DOCSLIB.ORG

Decomposition, Approximation, and Coloring of Odd-Minor-Free Graphs

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Decomposition, Approximation, and Coloring of Odd-Minor-Free Graphs Decomposition, approximation, and coloring of odd-minor-free graphs The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters. Citation Demaine, Erik D., Mohammad Taghi Hajiagharyi, Ken-ichi Kawarabayashi. "Decomposition, Approximation, and Coloring of Odd-Minor-Free Graphs" Proceedings of the Twenty-First Annual ACM-SIAM Symposium on Discrete Algorithms, Jan. 2010. 329-344. Copyright © 2010 Society for Industrial and Applied Mathematics. As Published http://www.siam.org/proceedings/soda/2010/ SODA10_028_demainee.pdf Publisher Association for Computing Machinery / Society for Industrial and Applied Mathematics Version Author's final manuscript Citable link http://hdl.handle.net/1721.1/62025 Terms of Use Creative Commons Attribution-Noncommercial-Share Alike 3.0 Detailed Terms http://creativecommons.org/licenses/by-nc-sa/3.0/ Decomposition, Approximation, and Coloring of Odd-Minor-Free Graphs Erik D. Demaine∗ MohammadTaghi Hajiaghayiy Ken-ichi Kawarabayashiz Abstract between the pieces. For example, many optimization prob- We prove two structural decomposition theorems about graphs ex- lems can be solved exactly on graphs of bounded treewidth; cluding a fixed odd minor H, and show how these theorems can what graphs can be partitioned into a small number k of be used to obtain approximation algorithms for several algorithmic bounded-treewidth pieces? In many cases, each piece gives problems in such graphs. Our decomposition results provide new a lower/upper bound on the optimal solution for the entire structural insights into odd-H-minor-free graphs, on the one hand graph, so solving the problem exactly in each piece gives a generalizing the central structural result from Graph Minor The- k-approximation to the problem. Such a decomposition into ory, and on the other hand providing an algorithmic decomposition bounded-treewidth graphs would also be practical, as many into two bounded-treewidth graphs, generalizing a similar result for NP-hard optimization problems are now solved in practice minors. As one example of how these structural results conquer dif- using dynamic programming on low-treewidth graphs; see, ficult problems, we obtain a polynomial-time 2-approximation for e.g., [Bod05, Ami01, Tho98]. Recently, this decomposi- vertex coloring in odd-H-minor-free graphs, improving on the pre- tion approach has been successfully used for graph coloring, 1−" vious O(jV (H)j)-approximation for such graphs and generalizing which is inapproximable within n for any " > 0 unless the previous 2-approximation for H-minor-free graphs. The class P = NP [Zuc07], yet has a 2-approximation in any minor- of odd-H-minor-free graphs is a vast generalization of the well- closed graph family by this approach [DHK05]. Refinements studied H-minor-free graph families and includes, for example, all of this approach have led to PTASs for many graph problems bipartite graphs plus a bounded number of apices. Odd-H-minor- [DHK05, Bak94, Epp00]. free graphs are particularly interesting from a structural graph the- Decomposition of graphs into computationally simpler ory perspective because they break away from the sparsity of H- pieces goes back to the work of Nash-Williams in 1964 minor-free graphs, permitting a quadratic number of edges. [NW64] and of Chartrand et al. [CGH71]. In particular, Chartrand et al. [CGH71] conjectured in 1971 that any 1 Introduction planar graph can have its edges partitioned into two pieces, each inducing an outerplanar graph. This conjecture was Decomposition or partitioning of graphs into smaller pieces only just proved by Goncalves [Gon05], together with a is a fundamental way to design graph algorithms. One linear-time algorithm for computing the decomposition. In of the most famous such decompositions is the divide- particular, this result establishes that any planar graph can be and-conquer separator decomposition for planar graphs of decomposed into two graphs of treewidth 2, a result proved Lipton and Tarjan [LT80], which has been generalized to earlier [Ked96]. Ding et al. [DOSV00] proved that every arbitrary graphs via sparsest cut [LR99, ARV04]. These bounded-genus graph can be decomposed into two graphs of decompositions are based on finding relatively small cuts in bounded treewidth (where the bound depends on the genus). the graph to minimize the interaction between the pieces. To Thomas [Tho95] conjectured that every graph excluding a make the pieces relatively small, the decompositions cut the fixed minor has such a decomposition. This conjecture has graph into many pieces. been proved in two papers [DDO+04, DHK05]; the latter A different kind of decomposition that has received proof is both simpler and algorithmic. A recent extension of more attention recently is to partition the graph into a this result [DHM07] allows the bounded-treewidth pieces to small number of computationally simpler (but not neces- be formed by contractions instead of deletions/subgraphs. sarily small) pieces, without much regard to the interaction In this paper, we generalize the decomposition result to odd-minor-free graphs: every graph excluding a fixed ∗MIT Computer Science and Artificial Intelligence Laboratory, 32 Vas- sar Street, Cambridge, MA 02139, USA. [email protected] odd minor can be decomposed into two bounded-treewidth yAT&T Labs — Research, 180 Park Ave., Florham Park, NJ 07932, graphs. The family of odd-H-minor-free graphs is strictly USA, [email protected]. This work was performed while the more general than H-minor-free graphs for any graph H; for author was at Carnegie Mellon University. example, it includes all complete bipartite graphs K . z n=2;n=2 National Institute for Informatics, 2-1-2 Hitotsubashi, Chiyoda-ku, In interesting contrast, the Lipton-Tarjan separator decom- Tokyo 101-8430, Japan. k [email protected] position (and thus its algorithmic consequences) cannot be ized to odd-H-minor-free graphs. Our results provide the generalized to odd-minor-free graphs, because small sep- first step in this direction by providing new algorithmic arators do not exist. Another contrast between odd-H- structural tools for odd-minor-free graphs. In particular, minor-free graphs and H-minor-free graphs is that the for- the minor-free analogs of our decomposition results have mer can have a quadratic number of edges, while the latter found many applications in just the past few years—see, e.g., are always sparse. Odd-minor-free graphs have been con- [BBCH07, APS07, ASS07, Kle05, DFHT05, DHM]—and sidered extensively in the graph theory literature (see, e.g., we analogously expect our new tools to enable several new [Gue01, GGG+04, Gue05, JT95]) and recently in theoreti- algorithmic results for odd-minor-free graphs. cal computer science [KM06]. Our immediate goal is not to develop practical algo- We prove our result by generalizing a structural result rithms: it is not clear where odd-minor-free graphs would that is the heart of Graph Minor Theory [RS03], which arise in practice, and even most algorithms for H-minor- has many algorithmic applications [Gro03, DFHT05, DH05, free graphs are currently impractical because of huge bounds DHK05], to the case of odd-minor-free graphs. Specifically, from Graph Minors. Rather, the goal of this paper is to im- we prove that every odd-minor-free graph can be decom- prove our algorithmic structural understanding, in particu- posed into a clique-sum of “almost-bipartite” graphs and lar powerful decompositions enabling efficient approxima- graphs that are “almost-embeddable” into bounded-genus tion algorithms, beyond standard graph families. surfaces. (In the original Graph Minor version, only graphs of the second type exist.) Our primary challenge in this result 1.1 Odd-Minor-Free Graphs and Their Importance. is that odd-minor-free graphs can be dense, and this density Recall that a graph H is a minor of G if H can be obtained may be equally spread throughout the graph, so we cannot by contracting and deleting edges in G. Equivalently, H is a hope to find small separations to split the graph into pieces as minor of G precisely if there are jV (H)j vertex-disjoint trees in previous decomposition theorems. Nonetheless, we show in G, one tree Tv for each vertex v of H, such that for every how to decompose odd-minor-free graphs, by showing a con- edge e = fv; wg in H there is an edge e^ in G connecting the nection to the existence of clique minors; this algorithmic two corresponding trees Tv and Tw. Now H is an odd minor technique may be useful for future structural algorithms on of G if, in addition, all the vertices of the trees can be two- dense graphs. colored in such a way that (1) the edges within each tree Tv in All of our results are algorithmic: the decompositions G are bichromatic, while (2) the edge e^ connecting trees Tv can be computed in polynomial time. Our decompositions and Tw in G corresponding to each edge e = fv; wg of H is can also be used to obtain a variety of approximation algo- monochromatic. In particular, the class of odd-H-minor-free rithms for problems on odd-minor-free graphs. In particular, graphs (excluding a fixed graph H as an odd minor) is more our decomposition into two bounded-treewidth graphs (The- general than the class of H-minor-free graphs (excluding a orem 1.1 below) immediately leads to a 2-approximation for fixed graph H as a minor). many NP-complete problems in odd-minor-free graphs. For Indeed, the class of odd-H-minor-free graphs is strictly example, our 2-approximation for graph coloring in odd-Kk- more general: the complete bipartite graph Kn=2;n=2 cer- minor-free graphs improves the recent O(k)-approximation tainly contains a Kk minor for k ≤ n=2, but on the other algorithm [KM06] and generalizes the 2-approximation for hand, it does not contain Kk as an oddp minor for k ≥ 3.
Load more

Recommended publications
  • Prizes and Awards Session
    PRIZES AND AWARDS SESSION Wednesday, July 12, 2021 9:00 AM EDT 2021 SIAM Annual Meeting July 19 – 23, 2021 Held in Virtual Format 1 Table of Contents AWM-SIAM Sonia Kovalevsky Lecture ................................................................................................... 3 George B. Dantzig Prize ............................................................................................................................. 5 George Pólya Prize for Mathematical Exposition .................................................................................... 7 George Pólya Prize in Applied Combinatorics ......................................................................................... 8 I.E. Block Community Lecture .................................................................................................................. 9 John von Neumann Prize ......................................................................................................................... 11 Lagrange Prize in Continuous Optimization .......................................................................................... 13 Ralph E. Kleinman Prize .......................................................................................................................... 15 SIAM Prize for Distinguished Service to the Profession ....................................................................... 17 SIAM Student Paper Prizes ....................................................................................................................
    [Show full text]
  • Decomposition, Approximation, and Coloring of Odd-Minor-Free Graphs
    Decomposition, Approximation, and Coloring of Odd-Minor-Free Graphs Erik D. Demaine∗ MohammadTaghi Hajiaghayiy Ken-ichi Kawarabayashiz Abstract between the pieces. For example, many optimization prob- We prove two structural decomposition theorems about graphs ex- lems can be solved exactly on graphs of bounded treewidth; cluding a fixed odd minor H, and show how these theorems can what graphs can be partitioned into a small number k of be used to obtain approximation algorithms for several algorithmic bounded-treewidth pieces? In many cases, each piece gives problems in such graphs. Our decomposition results provide new a lower/upper bound on the optimal solution for the entire structural insights into odd-H-minor-free graphs, on the one hand graph, so solving the problem exactly in each piece gives a generalizing the central structural result from Graph Minor The- k-approximation to the problem. Such a decomposition into ory, and on the other hand providing an algorithmic decomposition bounded-treewidth graphs would also be practical, as many into two bounded-treewidth graphs, generalizing a similar result for NP-hard optimization problems are now solved in practice minors. As one example of how these structural results conquer dif- using dynamic programming on low-treewidth graphs; see, ficult problems, we obtain a polynomial-time 2-approximation for e.g., [Bod05, Ami01, Tho98]. Recently, this decomposi- vertex coloring in odd-H-minor-free graphs, improving on the pre- tion approach has been successfully used for graph coloring, 1−" vious O(jV (H)j)-approximation for such graphs and generalizing which is inapproximable within n for any " > 0 unless the previous 2-approximation for H-minor-free graphs.
    [Show full text]
  • SPECIAL ISSUE in HONOR of GEOFF WHITTLE Geoff Whittle Turned 60 in 2010. to Mark
    SPECIAL ISSUE IN HONOR OF GEOFF WHITTLE DILLON MAYHEW, JAMES OXLEY, AND CHARLES SEMPLE Geoff Whittle turned 60 in 2010. To mark this occasion, this special issue (guest- edited by Dillon Mayhew and Charles Semple) aims to recognize and celebrate his mathematical contributions. Perhaps surprisingly, we can confidently say that Ge- off’s best work lies ahead of him, for some of the most significant results from his longstanding collaboration with Jim Geelen and Bert Gerards have been announced, but not yet published. This collaboration has produced matroid analogues of the well-quasi-ordering results from Robertson and Seymour’s monumental Graph Mi- nors Project. In addition, it is anticipated that Rota’s conjecture, a problem long considered to be one of the most difficult in discrete mathematics, will also be set- tled by Jim, Bert, and Geoff. It is already clear that their work deserves to be ranked alongside that of W. T. Tutte and Paul Seymour as the most significant ever done in matroid theory. In this preface, we will introduce matroids after giving some biographical details of Geoff’s life. Our main task is to give a brief description of Geoff’s contributions to the development of matroid theory. 1. Brief Biography Geoff was born in 1950 in Launceston, a city in northern Tasmania, Australia. He completed high school in Launceston in 1968, and, in the following year, he worked as a miner in Tasmania, Western Australia, and the Northern Territory before leaving to travel through Asia and Europe. In 1971, he enrolled at the University of Tasmania, graduating in 1973 with a bachelor’s degree in Philosophy and Mathematics.
    [Show full text]
  • An Introduction to the Discharging Method Via Graph Coloring
    An Introduction to the Discharging Method via Graph Coloring Daniel W. Cranston∗ Douglas B. West† October 10, 2016 Abstract We provide a “how-to” guide to the use and application of the Discharging Method. Our aim is not to exhaustively survey results proved by this technique, but rather to demystify the technique and facilitate its wider use, using applications in graph coloring as examples. Along the way, we present some new proofs and new problems. 1 Introduction Arguments that can be phrased in the language of the Discharging Method have been used in graph theory for more than 100 years, though that name is much more recent. The most famous application of the method is the proof of the Four Color Theorem, stating that graphs embeddable in the plane have chromatic number at most 4. However, the method remains mysterious to many. Our aim is to explain its use and make the method more widely accessible. Although we mention many applications, including stronger versions of results proved here, cataloguing applications is not our goal. Borodin [23] presents a survey of applications of discharging to coloring of plane graphs. Discharging is most commonly used as a tool in a two-pronged approach to inductive proofs, typically for sparse graphs. In this context, it is used to prove that a global sparse- ness hypothesis guarantees the existence of some desired local structure. The method has been applied to many types of problems (including graph embeddings and decompositions, spread of infections in networks, geometric problems, etc.). Nevertheless, we present only applications in graph coloring (where it has been used most often), in order to emphasize the discharging techniques.
    [Show full text]
  • Professor Shang-Hua Teng
    Professor Shang-Hua Teng Department of Computer Science, University of Southern California 1039 25 Street 3710 McClintock, RTH 505, Los Angeles, CA 90089 Santa Monica, CA 90403 (213) 740-3667, [email protected] (617) 440-4281 EMPLOYMENT: University Professor USC (2017 – ) Seeley G. Mudd Professor (Computer Science and Mathematics) USC (2012 – ) Chair and Seeley G. Mudd Professor (Computer Science) USC (2009 – 2012) Professor (Computer Science) Boston University (2002 – 2009) Research Affiliate Professor (Mathematics) MIT (1999 – Present) Visiting Research Scientist Microsoft Research Asia and New England (2004 – Present) Senior Research Scientist Akamai (1997 – 2009) Visiting Professor (Computer Science) Tsinghua University (2004 – 2008) Professor (Computer Science)[Associate Professor before 1999] UIUC (2000 – 2002) Research Scientist IBM Almaden Research Center (1997 – 1999) Assistant Professor (Computer Science) Univ. of Minnesota (1994 – 1997) Research Scientist Intel Corporation (Summer 1994, 1995) Instructor (Mathematics) MIT (1992 – 1994) Computer Scientist NASA Ames Research Center (Summer 1993) Research Scientist Xerox Palo Alto Research Center (1991 – 1992) EDUCATION: Ph.D. (Computer Science) Carnegie Mellon University (1991) Thesis: “A Unified Geometric Approach to Graph Partitioning.” Advisor: Gary Lee Miller M.S. (Computer Science) University of Southern California (1988) Thesis: “Security, Verifiablity, and Universality in Distributed Computing”. Advised by Leonard Adleman and Ming-Deh Huang B.S. (Computer Science) & B.A. (Electrical
    [Show full text]
  • Call for Nominations
    CORE Metadata, citation and similar papers at core.ac.uk Provided by Elsevier - Publisher Connector D.R. FULKERSON PRIZE Call for nominations The Fulkerson Prize Committee invites nominations for the Delbert Ray Fulkerson Prize, sponsored jointly by the Mathematical Programming Society and the American Mathematical Society. The Fulkerson Prize is for outstanding papers in the area of discrete mathematics. The Prize will be awarded at the XVIIIth International Symposium on Mathematical Programming to be held in Copenhagen, Denmark, August 18–22, 2003. Eligible papers should represent the final publication of the main result(s) and should have been published in a recognized journal, or in a comparable, well- refereed volume intended to publish final publications only, during the six calendar years preceding the year of the Symposium (thus, from January 1997 through December 2002). The prizes will be given for single papers, not series of papers or books, and in the event of joint authorship the prize will be divided. The term ‘discrete mathematics’ is interpreted broadly and is intended to include graph theory, networks, mathematical programming, applied combinatorics, applications of discrete mathematics to computer science, and related subjects. While research work in these areas is usually not far removed from practical applications, the judging of papers will only be based on their mathematical quality and significance. Previous winners of the Fulkerson Prize are listed below. Further information about the Fulkerson Prize can be found at www.mathprog.org/prz/fulkerson.htm. The Fulkerson Prize Committee consists of Gerard Cornuejols (Carnegie-Mellon University), Andrew M. Odlyzko (University of Minnesota), and David P.
    [Show full text]
  • Computational Complexity; Slides 7, HT 2019 Logarithmic Space
    Computational Complexity; slides 7, HT 2019 Logarithmic space Prof. Paul W. Goldberg (Dept. of Computer Science, University of Oxford) HT 2019 Paul Goldberg Logarithmic space 1 / 37 Logarithmic Space Polynomial space: seems more powerful than NP. Linear space: we noted is similar to polynomial space Sub-linear space? To be meaningful, we consider Turing machines with separate input tape and only count working space. LOGSPACE (or,L) Problems solvable by logarithmic space bounded TM NLOGSPACE (or,NL) Problems solvable by logarithmic space bounded NTM Not hard to show that L⊆NL⊆P (Sipser Chapter 8.4, Arora/Barak, p.80) Paul Goldberg Logarithmic space 2 / 37 Hence, LOGSPACE contains all problems requiring only a constant number of counters/pointers for solving. NLOGSPACE contains all problems requiring only a constant number of counters/pointers for verifying solutions. Problems in L and NL What sort of problems are in L and NL? In logarithmic space we can store a fixed number of counters (up to length of input) a fixed number of pointers to positions in the input string. Paul Goldberg Logarithmic space 3 / 37 Problems in L and NL What sort of problems are in L and NL? In logarithmic space we can store a fixed number of counters (up to length of input) a fixed number of pointers to positions in the input string. Hence, LOGSPACE contains all problems requiring only a constant number of counters/pointers for solving. NLOGSPACE contains all problems requiring only a constant number of counters/pointers for verifying solutions. Paul Goldberg Logarithmic space 3 / 37 Example.
    [Show full text]
  • On the Ellipsoid Method for Systems of Linear Inequalities
    Available online a t www.pelagiaresearchlibrary.com Pelagia Research Library Advances in Applied Science Research, 2012, 3 (5):3354-3361 ISSN: 0976-8610 CODEN (USA): AASRFC On the Ellipsoid Method for Systems of Linear Inequalities B. O. Adejo 1 and A. M. Okutachi 2 Department of Mathematical Sciences, Kogi State University, Anyigba, Nigeria _____________________________________________________________________________________________ ABSTRACT In this article, we give an overview of the basic ellipsoid method, its antecedents and modifications made to improve its rate of convergence. Furthermore, its importance is highlighted. Keywords: Polynomial time, feasible solution, perturbed system _____________________________________________________________________________________________ INTRODUCTION In 1947, G. B. Dantzig formulated the Linear Programming (LP) problem and provided the simplex method to solve it. The simplex method is still the widely used method to solve LP problems, although polynomial-time algorithms have emerged that work on the relative interior of the feasible region in their search for solutions to the LP problem. The simplex method was made available in 1951. It is an iterative ( i.e algebraic) procedure which either solves a LP problem in a finite number of steps or gives an indication that there is an unbounded solution or infeasible solution to the problem under consideration. The simplex method can be viewed as an exterior point method since it ‘crawls’ along the edges (vertices) of the feasible region in its search for solution to the LP problem. In 1979, Khachian L. G [11] resolved an open question of whether linear programming problems belonged to the P- class (i. e. class of problems solvable in polynomial time) or to the NP-class (i.
    [Show full text]
  • Algorithmic Graph Minors: Turning Combinatorics to Algorithms Page 1/1111 Antichain: an Infinite Sequence on Non- ≤-Comparable Elements
    Graph Minors and algorithms The irrelevant vertex technique Algorithmic Graph Minors: turning Combinatorics to Algorithms Dimitrios M. Thilikos Department of Mathematics { µQλ8 UoA - National and Kapodistrian University of Athens Laboratoire d'Informatique, de Robotique et de Micro´electroniquede Montpellier (LIRMM) Montpellier, February 2, 2012 Dimitrios M. Thilikos Department of Mathematics, UoA Algorithmic Graph Minors: turning Combinatorics to Algorithms Page 1/1111 Antichain: an infinite sequence on non- ≤-comparable elements. We say that X is Well-Quasi-Ordered by ≤ if it has no infinite antichain Examples: I 2N is not W.Q.O. by set inclusion. I N is not W.Q.O. by divisibility. k I N is W.Q.O. by by component-wise ordering. Graph Minors and algorithms The irrelevant vertex technique Well Quasi Ordering Theory Let X be a set and let \≤" be a partial ordering relation on X . Dimitrios M. Thilikos Department of Mathematics, UoA Algorithmic Graph Minors: turning Combinatorics to Algorithms Page 2/1112 We say that X is Well-Quasi-Ordered by ≤ if it has no infinite antichain Examples: I 2N is not W.Q.O. by set inclusion. I N is not W.Q.O. by divisibility. k I N is W.Q.O. by by component-wise ordering. Graph Minors and algorithms The irrelevant vertex technique Well Quasi Ordering Theory Let X be a set and let \≤" be a partial ordering relation on X . Antichain: an infinite sequence on non- ≤-comparable elements. Dimitrios M. Thilikos Department of Mathematics, UoA Algorithmic Graph Minors: turning Combinatorics to Algorithms Page 3/1112 Examples: I 2N is not W.Q.O.
    [Show full text]
  • Arxiv:2105.07884V3 [Math.HO] 20 Jun 2021
    Enumerative and Algebraic Combinatorics in the 1960's and 1970's Richard P. Stanley University of Miami (version of 17 June 2021) The period 1960{1979 was an exciting time for enumerative and alge- braic combinatorics (EAC). During this period EAC was transformed into an independent subject which is even stronger and more active today. I will not attempt a comprehensive analysis of the development of EAC but rather focus on persons and topics that were relevant to my own career. Thus the discussion will be partly autobiographical. There were certainly deep and important results in EAC before 1960. Work related to tree enumeration (including the Matrix-Tree theorem), parti- tions of integers (in particular, the Rogers-Ramanujan identities), the Redfield- P´olya theory of enumeration under group action, and especially the repre- sentation theory of the symmetric group, GL(n; C) and some related groups, featuring work by Georg Frobenius (1849{1917), Alfred Young (1873{1940), and Issai Schur (1875{1941), are some highlights. Much of this work was not concerned with combinatorics per se; rather, combinatorics was the nat- ural context for its development. For readers interested in the development of EAC, as well as combinatorics in general, prior to 1960, see Biggs [14], Knuth [77, §7.2.1.7], Stein [147], and Wilson and Watkins [153]. Before 1960 there are just a handful of mathematicians who did a sub- stantial amount of enumerative combinatorics. The most important and influential of these is Percy Alexander MacMahon (1854-1929). He was a arXiv:2105.07884v3 [math.HO] 20 Jun 2021 highly original pioneer, whose work was not properly appreciated during his lifetime except for his contributions to invariant theory and integer parti- tions.
    [Show full text]
  • Front Matter
    Cambridge University Press 978-0-521-19527-0 - The Design of Approximation Algorithms David P. Williamson and David B. Shmoys Frontmatter More information The Design of Approximation Algorithms Discrete optimization problems are everywhere, from traditional operations research plan- ning problems, such as scheduling, facility location, and network design; to computer science problems in databases; to advertising issues in viral marketing. Yet most such prob- lems are NP-hard. Thus unless P = NP, there are no efficient algorithms to find optimal solutions to such problems. This book shows how to design approximation algorithms: efficient algorithms that find provably near-optimal solutions. The book is organized around central algorithmic techniques for designing approxi- mation algorithms, including greedy and local search algorithms, dynamic programming, linear and semidefinite programming, and randomization. Each chapter in the first part of the book is devoted to a single algorithmic technique, which is then applied to several different problems. The second part revisits the techniques but offers more sophisticated treatments of them. The book also covers methods for proving that optimization problems are hard to approximate. Designed as a textbook for graduate-level algorithms courses, the book will also serve as a reference for researchers interested in the heuristic solution of discrete optimization problems. David P. Williamson is a Professor at Cornell University with a joint appointment in the School of Operations Research and Information Engineering and in the Department of Information Science. Prior to joining Cornell, he was a Research Staff Member at the IBM T.J. Watson Research Center and a Senior Manager at the IBM Almaden Research Center.
    [Show full text]
  • A Guide to the Discharging Method
    A Guide to the Discharging Method Daniel W. Cranston∗ Douglas B. West† July 23, 2013 Abstract We provide a “how-to” guide to the use and application of the Discharging Method. Our aim is not to exhaustively survey results that have been proved by this technique, but rather to demystify the technique and facilitate its wider use. Along the way, we present some new proofs and new problems. 1 Introduction The Discharging Method has been used in graph theory for more than 100 years. Its most famous application is the proof of the Four Color Theorem, stating that graphs embeddable in the plane have chromatic number at most 4. Nevetheless, the method remains a mystery to many graph theorists. Our purpose in this guide is to explain its use in order to make the method more widely accessible. Although we mention many applications, such as stronger versions of results proved here, cataloguing applications is not our aim. For a survey of applications in coloring of plane graphs we refer the reader to Borodin [45]. Discharging is a tool in a two-pronged approach to inductive proofs. It can be viewed as an amortized counting argument used to prove that a global hypothesis guarantees the existence of some desirable local configuration. In an application of the resulting structure theorem, one shows that each such local configuration cannot occur in a minimal counterexample to the desired conclusion. More precisely, a configuration is reducible for a graph property if it cannot occur in a minimal graph failing that property. This leads to the phrase “an unavoidable set of reducible configurations” to describe the overall process.
    [Show full text]