Ringing the Changes II

Similar documents
Campanology - Ringing the changes

arxiv: v2 [math.gr] 17 Dec 2017

Planar Graphs (1) Planar Graphs (3) Planar Graphs (2) Planar Graphs (4)

UNIQUENESS OF HIGHLY REPRESENTATIVE SURFACE EMBEDDINGS

ON HAMILTON CIRCUITS IN CAYLEY DIGRAPHS OVER GENERALIZED DIHEDRAL GROUPS

HAMILTONICITY IN CAYLEY GRAPHS AND DIGRAPHS OF FINITE ABELIAN GROUPS.

Hamiltonicity of digraphs for universal cycles of permutations

Three-coloring triangle-free graphs on surfaces VII. A linear-time algorithm

TOWARDS A SPLITTER THEOREM FOR INTERNALLY 4-CONNECTED BINARY MATROIDS VII

ARRANGEABILITY AND CLIQUE SUBDIVISIONS. Department of Mathematics and Computer Science Emory University Atlanta, GA and

Fine structure of 4-critical triangle-free graphs III. General surfaces

Dirac s Map-Color Theorem for Choosability

arxiv: v1 [cs.ds] 2 Oct 2018

8.5 Sequencing Problems

Imprimitive symmetric graphs with cyclic blocks

arxiv: v1 [cs.dm] 12 Jun 2016

Research Article Nonnormal Edge-Transitive Cubic Cayley Graphs of Dihedral Groups

Fine Structure of 4-Critical Triangle-Free Graphs II. Planar Triangle-Free Graphs with Two Precolored 4-Cycles

Hamiltonian circuits in Cayley digraphs. Dan Isaksen. Wayne State University

A well-quasi-order for tournaments

CS 301: Complexity of Algorithms (Term I 2008) Alex Tiskin Harald Räcke. Hamiltonian Cycle. 8.5 Sequencing Problems. Directed Hamiltonian Cycle

Paul Erdős and Graph Ramsey Theory

Odd Components of Co-Trees and Graph Embeddings 1

Laplacian Integral Graphs with Maximum Degree 3

arxiv: v2 [math.co] 7 Jan 2016

Polynomial-time Reductions

Massachusetts Institute of Technology 6.042J/18.062J, Fall 02: Mathematics for Computer Science Professor Albert Meyer and Dr.

1 Spaces and operations Continuity and metric spaces Topological spaces Compactness... 3

Hamiltonian Graphs Graphs

Some results on the reduced power graph of a group

Graphs and digraphs with all 2 factors isomorphic

Planar Ramsey Numbers for Small Graphs

The power graph of a finite group, II

Three-coloring triangle-free graphs on surfaces III. Graphs of girth five

HAMILTON CYCLES IN CAYLEY GRAPHS

We would like a theorem that says A graph G is hamiltonian if and only if G has property Q, where Q can be checked in polynomial time.

Three-coloring triangle-free graphs on surfaces IV. Bounding face sizes of 4-critical graphs

Chapter 34: NP-Completeness

3-coloring triangle-free planar graphs with a precolored 8-cycle

Selected Topics in AGT Lecture 4 Introduction to Schur Rings

A Family of One-regular Graphs of Valency 4

Graphs with large maximum degree containing no odd cycles of a given length

Applications of the Lopsided Lovász Local Lemma Regarding Hypergraphs

P versus NP. Math 40210, Fall November 10, Math (Fall 2015) P versus NP November 10, / 9

Physical Mapping. Restriction Mapping. Lecture 12. A physical map of a DNA tells the location of precisely defined sequences along the molecule.

6 Euler Circuits and Hamiltonian Cycles

CSCE 551 Final Exam, April 28, 2016 Answer Key

8. INTRACTABILITY I. Lecture slides by Kevin Wayne Copyright 2005 Pearson-Addison Wesley. Last updated on 2/6/18 2:16 AM

On non-hamiltonian circulant digraphs of outdegree three

5. Partitions and Relations Ch.22 of PJE.

Sergey Norin Department of Mathematics and Statistics McGill University Montreal, Quebec H3A 2K6, Canada. and

Constructing Linkages for Drawing Plane Curves

Basic definitions. Remarks

4/12/2011. Chapter 8. NP and Computational Intractability. Directed Hamiltonian Cycle. Traveling Salesman Problem. Directed Hamiltonian Cycle

P P P NP-Hard: L is NP-hard if for all L NP, L L. Thus, if we could solve L in polynomial. Cook's Theorem and Reductions

UNAVOIDABLE INDUCED SUBGRAPHS IN LARGE GRAPHS WITH NO HOMOGENEOUS SETS

Parity Versions of 2-Connectedness

Complete graphs whose topological symmetry groups are polyhedral

Data Structures and Algorithms

SUB-EXPONENTIALLY MANY 3-COLORINGS OF TRIANGLE-FREE PLANAR GRAPHS

Contact author address Dragan Marusic 3 IMFM, Oddelek za matematiko Univerza v Ljubljani Jadranska 19, 1111 Ljubljana Slovenija Tel.: F

P versus NP. Math 40210, Spring September 16, Math (Spring 2012) P versus NP September 16, / 9

The Manhattan Product of Digraphs

4 CONNECTED PROJECTIVE-PLANAR GRAPHS ARE HAMILTONIAN. Robin Thomas* Xingxing Yu**

Latin squares: Equivalents and equivalence

Chapter 8. NP and Computational Intractability. Slides by Kevin Wayne. Copyright 2005 Pearson-Addison Wesley. All rights reserved.

1 More finite deterministic automata

Algorithms and Theory of Computation. Lecture 22: NP-Completeness (2)

4. How to prove a problem is NPC

Combining the cycle index and the Tutte polynomial?

Claw-free Graphs. III. Sparse decomposition

MATH 22 HAMILTONIAN GRAPHS. Lecture V: 11/18/2003

On cyclic decompositions of K n+1,n+1 I into a 2-regular graph with at most 2 components

Malaya J. Mat. 2(3)(2014)

Ordering and Reordering: Using Heffter Arrays to Biembed Complete Graphs

arxiv: v1 [cs.ds] 20 Feb 2017

Fine structure of 4-critical triangle-free graphs III. General surfaces

arxiv: v3 [cs.dm] 18 Oct 2017

NP and Computational Intractability

arxiv:quant-ph/ v1 15 Apr 2005

Distance between two k-sets and Path-Systems Extendibility

Theory of Computation Chapter 9

A dynamic programming approach to generation of strong traces

Paths and cycles in extended and decomposable digraphs

How many randomly colored edges make a randomly colored dense graph rainbow hamiltonian or rainbow connected?

Folding graphs and applications, d après Stallings

5 Flows and cuts in digraphs

McGill University Faculty of Science. Solutions to Practice Final Examination Math 240 Discrete Structures 1. Time: 3 hours Marked out of 60

For a group G and a subgroup H G, denoteby[g:h] the set of right cosets of H in G, thatis

Compatible Circuit Decompositions of 4-Regular Graphs

8.5 Sequencing Problems. Chapter 8. NP and Computational Intractability. Hamiltonian Cycle. Hamiltonian Cycle

Hamilton Cycles in Digraphs of Unitary Matrices

Fixed Parameter Algorithms for Interval Vertex Deletion and Interval Completion Problems

Induced subgraphs of graphs with large chromatic number. IX. Rainbow paths

Nonnegative Matrices I

Automorphism groups of wreath product digraphs

Hamiltonian paths in tournaments A generalization of sorting DM19 notes fall 2006

The Inclusion Exclusion Principle

Spanning Paths in Infinite Planar Graphs

Simplification by Truth Table and without Truth Table

Transcription:

Ringing the Changes II

Cayley Color Graphs Let G be a group and H a subset of G. We define a Cayley digraph (directed graph) by: Each element of G is a vertex of the digraph Γ and an arc is drawn from vertex a to vertex b if there is an element h H so that b = ha in G. If H = H -1 (the inverse of every element in H is also in H), then for every arc from a to b there is also an arc from b to a. In this situation it is common to replace the pair of arcs by an undirected edge and the Cayley digraph is called a Cayley graph. When each element of H is its own inverse, we may label the edges of the Cayley graph by the elements of H which determine the edge. This labeled graph is called a Cayley color graph.

Cayley Color Graphs Let G be the symmetric group S 3 whose elements are (written in cyclic notation): G = {I, (12), (13), (23), (123), (132)}. Let H = {(12), (132)}. The Cayley digraph C G (H) is: I (12) (132) (13) = (132)(12) (23) = (12)(132) (123)

Cayley Color Graphs Let H' = {(123), (132)}. The Cayley graph C G (H') is: I (12) (132) (13) (23) (123)

Cayley Color Graphs Let H'' = {(12), (13)}. The Cayley color graph C G (H'') is: I (12) (123) (23) (12) (13) (13) (132)

Ringing the Changes We can use Cayley color graphs to model change ringing. For n bells, we will use the symmetric group S n which consists of all the permutations on n symbols. The transitions (permutations which either fix or interchange adjacent bells) form the set H. Since each transition is an involution (has order 2), we can form the Cayley color graph C Sn (H). To obtain an extent on n bells which satisfies rules 1, 2 and 3 we need only find a Hamiltonian cycle (which visits each vertex exactly once before returning to the starting vertex) in this color graph and starting at the identity vertex, trace the cycle using the edge labels to obtain each change.

Hamiltonian Cycles As an example consider an extent of 3-bells (singles). The Cayley color graph is on S 3 and H = {(12),(23)}. I (12) (123) (23) (12) = a (23) = b (13) (132) Hamiltonian: ababab = (ab) 3 quick six bababa = (ba) 3 slow six Quick Six 123 = I 213 = (12) 231 = (123) 321 = (13) 312 = (132) 132 = (23) 123 = I

Another Example Consider the Cayley color graph of S 4 with H = {(12), (23), (34)}. Starting with (12) this is Double Court Minimus.

Another Example Consider the Cayley color graph of S 4 with H = {(12), (23), (34)}. Starting at JT gives the Johnson-Trotter sequence. JT Starting with (12) this is Double Canterbury Minimus.

Finding Hamiltonian Cycles Graphs containing a Hamiltonian cycle are called Hamiltonian graphs. Deciding whether or not a graph is Hamiltonian is an NP-problem, meaning that there is no known fast algorithm that will answer this question for all graphs. There are some theorems which give sufficient (but not necessary) conditions for a graph to be Hamiltonian (Ore, Dirac, Chvátal) but none of these are applicable in our situation. For small graphs we can use a basic backtracking algorithm to find Hamiltonian cycles, but in our case there is a theoretical result which is useful.

Rapaport's Construction In Cayley colour groups and Hamilton lines [Scripta Mathematica 24(1959), pp.51-58], Elvira Strasser Rapaport proved that the Cayley color graphs we are interested in are in fact Hamiltonian. Her proof is based on the following lemma: Lemma: A connected graph, regular of degree three, has a Hamiltonian cycle if there exists a set P of cycles and a set Q of 4-cycles each partitioning the vertex set of the graph and such that no member of P contains every vertex of a member of Q.

Rapaport's Lemma Lemma: A connected graph, regular of degree three, has a Hamiltonian cycle if there exists a set P of cycles and a set Q of 4-cycles each partitioning the vertex set of the graph and such that no member of P contains every vertex of a member of Q. Proof: Repeat this operation until it can't be done again: P 1 Q 1 P 2 P 1, P 2 P, Q 1 Q P has become P', still a disjoint union of cycles covering the vertex set. If P' is not a single cycle, it contains two cycles C 1 and C 2 joined by an edge E = v 1 v 2, v 1 C 1, v 2 C 2 (since the graph is connected).

Rapaport's Lemma Lemma: A connected graph, regular of degree three, has a Hamiltonian cycle if there exists a set P of cycles and a set Q of 4-cycles each partitioning the vertex set of the graph and such that no member of P contains every vertex of a member of Q. Proof (cont.): If E is in some Q i, then the two other edges through v 1 must be in C 1 and one of them, say v 1 p 1 is in Q i. Similarly, an edge v 2 p 2 in C 2 is also in Q i. We could thus carry out the above operation again, a contradiction. Thus, E is in no Q i. The edges v 1 p 1 and v 1 p 3 are in C 1. As v 1 is in some Q 1, these two edges must also be in Q 1 (as E is not and valency is 3). C 1 thus contains 3 vertices and 2 consecutive edges of Q 1. The consecutive edges could not have arisen from the operation, so these edges must have been in some P i to start with. The 4 th vertex of Q 1 could not be in this P i by hypothesis.

Rapaport's Lemma Lemma: A connected graph, regular of degree three, has a Hamiltonian cycle if there exists a set P of cycles and a set Q of 4-cycles each partitioning the vertex set of the graph and such that no member of P contains every vertex of a member of Q. Proof (cont.): This 4 th vertex of Q 1, say V, can not be on any of the P j in P. If it were, it would have to be connected to two vertices of its cycle P j and also two vertices of cycle P i, but valency 3 would then force these two cycles to have a common vertex, a contradiction.

Rapaport's Theorem Theorem: For n 3, the Cayley color graph C sn (H) is Hamiltonian when H = {g 1 =(12), g 2 =(12)(34)(56)..., g 3 = (23)(45)(67)... }. Pf: The permutations (12) and (1 3 5 7...) {= g 2 g 3 } are known to generate S n, so the three permutations of H certainly do. This means that the Cayley color graph is connected. For n > 4, start at any vertex and follow the path g 1 g 2 g 1 g 2. Since this product is the identity transformation, the path determines a 4-cycle. If two of the 4-cycles generated this way had a vertex in common, they would have to be identical. Let Q be the set of all such 4- cycles. There are n!/4 such 4-cycles in Q.

Rapaport's Theorem Theorem: For n 3, the Cayley color graph C sn (H) is Hamiltonian when H = {g 1 =(12), g 2 =(12)(34)(56)..., g 3 = (23)(45)(67)... }. Pf(cont): The set P consists of the 12-cycles that are formed from the relation (g 1 g 3 ) 6 = id. There are n!/12 disjoint 12-cycles in P. As no product of alternating g 3 and g 1 can equal g 2 (consider what happens to 4 and 5) none of these 12-cycles can contain all (or even 3) vertices of one of Q's 4-cycles. Thus, the conditions of the lemma are satisfied and a Hamiltonian cycle exists in this graph. If n = 3, H = {g 1 = g 2 = (12), g 3 = (23)} and the quick six singles gives the result. If n = 4, replace the 12- cycles of P by the 6-cycles generated by (g 1 g 3 ) 3 =id.

Example when n = 4 α I β (12) (12)(34) (23) η λ η λ β α

Other Choices Although Rapaport did not use them, P could also be made up of 2n-cycles which come from (g 2 g 3 ) n = id. There are n!/2n = ½(n-1)! of these 2n-cycles. For n = 4 we would have 3 8-cycles forming P, and the algorithm of the lemma will still work.

Example when n = 4 α I β (12) (12)(34) (23) η λ η λ β α

Other Choices However, as the following examples show, not all Hamiltonian cycles come from this construction of Rapaport.

Example when n = 4 α I β (12) (12)(34) (23) η λ η λ β α

Example when n = 4 α I β (12) (12)(34) (23) η λ η λ Single Court starting with (12) β α

Example when n = 4 α I β (12) (12)(34) (23) η λ η λ β α Reverse Bob starts with (12)(34)

Euler Characteristic The Cayley graphs we have examined so far are not always planar graphs. We can embed a graph that we are interested in in a planar way (no edges crossing except at vertices) on a surface with a certain number of holes. If a connected graph with n vertices and e edges has a planar embedding in a plane or on a sphere then n e + f = 2 where f is the number of faces (minimal [having no chords] cycles) of the graph. Example: n = 4, e = 6, f = 4 (don't forget the outside face) 4 6 + 4 = 2

Euler Characteristic The quantity n e + f is called the Euler characteristic of the graph and is related to the type of surface in which the graph may have a planar embedding. The result here is: Theorem: Suppose that a graph with n vertices, e edges and f faces has a simple imbedding in a closed surface S, then (1) If S is a sphere with g handles then n-e+f = 2 2g (2) If S is a sphere with c cross-caps then n-e+f = 2 - c. A simple imbedding is one in which each face (including the infinite face) is topologically a disk. The surfaces of type (1) are orientable while those of type (2) are nonorientable.

Embedding Graphs Sphere with 1 handle = Torus (Doughnut)

Embedding Graphs Torus Möbius Strip

Projective Plane Embedding Graphs Cross Cap

Embedding Rapaport's Graph The Rapaport graph has n! vertices, 3n!/2 edges and its faces are the n!/4 4-cycles, n!/12 12-cycles and ½(n-1)! 2n-cycles for n > 4. Thus the Euler characteristic is n! - 3n!/2 + (n!/4 + n!/12 + n!/2n) = n! ( 1 3/2 + 1/4 + 1/12 + 1/2n) = n!(-1/6 + 1/2n) n! (3 n)/6n = (n-1)! (3 n)/6 = 2 [(n-1)(n-2)(n-3) 2 (n-4)! + 12]/6. For n = 4 we have 24 vertices, 36 edges, 6 + 4 + 3 faces, so n e + f = 24 36 + 13 = 1 = 2 1 and so, it may be embedded in a projective plane.