Clarkson s Algorithm for Violator Spaces. Yves Brise, ETH Zürich, CCCG Joint work with Bernd Gärtner
|
|
- Angela Fleming
- 5 years ago
- Views:
Transcription
1 Clarkson s Algorithm for Violator Spaces Yves Brise, ETH Zürich, CCCG Joint work with Bernd Gärtner
2 A Hierarchy of Optimization Frameworks Linear Programming Halfspaces (constraints), Optimization direction Optimal point LP-type Problems, Sharir, Welzl (1992) Set of constraints H, Total order on 2 H, Value function ϕ :2 H R Optimal subset of H Violator Spaces, Skovron (2002) Set of constraints H, Violator mapping V :2 H 2 H B H minimal and V (B) =
3 Smallest Enclosing Ball Given a set H of n points in R d, find the smallest enclosing ball. Consider R H. Basis: A minimal set of points that have the same smallest enclosing ball. R Violators: Points that lie outside of the smallest enclosing ball of R. H
4 Definition Violator Space The pair (H, V ) is a violator space if H is a finite set, and V :2 H 2 H satisfies the following conditions. Consistency: G H, V (G) G = Locality: F G H, if V (F ) G = then V (F )=V(G) Basis: A basis of G H is a minimal subset B of G, s.t. V (B) =V (G). The combinatorial dimension of (H, V ), denoted by dim(h, V ), is the size of the largest basis in (H, V ).
5 Sampling Idea choose R H u.a.r.; B basis(r); while (V (B) = ) { choose h V (B); B basis(b {h}); } return B;
6 Clarkson s Algorithm Main Idea: Sampling! But do it the right way. while (V (B) = ) { choose h V (B); B basis(b {h}); } return B; 1. Don t discard too many violators. German Algorithm 2. Adjust weights and do resampling in every round. Swiss Algorithm
7 Overview Clarkson (H, V ), H = n, dim(h, V )=δ German Algorithm (δ + 1) (H, V ), n = O(δ n) Brute (H, V ), O(δ ln n ) Swiss Force H = O(δ 2 ) Algorithm Las Vegas Algorithms for Linear and Integer Programming When the Dimension is Small, Clarkson, Initially developed for Linear Programming.
8 Swiss Algorithm µ h 1, h H; 1 do { choose R H according to µ; B basis(r); µ h 2µ h, h V (B); } until (V (B) = ) return B; µ h is the multiplicity of node h
9 Simplified Clarkson µ h 1, h H; while (V (B) = ) { choose R H according to µ; B basis(r); µ h 2µ h, h V (B); } return B; if µ(v (B)) 1 2d µ(h) then µ h 2µ h, h V (B); Not intuitive! Only there for the analysis.
10 Overview Clarkson (H, V ), H = n, dim(h, V )=δ German Algorithm (δ + 1) (H, V ), n = O(δ n) Brute (H, V ), O(δ ln n ) Swiss Force H = O(δ 2 ) Algorithm Theorem: Let (H, V ) be a violator space, H = n, dim(h, V )=δ. We can compute a basis of (H, V ) using an expected number of O(δ 2 n + f (δ)) unit cost operations.
11 Result Overview Gärtner, Welzl, 2001 Brise, Gärtner, 2009 German Swiss Brute Force LP Simplified Simplified Subexponential LP-type Simplified Simplified Subexponential VS Simplified Simplified Exponential Kalai, Matousek, Sharir, Welzl, 1992 Gärtner, 1995
12 Thank You For Your Attention! Yves Brise,
13 Runtime 2nd Stage Lemma: Let (H, V ) be a violator space, H = n, dim(h, V )=δ, B any basis of (H, V ), and k N. Then after kδ rounds of Stage II, 2 k E[µ (kδ) (B)] < ne k/2. By choosing k large enough we derive a contradiction. Proof: (lower bound) By a simple Lemma we have that B contains at least one violator of every round. After kδ rounds there must be an element in B that has doubled its weight at least k times, because B δ.
14 Runtime 2nd Stage Proof: (upper bound) E[µ () (H)] = E[µ (0) (H)] + i=1 E[ (i) (H)], where (i) is the increment in multiplicity in round i. i=1 i=1 E[ (i) (H)] = E[ (i) (H) µ (i 1) (H) =t] Pr[µ (i 1) (H) =t] t=0 Sampling Lemma: E[ (i) (H) µ (i 1) (H) =t] δ t r r +1
15 Runtime 2nd Stage Calculating, we get a recursive equation: E[µ () (H)] n + δ r +1 i=1 E[µ (i 1) (H)] δr r +1 Finally, plugging in r =2δ 2, we get E(µ () (H)) (1 + δ r ) n ( δ )kδ n < ne k/2 This concludes the proof.
16 Open Problems
17 Clarkson Dimension # rounds / points in R d Average of 10^3 trials Lemma: Let (H, V )bea violator space, dim(h, V )=δ. Stage 1 runs for at most δ + 1 rounds dimension d unit cube unit ball simplex
18 Clarkson Dimension Average of 10^3 trials Size of G / points in R d 22,000 18,400 14,800 11,200 7,600 Theorem: Let (H, V )bea violator space, dim(h, V )=δ, H = n, and G (δ) the working set of Stage 1 after δ rounds. Then, E[ G (δ) ] 2(δ + 1) n/2. 4, dimension d unit cube unit ball simplex
19 Cyclic VS, dim=2? The generalized linear complementarity problem with P-matrices gives rise to possibly cyclic unique sink orientations (USO) of hypercube grids. 123 GLCP Grid USO Violator Space Lemma (Gärtner, Morris, Rüst, 2005): 2-dimensional cyclic grid USO Grid USO cannot be a model for cyclic, non-degenerate violator spaces of dim = 2.
20 Other Open Problems Subexponential analysis for violator spaces Number of violator spaces Unified Clarkson algorithm
21 Junk
22 Sampling Lemma Let P be a set of size n, ϕ :2 P R, and R P, r = R. Violators: V (R) := {p P\R ϕ(r {p}) = ϕ(r)} Extreme: X (R) := {p R ϕ(r\{p}) = ϕ(r)} Observation: p violates R p is extreme in R {p} Lemma (Gärtner, Welzl, 2001): For 0 r n, E[ V (R) ] n r = E [ X (R) ] r+1
23 Sampling Lemma (proof) n r E[ V (R) ] = R ( S r) = = = R ( S r) Q ( S r+1) n r +1 s S\R s S\R s Q [s violates R] [s is extreme in R {s}] [s is extreme in Q] E[ X (R) ]
24 2nd Stage Calculation E[µ () (H)] n + = n + i=1 i=1 dr r +1 = n + d t=0 r +1 d t r r +1 Pr[µ(i 1) (H) =t] d r +1 t=0 t=0 t Pr[µ (i 1) (H) =t] Pr[µ (i 1) (H) =t] i=1 E[µ (i 1) (H)] dr r +1
25 History In 1992, the first subexponential algorithms for Linear Programming are developed by Matousek, Sharir, and Welzl, and independently by Kalai. In 1996, Sharir and Welzl extend the machinery to LP-type problems, which are a generalization of Linear Programming. In 2002, Skovron introduces Violator Spaces. Subsequently, it is showed by Gärtner, Matousek, Rüst, and Skovron that a lot of the machinery carries over. In LP-type there is a value function ϕ, which is used to define violators and extreme points. It turns out that ϕ is not necessary. The problem defines the violator mapping V. Extreme: X (R) := {p R V (R\{p}) = V (R)}
26 Smallest Enclosing Ball Nondegeneracy: Less than d + 2 points on any ball. P Regularity: Every smallest ball has exactly the same number of points on the boundary.
27 Definition Violator Space The pair (H, V ) is a violator space if H is a finite set, and V :2 H 2 H satisfies the following conditions. Consistency: G H, V (G) G = Locality: F G H, if V (F ) G = then V (F )=V(G) s Not everything is local! Consider the diameter problem: Given P, find the p D r F = {q, s} G = {p, q, s} Euclidean diameter of P. q = V (F ) = V (G) ={r}
28 Cyclic Violator Spaces f g G B(G) V (G) f, g, h f f h g g f h h g f, g f h f, h h g g, h g f f, g, h f, g, h h A point is locally smaller than another if it is further clockwise w.r.t. the center (measure the acute angle).
29 Some Examples Linear Programming Polytope Distance OPT s c OPT dim = d dim = d +1 s Largest Ellipsoid in Polytope Rectilinear 3-Center Problem dim = 43 dim d 2 Sharir, Welzl, 1996
30 German Algorithm choose R H u.a.r.; B basis(r); while (V (B) = ) { R R V (R); B basis(r); } return B;
31 Combinatorial Dimension Consider the violator space (H, V ) Basis: A basis of G H is a minimal subset B of G, s.t. V (B) =V (G). The combinatorial dimension of (H, V ), denoted by dim(h, V ), is the size of the largest basis in (H, V ). Smallest Enclosing Ball: It is well known that d + 1 points in R d uniquely determine a ball. Therefore dim d + 1.
Linear Programming Randomization and Abstract Frameworks
appeared in Proc. 13th Ann. Symp. on Theoret. Aspects of Comput. Sci. Lecture Notes in Computer Science 1046 (1996) 669-687. Linear Programming Randomization and Abstract Frameworks Bernd Gärtner and Emo
More informationpursued in order to overcome the exponential worst case behavior. First, the average complexity has been studied which assumes that the input is rando
Linear Programming { Randomization and Abstract Frameworks? Bernd Gartner and Emo Welzl Institut fur Informatik, Freie Universitat Berlin, Takustr. 9, D-14195 Berlin, Germany fgaertner,emog@inf.fu-berlin.de
More informationChapter 0 Introduction Suppose this was the abstract of a journal paper rather than the introduction to a dissertation. Then it would probably end wit
Chapter 0 Introduction Suppose this was the abstract of a journal paper rather than the introduction to a dissertation. Then it would probably end with some cryptic AMS subject classications and a few
More informationA Subexponential Bound for Linear Programming
appeared in Algorithmica 6 (996) 498-56. A Subexponential Bound for Linear Programming Jiří Matoušek Department of Applied Mathematics Charles University, Malostranské nám. 25 8 00 Praha, Czechoslovakia
More informationThe Random-Facet Simplex Algorithm on Combinatorial Cubes
The Random-Facet Simplex lgorithm on Combinatorial Cubes Bernd Gärtner, ETH Zürich, Switzerland January, 2002 bstract The Random-Facet algorithm is a randomized variant of the simplex method which is known
More informationON THE ALGORITHMIC ASPECTS OF DISCRETE AND LEXICOGRAPHIC HELLY-TYPE THEOREMS AND THE DISCRETE LP-TYPE MODEL
SIAM J. COMPUT. Vol. 0, No. 0, pp. 000 000 c 200X Society for Industrial and Applied Mathematics ON THE ALGORITHMIC ASPECTS OF DISCRETE AND LEXICOGRAPHIC HELLY-TYPE THEOREMS AND THE DISCRETE LP-TYPE MODEL
More informationThe Ellipsoid Algorithm
The Ellipsoid Algorithm John E. Mitchell Department of Mathematical Sciences RPI, Troy, NY 12180 USA 9 February 2018 Mitchell The Ellipsoid Algorithm 1 / 28 Introduction Outline 1 Introduction 2 Assumptions
More informationarxiv: v1 [cs.cg] 3 Dec 2014
Combinatorial Redundancy Detection Komei Fukuda 1, Bernd Gärtner 2, and May Szedlák 3 1 Department of Mathematics and Institute of Theoretical Computer Science ETH Zürich, CH-8092 Zürich, Switzerland komeifukuda@mathethzch
More informationMotivating examples Introduction to algorithms Simplex algorithm. On a particular example General algorithm. Duality An application to game theory
Instructor: Shengyu Zhang 1 LP Motivating examples Introduction to algorithms Simplex algorithm On a particular example General algorithm Duality An application to game theory 2 Example 1: profit maximization
More information15-850: Advanced Algorithms CMU, Spring 2017 Lecture #17: The Ellipsoid Algorithm March 3, 2017
15-850: Advanced Algorithms CMU, Spring 2017 Lecture #17: The Ellipsoid Algorithm March 3, 2017 Lecturer: Anupam Gupta Scribe: Benjamin Berg, David K. Isenberg In this lecture, we discuss some polynomial-time
More informationA subexponential lower bound for the Random Facet algorithm for Parity Games
A subexponential lower bound for the Random Facet algorithm for Parity Games Oliver Friedmann 1 Thomas Dueholm Hansen 2 Uri Zwick 3 1 Department of Computer Science, University of Munich, Germany. 2 Center
More informationCharles University, Malostranske nam. 25. Micha Sharir. Tel Aviv University, Tel Aviv 69978, Israel. and. Emo Welzl.
A Subexponential Bound for Linear Programming y Jiri Matousek z Department of Applied Mathematics Charles University, Malostranske nam. 25 8 00 Praha, Czechoslovakia e-mail: matousek@kam.mff.cuni.cz Micha
More informationAn efficient approximation for point-set diameter in higher dimensions
CCCG 2018, Winnipeg, Canada, August 8 10, 2018 An efficient approximation for point-set diameter in higher dimensions Mahdi Imanparast Seyed Naser Hashemi Ali Mohades Abstract In this paper, we study the
More informationTopics in Theoretical Computer Science April 08, Lecture 8
Topics in Theoretical Computer Science April 08, 204 Lecture 8 Lecturer: Ola Svensson Scribes: David Leydier and Samuel Grütter Introduction In this lecture we will introduce Linear Programming. It was
More informationWeek 8. 1 LP is easy: the Ellipsoid Method
Week 8 1 LP is easy: the Ellipsoid Method In 1979 Khachyan proved that LP is solvable in polynomial time by a method of shrinking ellipsoids. The running time is polynomial in the number of variables n,
More informationHelly-type Theorems and Generalized Linear Programming. Nina Amenta. Abstract
Helly-type Theorems and Generalized Linear Programming Nina Amenta Abstract Recent combinatorial algorithms for linear programming can also be applied to certain nonlinear problems. We call these Generalized
More informationFrom the Zonotope Construction to the Minkowski Addition of Convex Polytopes
From the Zonotope Construction to the Minkowski Addition of Convex Polytopes Komei Fukuda School of Computer Science, McGill University, Montreal, Canada Abstract A zonotope is the Minkowski addition of
More informationA Simple Proof of Optimal Epsilon Nets
A Simple Proof of Optimal Epsilon Nets Nabil H. Mustafa Kunal Dutta Arijit Ghosh Abstract Showing the existence of ɛ-nets of small size has been the subject of investigation for almost 30 years, starting
More informationJohnson-Lindenstrauss, Concentration and applications to Support Vector Machines and Kernels
Johnson-Lindenstrauss, Concentration and applications to Support Vector Machines and Kernels Devdatt Dubhashi Department of Computer Science and Engineering, Chalmers University, dubhashi@chalmers.se Functional
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 informationMinimizing Interference in Wireless Networks. Yves Brise, ETH Zürich, Joint work with Marek, Michael, and Tobias.
Minimizing Interference in Wireless Networks Yves Brise, ETH Zürich, 20100302 Joint work with Marek, Michael, and Tobias. The Problem The Problem Formal Description Input: Output: Set V of n vertices in
More informationTwo New Bounds on the Random-Edge Simplex Algorithm
Konrad-Zuse-Zentrum für Informationstechnik Berlin Takustraße 7 D-14195 Berlin-Dahlem Germany BERND GÄRTNER AND VOLKER KAIBEL Two New Bounds on the Random-Edge Simplex Algorithm ZIB-Report 05-14 (February
More informationThe polynomial method in combinatorics
AMS joint meetings 4 January 2012 Overview In the last five years, several challenging problems in combinatorics have been solved by introducing polynomials into the problem in an unexpected way. This
More informationCS675: Convex and Combinatorial Optimization Spring 2018 The Ellipsoid Algorithm. Instructor: Shaddin Dughmi
CS675: Convex and Combinatorial Optimization Spring 2018 The Ellipsoid Algorithm Instructor: Shaddin Dughmi History and Basics Originally developed in the mid 70s by Iudin, Nemirovski, and Shor for use
More informationCS 6820 Fall 2014 Lectures, October 3-20, 2014
Analysis of Algorithms Linear Programming Notes CS 6820 Fall 2014 Lectures, October 3-20, 2014 1 Linear programming The linear programming (LP) problem is the following optimization problem. We are given
More informationOptimization methods NOPT048
Optimization methods NOPT048 Jirka Fink https://ktiml.mff.cuni.cz/ fink/ Department of Theoretical Computer Science and Mathematical Logic Faculty of Mathematics and Physics Charles University in Prague
More informationDiscrete and Lexicographic Helly Theorems and their Relations to LP-Type Problems
TEL AVIV UNIVERSITY The Raymond and Beverly Sackler Faculty of Exact Sciences School of Mathematical Sciences Discrete and Lexicographic Helly Theorems and their Relations to LP-Type Problems Thesis submitted
More informationNash Equilibrium: Existence & Computation
: & IIIS, Tsinghua University zoy.blood@gmail.com March 28, 2016 Overview 1 Fixed-Point Theorems: Kakutani & Brouwer Proof of 2 A Brute Force Solution Lemke Howson Algorithm Overview Fixed-Point Theorems:
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 informationOptimization methods NOPT048
Optimization methods NOPT048 Jirka Fink https://ktiml.mff.cuni.cz/ fink/ Department of Theoretical Computer Science and Mathematical Logic Faculty of Mathematics and Physics Charles University in Prague
More informationRandomized Optimization by Simplex-Type Methods
Randomized Optimization by Simplex-Type Methods Dissertation zur Erlangung des Doktorgrades vorgelegt am Fachbereich Mathematik und Informatik der Freien Universität Berlin von Bernd Gärtner verteidigt
More informationSzemerédi-Trotter theorem and applications
Szemerédi-Trotter theorem and applications M. Rudnev December 6, 2004 The theorem Abstract These notes cover the material of two Applied post-graduate lectures in Bristol, 2004. Szemerédi-Trotter theorem
More informationApproximate Range Searching: The Absolute Model
Approximate Range Searching: The Absolute Model Guilherme D. da Fonseca Department of Computer Science University of Maryland College Park, Maryland 20742 fonseca@cs.umd.edu David M. Mount Department of
More informationChapter 11. Approximation Algorithms. Slides by Kevin Wayne Pearson-Addison Wesley. All rights reserved.
Chapter 11 Approximation Algorithms Slides by Kevin Wayne. Copyright @ 2005 Pearson-Addison Wesley. All rights reserved. 1 Approximation Algorithms Q. Suppose I need to solve an NP-hard problem. What should
More informationA non-linear lower bound for planar epsilon-nets
A non-linear lower bound for planar epsilon-nets Noga Alon Abstract We show that the minimum possible size of an ɛ-net for point objects and line (or rectangle)- ranges in the plane is (slightly) bigger
More informationGeometry. Covering Things with Things. Stefan Langerman 1 and Pat Morin Introduction
Discrete Comput Geom 33:717 729 (2005) DOI: 10.1007/s00454-004-1108-4 Discrete & Computational Geometry 2004 Springer-Verlag New York, LLC Covering Things with Things Stefan Langerman 1 and Pat Morin 2
More information1 Matroid intersection
CS 369P: Polyhedral techniques in combinatorial optimization Instructor: Jan Vondrák Lecture date: October 21st, 2010 Scribe: Bernd Bandemer 1 Matroid intersection Given two matroids M 1 = (E, I 1 ) and
More informationarxiv: v2 [cs.cg] 28 Sep 2012
No Dimension Independent Core-Sets for Containment under Homothetics arxiv:1010.49v [cs.cg] 8 Sep 01 René Brandenberg Zentrum Mathematik Technische Universität München Boltzmannstr. 3 85747 Garching bei
More informationLower Bounds on the Complexity of Simplex Range Reporting on a Pointer Machine
Lower Bounds on the Complexity of Simplex Range Reporting on a Pointer Machine Extended Abstract Bernard Chazelle Department of Computer Science Princeton University Burton Rosenberg Department of Mathematics
More informationSimilarity searching, or how to find your neighbors efficiently
Similarity searching, or how to find your neighbors efficiently Robert Krauthgamer Weizmann Institute of Science CS Research Day for Prospective Students May 1, 2009 Background Geometric spaces and techniques
More informationarxiv: v1 [cs.cg] 16 May 2011
Collinearities in Kinetic Point Sets arxiv:1105.3078v1 [cs.cg] 16 May 2011 Ben D. Lund George B. Purdy Justin W. Smith Csaba D. Tóth August 24, 2018 Abstract Let P be a set of n points in the plane, each
More informationLP-related Properties of Polytopes with Few Facets
DISS. ETH No. 579, 2003 LP-related Properties of Polytopes with Few Facets A dissertation submitted to the Swiss Federal Institute of Technology, ETH Zürich for the degree of Doctor of Technical Sciences
More informationIterative Rounding and Relaxation
Iterative Rounding and Relaxation James Davis Department of Computer Science Rutgers University Camden jamesdav@camden.rutgers.edu March 11, 2010 James Davis (Rutgers Camden) Iterative Rounding 1 / 58
More informationCS 372: Computational Geometry Lecture 14 Geometric Approximation Algorithms
CS 372: Computational Geometry Lecture 14 Geometric Approximation Algorithms Antoine Vigneron King Abdullah University of Science and Technology December 5, 2012 Antoine Vigneron (KAUST) CS 372 Lecture
More informationDiscrete Geometry. Problem 1. Austin Mohr. April 26, 2012
Discrete Geometry Austin Mohr April 26, 2012 Problem 1 Theorem 1 (Linear Programming Duality). Suppose x, y, b, c R n and A R n n, Ax b, x 0, A T y c, and y 0. If x maximizes c T x and y minimizes b T
More informationWeek 2. The Simplex method was developed by Dantzig in the late 40-ties.
1 The Simplex method Week 2 The Simplex method was developed by Dantzig in the late 40-ties. 1.1 The standard form The simplex method is a general description algorithm that solves any LPproblem instance.
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 informationThe Traveling Salesman Problem with Few Inner Points
The Traveling Salesman Problem with Few Inner Points Vladimir G. Deĭneko 1,, Michael Hoffmann 2, Yoshio Okamoto 2,, and Gerhard J. Woeginger 3, 1 Warwick Business School, The University of Warwick, Conventry
More informationCombinatorial Generalizations of Jung s Theorem
Discrete Comput Geom (013) 49:478 484 DOI 10.1007/s00454-013-9491-3 Combinatorial Generalizations of Jung s Theorem Arseniy V. Akopyan Received: 15 October 011 / Revised: 11 February 013 / Accepted: 11
More informationAsymptotic Analysis. Slides by Carl Kingsford. Jan. 27, AD Chapter 2
Asymptotic Analysis Slides by Carl Kingsford Jan. 27, 2014 AD Chapter 2 Independent Set Definition (Independent Set). Given a graph G = (V, E) an independent set is a set S V if no two nodes in S are joined
More informationHypergraph Matching by Linear and Semidefinite Programming. Yves Brise, ETH Zürich, Based on 2010 paper by Chan and Lau
Hypergraph Matching by Linear and Semidefinite Programming Yves Brise, ETH Zürich, 20110329 Based on 2010 paper by Chan and Lau Introduction Vertex set V : V = n Set of hyperedges E Hypergraph matching:
More informationHamming codes and simplex codes ( )
Chapter 6 Hamming codes and simplex codes (2018-03-17) Synopsis. Hamming codes are essentially the first non-trivial family of codes that we shall meet. We start by proving the Distance Theorem for linear
More informationSolving the MWT. Recall the ILP for the MWT. We can obtain a solution to the MWT problem by solving the following ILP:
Solving the MWT Recall the ILP for the MWT. We can obtain a solution to the MWT problem by solving the following ILP: max subject to e i E ω i x i e i C E x i {0, 1} x i C E 1 for all critical mixed cycles
More informationGeneralized Ham-Sandwich Cuts
DOI 10.1007/s00454-009-9225-8 Generalized Ham-Sandwich Cuts William Steiger Jihui Zhao Received: 17 March 2009 / Revised: 4 August 2009 / Accepted: 14 August 2009 Springer Science+Business Media, LLC 2009
More informationThe Maximum Likelihood Threshold of a Graph
The Maximum Likelihood Threshold of a Graph Elizabeth Gross and Seth Sullivant San Jose State University, North Carolina State University August 28, 2014 Seth Sullivant (NCSU) Maximum Likelihood Threshold
More informationImproved Bounds on the Union Complexity of Fat Objects
Discrete Comput Geom (2008) 40: 127 140 DOI 10.1007/s00454-007-9029-7 Improved Bounds on the Union Complexity of Fat Objects Mark de Berg Received: 20 October 2006 / Revised: 9 July 2007 / Published online:
More information1 The linear algebra of linear programs (March 15 and 22, 2015)
1 The linear algebra of linear programs (March 15 and 22, 2015) Many optimization problems can be formulated as linear programs. The main features of a linear program are the following: Variables are real
More informationa 1 a 2 a 3 a 4 v i c i c(a 1, a 3 ) = 3
AM 221: Advanced Optimization Spring 2016 Prof. Yaron Singer Lecture 17 March 30th 1 Overview In the previous lecture, we saw examples of combinatorial problems: the Maximal Matching problem and the Minimum
More informationCS599: Convex and Combinatorial Optimization Fall 2013 Lecture 17: Combinatorial Problems as Linear Programs III. Instructor: Shaddin Dughmi
CS599: Convex and Combinatorial Optimization Fall 2013 Lecture 17: Combinatorial Problems as Linear Programs III Instructor: Shaddin Dughmi Announcements Today: Spanning Trees and Flows Flexibility awarded
More informationConflict-Free Coloring and its Applications
Conflict-Free Coloring and its Applications Shakhar Smorodinsky September 12, 2011 Abstract Let H = (V, E) be a hypergraph. A conflict-free coloring of H is an assignment of colors to V such that, in each
More informationLinear Programming. Jie Wang. University of Massachusetts Lowell Department of Computer Science. J. Wang (UMass Lowell) Linear Programming 1 / 47
Linear Programming Jie Wang University of Massachusetts Lowell Department of Computer Science J. Wang (UMass Lowell) Linear Programming 1 / 47 Linear function: f (x 1, x 2,..., x n ) = n a i x i, i=1 where
More informationd _< lim < d. d + 2 ~o (1/~)log(1/~)
Discrete Comput Geom 7:163-173 (1992) Discrete & Computational Geometry ~) 1992 Springer-Verlag New York Inc. Almost Tight Bounds for e-nets Jfinos Koml6s, 1 Jfinos Pach, 2'3 and Gerhard Woeginger 4 t
More informationA Note on Smaller Fractional Helly Numbers
A Note on Smaller Fractional Helly Numbers Rom Pinchasi October 4, 204 Abstract Let F be a family of geometric objects in R d such that the complexity (number of faces of all dimensions on the boundary
More informationStreaming Algorithms for Extent Problems in High Dimensions
Streaming Algorithms for Extent Problems in High Dimensions Pankaj K. Agarwal R. Sharathkumar Department of Computer Science Duke University {pankaj, sharath}@cs.duke.edu Abstract We develop single-pass)
More informationLinear Programming. Chapter Introduction
Chapter 3 Linear Programming Linear programs (LP) play an important role in the theory and practice of optimization problems. Many COPs can directly be formulated as LPs. Furthermore, LPs are invaluable
More informationEuropean Journal of Combinatorics
European Journal of Combinatorics 30 (2009) 1686 1695 Contents lists available at ScienceDirect European Journal of Combinatorics ournal homepage: www.elsevier.com/locate/ec Generalizations of Heilbronn
More informationIs submodularity testable?
Is submodularity testable? C. Seshadhri IBM Almaden Research Center 650 Harry Road, San Jose, CA 95120 csesha@us.ibm.com Jan Vondrák IBM Almaden Research Center 650 Harry Road, San Jose, CA 95120 jvondrak@us.ibm.com
More informationChoosing Subsets with Maximum Weighted Average
Choosing Subsets with Maximum Weighted Average David Eppstein Daniel S. Hirschberg Department of Information and Computer Science University of California, Irvine, CA 92717 October 4, 1996 Abstract Given
More informationChapter 11. Approximation Algorithms. Slides by Kevin Wayne Pearson-Addison Wesley. All rights reserved.
Chapter 11 Approximation Algorithms Slides by Kevin Wayne. Copyright @ 2005 Pearson-Addison Wesley. All rights reserved. 1 P and NP P: The family of problems that can be solved quickly in polynomial time.
More informationInterference Minimization in Asymmetric Sensor Networks
Interference Minimization in Asymmetric Sensor Networks Yves Brise 1, Kevin Buchin 2, Dustin Eversmann 3, Michael Hoffmann 1, and Wolfgang Mulzer 3 1 ETH Zürich, Switzerland, hoffmann@inf.ethz.ch 2 TU
More informationOptimal Data-Dependent Hashing for Approximate Near Neighbors
Optimal Data-Dependent Hashing for Approximate Near Neighbors Alexandr Andoni 1 Ilya Razenshteyn 2 1 Simons Institute 2 MIT, CSAIL April 20, 2015 1 / 30 Nearest Neighbor Search (NNS) Let P be an n-point
More informationIntroduction to Integer Programming
Lecture 3/3/2006 p. /27 Introduction to Integer Programming Leo Liberti LIX, École Polytechnique liberti@lix.polytechnique.fr Lecture 3/3/2006 p. 2/27 Contents IP formulations and examples Total unimodularity
More informationCOS597D: Information Theory in Computer Science October 19, Lecture 10
COS597D: Information Theory in Computer Science October 9, 20 Lecture 0 Lecturer: Mark Braverman Scribe: Andrej Risteski Kolmogorov Complexity In the previous lectures, we became acquainted with the concept
More informationLecture 10 February 4, 2013
UBC CPSC 536N: Sparse Approximations Winter 2013 Prof Nick Harvey Lecture 10 February 4, 2013 Scribe: Alexandre Fréchette This lecture is about spanning trees and their polyhedral representation Throughout
More informationGrothendieck s Inequality
Grothendieck s Inequality Leqi Zhu 1 Introduction Let A = (A ij ) R m n be an m n matrix. Then A defines a linear operator between normed spaces (R m, p ) and (R n, q ), for 1 p, q. The (p q)-norm of A
More informationThe Ellipsoid (Kachiyan) Method
Yinyu Ye, MS&E, Stanford MS&E310 Lecture Note: Ellipsoid Method 1 The Ellipsoid (Kachiyan) Method Yinyu Ye Department of Management Science and Engineering Stanford University Stanford, CA 94305, U.S.A.
More informationarxiv: v2 [math.co] 11 Jul 2014
Enumeration of PLCP-orientations of the 4-cube arxiv:1309.7225v2 [math.co] 11 Jul 2014 Lorenz Klaus National Institute of Informatics, Japan and JST, ERATO, Kawarabayashi Large Graph Project, Japan lorenz@nii.ac.jp
More information6.854J / J Advanced Algorithms Fall 2008
MIT OpenCourseWare http://ocw.mit.edu 6.854J / 18.415J Advanced Algorithms Fall 2008 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms. 18.415/6.854 Advanced
More informationTheory and Internet Protocols
Game Lecture 2: Linear Programming and Zero Sum Nash Equilibrium Xiaotie Deng AIMS Lab Department of Computer Science Shanghai Jiaotong University September 26, 2016 1 2 3 4 Standard Form (P) Outline
More informationLinear Programming. Scheduling problems
Linear Programming Scheduling problems Linear programming (LP) ( )., 1, for 0 min 1 1 1 1 1 11 1 1 n i x b x a x a b x a x a x c x c x z i m n mn m n n n n! = + + + + + + = Extreme points x ={x 1,,x n
More informationA Redundant Klee-Minty Construction with All the Redundant Constraints Touching the Feasible Region
A Redundant Klee-Minty Construction with All the Redundant Constraints Touching the Feasible Region Eissa Nematollahi Tamás Terlaky January 5, 2008 Abstract By introducing some redundant Klee-Minty constructions,
More information3.7 Cutting plane methods
3.7 Cutting plane methods Generic ILP problem min{ c t x : x X = {x Z n + : Ax b} } with m n matrix A and n 1 vector b of rationals. According to Meyer s theorem: There exists an ideal formulation: conv(x
More informationEssential Spectra of complete manifolds
Essential Spectra of complete manifolds Zhiqin Lu Analysis, Complex Geometry, and Mathematical Physics: A Conference in Honor of Duong H. Phong May 7, 2013 Zhiqin Lu, Dept. Math, UCI Essential Spectra
More informationEECS 495: Combinatorial Optimization Lecture Manolis, Nima Mechanism Design with Rounding
EECS 495: Combinatorial Optimization Lecture Manolis, Nima Mechanism Design with Rounding Motivation Make a social choice that (approximately) maximizes the social welfare subject to the economic constraints
More informationNew Conjectures in the Geometry of Numbers
New Conjectures in the Geometry of Numbers Daniel Dadush Centrum Wiskunde & Informatica (CWI) Oded Regev New York University Talk Outline 1. A Reverse Minkowski Inequality & its conjectured Strengthening.
More informationInput: System of inequalities or equalities over the reals R. Output: Value for variables that minimizes cost function
Linear programming Input: System of inequalities or equalities over the reals R A linear cost function Output: Value for variables that minimizes cost function Example: Minimize 6x+4y Subject to 3x + 2y
More informationHomework #2 Solutions Due: September 5, for all n N n 3 = n2 (n + 1) 2 4
Do the following exercises from the text: Chapter (Section 3):, 1, 17(a)-(b), 3 Prove that 1 3 + 3 + + n 3 n (n + 1) for all n N Proof The proof is by induction on n For n N, let S(n) be the statement
More informationPOLARS AND DUAL CONES
POLARS AND DUAL CONES VERA ROSHCHINA Abstract. The goal of this note is to remind the basic definitions of convex sets and their polars. For more details see the classic references [1, 2] and [3] for polytopes.
More informationApplied Mathematics &Optimization
Appl Math Optim 29: 211-222 (1994) Applied Mathematics &Optimization c 1994 Springer-Verlag New Yor Inc. An Algorithm for Finding the Chebyshev Center of a Convex Polyhedron 1 N.D.Botin and V.L.Turova-Botina
More informationIs submodularity testable?
Is submodularity testable? C. Seshadhri 1 Jan Vondrák 2 1 Sandia National Labs, Livermore, CA 94551 2 IBM Almaden Research Center, 650 Harry Road, San Jose, CA 95120 scomand@sandia.gov jvondrak@us.ibm.com
More informationCombinatorial Optimization Spring Term 2015 Rico Zenklusen. 2 a = ( 3 2 ) 1 E(a, A) = E(( 3 2 ), ( 4 0
3 2 a = ( 3 2 ) 1 E(a, A) = E(( 3 2 ), ( 4 0 0 1 )) 0 0 1 2 3 4 5 Figure 9: An example of an axis parallel ellipsoid E(a, A) in two dimensions. Notice that the eigenvectors of A correspond to the axes
More information1 Overview. 2 Extreme Points. AM 221: Advanced Optimization Spring 2016
AM 22: Advanced Optimization Spring 206 Prof. Yaron Singer Lecture 7 February 7th Overview In the previous lectures we saw applications of duality to game theory and later to learning theory. In this lecture
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 informationNotation. 0,1,2,, 1 with addition and multiplication modulo
Notation Q,, The set of all natural numbers 1,2,3, The set of all integers The set of all rational numbers The set of all real numbers The group of permutations of distinct symbols 0,1,2,,1 with addition
More informationA Strongly Polynomial Simplex Method for Totally Unimodular LP
A Strongly Polynomial Simplex Method for Totally Unimodular LP Shinji Mizuno July 19, 2014 Abstract Kitahara and Mizuno get new bounds for the number of distinct solutions generated by the simplex method
More informationThis article was published in an Elsevier journal. The attached copy is furnished to the author for non-commercial research and education use, including for instruction at the author s institution, sharing
More informationA Combinatorial Active Set Algorithm for Linear and Quadratic Programming
A Combinatorial Active Set Algorithm for Linear and Quadratic Programming Andrew J Miller University of Wisconsin-Madison 1513 University Avenue, ME 34 Madison, WI, USA 53706-157 ajmiller5@wiscedu Abstract
More informationACO Comprehensive Exam 19 March Graph Theory
1. Graph Theory Let G be a connected simple graph that is not a cycle and is not complete. Prove that there exist distinct non-adjacent vertices u, v V (G) such that the graph obtained from G by deleting
More informationA Randomized Algorithm for Large Scale Support Vector Learning
A Randomized Algorithm for Large Scale Support Vector Learning Krishnan S. Department of Computer Science and Automation Indian Institute of Science Bangalore-12 rishi@csa.iisc.ernet.in Chiranjib Bhattacharyya
More informationA Polynomial-Time Approximation Scheme for Euclidean Steiner Forest
A Polynomial-Time Approximation Scheme for Euclidean Steiner Forest Borradaile, G., Klein, P. N., & Mathieu, C. (25). A polynomial-time approximation scheme for Euclidean Steiner forest. ACM Transactions
More information