Non-Uniform Graph Partitioning
|
|
- Shonda Joleen Thornton
- 5 years ago
- Views:
Transcription
1 Robert Seffi Roy Kunal Krauthgamer Naor Schwartz Talwar Weizmann Institute Technion Microsoft Research Microsoft Research
2 Problem Definition Introduction Definitions Related Work Our Result Input: G = (V, E), w : E R +. Capacities n 1, n 2,..., n k s.t. k j=1 n j n.
3 Problem Definition Introduction Definitions Related Work Our Result Input: Output: G = (V, E), w : E R +. Capacities n 1, n 2,..., n k s.t. k j=1 n j n. A partition S 1,..., S k of V where each S j n j minimizing: 1 2 k δ(s j ). j=1
4 Problem Definition Introduction Definitions Related Work Our Result Input: Output: Note: G = (V, E), w : E R +. Capacities n 1, n 2,..., n k s.t. k j=1 n j n. A partition S 1,..., S k of V where each S j n j minimizing: 1 2 k δ(s j ). j=1 Number of parts k might depend on n. Capacities might be of different magnitudes.
5 Example - Cloud Computing Definitions Related Work Our Result processes
6 Example - Cloud Computing Definitions Related Work Our Result processes
7 Example - Cloud Computing Definitions Related Work Our Result processes servers n 1 n 2 n 3 n 4 n 5 n 6 n 7 n 8 n 9
8 Example - Cloud Computing Definitions Related Work Our Result processes servers n 1 n 2 n 3 n 4 n 5 n 6 n 7 n 8 n 9
9 Motivation Introduction Definitions Related Work Our Result Theoretical: Captures well studied problems: Min-Bisection, Min b-balanced-cut, Min k-partitioning.
10 Motivation Introduction Definitions Related Work Our Result Theoretical: Captures well studied problems: Min-Bisection, Min b-balanced-cut, Min k-partitioning. Practical: Cloud and Parallel Computing: parallelism. Hardware design: VLSI layout, circuit testing. Data mining: clustering. Social network analysis: community discovery. Vision: pattern recognition. Scientific Computing: linear systems.
11 Related Work Introduction Definitions Related Work Our Result Heuristics: [Barnes-82], [Barnes-Vanneli-Walker-88], [Sanchis-89], [Hadley-Mark-Vanneli-92], [Rendl-Wolkowicz-95]... Mainly use spectral theory, local search and quadratic programming.
12 Related Work Introduction Definitions Related Work Our Result Heuristics: [Barnes-82], [Barnes-Vanneli-Walker-88], [Sanchis-89], [Hadley-Mark-Vanneli-92], [Rendl-Wolkowicz-95]... Mainly use spectral theory, local search and quadratic programming. Worst Case Guarantee No meaningful known bounds.
13 Related Work (Cont.) Introduction Definitions Related Work Our Result Balanced Graph Partitioning: (n j n /k)
14 Related Work (Cont.) Introduction Definitions Related Work Our Result Balanced Graph Partitioning: (n j n /k) Min-Bisection: (k = 2) { O(log 3/2 n) [Feige-Krauthgamer-02] True Approx. O (log n) [Räcke-08]
15 Related Work (Cont.) Introduction Definitions Related Work Our Result Balanced Graph Partitioning: (n j n /k) Min-Bisection: (k = 2) { O(log 3/2 n) [Feige-Krauthgamer-02] True Approx. O (log n) [Räcke-08] Bicriteria Approx. Related to Sparsest-Cut and Min b-balanced-cut. { O (log n) [Leighton-Rao-99] O ( log n ) [Arora-Rao-Vazirani-08]
16 Related Work (Cont.) Introduction Definitions Related Work Our Result Balanced Graph Partitioning: (n j n /k) Min-Bisection: (k = 2) { O(log 3/2 n) [Feige-Krauthgamer-02] True Approx. O (log n) [Räcke-08] Bicriteria Approx. Related to Sparsest-Cut and Min b-balanced-cut. { O (log n) [Leighton-Rao-99] O ( log n ) [Arora-Rao-Vazirani-08] Min k-partitioning: (general k) NP-Hardness no true approximation [Andreev-Räcke-06]
17 Related Work (Cont.) Introduction Definitions Related Work Our Result Balanced Graph Partitioning: (n j n /k) Min-Bisection: (k = 2) { O(log 3/2 n) [Feige-Krauthgamer-02] True Approx. O (log n) [Räcke-08] Bicriteria Approx. Related to Sparsest-Cut and Min b-balanced-cut. { O (log n) [Leighton-Rao-99] O ( log n ) [Arora-Rao-Vazirani-08] Min k-partitioning: (general k) NP-Hardness no true approximation [Andreev-Räcke-06] Bicriteria Approx. { (O (log n), 2) [Even-Naor-Rao-Schieber-99] ( O ( log n log k ), 2 ) [Krauthgamer-Naor-S-09] { (O(ε 2 log 3/2 n), 1 + ε) [Andreev-Räcke-06] (O(log n), 1 + ε) [Feldmann-Forschini-12]
18 Related Work (Cont.) Introduction Definitions Related Work Our Result Capacitated Metric Labeling: The same as with additional assignment costs. (O (log n), O((log k))) n j n /k [Naor-S-05] (O (log n), 1) constant k [Andrews-Hajiaghayi-Karloff-Moitra-11] NP ZP T IME ( n polylog(n)) No finite approximation that violates capacities by O(log 1/2 ε k), ε > 0. [Andrews-Hajiaghayi-Karloff-Moitra-11]
19 Our Result Introduction Definitions Related Work Our Result Theorem [Krauthgamer-Naor-S-Talwar-13] There is a bicriteria approximation algorithm achieving a guarantee of: (O (log n), O(1)) for.
20 Known - Issues 1 Recursive Partitioning: How many vertices to cut in each step?
21 Known - Issues 1 Recursive Partitioning: How many vertices to cut in each step? 2 Spreading Metrics: Only spreading with average capacity - insufficient.
22 Known - Issues 1 Recursive Partitioning: How many vertices to cut in each step? 2 Spreading Metrics: Only spreading with average capacity - insufficient. 3 Räcke s Tree Decomposition: Dynamic programming does not seem to yield poly running time.
23 Known - Issues 1 Recursive Partitioning: How many vertices to cut in each step? 2 Spreading Metrics: Only spreading with average capacity - insufficient. 3 Räcke s Tree Decomposition: Dynamic programming does not seem to yield poly running time. Our Approach Configuration LP. Randomized rounding + concentration via stopping times.
24 Configuration LP Introduction F j {S : S V, S n j }
25 Configuration LP Introduction F j {S : S V, S n j } (P) min s.t. 1 2 k j=1 k S F j δ(s) x S,j j=1 S F j:u S S F j x S,j 1 x S,j 0 x S,j 1 u V j = 1,..., k j = 1,..., k, S F j
26 Configuration LP (Cont.) Theorem (P) can be efficiently solved up to a loss of O(log n) in the objective.
27 Configuration LP (Cont.) Theorem (P) can be efficiently solved up to a loss of O(log n) in the objective. Proof Outline: Dual separation oracle of (P) relates to Min ρ-unbalanced-cut. from [Räcke-08] give an O(log n) approximation.
28 Configuration LP (Cont.) Theorem (P) can be efficiently solved up to a loss of O(log n) in the objective. Proof Outline: Dual separation oracle of (P) relates to Min ρ-unbalanced-cut. from [Räcke-08] give an O(log n) approximation. Difficulty: Applying the above in algorithms for solving mixed packing/covering LPs yields O(log n) loss in cost and capacity.
29 Configuration LP (Cont.) Theorem (P) can be efficiently solved up to a loss of O(log n) in the objective. Proof Outline: Dual separation oracle of (P) relates to Min ρ-unbalanced-cut. from [Räcke-08] give an O(log n) approximation. Difficulty: Applying the above in algorithms for solving mixed packing/covering LPs yields O(log n) loss in cost and capacity. Solution: Scaling constraints differently.
30 Randomized Introduction Assumption: n 1 n 2... n k and each n j is a power of 2.
31 Randomized Introduction Assumption: n 1 n 2... n k and each n j is a power of 2. Notation: n 1 = n 2 = n 3 > n 4 = n 5 > n 6 = n 7 = n 8 > W 1 W 2 W 3
32 Randomized Introduction Assumption: n 1 n 2... n k and each n j is a power of 2. Notation: n 1 = n 2 = n 3 > n 4 = n 5 > n 6 = n 7 = n 8 > W 1 W 2 W 3 } W i {j : n j = 2 (i 1) n 1 i = 1, 2,..., l k i W i
33 Randomized (Cont.) Idea: Covering vertices by random cuts.
34 Randomized (Cont.) Idea: Covering vertices by random cuts. - General Approach 1 H i for every i = 1,..., l. 2 While V : Choose j Unif[1,..., k]. Choose S F j w.p. x S,j. Let r be the mega-bucket s.t. j W r. H r H r {S V }. V V \ S.
35 Randomized - Example H 1 = φ H 2 = φ H 3 = φ H l = φ
36 Randomized - Example H 1 = φ H 2 = φ H 3 = φ S 1 H l = φ S 1, j 1 j 1 W 3
37 Randomized - Example H 1 = φ H 2 = φ H 3 = S 1 S 1 H l = φ S 1, j 1 j 1 W 3
38 Randomized - Example H 1 = φ H 2 = φ H 3 = S 1 S 2 S 1 H l = φ S 2, j 2 j 2 W 1
39 Randomized - Example H 1 = S 2 \S 1 H 2 = φ H 3 = S 1 S 2 S 1 H l = φ S 2, j 2 j 2 W 1
40 Randomized - Example H 1 = S 2 \S 1 H 2 = φ H 3 = S 1 S 2 S 1 H l = φ S 3 S 3, j 3 j 3 W 3
41 Randomized - Example H 1 = S 2 \S 1 H 2 = φ H 3 = S 1, S 3 \ S 1 S 2 S 2 S 1 H l = φ S 3 S 3, j 3 j 3 W 3
42 Randomized - Example H 1 = S 2 \S 1 H 2 = φ H 3 = S 1, S 3 \ S 1 S 2 S 2 S 1 S 4 H l = φ S 3 S 4, j 4 j 4 W 2
43 Randomized - Example H 1 = S 2 \S 1 H 2 = S 4 \ S 2 S 3 H 3 = S 1, S 3 \ S 1 S 2 S 2 S 1 S 4 H l = φ S 3 S 4, j 4 j 4 W 2
44 Randomized - Example H 1 = S 2 \S 1 H 2 = S 4 \ S 2 S 3 H 3 = S 1, S 3 \ S 1 S 2 S 5 S 2 S 1 S 4 H l = φ S 3 S 5, j 5 j 5 W l
45 Randomized - Example H 1 = S 2 \S 1 H 2 = S 4 \ S 2 S 3 H 3 = S 1, S 3 \ S 1 S 2 S 5 S 2 S 1 S 4 H l = S 5 \ S 1 S 2 S 3 S 5, j 5 j 5 W l
46 Randomized (Cont.) Question: What to do when all vertices are covered?
47 Randomized (Cont.) Question: What to do when all vertices are covered? Merge: While H i > k i merge smallest two cuts in H i.
48 Randomized (Cont.) Question: What to do when all vertices are covered? Merge: While H i > k i merge smallest two cuts in H i. Theorem - Cost E [cost(alg)] 2 cost (P). Proof Outline: k Pr [u and v covered in different iterations] x S,j. j=1 S F j:(u,v) δ(s)
49 Capacity - Attempt I Observation - Interchanging Cuts Within W i It suffices to upper bound: N i number of vertices covered by H i at the end.
50 Capacity - Attempt I Observation - Interchanging Cuts Within W i It suffices to upper bound: N i number of vertices covered by H i at the end. Note: E [N i ] k i 2 (i 1) n 1.
51 Capacity - Attempt I Observation - Interchanging Cuts Within W i It suffices to upper bound: N i number of vertices covered by H i at the end. Note: E [N i ] k i 2 (i 1) n 1. l = O(log k) by merging every W i with k i 2 1 /2(i 1) into W 1. (l is number of mega-buckets)
52 Capacity - Attempt I Observation - Interchanging Cuts Within W i It suffices to upper bound: N i number of vertices covered by H i at the end. Note: E [N i ] k i 2 (i 1) n 1. l = O(log k) by merging every W i with k i 2 1 /2(i 1) into W 1. (l is number of mega-buckets) Conclusion: Markov + union bound O(log k) capacity violation.
53 Capacity - Attempt II Martingale M i,t E [N i (S 1, j 1 ),..., (S t, j t )] {M i,t } t=0 is a martingale with respect to {(S t, j t )} t=1.
54 Capacity - Attempt II Martingale M i,t E [N i (S 1, j 1 ),..., (S t, j t )] {M i,t } t=0 is a martingale with respect to {(S t, j t )} t=1. Note: M i,0 = E [N i ].
55 Capacity - Attempt II Martingale M i,t E [N i (S 1, j 1 ),..., (S t, j t )] {M i,t } t=0 is a martingale with respect to {(S t, j t )} t=1. Note: M i,0 = E [N i ]. M i,t M i,t 1 k i 2 (i 1) n 1. (Lipschitz)
56 Capacity - Attempt II Martingale M i,t E [N i (S 1, j 1 ),..., (S t, j t )] {M i,t } t=0 is a martingale with respect to {(S t, j t )} t=1. Note: M i,0 = E [N i ]. M i,t M i,t 1 k i 2 (i 1) n 1. (Lipschitz) Number of iterations to cover V : T = Θ(k log n).
57 Capacity - Attempt II Martingale M i,t E [N i (S 1, j 1 ),..., (S t, j t )] {M i,t } t=0 is a martingale with respect to {(S t, j t )} t=1. Note: M i,0 = E [N i ]. M i,t M i,t 1 k i 2 (i 1) n 1. (Lipschitz) Number of iterations to cover V : T = Θ(k log n). Conclusion: Azuma + union bound O( k log n log log k) capacity violation.
58 Capacity - Attempt III Worst Conditional Variance: Let (T 1, r 1 ),..., (T t 1, r t 1 ) be the realization that maximizes: Var [ M i,t M i,t 1 (S 1, j 1 ) = (T 1, r 1 ),..., (S t 1, j t 1 ) = (T t 1, r t 1 ) ]. v i,t is the worst variance value.
59 Capacity - Attempt III Worst Conditional Variance: Let (T 1, r 1 ),..., (T t 1, r t 1 ) be the realization that maximizes: Var [ M i,t M i,t 1 (S 1, j 1 ) = (T 1, r 1 ),..., (S t 1, j t 1 ) = (T t 1, r t 1 ) ]. v i,t is the worst variance value. Note: For every t a different conditioning might be chosen. Martingale concentration via bounded variances (Bernstein) is not sufficient: t 1 v i,t might be too big!
60 Capacity - Attempt III (Cont.) Pr Var [ M i,t M i,t 1 (S 1, j 1 ),..., (S t 1, j t 1 ) ] is small? t 1
61 Capacity - Attempt III (Cont.) Pr Var [ M i,t M i,t 1 (S 1, j 1 ),..., (S t 1, j t 1 ) ] is small? t 1 Theorem - Conditional Variances Sum [ Pr Var [ M i,t M i,t 1 (S 1, j 1 ),..., (S t 1, j t 1 ) ] t 1 ] 2α k i 2 (i 1) n /α.
62 Capacity - Attempt III (Cont.) Pr Var [ M i,t M i,t 1 (S 1, j 1 ),..., (S t 1, j t 1 ) ] is small? t 1 Theorem - Conditional Variances Sum [ Pr Var [ M i,t M i,t 1 (S 1, j 1 ),..., (S t 1, j t 1 ) ] t 1 ] 2α k i 2 (i 1) n /α. Intuition: In later iterations the changes are smaller in expectation.
63 Capacity - Attempt III (Cont.) Martingale {M i,t } t=0 Good Event Variances are small.
64 Capacity - Attempt III (Cont.) Martingale {M i,t } t=0 Good Event Variances are small.???
65 Capacity - Attempt III (Cont.) Martingale {M i,t } t=0 Good Event Variances are small.??? Freedman s Inequality (stopping-time based concentration)
66 Capacity - Attempt III (Cont.) Immediate Conclusion: Freedman s inequality yields O(log log k) capacity violation.
67 Capacity - Attempt III (Cont.) Immediate Conclusion: Freedman s inequality yields O(log log k) capacity violation. Question How do we get O(1) capacity violation?
68 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
69 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
70 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
71 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
72 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
73 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
74 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
75 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
76 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
77 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
78 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
79 Instance Transformation W 1 W 2 W 3 W 4 W 5 W 6 W 7 W 8 W 9
80 Instance Transformation (Cont.) What is an instance transformation?
81 Instance Transformation (Cont.) What is an instance transformation? Total capacity of non-empty W i grows by a constant c. Inverse transformation moves cuts from W i to some W j, j i. Inverse transformation incurs a c capacity violation.
82 Instance Transformation (Cont.) What is an instance transformation? Total capacity of non-empty W i grows by a constant c. Inverse transformation moves cuts from W i to some W j, j i. Inverse transformation incurs a c capacity violation. What does an instance transformation provide?
83 Instance Transformation (Cont.) What is an instance transformation? Total capacity of non-empty W i grows by a constant c. Inverse transformation moves cuts from W i to some W j, j i. Inverse transformation incurs a c capacity violation. What does an instance transformation provide? Large total capacity of W i Better concentration as the total capacity increases
84 Instance Transformation (Cont.) What is an instance transformation? Total capacity of non-empty W i grows by a constant c. Inverse transformation moves cuts from W i to some W j, j i. Inverse transformation incurs a c capacity violation. What does an instance transformation provide? Large total capacity of W i Better concentration as the total capacity increases Sum of conditional variances might be higher
85 Instance Transformation (Cont.) What is an instance transformation? Total capacity of non-empty W i grows by a constant c. Inverse transformation moves cuts from W i to some W j, j i. Inverse transformation incurs a c capacity violation. What does an instance transformation provide? Large total capacity of W i Better concentration as the total capacity increases Sum of conditional variances might be higher Probability that sum of conditional variances is small increases
86 Instance Transformation (Cont.) What is an instance transformation? Total capacity of non-empty W i grows by a constant c. Inverse transformation moves cuts from W i to some W j, j i. Inverse transformation incurs a c capacity violation. What does an instance transformation provide? Large total capacity of W i Better concentration as the total capacity increases Sum of conditional variances might be higher Probability that sum of conditional variances is small increases Conclusion: O(1) capacity violation.
87 Thank You! Questions?
Multiply Balanced k Partitioning
Multiply Balanced k Partitioning Amihood Amir 1,2,, Jessica Ficler 1,, Robert Krauthgamer 3, Liam Roditty 1, and Oren Sar Shalom 1 1 Department of Computer Science, Bar-Ilan University, Ramat-Gan 52900,
More informationCapacitated Metric Labeling
Capacitated Metric Labeling Matthew Andrews Mohammad Taghi Hajiaghayi Howard Karloff Ankur Moitra Abstract We introduce Capacitated Metric Labeling. As in Metric Labeling, we are given a weighted graph
More informationTopic: Balanced Cut, Sparsest Cut, and Metric Embeddings Date: 3/21/2007
CS880: Approximations Algorithms Scribe: Tom Watson Lecturer: Shuchi Chawla Topic: Balanced Cut, Sparsest Cut, and Metric Embeddings Date: 3/21/2007 In the last lecture, we described an O(log k log D)-approximation
More informationKey words. sparsification, approximation algorithm, graph partitioning, routing, metric space
VERTEX SPARSIFICATION AND OBLIVIOUS REDUCTIONS ANKUR MOITRA Abstract. Given an undirected, capacitated graph G = (V, E) and a set K V of terminals of size k, we construct an undirected, capacitated graph
More informationPartitioning Metric Spaces
Partitioning Metric Spaces Computational and Metric Geometry Instructor: Yury Makarychev 1 Multiway Cut Problem 1.1 Preliminaries Definition 1.1. We are given a graph G = (V, E) and a set of terminals
More informationHierarchical Community Decomposition Via Oblivious Routing Techniques
Hierarchical Community Decomposition Via Oblivious Routing Techniques Sean Kennedy Bell Labs Research (Murray Hill, NJ, SA) Oct 8 th, 2013 Joint work Jamie Morgenstern (CM), Gordon Wilfong (BLR), Liza
More informationORIE 6334 Spectral Graph Theory November 22, Lecture 25
ORIE 64 Spectral Graph Theory November 22, 206 Lecture 25 Lecturer: David P. Williamson Scribe: Pu Yang In the remaining three lectures, we will cover a prominent result by Arora, Rao, and Vazirani for
More informationHierarchical Clustering via Spreading Metrics
Journal of Machine Learning Research 18 2017) 1-35 Submitted 2/17; Revised 5/17; Published 8/17 Hierarchical Clustering via Spreading Metrics Aurko Roy College of Computing Georgia Institute of Technology
More informationHardness of Embedding Metric Spaces of Equal Size
Hardness of Embedding Metric Spaces of Equal Size Subhash Khot and Rishi Saket Georgia Institute of Technology {khot,saket}@cc.gatech.edu Abstract. We study the problem embedding an n-point metric space
More informationMinimum d-dimensional arrangement with fixed points
Abstract Minimum d-dimensional arrangement with fixed points Anupam Gupta Anastasios Sidiropoulos October 8, 2013 In the Minimum d-dimensional Arrangement Problem d-dimap) we are given a graph with edge
More informationFréchet embeddings of negative type metrics
Fréchet embeddings of negative type metrics Sanjeev Arora James R Lee Assaf Naor Abstract We show that every n-point metric of negative type (in particular, every n-point subset of L 1 ) admits a Fréchet
More informationNew algorithms for Disjoint Paths and Routing Problems
New algorithms for Disjoint Paths and Routing Problems Chandra Chekuri Dept. of Computer Science Univ. of Illinois (UIUC) Menger s Theorem Theorem: The maximum number of s-t edgedisjoint paths in a graph
More informationU.C. Berkeley CS294: Spectral Methods and Expanders Handout 11 Luca Trevisan February 29, 2016
U.C. Berkeley CS294: Spectral Methods and Expanders Handout Luca Trevisan February 29, 206 Lecture : ARV In which we introduce semi-definite programming and a semi-definite programming relaxation of sparsest
More informationAn Improved Approximation Algorithm for Requirement Cut
An Improved Approximation Algorithm for Requirement Cut Anupam Gupta Viswanath Nagarajan R. Ravi Abstract This note presents improved approximation guarantees for the requirement cut problem: given an
More informationUnbalanced Graph Cuts
Unbalanced Graph Cuts Ara Hayrapetyan 1, David Kempe 2, Martin Pál 3, and Zoya Svitkina 1 1 Dept. of Computer Science, Cornell University. {ara,zoya}@cs.cornell.edu 2 Dept. of Computer Science, University
More informationInteger Linear Programs
Lecture 2: Review, Linear Programming Relaxations Today we will talk about expressing combinatorial problems as mathematical programs, specifically Integer Linear Programs (ILPs). We then see what happens
More informationA Primal-Dual Randomized Algorithm for Weighted Paging
A Primal-Dual Randomized Algorithm for Weighted Paging Nikhil Bansal Niv Buchbinder Joseph (Seffi) Naor April 2, 2012 Abstract The study the weighted version of classic online paging problem where there
More informationPartitioning Algorithms that Combine Spectral and Flow Methods
CS369M: Algorithms for Modern Massive Data Set Analysis Lecture 15-11/11/2009 Partitioning Algorithms that Combine Spectral and Flow Methods Lecturer: Michael Mahoney Scribes: Kshipra Bhawalkar and Deyan
More informationOn Stochastic Decompositions of Metric Spaces
On Stochastic Decompositions of Metric Spaces Scribe of a talk given at the fall school Metric Embeddings : Constructions and Obstructions. 3-7 November 204, Paris, France Ofer Neiman Abstract In this
More informationRandomized Algorithms
Randomized Algorithms 南京大学 尹一通 Martingales Definition: A sequence of random variables X 0, X 1,... is a martingale if for all i > 0, E[X i X 0,...,X i1 ] = X i1 x 0, x 1,...,x i1, E[X i X 0 = x 0, X 1
More informationCS Communication Complexity: Applications and New Directions
CS 2429 - Communication Complexity: Applications and New Directions Lecturer: Toniann Pitassi 1 Introduction In this course we will define the basic two-party model of communication, as introduced in the
More informationd-dimensional Arrangement Revisited
-Dimensional Arrangement Revisite Daniel Rotter Jens Vygen Research Institute for Discrete Mathematics University of Bonn Revise version: April 5, 013 Abstract We revisit the -imensional arrangement problem
More informationWelfare Maximization with Friends-of-Friends Network Externalities
Welfare Maximization with Friends-of-Friends Network Externalities Extended version of a talk at STACS 2015, Munich Wolfgang Dvořák 1 joint work with: Sayan Bhattacharya 2, Monika Henzinger 1, Martin Starnberger
More informationApproximating Graph Spanners
Approximating Graph Spanners Michael Dinitz Johns Hopkins University Joint work with combinations of Robert Krauthgamer (Weizmann), Eden Chlamtáč (Ben Gurion), Ran Raz (Weizmann), Guy Kortsarz (Rutgers-Camden)
More informationLecture 20: Course summary and open problems. 1 Tools, theorems and related subjects
CSE 533: The PCP Theorem and Hardness of Approximation (Autumn 2005) Lecture 20: Course summary and open problems Dec. 7, 2005 Lecturer: Ryan O Donnell Scribe: Ioannis Giotis Topics we discuss in this
More informationApproximation Algorithms and Hardness of Integral Concurrent Flow
Approximation Algorithms and Hardness of Integral Concurrent Flow Parinya Chalermsook Julia Chuzhoy Alina Ene Shi Li May 16, 2012 Abstract We study an integral counterpart of the classical Maximum Concurrent
More informationLocal Search Based Approximation Algorithms. Vinayaka Pandit. IBM India Research Laboratory
Local Search Based Approximation Algorithms The k-median problem Vinayaka Pandit IBM India Research Laboratory joint work with Naveen Garg, Rohit Khandekar, and Vijay Arya The 2011 School on Approximability,
More informationNetwork Design and Game Theory Spring 2008 Lecture 6
Network Design and Game Theory Spring 2008 Lecture 6 Guest Lecturer: Aaron Archer Instructor: Mohammad T. Hajiaghayi Scribe: Fengming Wang March 3, 2008 1 Overview We study the Primal-dual, Lagrangian
More informationCMPUT 675: Approximation Algorithms Fall 2014
CMPUT 675: Approximation Algorithms Fall 204 Lecture 25 (Nov 3 & 5): Group Steiner Tree Lecturer: Zachary Friggstad Scribe: Zachary Friggstad 25. Group Steiner Tree In this problem, we are given a graph
More informationLecture 5. Max-cut, Expansion and Grothendieck s Inequality
CS369H: Hierarchies of Integer Programming Relaxations Spring 2016-2017 Lecture 5. Max-cut, Expansion and Grothendieck s Inequality Professor Moses Charikar Scribes: Kiran Shiragur Overview Here we derive
More informationOn a Cut-Matching Game for the Sparsest Cut Problem
On a Cut-Matching Game for the Sparsest Cut Problem Rohit Khandekar Subhash A. Khot Lorenzo Orecchia Nisheeth K. Vishnoi Electrical Engineering and Computer Sciences University of California at Berkeley
More informationCapacitated Network Design on Undirected Graphs
Capacitated Network Design on Undirected Graphs Deeparnab Chakrabarty, Ravishankar Krishnaswamy, Shi Li, and Srivatsan Narayanan Abstract. In this paper, we study the approximability of the capacitated
More information- Well-characterized problems, min-max relations, approximate certificates. - LP problems in the standard form, primal and dual linear programs
LP-Duality ( Approximation Algorithms by V. Vazirani, Chapter 12) - Well-characterized problems, min-max relations, approximate certificates - LP problems in the standard form, primal and dual linear programs
More informationCOS597D: Information Theory in Computer Science October 5, Lecture 6
COS597D: Information Theory in Computer Science October 5, 2011 Lecture 6 Lecturer: Mark Braverman Scribe: Yonatan Naamad 1 A lower bound for perfect hash families. In the previous lecture, we saw that
More informationA Unified Theorem on SDP Rank Reduction. yyye
SDP Rank Reduction Yinyu Ye, EURO XXII 1 A Unified Theorem on SDP Rank Reduction Yinyu Ye Department of Management Science and Engineering and Institute of Computational and Mathematical Engineering Stanford
More informationApproximating some network design problems with node costs
Approximating some network design problems with node costs Guy Kortsarz Zeev Nutov July 3, 2009 Abstract We study several multi-criteria undirected network design problems with node costs and lengths.
More informationCS 344 Design and Analysis of Algorithms. Tarek El-Gaaly Course website:
CS 344 Design and Analysis of Algorithms Tarek El-Gaaly tgaaly@cs.rutgers.edu Course website: www.cs.rutgers.edu/~tgaaly/cs344.html Course Outline Textbook: Algorithms by S. Dasgupta, C.H. Papadimitriou,
More informationON THE NP-COMPLETENESS OF SOME GRAPH CLUSTER MEASURES
ON THE NP-COMPLETENESS OF SOME GRAPH CLUSTER MEASURES JIŘÍ ŠÍMA AND SATU ELISA SCHAEFFER Academy of Sciences of the Czech Republic Helsinki University of Technology, Finland elisa.schaeffer@tkk.fi SOFSEM
More informationProject Summary. Approximating Network Design Problems on Directed and Undirected Graphs
Project Summary Approximating Network Design Problems on Directed and Undirected Graphs Network design on directed graphs are problems that require to select a minimum cost directed subgraph of a given
More information15-854: Approximation Algorithms Lecturer: Anupam Gupta Topic: Approximating Metrics by Tree Metrics Date: 10/19/2005 Scribe: Roy Liu
15-854: Approximation Algorithms Lecturer: Anupam Gupta Topic: Approximating Metrics by Tree Metrics Date: 10/19/2005 Scribe: Roy Liu 121 Introduction In this lecture, we show how to embed metric weighted
More informationApproximating Sparsest Cut in Graphs of Bounded Treewidth
Approximating Sparsest Cut in Graphs of Bounded Treewidth Eden Chlamtac 1, Robert Krauthgamer 1, and Prasad Raghavendra 2 1 Weizmann Institute of Science, Rehovot, Israel. {eden.chlamtac,robert.krauthgamer}@weizmann.ac.il
More informationFinite Metric Spaces & Their Embeddings: Introduction and Basic Tools
Finite Metric Spaces & Their Embeddings: Introduction and Basic Tools Manor Mendel, CMI, Caltech 1 Finite Metric Spaces Definition of (semi) metric. (M, ρ): M a (finite) set of points. ρ a distance function
More informationORIE 6334 Spectral Graph Theory December 1, Lecture 27 Remix
ORIE 6334 Spectral Graph Theory December, 06 Lecturer: David P. Williamson Lecture 7 Remix Scribe: Qinru Shi Note: This is an altered version of the lecture I actually gave, which followed the structure
More informationLabel Cover Algorithms via the Log-Density Threshold
Label Cover Algorithms via the Log-Density Threshold Jimmy Wu jimmyjwu@stanford.edu June 13, 2017 1 Introduction Since the discovery of the PCP Theorem and its implications for approximation algorithms,
More informationMetric Extension Operators, Vertex Sparsifiers and Lipschitz Extendability
Metric Extension Operators, Vertex Sparsifiers and Lipschitz Extendability Konstantin Maarychev IBM T.J. Watson Research Center Yury Maarychev Toyota Technological Institute at Chicago Abstract We study
More informationDesigning Competitive Online Algorithms via a Primal-Dual Approach. Niv Buchbinder
Designing Competitive Online Algorithms via a Primal-Dual Approach Niv Buchbinder Designing Competitive Online Algorithms via a Primal-Dual Approach Research Thesis Submitted in Partial Fulfillment of
More informationEfficient Primal- Dual Graph Algorithms for Map Reduce
Efficient Primal- Dual Graph Algorithms for Map Reduce Joint work with Bahman Bahmani Ashish Goel, Stanford Kamesh Munagala Duke University Modern Data Models Over the past decade, many commodity distributed
More informationThe Steiner k-cut Problem
The Steiner k-cut Problem Chandra Chekuri Sudipto Guha Joseph (Seffi) Naor September 23, 2005 Abstract We consider the Steiner k-cut problem which generalizes both the k-cut problem and the multiway cut
More informationOn Fixed Cost k-flow Problems
On Fixed Cost k-flow Problems MohammadTaghi Hajiaghayi 1, Rohit Khandekar 2, Guy Kortsarz 3, and Zeev Nutov 4 1 University of Maryland, College Park, MD. hajiagha@cs.umd.edu. 2 Knight Capital Group, Jersey
More informationLecture notes on the ellipsoid algorithm
Massachusetts Institute of Technology Handout 1 18.433: Combinatorial Optimization May 14th, 007 Michel X. Goemans Lecture notes on the ellipsoid algorithm The simplex algorithm was the first algorithm
More informationNotations. For any hypergraph H 2 V we have. H + H = and H + H = 2 V.
Notations For any hypergraph H 2 V we have H + H = and H + H = 2 V. Notations For any hypergraph H 2 V we have H + H = and H + H = 2 V. Given a Sperner hypergraph H, we can view it as the family of minimal
More informationMachine Minimization for Scheduling Jobs with Interval Constraints
Machine Minimization for Scheduling Jobs with Interval Constraints Julia Chuzhoy Sudipto Guha Sanjeev Khanna Joseph (Seffi) Naor Abstract The problem of scheduling jobs with interval constraints is a well-studied
More informationPrize-collecting Survivable Network Design in Node-weighted Graphs. Chandra Chekuri Alina Ene Ali Vakilian
Prize-collecting Survivable Network Design in Node-weighted Graphs Chandra Chekuri Alina Ene Ali Vakilian Survivable Network Design (SNDP) Collection of l pairs: s 1, t 1,, (s l, t l ) r(s i, t i ): requirement
More informationCutting corners cheaply, or how to remove Steiner points
Cutting corners cheaply, or how to remove Steiner points Lior Kamma The Weizmann Institute Robert Krauthgamer The Weizmann Institute Huy L. Nguy ên Princeton University Abstract Our main result is that
More informationLecture 14: Random Walks, Local Graph Clustering, Linear Programming
CSE 521: Design and Analysis of Algorithms I Winter 2017 Lecture 14: Random Walks, Local Graph Clustering, Linear Programming Lecturer: Shayan Oveis Gharan 3/01/17 Scribe: Laura Vonessen Disclaimer: These
More informationarxiv: v1 [cs.lg] 5 Jun 2018
A PROJECTION METHOD FOR METRIC-CONSTRAINED OPTIMIZATION NATE VELDT 1, DAVID F. GLEICH 2, ANTHONY WIRTH 3, AND JAMES SAUNDERSON 4 arxiv:1806.01678v1 [cs.lg] 5 Jun 2018 Abstract. We outline a new approach
More informationCutting corners cheaply, or how to remove Steiner points
Cutting corners cheaply, or how to remove Steiner points Lior Kamma Robert Krauthgamer Huy L. Nguy ên April 4, 2013 Abstract Our main result is that the Steiner Point Removal (SPR) problem can always be
More informationSanta Claus Schedules Jobs on Unrelated Machines
Santa Claus Schedules Jobs on Unrelated Machines Ola Svensson (osven@kth.se) Royal Institute of Technology - KTH Stockholm, Sweden March 22, 2011 arxiv:1011.1168v2 [cs.ds] 21 Mar 2011 Abstract One of the
More informationeigenvalue bounds and metric uniformization Punya Biswal & James R. Lee University of Washington Satish Rao U. C. Berkeley
eigenvalue bounds and metric uniformization Punya Biswal & James R. Lee University of Washington Satish Rao U. C. Berkeley separators in planar graphs Lipton and Tarjan (1980) showed that every n-vertex
More informationRandomized Competitive Algorithms for Generalized Caching
Randomized Competitive Algorithms for Generalized Caching Nikhil Bansal IBM T. J. Watson Research Yorktown Heights, USA nikhil@us.ibm.com Niv Buchbinder Computer Science Dept. Technion, Haifa, Israel nivb@cs.technion.ac.il
More informationLecture: Flow-based Methods for Partitioning Graphs (1 of 2)
Stat260/CS294: Spectral Graph Methods Lecture 10-02/24/2015 Lecture: Flow-based Methods for Partitioning Graphs (1 of 2) Lecturer: Michael Mahoney Scribe: Michael Mahoney Warning: these notes are still
More informationLecture 13: Spectral Graph Theory
CSE 521: Design and Analysis of Algorithms I Winter 2017 Lecture 13: Spectral Graph Theory Lecturer: Shayan Oveis Gharan 11/14/18 Disclaimer: These notes have not been subjected to the usual scrutiny reserved
More informationarxiv: v1 [cs.ds] 20 Nov 2016
Algorithmic and Hardness Results for the Hub Labeling Problem Haris Angelidakis 1, Yury Makarychev 1 and Vsevolod Oparin 2 1 Toyota Technological Institute at Chicago 2 Saint Petersburg Academic University
More informationImproved Approximating Algorithms for Directed Steiner Forest
Improved Approximating Algorithms for Directed Steiner Forest Moran Feldman Guy Kortsarz Zeev Nutov Abstract We consider the k-directed Steiner Forest (k-dsf) problem: Given a directed graph G = (V, E)
More informationNotes on MapReduce Algorithms
Notes on MapReduce Algorithms Barna Saha 1 Finding Minimum Spanning Tree of a Dense Graph in MapReduce We are given a graph G = (V, E) on V = N vertices and E = m N 1+c edges for some constant c > 0. Our
More informationarxiv:cs/ v1 [cs.cc] 7 Sep 2006
Approximation Algorithms for the Bipartite Multi-cut problem arxiv:cs/0609031v1 [cs.cc] 7 Sep 2006 Sreyash Kenkre Sundar Vishwanathan Department Of Computer Science & Engineering, IIT Bombay, Powai-00076,
More informationApproximation Basics
Approximation Basics, Concepts, and Examples Xiaofeng Gao Department of Computer Science and Engineering Shanghai Jiao Tong University, P.R.China Fall 2012 Special thanks is given to Dr. Guoqiang Li for
More informationGraph Partitioning Algorithms and Laplacian Eigenvalues
Graph Partitioning Algorithms and Laplacian Eigenvalues Luca Trevisan Stanford Based on work with Tsz Chiu Kwok, Lap Chi Lau, James Lee, Yin Tat Lee, and Shayan Oveis Gharan spectral graph theory Use linear
More informationAn LP Formulation and Approximation Algorithms for the Metric Labeling Problem
An LP Formulation and Approximation Algorithms for the Metric Labeling Problem C. Chekuri y S. Khanna z J. Naor x L. Zosin { August 4, 2004 Abstract We consider approximation algorithms for the metric
More informationDistributed Distance-Bounded Network Design Through Distributed Convex Programming
Distributed Distance-Bounded Network Design Through Distributed Convex Programming OPODIS 2017 Michael Dinitz, Yasamin Nazari Johns Hopkins University December 18, 2017 Distance Bounded Network Design
More informationAlgorithms for 2-Route Cut Problems
Algorithms for 2-Route Cut Problems Chandra Chekuri 1 and Sanjeev Khanna 2 1 Dept. of Computer Science, University of Illinois, Urbana, IL 61801. chekuri@cs.uiuc.edu 2 Dept. of CIS, Univ. of Pennsylvania,
More informationPREEMPTIVE RESOURCE CONSTRAINED SCHEDULING WITH TIME-WINDOWS
PREEMPTIVE RESOURCE CONSTRAINED SCHEDULING WITH TIME-WINDOWS Kanthi Sarpatwar IBM Research Joint Work With: Baruch Schieber (IBM Research) Hadas Shachnai (Technion) Kanthi Kiran Sarpatwar 1 / 22 Introduction
More informationHierarchies. 1. Lovasz-Schrijver (LS), LS+ 2. Sherali Adams 3. Lasserre 4. Mixed Hierarchy (recently used) Idea: P = conv(subset S of 0,1 n )
Hierarchies Today 1. Some more familiarity with Hierarchies 2. Examples of some basic upper and lower bounds 3. Survey of recent results (possible material for future talks) Hierarchies 1. Lovasz-Schrijver
More informationApproximation & Complexity
Summer school on semidefinite optimization Approximation & Complexity David Steurer Cornell University Part 1 September 6, 2012 Overview Part 1 Unique Games Conjecture & Basic SDP Part 2 SDP Hierarchies:
More informationThe Fast Optimal Voltage Partitioning Algorithm For Peak Power Density Minimization
The Fast Optimal Voltage Partitioning Algorithm For Peak Power Density Minimization Jia Wang, Shiyan Hu Department of Electrical and Computer Engineering Michigan Technological University Houghton, Michigan
More informationApproximating maximum satisfiable subsystems of linear equations of bounded width
Approximating maximum satisfiable subsystems of linear equations of bounded width Zeev Nutov The Open University of Israel Daniel Reichman The Open University of Israel Abstract We consider the problem
More information7. Lecture notes on the ellipsoid algorithm
Massachusetts Institute of Technology Michel X. Goemans 18.433: Combinatorial Optimization 7. Lecture notes on the ellipsoid algorithm The simplex algorithm was the first algorithm proposed for linear
More informationHallucination Helps: Energy Efficient Virtual Circuit Routing
Hallucination Helps: Energy Efficient Virtual Circuit Routing Antonios Antoniadis Sungjin Im Ravishankar Krishnaswamy Benjamin Moseley Abstract Viswanath Nagarajan Kirk Pruhs Cliff Stein We consider virtual
More informationChapter 8 The Disjoint Sets Class
Chapter 8 The Disjoint Sets Class 2 Introduction equivalence problem can be solved fairly simply simple data structure each function requires only a few lines of code two operations: and can be implemented
More information16 Embeddings of the Euclidean metric
16 Embeddings of the Euclidean metric In today s lecture, we will consider how well we can embed n points in the Euclidean metric (l 2 ) into other l p metrics. More formally, we ask the following question.
More information3.4 Relaxations and bounds
3.4 Relaxations and bounds Consider a generic Discrete Optimization problem z = min{c(x) : x X} with an optimal solution x X. In general, the algorithms generate not only a decreasing sequence of upper
More informationMulticriteria approximation through decomposition
Multicriteria approximation through decomposition Carl Burch Sven Krumke y Madhav Marathe z Cynthia Phillips x Eric Sundberg { Abstract We propose a general technique called solution decomposition to devise
More informationNP Completeness and Approximation Algorithms
Chapter 10 NP Completeness and Approximation Algorithms Let C() be a class of problems defined by some property. We are interested in characterizing the hardest problems in the class, so that if we can
More informationGRAPH PARTITIONING USING SINGLE COMMODITY FLOWS [KRV 06] 1. Preliminaries
GRAPH PARTITIONING USING SINGLE COMMODITY FLOWS [KRV 06] notes by Petar MAYMOUNKOV Theme The algorithmic problem of finding a sparsest cut is related to the combinatorial problem of building expander graphs
More informationLearning Theory. Ingo Steinwart University of Stuttgart. September 4, 2013
Learning Theory Ingo Steinwart University of Stuttgart September 4, 2013 Ingo Steinwart University of Stuttgart () Learning Theory September 4, 2013 1 / 62 Basics Informal Introduction Informal Description
More informationAlgorithmic Primitives for Network Analysis: Through the Lens of the Laplacian Paradigm
Algorithmic Primitives for Network Analysis: Through the Lens of the Laplacian Paradigm Shang-Hua Teng Computer Science, Viterbi School of Engineering USC Massive Data and Massive Graphs 500 billions web
More informationUnit 1A: Computational Complexity
Unit 1A: Computational Complexity Course contents: Computational complexity NP-completeness Algorithmic Paradigms Readings Chapters 3, 4, and 5 Unit 1A 1 O: Upper Bounding Function Def: f(n)= O(g(n)) if
More informationAssortment Optimization under the Multinomial Logit Model with Nested Consideration Sets
Assortment Optimization under the Multinomial Logit Model with Nested Consideration Sets Jacob Feldman School of Operations Research and Information Engineering, Cornell University, Ithaca, New York 14853,
More informationApproximability of Pricing Problems
joint work with Patrick Briest Department of Computer Science University of Liverpool U.K. 28-03-2007 Multi-Product Pricing Problems Websites comparing available products help customers make optimal buying
More informationA better approximation ratio for the Vertex Cover problem
A better approximation ratio for the Vertex Cover problem George Karakostas Dept. of Computing and Software McMaster University October 5, 004 Abstract We reduce the approximation factor for Vertex Cover
More informationOblivious randomized rounding
Oblivious randomized rounding Neal E. Young April 28, 2008 What would the world be like if... SAT is hard in the worst case, BUT... generating hard random instances of SAT is hard? Lipton, 1993 worst-case
More informationAlgorithms. Outline! Approximation Algorithms. The class APX. The intelligence behind the hardware. ! Based on
6117CIT - Adv Topics in Computing Sci at Nathan 1 Algorithms The intelligence behind the hardware Outline! Approximation Algorithms The class APX! Some complexity classes, like PTAS and FPTAS! Illustration
More informationConcept Drift. Albert Bifet. March 2012
Concept Drift Albert Bifet March 2012 COMP423A/COMP523A Data Stream Mining Outline 1. Introduction 2. Stream Algorithmics 3. Concept drift 4. Evaluation 5. Classification 6. Ensemble Methods 7. Regression
More informationRamsey partitions and proximity data structures
Ramsey partitions and proximity data structures Manor Mendel The Open University of Israel Assaf Naor Microsoft Research Abstract This paper addresses two problems lying at the intersection of geometric
More informationCS 6110 S16 Lecture 33 Testing Equirecursive Equality 27 April 2016
CS 6110 S16 Lecture 33 Testing Equirecursive Equality 27 April 2016 1 Equirecursive Equality In the equirecursive view of recursive types, types are regular labeled trees, possibly infinite. However, we
More informationRamsey partitions and proximity data structures
Ramsey partitions and proximity data structures Manor Mendel Assaf Naor Abstract This paper addresses two problems lying at the intersection of geometric analysis and theoretical computer science: The
More informationCapacitated Network Design Problems: Hardness, Approximation Algorithms, and Connections to Group Steiner Tree
Capacitated Network Design Problems: Hardness, Approximation Algorithms, and Connections to Group Steiner Tree MohammadTaghi Hajiaghayi 1, Rohit Khandekar 2, GuyKortsarz 3, and Zeev Nutov 4 1 University
More informationInderjit Dhillon The University of Texas at Austin
Inderjit Dhillon The University of Texas at Austin ( Universidad Carlos III de Madrid; 15 th June, 2012) (Based on joint work with J. Brickell, S. Sra, J. Tropp) Introduction 2 / 29 Notion of distance
More informationThe Steiner Network Problem
The Steiner Network Problem Pekka Orponen T-79.7001 Postgraduate Course on Theoretical Computer Science 7.4.2008 Outline 1. The Steiner Network Problem Linear programming formulation LP relaxation 2. The
More informationFINAL EXAM PRACTICE PROBLEMS CMSC 451 (Spring 2016)
FINAL EXAM PRACTICE PROBLEMS CMSC 451 (Spring 2016) The final exam will be on Thursday, May 12, from 8:00 10:00 am, at our regular class location (CSI 2117). It will be closed-book and closed-notes, except
More information