) is generated by the others, S is convexlydependent.
Figure: Convexly dependent set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 15 / 36
![Page 41: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/41.jpg)
The rank and the Convexly Independent Setproblem
Convexly independent sets
Definition: convexly independent set
S ⊆ V (G ) is a convexly independent set if, for all v ∈ S , v /∈ H(S − v).Equivalently, no vertex of S is generated by the others.
If a vertex of S ⊆ V (G ) is generated by the others, S is convexlydependent.
Figure: Convexly dependent set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 15 / 36
![Page 42: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/42.jpg)
The rank and the Convexly Independent Setproblem
Convexly independent sets
Definition: convexly independent set
S ⊆ V (G ) is a convexly independent set if, for all v ∈ S , v /∈ H(S − v).Equivalently, no vertex of S is generated by the others.
If a vertex of S ⊆ V (G ) is generated by the others, S is convexlydependent.
Figure: v4 ∈ H({v1, v13}).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 15 / 36
![Page 43: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/43.jpg)
The rank and the Convexly Independent Setproblem
Convexly independent sets
Definition: convexly independent set
S ⊆ V (G ) is a convexly independent set if, for all v ∈ S , v /∈ H(S − v).Equivalently, no vertex of S is generated by the others.
If a vertex of S ⊆ V (G ) is generated by the others, S is convexlydependent.
Figure: Convexly independent set.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 15 / 36
![Page 44: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/44.jpg)
The rank and the Convexly Independent Setproblem
Convexly independent sets
Definition: convexly independent set
S ⊆ V (G ) is a convexly independent set if, for all v ∈ S , v /∈ H(S − v).Equivalently, no vertex of S is generated by the others.
If a vertex of S ⊆ V (G ) is generated by the others, S is convexlydependent.
Figure: v1 /∈ H({v7, v12, v13}).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 15 / 36
![Page 45: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/45.jpg)
The rank and the Convexly Independent Setproblem
Convexly independent sets
Definition: convexly independent set
S ⊆ V (G ) is a convexly independent set if, for all v ∈ S , v /∈ H(S − v).Equivalently, no vertex of S is generated by the others.
If a vertex of S ⊆ V (G ) is generated by the others, S is convexlydependent.
Figure: v13 /∈ H({v1, v7, v12}).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 15 / 36
![Page 46: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/46.jpg)
The rank and the Convexly Independent Setproblem
Convexly independent sets
Definition: convexly independent set
S ⊆ V (G ) is a convexly independent set if, for all v ∈ S , v /∈ H(S − v).Equivalently, no vertex of S is generated by the others.
If a vertex of S ⊆ V (G ) is generated by the others, S is convexlydependent.
Figure: v12 /∈ H({v1, v7, v13}).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 15 / 36
![Page 47: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/47.jpg)
The rank and the Convexly Independent Setproblem
Convexly independent sets
Definition: convexly independent set
S ⊆ V (G ) is a convexly independent set if, for all v ∈ S , v /∈ H(S − v).Equivalently, no vertex of S is generated by the others.
If a vertex of S ⊆ V (G ) is generated by the others, S is convexlydependent.
Figure: v7 /∈ H({v1, v12, v13}).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 15 / 36
![Page 48: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/48.jpg)
The rank and the Convexly Independent Setproblem
The rank of a graph
Definition: rankThe rank of a graph, denoted by rk(G ), is the cardinality of a maximumconvexly independent set of G .
Convexly Independent SetInput: A graph G and an integer k.Question: rk(G ) ≥ k?
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 16 / 36
![Page 49: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/49.jpg)
The rank and the Convexly Independent Setproblem
The rank of a graph
Definition: rankThe rank of a graph, denoted by rk(G ), is the cardinality of a maximumconvexly independent set of G .
Convexly Independent SetInput: A graph G and an integer k.Question: rk(G ) ≥ k?
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 16 / 36
![Page 50: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/50.jpg)
The rank and the Convexly Independent Setproblem
State of the art
Parameter Geodetic Monophonic P3 P∗3
Interval of a set Polynomial NP-complete Polynomial PolynomialHull NP-complete Polynomial NP-complete NP-complete
Convexity NP-complete NP-complete NP-complete NP-completeRadon NP-complete NP-complete NP-hard NP-hard
Caratheodory NP-complete Polynomial NP-complete NP-completeHelly coNP-complete Polynomial Open OpenRank NP-complete NP-complete NP-complete NP-complete
Partition NP-complete Open NP-complete OpenPercolation time NP-complete NP-complete NP-complete NP-complete
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 17 / 36
![Page 51: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/51.jpg)
The rank and the Convexly Independent Setproblem
Connections with other graph parameters
k-tuple domination.
Conversion/percolation problems.
Feedback vertex set.
Open packing number.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 18 / 36
![Page 52: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/52.jpg)
The rank and the Convexly Independent Setproblem
Connections with other graph parameters
k-tuple domination.
Conversion/percolation problems.
Feedback vertex set.
Open packing number.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 18 / 36
![Page 53: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/53.jpg)
The rank and the Convexly Independent Setproblem
Connections with other graph parameters
k-tuple domination.
Conversion/percolation problems.
Feedback vertex set.
Open packing number.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 18 / 36
![Page 54: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/54.jpg)
The rank and the Convexly Independent Setproblem
Connections with other graph parameters
k-tuple domination.
Conversion/percolation problems.
Feedback vertex set.
Open packing number.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 18 / 36
![Page 55: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/55.jpg)
The rank and the Convexly Independent Setproblem
The open packing number
Definition: open packing
An open packing of a graph G is a set S such that, for every pair u, v ∈ S ,N(u) ∩ N(v) = ∅.
Definition: open packing number
The open packing number of a graph, denoted by ρo(G ), is the cardinalityof a maximum open packing of G .
Open Packing NumberInput: A graph G and an integer k.Question: ρo(G ) ≥ k?
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 19 / 36
![Page 56: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/56.jpg)
The rank and the Convexly Independent Setproblem
The open packing number
Definition: open packing
An open packing of a graph G is a set S such that, for every pair u, v ∈ S ,N(u) ∩ N(v) = ∅.
Definition: open packing number
The open packing number of a graph, denoted by ρo(G ), is the cardinalityof a maximum open packing of G .
Open Packing NumberInput: A graph G and an integer k.Question: ρo(G ) ≥ k?
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 19 / 36
![Page 57: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/57.jpg)
The rank and the Convexly Independent Setproblem
The open packing number
Definition: open packing
An open packing of a graph G is a set S such that, for every pair u, v ∈ S ,N(u) ∩ N(v) = ∅.
Definition: open packing number
The open packing number of a graph, denoted by ρo(G ), is the cardinalityof a maximum open packing of G .
Open Packing NumberInput: A graph G and an integer k.Question: ρo(G ) ≥ k?
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 19 / 36
![Page 58: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/58.jpg)
The rank and the Convexly Independent Setproblem
ρo(G ) and rk(G )
If S is an open packing, H(S) = S . Hence, every open packing isconvexly independent.
Not every convexly independent set is an open packing.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 20 / 36
![Page 59: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/59.jpg)
The rank and the Convexly Independent Setproblem
ρo(G ) and rk(G )
If S is an open packing, H(S) = S . Hence, every open packing isconvexly independent.
Not every convexly independent set is an open packing.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 20 / 36
![Page 60: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/60.jpg)
The rank and the Convexly Independent Setproblem
ρo(G ) and rk(G )
If S is an open packing, H(S) = S . Hence, every open packing isconvexly independent.
Not every convexly independent set is an open packing.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 20 / 36
![Page 61: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/61.jpg)
The rank and the Convexly Independent Setproblem
ρo(G ) and rk(G )
If S is an open packing, H(S) = S . Hence, every open packing isconvexly independent.
Not every convexly independent set is an open packing.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 20 / 36
![Page 62: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/62.jpg)
Results NP-completeness
Split graphs on the P3 convexity
TheoremThe Convexly Independent Set problem is NP-complete on the P3
convexity, even for split graphs with δ(G ) ≥ 2.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 21 / 36
![Page 63: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/63.jpg)
Results NP-completeness
Proof sketch
Reduction from Set Packing.
Set PackingInput: A family S of non-empty subsets Si ∈ S of a ground set and aninteger k .Question: S has at least k pairwise disjoint sets?
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 22 / 36
![Page 64: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/64.jpg)
Results NP-completeness
Proof sketch
S1 = {1, 6}, S2 = {1, 2},S3 = {2, 3},S4 = {3, 4}, S5 = {4, 5},S6 = {5, 6}.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 23 / 36
![Page 65: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/65.jpg)
Results NP-completeness
Proof sketch
S1 = {1, 6}, S2 = {1, 2},S3 = {2, 3},S4 = {3, 4}, S5 = {4, 5},S6 = {5, 6}.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 23 / 36
![Page 66: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/66.jpg)
Results NP-completeness
Proof sketch
S1 = {1, 6}, S2 = {1, 2},S3 = {2, 3},S4 = {3, 4}, S5 = {4, 5},S6 = {5, 6}.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 23 / 36
![Page 67: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/67.jpg)
Results NP-completeness
Proof sketch
S1 = {1, 6}, S2 = {1, 2},S3 = {2, 3},S4 = {3, 4}, S5 = {4, 5},S6 = {5, 6}.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 23 / 36
![Page 68: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/68.jpg)
Results NP-completeness
Proof sketch
S1 = {1, 6}, S2 = {1, 2},S3 = {2, 3},S4 = {3, 4}, S5 = {4, 5},S6 = {5, 6}.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 23 / 36
![Page 69: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/69.jpg)
Results NP-completeness
Proof sketch
S1 = {1, 6}, S2 = {1, 2},S3 = {2, 3},S4 = {3, 4}, S5 = {4, 5},S6 = {5, 6}.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 23 / 36
![Page 70: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/70.jpg)
Results NP-completeness
Open packing number
Corollary
The Open Packing Number is NP-complete for split graphs withδ(G ) ≥ 2.
It was known that the problem was NP-hard for chordal graphs [Henningand Slater, 1999].
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 24 / 36
![Page 71: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/71.jpg)
Results NP-completeness
Open packing number
Corollary
The Open Packing Number is NP-complete for split graphs withδ(G ) ≥ 2.
It was known that the problem was NP-hard for chordal graphs [Henningand Slater, 1999].
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 24 / 36
![Page 72: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/72.jpg)
Results NP-completeness
Bipartite graphs on the P3 convexity
TheoremThe Convexly Independent Set problem is NP-complete on the P3
convexity, even for bipartite graphs with diameter at most 3.
Reduction from Convexly Independent Set for split graphs withδ(G ) ≥ 2.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 25 / 36
![Page 73: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/73.jpg)
Results NP-completeness
Bipartite graphs on the P3 convexity
TheoremThe Convexly Independent Set problem is NP-complete on the P3
convexity, even for bipartite graphs with diameter at most 3.
Reduction from Convexly Independent Set for split graphs withδ(G ) ≥ 2.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 25 / 36
![Page 74: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/74.jpg)
Results Polynomial time algorithms
Threshold graphs on the P3 convexity
TheoremThe Convexly Independent Set problem can be solved in linear timefor threshold graphs on the P3 convexity.
TheoremIf G is a connected threshold graph with |V (G )| ≥ 3 and D ⊆ V (G ) is aset with all vertices of minimum degree G , then:
(i) if G is a star, then rk(G ) = |V (G )| − 1;
(ii) otherwise, if d(v) = 1 for all v ∈ D, then rk(G ) = |D|+ 1;
(iii) otherwise, rk(G ) = 2.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 26 / 36
![Page 75: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/75.jpg)
Results Polynomial time algorithms
Threshold graphs on the P3 convexity
TheoremThe Convexly Independent Set problem can be solved in linear timefor threshold graphs on the P3 convexity.
TheoremIf G is a connected threshold graph with |V (G )| ≥ 3 and D ⊆ V (G ) is aset with all vertices of minimum degree G , then:
(i) if G is a star, then rk(G ) = |V (G )| − 1;
(ii) otherwise, if d(v) = 1 for all v ∈ D, then rk(G ) = |D|+ 1;
(iii) otherwise, rk(G ) = 2.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 26 / 36
![Page 76: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/76.jpg)
Results Polynomial time algorithms
Threshold graphs on the P3 convexity
TheoremThe Convexly Independent Set problem can be solved in linear timefor threshold graphs on the P3 convexity.
TheoremIf G is a connected threshold graph with |V (G )| ≥ 3 and D ⊆ V (G ) is aset with all vertices of minimum degree G , then:
(i) if G is a star, then rk(G ) = |V (G )| − 1;
(ii) otherwise, if d(v) = 1 for all v ∈ D, then rk(G ) = |D|+ 1;
(iii) otherwise, rk(G ) = 2.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 26 / 36
![Page 77: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/77.jpg)
Results Polynomial time algorithms
Trees on the P3 convexity
TheoremThe Convexly Independent Set problem can be solved in timeO(n log ∆(T )) for trees on the P3 convexity.
Dynamic Programming.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 27 / 36
![Page 78: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/78.jpg)
Results Polynomial time algorithms
Trees on the P3 convexity
TheoremThe Convexly Independent Set problem can be solved in timeO(n log ∆(T )) for trees on the P3 convexity.
Dynamic Programming.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 27 / 36
![Page 79: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/79.jpg)
Results Polynomial time algorithms
Algorithm idea
Given a tree T .
Select a root r ∈ V (T ).
Consider that u ∈ V (T ) sends charge to v ∈ V (T ) if u and v areadjacent, u ∈ H(S) and u does not depend on v to be in H(S).
Pv (i , j , k) is the contribuition of v : maximum number of vertices ofthe subtree rooted on v that can be on the maximum convexlyindependent set under the condition given by i , j e k :
i = 1: the parent of v sends charge to v .j = 1: v is in the convexly independent set being considered.k: number of children sending charge to v .
Define:
f (v , i) = max{Pv (i , 0, 0),Pv (i , 0, 1)}.h(v , i) = max{max2≤k<d(v){P(i , 0, k)},max0≤k≤d(v) Pv (i , 1, k)}.g(v , i1, i2) = h(v , i1)− f (v , i2).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 28 / 36
![Page 80: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/80.jpg)
Results Polynomial time algorithms
Algorithm idea
Given a tree T .
Select a root r ∈ V (T ).
Consider that u ∈ V (T ) sends charge to v ∈ V (T ) if u and v areadjacent, u ∈ H(S) and u does not depend on v to be in H(S).
Pv (i , j , k) is the contribuition of v : maximum number of vertices ofthe subtree rooted on v that can be on the maximum convexlyindependent set under the condition given by i , j e k :
i = 1: the parent of v sends charge to v .j = 1: v is in the convexly independent set being considered.k: number of children sending charge to v .
Define:
f (v , i) = max{Pv (i , 0, 0),Pv (i , 0, 1)}.h(v , i) = max{max2≤k<d(v){P(i , 0, k)},max0≤k≤d(v) Pv (i , 1, k)}.g(v , i1, i2) = h(v , i1)− f (v , i2).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 28 / 36
![Page 81: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/81.jpg)
Results Polynomial time algorithms
Algorithm idea
Given a tree T .
Select a root r ∈ V (T ).
Consider that u ∈ V (T ) sends charge to v ∈ V (T ) if u and v areadjacent, u ∈ H(S) and u does not depend on v to be in H(S).
Pv (i , j , k) is the contribuition of v : maximum number of vertices ofthe subtree rooted on v that can be on the maximum convexlyindependent set under the condition given by i , j e k :
i = 1: the parent of v sends charge to v .j = 1: v is in the convexly independent set being considered.k: number of children sending charge to v .
Define:
f (v , i) = max{Pv (i , 0, 0),Pv (i , 0, 1)}.h(v , i) = max{max2≤k<d(v){P(i , 0, k)},max0≤k≤d(v) Pv (i , 1, k)}.g(v , i1, i2) = h(v , i1)− f (v , i2).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 28 / 36
![Page 82: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/82.jpg)
Results Polynomial time algorithms
Algorithm idea
Given a tree T .
Select a root r ∈ V (T ).
Consider that u ∈ V (T ) sends charge to v ∈ V (T ) if u and v areadjacent, u ∈ H(S) and u does not depend on v to be in H(S).
Pv (i , j , k) is the contribuition of v : maximum number of vertices ofthe subtree rooted on v that can be on the maximum convexlyindependent set under the condition given by i , j e k :
i = 1: the parent of v sends charge to v .j = 1: v is in the convexly independent set being considered.k: number of children sending charge to v .
Define:
f (v , i) = max{Pv (i , 0, 0),Pv (i , 0, 1)}.h(v , i) = max{max2≤k<d(v){P(i , 0, k)},max0≤k≤d(v) Pv (i , 1, k)}.g(v , i1, i2) = h(v , i1)− f (v , i2).
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 28 / 36
![Page 83: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/83.jpg)
Results Polynomial time algorithms
Recurrence relation I
Pv (0, 0, 0) =∑
u∈N′(v)
f (u, 0); (1)
Pv (0, 0, 1) =
−∞, if v have no children,∑
u∈N′(v)
f (u, 0) + maxu∈N′(v)
g(u, 0, 0), otherwise;(2)
Pv (0, 0, 2) =
−∞, if v has at most 1 child,∑
u∈N′(v)
f (u, 1) + max∀X⊆N′(v)
|X|=2
∑u∈X
g(u, 0, 1), otherwise;(3)
Pv (0, 0, k)k≥3
=
−∞, if v less than k children,∑
u∈N′(v)
f (u, 1) + max∀X⊆N′(v)
|X|=k
∑u∈X
g(u, 1, 1), otherwise;(4)
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 29 / 36
![Page 84: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/84.jpg)
Results Polynomial time algorithms
Recurrence relation II
Pv (0, 1, 0) =∑
u∈N′(v)
f (u, 1) + 1; (5)
Pv (0, 1, 1) =
−∞, if v have no children,∑
u∈N′(v)
f (u, 1) + maxu∈N′(v)
g(u, 1, 1) + 1, otherwise;(6)
Pv (0, 1, k)k≥2
= −∞; (7)
Pv (1, 0, 0) =
−∞, if v = r ,∑
u∈N′(v)
f (u, 0), otherwise;(8)
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 30 / 36
![Page 85: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/85.jpg)
Results Polynomial time algorithms
Recurrence relation III
Pv (1, 0, 1) =
−∞, if v have no children ou v = r ,∑
u∈N′(v)
f (u, 1) + maxu∈N′(v)
g(u, 0, 1), otherwise;(9)
Pv (1, 0, k)k≥2
=
−∞, if v has less than k children or v = r ,∑
u∈N′(v)
f (u, 1) + max∀S⊆N′(v)
|S|=k
∑u∈S
g(u, 1, 1), otherwise;
(10)
Pv (1, 1, 0) =
−∞, if v = r ,∑
u∈N′(v)
f (u, 1) + 1, otherwise;(11)
Pv (1, 1, k)k≥1
= −∞. (12)
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 31 / 36
![Page 86: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/86.jpg)
Results Polynomial time algorithms
Graphs without separating cliques on the monophonicconvexity
TheoremThe Convexly Independent Set problem is NP-complete on themonophonic convexity, even for graphs without separating clique.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 32 / 36
![Page 87: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/87.jpg)
Results Polynomial time algorithms
Graphs without separating cliques on the monophonicconvexity
Theorem [Dourado, Protti, Szwarcfiter, 2010]
If G is a graph with no separating clique, but is not a complete graph,then every pair of non-adjacent vertices is a hull set of G on themonophonic convexity.
LemmaThe Clique problem is NP-complete, even for graphs without separatingclique.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 33 / 36
![Page 88: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/88.jpg)
Results Polynomial time algorithms
Graphs without separating cliques on the monophonicconvexity
Theorem [Dourado, Protti, Szwarcfiter, 2010]
If G is a graph with no separating clique, but is not a complete graph,then every pair of non-adjacent vertices is a hull set of G on themonophonic convexity.
LemmaThe Clique problem is NP-complete, even for graphs without separatingclique.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 33 / 36
![Page 89: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/89.jpg)
Results Polynomial time algorithms
Open problems
Parameter Geodetic Monophonic P3 P∗3
Interval of a set Polynomial NP-complete Polynomial PolynomialHull NP-complete Polynomial NP-complete NP-complete
Convexity NP-complete NP-complete NP-complete NP-completeRadon NP-complete NP-complete NP-hard NP-hard
Caratheodory NP-complete Polynomial NP-complete NP-completeHelly coNP-complete Polynomial Open OpenRank NP-complete NP-complete NP-complete NP-complete
Partition NP-complete Open NP-complete OpenPercolation time NP-complete NP-complete NP-complete NP-complete
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 34 / 36
![Page 90: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/90.jpg)
Results Polynomial time algorithms
Open problems
Fill the gaps on the table.
New parameters:
Generating Degree.Exchange number.
New convexities.
Refine results:
Finding tractable cases.Exact (FPT?) algorithms.Strengthening hardness results.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 35 / 36
![Page 91: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/91.jpg)
Results Polynomial time algorithms
Open problems
Fill the gaps on the table.
New parameters:
Generating Degree.Exchange number.
New convexities.
Refine results:
Finding tractable cases.Exact (FPT?) algorithms.Strengthening hardness results.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 35 / 36
![Page 92: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/92.jpg)
Results Polynomial time algorithms
Open problems
Fill the gaps on the table.
New parameters:
Generating Degree.Exchange number.
New convexities.
Refine results:
Finding tractable cases.Exact (FPT?) algorithms.Strengthening hardness results.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 35 / 36
![Page 93: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/93.jpg)
Results Polynomial time algorithms
Open problems
Fill the gaps on the table.
New parameters:
Generating Degree.Exchange number.
New convexities.
Refine results:
Finding tractable cases.Exact (FPT?) algorithms.Strengthening hardness results.
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 35 / 36
![Page 94: Algorithms and complexity of graph convexity problemsyoshi/FoCM2014/talks/santos.pdf · Introduction Main path convexities Geodetic convexity Family of paths F G consists of all the](https://reader034.fdocuments.us/reader034/viewer/2022050513/5f9dd694f57cc67c73497a4a/html5/thumbnails/94.jpg)
Results Polynomial time algorithms
Thanks
Thank you for your attention!
¡Gracias por su atencion!
Vinıcius F. dos Santos (CEFET-MG) Graph convexity problems December 2014 36 / 36