Benjamin Doerr Partial Colorings of Unimodular Hypergraphs.

Slides:



Advertisements
Similar presentations
Great Theoretical Ideas in Computer Science
Advertisements

Coloring Warm-Up. A graph is 2-colorable iff it has no odd length cycles 1: If G has an odd-length cycle then G is not 2- colorable Proof: Let v 0, …,
1 Decomposing Hypergraphs with Hypertrees Raphael Yuster University of Haifa - Oranim.
C&O 355 Lecture 23 N. Harvey TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: A A A A A A A A A A.
Walks, Paths and Circuits Walks, Paths and Circuits Sanjay Jain, Lecturer, School of Computing.
C&O 355 Mathematical Programming Fall 2010 Lecture 22 N. Harvey TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: A A A.
Totally Unimodular Matrices
PROCESS MODELLING AND MODEL ANALYSIS © CAPE Centre, The University of Queensland Hungarian Academy of Sciences Analysis of Dynamic Process Models C13.
Shakhar Smorodinsky Courant Institute, New-York University (NYU) On the Chromatic Number of Some Geometric Hypergraphs.
Partial Colorings of Unimodular Hypergraphs Benjamin Doerr (MPI Saarbrücken)
1/17 Optimal Long Test with One Free Bit Nikhil Bansal (IBM) Subhash Khot (NYU)
Every H-decomposition of K n has a nearly resolvable alternative Wilson: e(H) | n(n-1)/2 and gcd(H) | n-1 n>> then there exists an H-decomposition of K.
The Mathematics of Sudoku
1 Discrete Structures & Algorithms Graphs and Trees: II EECE 320.
Symmetric Groups and Ramanujan Graphs Mike Krebs, Cal State LA (joint work with A. Shaheen)
Integer Polytope Sampling and Network Tomography Bertrand Haas Edoardo Airoldi Harvard University 06/21/2010 1Whistler Workshop June 2010.
Online Ramsey Games in Random Graphs Reto Spöhel Joint work with Martin Marciniszyn and Angelika Steger.
Discrete Structures Chapter 7B Graphs Nurul Amelina Nasharuddin Multimedia Department.
. Perfect Phylogeny Tutorial #11 © Ilan Gronau Original slides by Shlomo Moran.
EXPANDER GRAPHS Properties & Applications. Things to cover ! Definitions Properties Combinatorial, Spectral properties Constructions “Explicit” constructions.
Using Homogeneous Weights for Approximating the Partial Cover Problem
MCA 520: Graph Theory Instructor Neelima Gupta
(work appeared in SODA 10’) Yuk Hei Chan (Tom)
9.8 Graph Coloring. Coloring Goal: Pick as few colors as possible so that two adjacent regions never have the same color. See handout.
C&O 355 Mathematical Programming Fall 2010 Lecture 17 N. Harvey TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: AA A.
GRAPH Learning Outcomes Students should be able to:
GROUPS & THEIR REPRESENTATIONS: a card shuffling approach Wayne Lawton Department of Mathematics National University of Singapore S ,
Wei Wang Xi’an Jiaotong University Generalized Spectral Characterization of Graphs: Revisited Shanghai Conference on Algebraic Combinatorics (SCAC), Shanghai,
The Effect of Induced Subgraphs on Quasi-randomness Asaf Shapira & Raphael Yuster.
C&O 355 Mathematical Programming Fall 2010 Lecture 19 N. Harvey TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: AA A.
Perfect Gaussian Elimination and Chordality By Shashank Rao.
Chapter 9. Chapter Summary Relations and Their Properties n-ary Relations and Their Applications (not currently included in overheads) Representing Relations.
Edge-disjoint induced subgraphs with given minimum degree Raphael Yuster 2012.
Chapter 9: Geometric Selection Theorems 11/01/2013
Addressing and hashing Question: For a network of computers it is desirable to be able to send a message from A to B without B knowing that a message is.
Monochromatic Boxes in Colored Grids Joshua Cooper, USC Math Steven Fenner, USC CS Semmy Purewal, College of Charleston Math.
Testing the independence number of hypergraphs
Graph Colouring L09: Oct 10. This Lecture Graph coloring is another important problem in graph theory. It also has many applications, including the famous.
Chapter 1 Fundamental Concepts Introduction to Graph Theory Douglas B. West July 11, 2002.
Graph.
1 Decomposition into bipartite graphs with minimum degree 1. Raphael Yuster.
Chapter 9. Chapter Summary Relations and Their Properties n-ary Relations and Their Applications (not currently included in overheads) Representing Relations.
Latin squares Def: A Latin square of order n is a quadruple (R, C, S; L) where R, C and S are sets of cardinality n and L is a mapping L: R × C → S such.
Artur Czumaj DIMAP DIMAP (Centre for Discrete Maths and it Applications) Computer Science & Department of Computer Science University of Warwick Testing.
. Perfect Phylogeny Tutorial #10 © Ilan Gronau Original slides by Shlomo Moran.
Chapter 1 Systems of Linear Equations Linear Algebra.
Introduction to Graph Theory
3.6 Multiplying Matrices Homework 3-17odd and odd.
Optimization - Lecture 5, Part 1 M. Pawan Kumar Slides available online
C&O 355 Lecture 19 N. Harvey TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: A A A A A A A A A A.
1 GRAPH Learning Outcomes Students should be able to: Explain basic terminology of a graph Identify Euler and Hamiltonian cycle Represent graphs using.
Computational Molecular Biology Pooling Designs – Inhibitor Models.
Fundamental Graph Theory (Lecture 1) Lectured by Hung-Lin Fu 傅 恆 霖 Department of Applied Mathematics National Chiao Tung University.
An Introduction to Graph Theory
Algorithms and Networks
Proof technique (pigeonhole principle)
Anticoloring Problems on Graphs
PC trees and Circular One Arrangements
CSE 421: Introduction to Algorithms
GROUPS & THEIR REPRESENTATIONS: a card shuffling approach
Analysis of Algorithms
On the effect of randomness on planted 3-coloring models
Memoryless Determinacy of Parity Games
Problem Solving 4.
CSE 421: Introduction to Algorithms
Systems of distinct representations
Parameterized Complexity of Even Set (and others)
Chromatic Number of the Odd-Distance Graph
Perfect Phylogeny Tutorial #10
Integer and fractional packing of graph families
Presentation transcript:

Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Overview Partial Colorings of Unimodular Hypergraphs Introduction Coloring hypergraphs (discrepancy) Unimodular hypergraphs Partial coloring Partially coloring unimodular hypergraphs Motivation Result Application Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Hypergraphs Introduction Hypergraph: H = ( V ; E ) V E µ 2 j V = 5 j E : finite set of vertices : set of hyperedges H = ( V ; E ) V E µ 2 V j V = 5 vertices j E = 4 hyperedges Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Hypergraphs Introduction Hypergraph: Induced subhypergraph: H = ( V ; : finite set of vertices : set of hyperedges Induced subhypergraph: H = ( V ; E ) V E µ 2 V H V = ( ; f E \ j 2 g ) ) Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Discrepancy of Hypergraphs Introduction Discrepancy of Hypergraphs Color vertices s.t. all hyperedges are balanced: “2-coloring” “imbalance of hyperedge E”  : V ! f ¡ 1 ; + g  ( E ) : = P v 2 d i s c ( H ;  ) : = m a x E 2 j d i s c ( H ) : = m n  ; +1 -1  ( E ) = 1 + ¡ d i s c ( H ;  ) = j E 1 + 2 d i s c ( H ) = ;  1 Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Unimodular Hypergraphs Introduction Unimodular Hypergraphs Def: unimodular iff each induced subhypergraph has discrepancy at most one. Remark: means even “perfectly balanced” odd “almost perfect”, “1” cannot be avoided H … H = ( [ n ] ; f i : j 1 · g ) d i s c ( H ) · 1 j E )  ( E = j E ) j  ( E = 1 The queen of low-discrepancy hypergraphs! Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Unimodular Hypergraphs: Examples Introduction Unimodular Hypergraphs: Examples Intervals in . Rows/Columns in a grid: Bipartite graphs. [ n ] : = f 1 ; g V = [ m ] £ n E = f i g £ [ n ] j 2 m Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Partial Coloring Introduction Observe: is “caused” by the “odd” vertex in odd-cardinality hyperedges. Plan: Don’t color all vertices! “partial coloring” vertices with are “uncolored” , ... as before Aim: , but doesn’t count! “Nice partial coloring” d i s c ( H ) = 1 +1 -1 ?  : V ! f ¡ 1 ; + g +1 -1  ( v ) = v  ( E ) = P v 2 d i s c ( H ;  ) =  : V ! f g [Beck’s partial coloring method (1981)] Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Existence of Nice Partial Colorings? Partial Colorings of Unimodular Hypergraphs Existence of Nice Partial Colorings? Clearly, not all hypergraphs have nice partial colorings: Complete hypergraphs Projective planes, hypergraphs constructed from Hadamard matrices (proof: the Eigenvalue argument works also for partial colorings) Topic of this talk: Do at least unimodular hypergraphs have nice partial colorings? Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Unimodular hgs with no nice partial coloring Partial Colorings of Unimodular Hypergraphs Unimodular hgs with no nice partial coloring “singletons” “initial intervals” “intervals of length 3 and 5” H = ( [ n ] ; f i g j 2 ) H = ( [ n ] ; f i j 2 g ) H = ( [ n ] ; f i : j ¡ 2 _ 4 g ) No hope for partial coloring?  Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Sometimes it works: Partial Colorings of Unimodular Hypergraphs “length 3 intervals” Rows and columns in the grid. Uniform unimodular hypergraphs: All hyperedges contain the same number of vertices (needs proof). H = ( [ n ] ; f i + 1 2 g j ¡ ) +1 -1 Question: When are there nice partial colorings? Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Result Partial Colorings of Unimodular Hypergraphs The following two properties are equivalent: (i) has a perfectly balanced non-trivial (“nice”) partial coloring; (ii) there are an integer k and non-trivial vertex weights such that all hyperedges have integral weight . H w : V ! f ; 1 = k ( ¡ ) g w ( E ) = P v 2 3/5 Remark: The partial coloring in (i) colors at least half of the vertices with in (ii). Application: “Randomly rounding rationals is as easy as rounding half-integers” [STACS 2007 ] 1/5 w ( u ) 6 = 1/5 2/5 3/5 Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

IF: For all there is a RR such that Application IF: For all there is a RR such that [low rounding errors w.r.t. matrix A] [no rounding error w.r.t. totally unimodular matrix B] THEN: For all rational there is a RR s.t. “Proof”: such that integral Partial coloring: Exists such that such that iff RR of as above, Repeat until . x 2 f ; 1 = g n y 2 f ; 1 g n A x ¼ y B x = y y 2 f ; 1 g n x A x ¼ y B x = y x 2 f ; 1 = k : g n B x  2 f ¡ 1 ; g n B  = ~ x 2 f ; 1 = g n ~ x i = 1 2  i 6 = ~ y ~ x x : = ¡ ( 2 k ) ~ + y x 2 f ; 1 g n Benjamin Doerr Partial Colorings of Unimodular Hypergraphs

Thanks! Summary Partial Colorings of Unimodular Hypergraphs The following two properties are equivalent: (i) has a perfectly balanced non-trivial partial coloring; (ii) there are an integer k and non-trivial vertex weights such that all hyperedges have integral weight . H w : V ! f ; 1 = k ( ¡ ) g w ( E ) = P v 2 3/5 Remark: The partial coloring in (i) colors at least half of the vertices with in (ii). [Open Problem: How many?] Author claims an application. 1/5 w ( u ) 6 = 1/5 2/5 Thanks! 3/5 Benjamin Doerr Partial Colorings of Unimodular Hypergraphs