SlideShare ist ein Scribd-Unternehmen logo
1 von 17
Downloaden Sie, um offline zu lesen
Feedback Vertex Set
AAGA Project
University of Pierre et Marie Curie
NADARAJAH Mag-Stellon
Contents
1 Introduction 1
2 Strategies 2
2.1 Naive Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2 Fixed Parameter Tractable . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2.1 Definitions and notations . . . . . . . . . . . . . . . . . . . . . . . . . 2
2.2.2 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2.3 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3 Disjoint-FVS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3.1 Definitions et notations . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3.2 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4 2-Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.4.1 Definitions and notations . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.4.2 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.4.3 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.5 Randomized Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.5.1 Definitions et notations . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.5.2 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3 Conclusion 13
4 References 14
1. Introduction
This paper deals with one of the first NP complete problems : FeedBack Vertex Set.
NP complete category problems are hard to solve. In general, algorithms providing a solution
to a problem of these are rather slow. That’s why, these solutions are unusable in practice
(although on a reduced instance of a problem).
However, it is possible to verify a solution to a NP complete problems in polynomial time.
At present, we know that it is sufficient to find a polynomial solution to a NP-complete
problem in order to solve them all. This is the subject of the biggest computer science prob-
lem : P versus NP.
In the rest of this paper, we will analyze different algorithms to solve this problem. FVS
is the subject of much research (still ongoing), solutions provide an improvement in time
complexity (they tend to Θ(polynomial)).
What is FVS ? This is a problem related to cyclic graph. The goal is to find a mini-
mum number of nodes "playing a role" in cycles of a graph. By removing these nodes of the
graph is obtained an acyclic graph. More formally, FVS is the set of N nodes of a cyclic
graph G whose deletion makes the graph G acyclic.
This problem is a major problem in combinatorial optimization. It is frequently encoun-
tered in operating systems, databases (management of deadlocks), creation of printed circuits
(VLSI), compilers optimizing ...
2. Strategies
2.1 Naive Algorithm
A naive algorithm is often an intuitive solution. Here, it would be to test all combinations
C of vertices of graph G. If G − C makes G acyclic then C is a FV S of G.
This approach is set aside in view of its time complexity elevated.
2.2 Fixed Parameter Tractable
2.2.1 Definitions and notations
u
a
b
c
d
e
f
Figure 2.1: u is an u-flower of order 2 : u has two disjoints cycles but u in common. This
cycles are petals and u is a "flower with two petals".
2.2 Fixed Parameter Tractable 3
ts
a
b
c
d
e
f
Figure 2.2: u-flower computation : We split u into two nodes t and s. The subgraph s
contains all neighbors nodes v of s such that (vs) is a directed "incoming" arc (there is a
oriented path from v to s). The subgraph t contains all neighbors nodes with "outgoing"
arc. The maximal number of flower is the number of disjoint path (st), called petal(u).
2.2.2 Algorithm
A Fixed Parameter Tractable algorithm (FPT) solves a problem in Θ(f(k)polynomial(n)).
Therefore, It does not only depend on the input size of the problem. Thus, for relatively
small values of k, we hope to obtain a solution in Θ(polynomial(n)).
FPT Algorithm
-----
Input : Graph G = (V,E)
Output : k-FVS of G
Choose a (k+1) nodes subgraph H of G
Choose a (k) nodes subgraph I of H
while H != G
Add a node of G into H and I
Compute FVS -Reduction on I
if FVS -Reduction fails
return G does not have k-FVS
return I
To improve the quality of this algorithm, we can define heuristics in order to get better
subgraph H and I. Let’s give some heuristic ideas :
• Select by greedy method, maximum degree nodes which make up G acyclic.
• We compute an -approximation (see 2.4) of G. Then, we use previous heuristic on
-approximation nodes.
2.2 Fixed Parameter Tractable 4
FV S − Reduction is a set of rules to apply on a graph. This algorithm performs
FV S − Reduction on I (I is a (k + 1)FV S of H) to reduce it to k.
Let I = ∅ and we going to give rules to carry out during FV S − Reduction :
u
Figure 2.3: Add u to I, decrease k and remove u and adjacent edges of u from the graph.
u v
u v
Figure 2.4: Remove all edge of (uv) but one.
u Graphe
Graphe
Figure 2.5: If u does not have any incoming edge or outgoing edge then remove u and its
incident edges from the graph.
u v
merge
Figure 2.6: If there is only single path between u and v then merge u and v.
uGraphe
a
b
c
2.2 Fixed Parameter Tractable 5
Graphe
a
b
c
Figure 2.7: If petal(u) = 1, remove u and its incident edges from Graphe. Create an edge
between u neighbors in Graphe and nodes of the cycle.
u
Graphe
a
b
c
d
e
f
Figure 2.8: If petal(u) > |k|, add u to I, decrease k, remove u and its incident edges from
Graphe.
2.2.3 Example
A
B
C
D
E F
Figure 2.9: We set FTP to k = 2 for this graph G.
2.2 Fixed Parameter Tractable 6
A
B
C
Figure 2.10: We choose a k + 1 = 3 nodes subgraph H of G following the first heuristic (see
2.2.2).
A
C
Figure 2.11: We choose a k = 2 nodes subgraph I of H and we execute rules 4 of FV S −
Reduction, we merge C and A (see 2.6).
B
A
D
E F
Figure 2.12: The graph G is equals to H and FV S − Reduction has succeeded on I. Then,
I is a 2 − FV S of G.
2.3 Disjoint-FVS 7
2.3 Disjoint-FVS
2.3.1 Definitions et notations
Let a graph G(V, E) and FV S F of G. If we can compute a k nodes FV S L of G such as
L V − F then F and L are Disjoint-FV S.
Node set of graph G
FVS F k-FVS L
2.3.2 Algorithm
This algorithm is based on nodes swap. Let’s explain how it works. We find two sets of
nodes which are disjoint-FV S on G and we call them V 1 and V 2.
First of all, we remove all degree-0 and degree-1 nodes. It’s a kind of clean up in order to
compute FV S faster. Then, for all degree-2 v in V 1 nodes, we have to follow two rules.
B
A
C
Figure 2.13: Let B ∈ V 1 and (C, A) ∈ V 22
. B is a degree-2 and its neighbors A, C are into
V 2. So, we include B into V 1 − FV S, we remove B from G and we decrease k.
2.3 Disjoint-FVS 8
B
A
C
Figure 2.14: Let (B, A) ∈ V 12
and C ∈ V 2. B is a degree-2 and only C is into V 2. So, we
move B from V 1 to V 2.
The complete algorithm is written below :
2.4 2-Approximation 9
2.4 2-Approximation
2.4.1 Definitions and notations
Let the function w : V (G) → N denotes the weight of a node in the graph G.
A graph G is clean if it contains no vertex of degree less than two.
A cycle C is semidisjoint if ∀u ∈ C, d(u) = 2 with at most one exception.
2.4.2 Algorithm
An approximation algorithm gives an approximate solution to a complex problem. The qual-
ity of an approximation algorithm is given by the ratio r of the solution approached on the
optimal solution.
We will use a 2-approximation algorithm that applies only to weighted and undirected graphs.
This algorithm will be based on the principle of local searches. This principle focuses on a
subgraph H of G and attempts to resolve the problem on H. Thus, if we can make H acyclic
that means that we can remove cycles on G. This process is repeated until G has no more
cycles. The 2-approximation algorithm is bounded by Θ(min|E|log|V |, |V |2
).
2.4 2-Approximation 10
2.4.3 Example
A
B
C
D
E F
Figure 2.15: A graph G whose nodes are weighted by 1
A
B
C
Figure 2.16: The CleanUp remove all degree-1 nodes. We move A, B, C from G to the Stack
and F. At step 1 et 2 (pop of C and B), F − u is a FV S of G. A the third step (pop of A),
F − u is not a FV S of G then F = A is a FV S of G.
B
C
D
E F
Figure 2.17: The graph G after 2-approximation
2.5 Randomized Algorithm 11
2.5 Randomized Algorithm
2.5.1 Definitions et notations
A leaf is degree-1 node. A link point is a degree-2 node. A branch point is a node of degree
more than three. A graph is rich if it contains only branch point and no node with edges
on itself (self node)
2.5.2 Algorithm
This algorithm is based on the fact that in a rich graph, if we randomly take a node s, there
is a probability of at least 1/2 that s is a neighbor of a FV S node.
Indeed, let a graph G(V, E) and a graph F (FV S of G), then G − F is a forest. Let,
X = V − F. We have, |E(X)| < |X|. Every node of X is a branch point and 3|X| <
w∈X
deg(v) = |E(F, X)| + 2|E(X)|.
We use this property with the algorithm below :
We can notice that this algorithm achieves the FV S with the minimum of probability
(1/4j
). But we can, however, increase artificially the probability of this algorithm.
2.5 Randomized Algorithm 12
We have a FV S with better probability than Single Guess due to the product of proba-
bilities.
Nevertheless, the worries of the Repeat Guess algorithm is that each call to Single Guess is
made independently. So we may have already calculating the Single Guess that one is trying
to achieve. It was a small but real probability that the calculation of FV S ends in a very
long time and that even with a very small FV S.
We will change this algorithm to improve this weakness.
A theorem on branchy graphs states that
w∈X
deg(v) 6x
w∈X
deg(v).
So we will limit our iterations by Max given to arbitrarily set the number of iteration and
c6w(F)
where w(F) is the weight of FV S lowest found by the algorithm.
3. Conclusion
In this paper, I have presented some great algorithms for the FV S problem which are re-
ally tricky. All of those algorithms give a solution with a great time complexity for the
Feedback V ertex Set problem.
As we can see on this paper On Feedback Vertex Set New Measure and New Structures, this
problem is far from being closed and is still relevant for a good time.
Figure 3.1: History of parameterized algorithms for FV S problem.
Acknowledgment. Thank Binh Minh Bui Xuan (see 4) for the help provided through-
out this project.
4. References
• Binh-Minh Bui-Xuan | Algorithms, Programs and Resolution Team, LIP6 | http://www-
apr.lip6.fr/ buixuan/
• Irit Dinur, Samuel Safra, On the hardness of approximating vertex cover, 2005 Annals
of Mathematics, Pages 439-485 from Volume 162
• Fedor V. Fomin, Serge Gaspers, Artem V. Pyatkin, Igor Razgon, On the Minimum
Feedback Vertex Set Problem: Exact and Enumeration Algorithms, 2007 , Algorith-
mica, Volume 52, Issue 2 , pp 293-307
• Igor Razgon , Computing Minimum Directed Feedback Vertex Set in O(1.9977n) , 2007
, WSPC
• Jianer Chen, Yang Liu, Songjian Lu, Barry O’sullivan, Igor Razgon , A fixed-parameter
algorithm for the directed feedback vertex set problem , Volume 55 Issue 5, October
2008 Article No. 21 , Journal of the ACM (JACM)
• Jianer Chena, Fedor V. Fominb, Yang Liua, Songjian Lua, Yngve Villangerb , Improved
algorithms for feedback vertex set problems , 2008 , Journal of Computer and System
Sciences, Volume 74, Issue 7
• Ann Becker, Dan Geiger , Optimization of Pearl’s method of conditioning and greedy-
like approximation algorithms for the vertex feedback set problem , 1996 , Artificial
Intelligence Volume 83, Issue 1
• Bafna, V., Berman, P., and Fujito, T. (1995). Constant ratio approximations of the
weighted feedback vertex set problem for undirected graphs. In Proceedings of the
Sixth Annual Symposium on Algorithms and Computation (ISAAC95)
• Bar-Yehuda, R., Geiger, D., Naor, J., Roth, R. (1994). Approximation algorithms
for the feedback vertex set problems with applications to constraint satisfaction and
Bayesian inference. In Proceedings of the 5th Annual ACM-Siam Symposium On
Discrete Algorithms, pp. 344,354.
• Fedor V. Fomin Serge Gaspers Artem V. Pyatkin Igor Razgon, On the minimum feed-
back vertex set problem : Exact and enumeration algorithms. (2007).
15
• Rudolf Fleischer, Xi Wu, Liwei Yuan, Experimental Study of FPT Algorithms for the
Directed Feedback Vertex Set Problem (2009). 17th Annual European Symposium,
Copenhagen, Denmark, September 7-9, 2009. Proceedings
• A. Becker, R. Bar-Yehuda and D. Geiger (2000) "Randomized Algorithms for the Loop
Cutset Problem", Volume 12, pages 219-234
• Yixin Cao, Jianer Chen, Yang Liu, On Feedback Vertex Set: New Measure and New
Structures, Algorithm Theory - SWAT 2010, 12th Scandinavian Symposium and Work-
shops on Algorithm Theory, Bergen, Norway, June 21-23, 2010. Proceedings
• Vineet Bafna, Piotr Berman, Toshihiro Fujito, A 2-Approximation Algorithm for the
Undirected Feedback Vertex Set Problem, SIAM Journal on Discrete Mathematics,
Volume 12 Issue 3, Sept. 1999, Pages 289-297

Weitere ähnliche Inhalte

Was ist angesagt?

Image compression using discrete wavelet transform
Image compression using discrete wavelet transformImage compression using discrete wavelet transform
Image compression using discrete wavelet transform
Harshal Ladhe
 
Image trnsformations
Image trnsformationsImage trnsformations
Image trnsformations
John Williams
 

Was ist angesagt? (20)

Deep Learning Explained: The future of Artificial Intelligence and Smart Netw...
Deep Learning Explained: The future of Artificial Intelligence and Smart Netw...Deep Learning Explained: The future of Artificial Intelligence and Smart Netw...
Deep Learning Explained: The future of Artificial Intelligence and Smart Netw...
 
Predicting Diabetes Using Machine Learning
Predicting Diabetes Using Machine LearningPredicting Diabetes Using Machine Learning
Predicting Diabetes Using Machine Learning
 
Image compression using discrete wavelet transform
Image compression using discrete wavelet transformImage compression using discrete wavelet transform
Image compression using discrete wavelet transform
 
04 image enhancement edge detection
04 image enhancement edge detection04 image enhancement edge detection
04 image enhancement edge detection
 
Forged authenticity: the case of deepfakes
Forged authenticity: the case of deepfakesForged authenticity: the case of deepfakes
Forged authenticity: the case of deepfakes
 
Fuzzy c-means clustering for image segmentation
Fuzzy c-means  clustering for image segmentationFuzzy c-means  clustering for image segmentation
Fuzzy c-means clustering for image segmentation
 
Graphs
GraphsGraphs
Graphs
 
Lecture 8 (Stereo imaging) (Digital Image Processing)
Lecture 8 (Stereo imaging) (Digital Image Processing)Lecture 8 (Stereo imaging) (Digital Image Processing)
Lecture 8 (Stereo imaging) (Digital Image Processing)
 
Image trnsformations
Image trnsformationsImage trnsformations
Image trnsformations
 
FUZZY LOGIC
FUZZY LOGIC FUZZY LOGIC
FUZZY LOGIC
 
Signals and system
Signals and systemSignals and system
Signals and system
 
IRJET- Diabetes Prediction using Machine Learning
IRJET- Diabetes Prediction using Machine LearningIRJET- Diabetes Prediction using Machine Learning
IRJET- Diabetes Prediction using Machine Learning
 
U-Net: Convolutional Networks for Biomedical Image Segmentation
U-Net: Convolutional Networks for Biomedical Image SegmentationU-Net: Convolutional Networks for Biomedical Image Segmentation
U-Net: Convolutional Networks for Biomedical Image Segmentation
 
Introduction to Deep learning
Introduction to Deep learningIntroduction to Deep learning
Introduction to Deep learning
 
Bias variance dichotomy
Bias variance dichotomyBias variance dichotomy
Bias variance dichotomy
 
Fuzzy logic
Fuzzy logicFuzzy logic
Fuzzy logic
 
Social media e (big) Data
Social media e (big) DataSocial media e (big) Data
Social media e (big) Data
 
Machine learning seminar presentation
Machine learning seminar presentationMachine learning seminar presentation
Machine learning seminar presentation
 
Jpeg dct
Jpeg dctJpeg dct
Jpeg dct
 
Spatial enhancement techniques
Spatial enhancement techniquesSpatial enhancement techniques
Spatial enhancement techniques
 

Ähnlich wie Feedback Vertex Set

Introdution to differential forms
Introdution to differential formsIntrodution to differential forms
Introdution to differential forms
Dunga Pessoa
 
study Streaming Multigrid For Gradient Domain Operations On Large Images
study Streaming Multigrid For Gradient Domain Operations On Large Imagesstudy Streaming Multigrid For Gradient Domain Operations On Large Images
study Streaming Multigrid For Gradient Domain Operations On Large Images
Chiamin Hsu
 
Independence Complexes
Independence ComplexesIndependence Complexes
Independence Complexes
Rickard Fors
 
Comparative Report Ed098
Comparative Report Ed098Comparative Report Ed098
Comparative Report Ed098
mikebrowl
 

Ähnlich wie Feedback Vertex Set (20)

1406
14061406
1406
 
Shor's discrete logarithm quantum algorithm for elliptic curves
 Shor's discrete logarithm quantum algorithm for elliptic curves Shor's discrete logarithm quantum algorithm for elliptic curves
Shor's discrete logarithm quantum algorithm for elliptic curves
 
Introdution to differential forms
Introdution to differential formsIntrodution to differential forms
Introdution to differential forms
 
study Streaming Multigrid For Gradient Domain Operations On Large Images
study Streaming Multigrid For Gradient Domain Operations On Large Imagesstudy Streaming Multigrid For Gradient Domain Operations On Large Images
study Streaming Multigrid For Gradient Domain Operations On Large Images
 
RiemannTEX
RiemannTEXRiemannTEX
RiemannTEX
 
Solving the Poisson Equation
Solving the Poisson EquationSolving the Poisson Equation
Solving the Poisson Equation
 
A Fast Near Optimal Vertex Cover Algorithm (NOVCA)
A Fast Near Optimal Vertex Cover Algorithm (NOVCA)A Fast Near Optimal Vertex Cover Algorithm (NOVCA)
A Fast Near Optimal Vertex Cover Algorithm (NOVCA)
 
Optimisation random graph presentation
Optimisation random graph presentationOptimisation random graph presentation
Optimisation random graph presentation
 
Independence Complexes
Independence ComplexesIndependence Complexes
Independence Complexes
 
Comparative Report Ed098
Comparative Report Ed098Comparative Report Ed098
Comparative Report Ed098
 
METRIC DIMENSION AND UNCERTAINTY OF TRAVERSING ROBOTS IN A NETWORK
METRIC DIMENSION AND UNCERTAINTY OF TRAVERSING ROBOTS IN A NETWORKMETRIC DIMENSION AND UNCERTAINTY OF TRAVERSING ROBOTS IN A NETWORK
METRIC DIMENSION AND UNCERTAINTY OF TRAVERSING ROBOTS IN A NETWORK
 
METRIC DIMENSION AND UNCERTAINTY OF TRAVERSING ROBOTS IN A NETWORK
METRIC DIMENSION AND UNCERTAINTY OF TRAVERSING ROBOTS IN A NETWORKMETRIC DIMENSION AND UNCERTAINTY OF TRAVERSING ROBOTS IN A NETWORK
METRIC DIMENSION AND UNCERTAINTY OF TRAVERSING ROBOTS IN A NETWORK
 
Lego like spheres and tori, enumeration and drawings
Lego like spheres and tori, enumeration and drawingsLego like spheres and tori, enumeration and drawings
Lego like spheres and tori, enumeration and drawings
 
Algorithms explained
Algorithms explainedAlgorithms explained
Algorithms explained
 
Spectral methods for solving differential equations
Spectral methods for solving differential equationsSpectral methods for solving differential equations
Spectral methods for solving differential equations
 
Equivalent condition and two algorithms for hamiltonian graphs
Equivalent condition and two algorithms for hamiltonian graphsEquivalent condition and two algorithms for hamiltonian graphs
Equivalent condition and two algorithms for hamiltonian graphs
 
Differential Equations Assignment Help
Differential Equations Assignment HelpDifferential Equations Assignment Help
Differential Equations Assignment Help
 
Differential Equations Assignment Help
Differential Equations Assignment HelpDifferential Equations Assignment Help
Differential Equations Assignment Help
 
Algorithmic Mathematics.
Algorithmic Mathematics.Algorithmic Mathematics.
Algorithmic Mathematics.
 
Some Engg. Applications of Matrices and Partial Derivatives
Some Engg. Applications of Matrices and Partial DerivativesSome Engg. Applications of Matrices and Partial Derivatives
Some Engg. Applications of Matrices and Partial Derivatives
 

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...
Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...
Try MyIntelliAccount Cloud Accounting Software As A Service Solution Risk Fre...
 
Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...
Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...
Reassessing the Bedrock of Clinical Function Models: An Examination of Large ...
 
W01_panagenda_Navigating-the-Future-with-The-Hitchhikers-Guide-to-Notes-and-D...
W01_panagenda_Navigating-the-Future-with-The-Hitchhikers-Guide-to-Notes-and-D...W01_panagenda_Navigating-the-Future-with-The-Hitchhikers-Guide-to-Notes-and-D...
W01_panagenda_Navigating-the-Future-with-The-Hitchhikers-Guide-to-Notes-and-D...
 
Vip Call Girls Noida ➡️ Delhi ➡️ 9999965857 No Advance 24HRS Live
Vip Call Girls Noida ➡️ Delhi ➡️ 9999965857 No Advance 24HRS LiveVip Call Girls Noida ➡️ Delhi ➡️ 9999965857 No Advance 24HRS Live
Vip Call Girls Noida ➡️ Delhi ➡️ 9999965857 No Advance 24HRS Live
 
How To Troubleshoot Collaboration Apps for the Modern Connected Worker
How To Troubleshoot Collaboration Apps for the Modern Connected WorkerHow To Troubleshoot Collaboration Apps for the Modern Connected Worker
How To Troubleshoot Collaboration Apps for the Modern Connected Worker
 
Steps To Getting Up And Running Quickly With MyTimeClock Employee Scheduling ...
Steps To Getting Up And Running Quickly With MyTimeClock Employee Scheduling ...Steps To Getting Up And Running Quickly With MyTimeClock Employee Scheduling ...
Steps To Getting Up And Running Quickly With MyTimeClock Employee Scheduling ...
 
Diamond Application Development Crafting Solutions with Precision
Diamond Application Development Crafting Solutions with PrecisionDiamond Application Development Crafting Solutions with Precision
Diamond Application Development Crafting Solutions with Precision
 
SyndBuddy AI 2k Review 2024: Revolutionizing Content Syndication with AI
SyndBuddy AI 2k Review 2024: Revolutionizing Content Syndication with AISyndBuddy AI 2k Review 2024: Revolutionizing Content Syndication with AI
SyndBuddy AI 2k Review 2024: Revolutionizing Content Syndication with AI
 
Unveiling the Tech Salsa of LAMs with Janus in Real-Time Applications
Unveiling the Tech Salsa of LAMs with Janus in Real-Time ApplicationsUnveiling the Tech Salsa of LAMs with Janus in Real-Time Applications
Unveiling the Tech Salsa of LAMs with Janus in Real-Time Applications
 
Microsoft AI Transformation Partner Playbook.pdf
Microsoft AI Transformation Partner Playbook.pdfMicrosoft AI Transformation Partner Playbook.pdf
Microsoft AI Transformation Partner Playbook.pdf
 
Hand gesture recognition PROJECT PPT.pptx
Hand gesture recognition PROJECT PPT.pptxHand gesture recognition PROJECT PPT.pptx
Hand gesture recognition PROJECT PPT.pptx
 
Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...
Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...
Short Story: Unveiling the Reasoning Abilities of Large Language Models by Ke...
 
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
 
The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...
The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...
The Real-World Challenges of Medical Device Cybersecurity- Mitigating Vulnera...
 
Right Money Management App For Your Financial Goals
Right Money Management App For Your Financial GoalsRight Money Management App For Your Financial Goals
Right Money Management App For Your Financial Goals
 
Software Quality Assurance Interview Questions
Software Quality Assurance Interview QuestionsSoftware Quality Assurance Interview Questions
Software Quality Assurance Interview Questions
 
Optimizing AI for immediate response in Smart CCTV
Optimizing AI for immediate response in Smart CCTVOptimizing AI for immediate response in Smart CCTV
Optimizing AI for immediate response in Smart CCTV
 
Tech Tuesday-Harness the Power of Effective Resource Planning with OnePlan’s ...
Tech Tuesday-Harness the Power of Effective Resource Planning with OnePlan’s ...Tech Tuesday-Harness the Power of Effective Resource Planning with OnePlan’s ...
Tech Tuesday-Harness the Power of Effective Resource Planning with OnePlan’s ...
 
Shapes for Sharing between Graph Data Spaces - and Epistemic Querying of RDF-...
Shapes for Sharing between Graph Data Spaces - and Epistemic Querying of RDF-...Shapes for Sharing between Graph Data Spaces - and Epistemic Querying of RDF-...
Shapes for Sharing between Graph Data Spaces - and Epistemic Querying of RDF-...
 
Learn the Fundamentals of XCUITest Framework_ A Beginner's Guide.pdf
Learn the Fundamentals of XCUITest Framework_ A Beginner's Guide.pdfLearn the Fundamentals of XCUITest Framework_ A Beginner's Guide.pdf
Learn the Fundamentals of XCUITest Framework_ A Beginner's Guide.pdf
 

Feedback Vertex Set

  • 1. Feedback Vertex Set AAGA Project University of Pierre et Marie Curie NADARAJAH Mag-Stellon
  • 2. Contents 1 Introduction 1 2 Strategies 2 2.1 Naive Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.2 Fixed Parameter Tractable . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.2.1 Definitions and notations . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.2.2 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.2.3 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.3 Disjoint-FVS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.3.1 Definitions et notations . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.3.2 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.4 2-Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.4.1 Definitions and notations . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.4.2 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.4.3 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.5 Randomized Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.5.1 Definitions et notations . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.5.2 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3 Conclusion 13 4 References 14
  • 3. 1. Introduction This paper deals with one of the first NP complete problems : FeedBack Vertex Set. NP complete category problems are hard to solve. In general, algorithms providing a solution to a problem of these are rather slow. That’s why, these solutions are unusable in practice (although on a reduced instance of a problem). However, it is possible to verify a solution to a NP complete problems in polynomial time. At present, we know that it is sufficient to find a polynomial solution to a NP-complete problem in order to solve them all. This is the subject of the biggest computer science prob- lem : P versus NP. In the rest of this paper, we will analyze different algorithms to solve this problem. FVS is the subject of much research (still ongoing), solutions provide an improvement in time complexity (they tend to Θ(polynomial)). What is FVS ? This is a problem related to cyclic graph. The goal is to find a mini- mum number of nodes "playing a role" in cycles of a graph. By removing these nodes of the graph is obtained an acyclic graph. More formally, FVS is the set of N nodes of a cyclic graph G whose deletion makes the graph G acyclic. This problem is a major problem in combinatorial optimization. It is frequently encoun- tered in operating systems, databases (management of deadlocks), creation of printed circuits (VLSI), compilers optimizing ...
  • 4. 2. Strategies 2.1 Naive Algorithm A naive algorithm is often an intuitive solution. Here, it would be to test all combinations C of vertices of graph G. If G − C makes G acyclic then C is a FV S of G. This approach is set aside in view of its time complexity elevated. 2.2 Fixed Parameter Tractable 2.2.1 Definitions and notations u a b c d e f Figure 2.1: u is an u-flower of order 2 : u has two disjoints cycles but u in common. This cycles are petals and u is a "flower with two petals".
  • 5. 2.2 Fixed Parameter Tractable 3 ts a b c d e f Figure 2.2: u-flower computation : We split u into two nodes t and s. The subgraph s contains all neighbors nodes v of s such that (vs) is a directed "incoming" arc (there is a oriented path from v to s). The subgraph t contains all neighbors nodes with "outgoing" arc. The maximal number of flower is the number of disjoint path (st), called petal(u). 2.2.2 Algorithm A Fixed Parameter Tractable algorithm (FPT) solves a problem in Θ(f(k)polynomial(n)). Therefore, It does not only depend on the input size of the problem. Thus, for relatively small values of k, we hope to obtain a solution in Θ(polynomial(n)). FPT Algorithm ----- Input : Graph G = (V,E) Output : k-FVS of G Choose a (k+1) nodes subgraph H of G Choose a (k) nodes subgraph I of H while H != G Add a node of G into H and I Compute FVS -Reduction on I if FVS -Reduction fails return G does not have k-FVS return I To improve the quality of this algorithm, we can define heuristics in order to get better subgraph H and I. Let’s give some heuristic ideas : • Select by greedy method, maximum degree nodes which make up G acyclic. • We compute an -approximation (see 2.4) of G. Then, we use previous heuristic on -approximation nodes.
  • 6. 2.2 Fixed Parameter Tractable 4 FV S − Reduction is a set of rules to apply on a graph. This algorithm performs FV S − Reduction on I (I is a (k + 1)FV S of H) to reduce it to k. Let I = ∅ and we going to give rules to carry out during FV S − Reduction : u Figure 2.3: Add u to I, decrease k and remove u and adjacent edges of u from the graph. u v u v Figure 2.4: Remove all edge of (uv) but one. u Graphe Graphe Figure 2.5: If u does not have any incoming edge or outgoing edge then remove u and its incident edges from the graph. u v merge Figure 2.6: If there is only single path between u and v then merge u and v. uGraphe a b c
  • 7. 2.2 Fixed Parameter Tractable 5 Graphe a b c Figure 2.7: If petal(u) = 1, remove u and its incident edges from Graphe. Create an edge between u neighbors in Graphe and nodes of the cycle. u Graphe a b c d e f Figure 2.8: If petal(u) > |k|, add u to I, decrease k, remove u and its incident edges from Graphe. 2.2.3 Example A B C D E F Figure 2.9: We set FTP to k = 2 for this graph G.
  • 8. 2.2 Fixed Parameter Tractable 6 A B C Figure 2.10: We choose a k + 1 = 3 nodes subgraph H of G following the first heuristic (see 2.2.2). A C Figure 2.11: We choose a k = 2 nodes subgraph I of H and we execute rules 4 of FV S − Reduction, we merge C and A (see 2.6). B A D E F Figure 2.12: The graph G is equals to H and FV S − Reduction has succeeded on I. Then, I is a 2 − FV S of G.
  • 9. 2.3 Disjoint-FVS 7 2.3 Disjoint-FVS 2.3.1 Definitions et notations Let a graph G(V, E) and FV S F of G. If we can compute a k nodes FV S L of G such as L V − F then F and L are Disjoint-FV S. Node set of graph G FVS F k-FVS L 2.3.2 Algorithm This algorithm is based on nodes swap. Let’s explain how it works. We find two sets of nodes which are disjoint-FV S on G and we call them V 1 and V 2. First of all, we remove all degree-0 and degree-1 nodes. It’s a kind of clean up in order to compute FV S faster. Then, for all degree-2 v in V 1 nodes, we have to follow two rules. B A C Figure 2.13: Let B ∈ V 1 and (C, A) ∈ V 22 . B is a degree-2 and its neighbors A, C are into V 2. So, we include B into V 1 − FV S, we remove B from G and we decrease k.
  • 10. 2.3 Disjoint-FVS 8 B A C Figure 2.14: Let (B, A) ∈ V 12 and C ∈ V 2. B is a degree-2 and only C is into V 2. So, we move B from V 1 to V 2. The complete algorithm is written below :
  • 11. 2.4 2-Approximation 9 2.4 2-Approximation 2.4.1 Definitions and notations Let the function w : V (G) → N denotes the weight of a node in the graph G. A graph G is clean if it contains no vertex of degree less than two. A cycle C is semidisjoint if ∀u ∈ C, d(u) = 2 with at most one exception. 2.4.2 Algorithm An approximation algorithm gives an approximate solution to a complex problem. The qual- ity of an approximation algorithm is given by the ratio r of the solution approached on the optimal solution. We will use a 2-approximation algorithm that applies only to weighted and undirected graphs. This algorithm will be based on the principle of local searches. This principle focuses on a subgraph H of G and attempts to resolve the problem on H. Thus, if we can make H acyclic that means that we can remove cycles on G. This process is repeated until G has no more cycles. The 2-approximation algorithm is bounded by Θ(min|E|log|V |, |V |2 ).
  • 12. 2.4 2-Approximation 10 2.4.3 Example A B C D E F Figure 2.15: A graph G whose nodes are weighted by 1 A B C Figure 2.16: The CleanUp remove all degree-1 nodes. We move A, B, C from G to the Stack and F. At step 1 et 2 (pop of C and B), F − u is a FV S of G. A the third step (pop of A), F − u is not a FV S of G then F = A is a FV S of G. B C D E F Figure 2.17: The graph G after 2-approximation
  • 13. 2.5 Randomized Algorithm 11 2.5 Randomized Algorithm 2.5.1 Definitions et notations A leaf is degree-1 node. A link point is a degree-2 node. A branch point is a node of degree more than three. A graph is rich if it contains only branch point and no node with edges on itself (self node) 2.5.2 Algorithm This algorithm is based on the fact that in a rich graph, if we randomly take a node s, there is a probability of at least 1/2 that s is a neighbor of a FV S node. Indeed, let a graph G(V, E) and a graph F (FV S of G), then G − F is a forest. Let, X = V − F. We have, |E(X)| < |X|. Every node of X is a branch point and 3|X| < w∈X deg(v) = |E(F, X)| + 2|E(X)|. We use this property with the algorithm below : We can notice that this algorithm achieves the FV S with the minimum of probability (1/4j ). But we can, however, increase artificially the probability of this algorithm.
  • 14. 2.5 Randomized Algorithm 12 We have a FV S with better probability than Single Guess due to the product of proba- bilities. Nevertheless, the worries of the Repeat Guess algorithm is that each call to Single Guess is made independently. So we may have already calculating the Single Guess that one is trying to achieve. It was a small but real probability that the calculation of FV S ends in a very long time and that even with a very small FV S. We will change this algorithm to improve this weakness. A theorem on branchy graphs states that w∈X deg(v) 6x w∈X deg(v). So we will limit our iterations by Max given to arbitrarily set the number of iteration and c6w(F) where w(F) is the weight of FV S lowest found by the algorithm.
  • 15. 3. Conclusion In this paper, I have presented some great algorithms for the FV S problem which are re- ally tricky. All of those algorithms give a solution with a great time complexity for the Feedback V ertex Set problem. As we can see on this paper On Feedback Vertex Set New Measure and New Structures, this problem is far from being closed and is still relevant for a good time. Figure 3.1: History of parameterized algorithms for FV S problem. Acknowledgment. Thank Binh Minh Bui Xuan (see 4) for the help provided through- out this project.
  • 16. 4. References • Binh-Minh Bui-Xuan | Algorithms, Programs and Resolution Team, LIP6 | http://www- apr.lip6.fr/ buixuan/ • Irit Dinur, Samuel Safra, On the hardness of approximating vertex cover, 2005 Annals of Mathematics, Pages 439-485 from Volume 162 • Fedor V. Fomin, Serge Gaspers, Artem V. Pyatkin, Igor Razgon, On the Minimum Feedback Vertex Set Problem: Exact and Enumeration Algorithms, 2007 , Algorith- mica, Volume 52, Issue 2 , pp 293-307 • Igor Razgon , Computing Minimum Directed Feedback Vertex Set in O(1.9977n) , 2007 , WSPC • Jianer Chen, Yang Liu, Songjian Lu, Barry O’sullivan, Igor Razgon , A fixed-parameter algorithm for the directed feedback vertex set problem , Volume 55 Issue 5, October 2008 Article No. 21 , Journal of the ACM (JACM) • Jianer Chena, Fedor V. Fominb, Yang Liua, Songjian Lua, Yngve Villangerb , Improved algorithms for feedback vertex set problems , 2008 , Journal of Computer and System Sciences, Volume 74, Issue 7 • Ann Becker, Dan Geiger , Optimization of Pearl’s method of conditioning and greedy- like approximation algorithms for the vertex feedback set problem , 1996 , Artificial Intelligence Volume 83, Issue 1 • Bafna, V., Berman, P., and Fujito, T. (1995). Constant ratio approximations of the weighted feedback vertex set problem for undirected graphs. In Proceedings of the Sixth Annual Symposium on Algorithms and Computation (ISAAC95) • Bar-Yehuda, R., Geiger, D., Naor, J., Roth, R. (1994). Approximation algorithms for the feedback vertex set problems with applications to constraint satisfaction and Bayesian inference. In Proceedings of the 5th Annual ACM-Siam Symposium On Discrete Algorithms, pp. 344,354. • Fedor V. Fomin Serge Gaspers Artem V. Pyatkin Igor Razgon, On the minimum feed- back vertex set problem : Exact and enumeration algorithms. (2007).
  • 17. 15 • Rudolf Fleischer, Xi Wu, Liwei Yuan, Experimental Study of FPT Algorithms for the Directed Feedback Vertex Set Problem (2009). 17th Annual European Symposium, Copenhagen, Denmark, September 7-9, 2009. Proceedings • A. Becker, R. Bar-Yehuda and D. Geiger (2000) "Randomized Algorithms for the Loop Cutset Problem", Volume 12, pages 219-234 • Yixin Cao, Jianer Chen, Yang Liu, On Feedback Vertex Set: New Measure and New Structures, Algorithm Theory - SWAT 2010, 12th Scandinavian Symposium and Work- shops on Algorithm Theory, Bergen, Norway, June 21-23, 2010. Proceedings • Vineet Bafna, Piotr Berman, Toshihiro Fujito, A 2-Approximation Algorithm for the Undirected Feedback Vertex Set Problem, SIAM Journal on Discrete Mathematics, Volume 12 Issue 3, Sept. 1999, Pages 289-297