Groups, Graphs & Isospectrality Rami Band Ori Parzanchevski Gilad Ben-Shach Uzy Smilansky.

Slides:



Advertisements
Similar presentations
Group Theory II.
Advertisements

Great Theoretical Ideas in Computer Science for Some.
Adam Sawicki, Rami Band, Uzy Smilansky Scattering from isospectral graphs.
Chapter 8 Topics in Graph Theory
10 October 2006 Foundations of Logic and Constraint Programming 1 Unification ­An overview Need for Unification Ranked alfabeths and terms. Substitutions.
General Linear Model With correlated error terms  =  2 V ≠  2 I.
An Ω(n 1/3 ) Lower Bound for Bilinear Group Based Private Information Retrieval Alexander Razborov Sergey Yekhanin.
Graph-02.
Bangalore conference, December, Rank 3-4 Coxeter Groups, Quaternions and Quasicrystals Mehmet Koca Department of Physics College of Science.
Recursive Definitions and Structural Induction
Symmetric Group Sym(n) As we know a permutation  is a bijective mapping of a set A onto itself:  : A  A. Permutations may be multiplied and form the.
What one cannot hear? On drums which sound the same
Complexity 16-1 Complexity Andrei Bulatov Non-Approximability.
1 2 Introduction In last chapter we saw a few consistency tests. In this chapter we are going to prove the properties of Plane-vs.- Plane test: Thm[RaSa]:
Tracing periodic orbits and cycles in the nodal count Amit Aronovitch, Ram Band, Yehonatan Elon, Idan Oren, Uzy Smilansky xz ( stands for the integral.
Computational Geometry Seminar Lecture 1
Validating Streaming XML Documents Luc Segoufin & Victor Vianu Presented by Harel Paz.
Contents Introduction Related problems Constructions –Welch construction –Lempel construction –Golomb construction Special properties –Periodicity –Nonattacking.
Definition Dual Graph G* of a Plane Graph:
Resolving the isospectrality of the dihedral graphs Ram Band, Uzy Smilansky A graph G is made of V vertices and B bonds and a connectivity matrix C ij.
Curve Curve: The image of a continous map from [0,1] to R 2. Polygonal curve: A curve composed of finitely many line segments. Polygonal u,v-curve: A polygonal.
Group theory 101 Suggested reading: Landau & Lifshits, Quantum Mechanics, Ch. 12 Tinkham, Group Theory and Quantum Mechanics Dresselhaus, Dresselhaus,
Graph Theory Chapter 6 Planar Graphs Ch. 6. Planar Graphs.
Wei Wang Xi’an Jiaotong University Generalized Spectral Characterization of Graphs: Revisited Shanghai Conference on Algebraic Combinatorics (SCAC), Shanghai,
Matrix Completion Problems for Various Classes of P-Matrices Leslie Hogben Department of Mathematics, Iowa State University, Ames, IA 50011
Multiway Trees. Trees with possibly more than two branches at each node are know as Multiway trees. 1. Orchards, Trees, and Binary Trees 2. Lexicographic.
Chapter 9. Chapter Summary Relations and Their Properties n-ary Relations and Their Applications (not currently included in overheads) Representing Relations.
© by Kenneth H. Rosen, Discrete Mathematics & its Applications, Sixth Edition, Mc Graw-Hill, 2007 Chapter 9 (Part 2): Graphs  Graph Terminology (9.2)
Quantum One: Lecture Representation Independent Properties of Linear Operators 3.
Chapter 6 Graph Theory R. Johnsonbaugh Discrete Mathematics 5 th edition, 2001.
1 CS104 : Discrete Structures Chapter V Graph Theory.
Three different ways There are three different ways to show that ρ(A) is a simple eigenvalue of an irreducible nonnegative matrix A:
With Jesper Lykke Jacobsen Hubert Saleur Alan D. Sokal Andrea Sportiello Fields for Trees and Forests Physical Review Letters 93 (2004) Bari, September.
17. Group Theory 1.Introduction to Group Theory 2.Representation of Groups 3.Symmetry & Physics 4.Discrete Groups 5.Direct Products 6.Symmetric Groups.
Petrie maps and Petrie configurations Jurij Kovič, University of Primorska, Slovenia EuroGIGA Final Conference Berlin, February 20, 2014.
5.2 Trees  A tree is a connected graph without any cycles.
Leszek Sirko, Oleh Hul Michał Ławniczak, Szymon Bauch
Planar Graphs Graph Coloring
Theory of electric networks: The two-point resistance and impedance F. Y. Wu Northeastern University Boston, Massachusetts USA.
Chapter 5 Graphs  the puzzle of the seven bridge in the Königsberg,  on the Pregel.
Mathematical Preliminaries
Unit – V Graph theory. Representation of Graphs Graph G (V, E,  ) V Set of vertices ESet of edges  Function that assigns vertices {v, w} to each edge.
An Introduction to Graph Theory
Math 3121 Abstract Algebra I Lecture 7: Finish Section 7 Sections 8.
Chapter 9. Chapter Summary Relations and Their Properties n-ary Relations and Their Applications (not currently included in overheads) Representing Relations.
4. Relations and Digraphs Binary Relation Geometric and Algebraic Representation Method Properties Equivalence Relations Operations.
 Quotient graph  Definition 13: Suppose G(V,E) is a graph and R is a equivalence relation on the set V. We construct the quotient graph G R in the follow.
Graphs Basic properties.
8.4 Closures of Relations Definition: The closure of a relation R with respect to property P is the relation obtained by adding the minimum number of.
Great Theoretical Ideas in Computer Science for Some.
COMPSCI 102 Introduction to Discrete Mathematics.
Week 11 - Wednesday.  What did we talk about last time?  Graphs  Paths and circuits.
1 Lecture 5 (part 2) Graphs II (a) Circuits; (b) Representation Reading: Epp Chp 11.2, 11.3
De Bruijn sequences 陳柏澍 Novembers Each of the segments is one of two types, denoted by 0 and 1. Any four consecutive segments uniquely determine.
Theory of Computational Complexity Probability and Computing Chapter Hikaru Inada Iwama and Ito lab M1.
The Language of Sets If S is a set, then
37TH ANNUAL AMATYC CONFERENCE A Modern Existence Proof Through Graphs
Applied Discrete Mathematics Week 13: Graphs
Creating a cell complex = CW complex
Computability and Complexity
CSNB 143 Discrete Mathematical Structures
Computability and Complexity
Quantum One.
Complexity 6-1 The Class P Complexity Andrei Bulatov.
G-v, or G-{v} When we remove a vertex v from a graph, we must remove all edges incident with the vertex v. When a edge is removed from a graph, without.
Systems of distinct representations
Quantum Foundations Lecture 3
Graphs G = (V, E) V are the vertices; E are the edges.
Definition 8: Graphs that have a number assigned to each edge or each vertex are called weighted graphs weighted digraphs.
Vector Spaces RANK © 2012 Pearson Education, Inc..
Presentation transcript:

Groups, Graphs & Isospectrality Rami Band Ori Parzanchevski Gilad Ben-Shach Uzy Smilansky

What is a graph ? The graph’s spectrum is the sequence of basic frequencies at which the graph can vibrate. The graph’s spectrum is the sequence of basic frequencies at which the graph can vibrate. The spectrum depends on the shape of the graph. The spectrum depends on the shape of the graph.

‘Can one hear the shape of a drum ?’ was first asked by Marc Kac (1966). ‘Can one hear the shape of a drum ?’ was first asked by Marc Kac (1966). ‘Can one hear the shape of a graph?’ ‘Can one hear the shape of a graph?’ Can one deduce the shape from the spectrum ? Can one deduce the shape from the spectrum ? Is it possible to have two different graphs with the same spectrum (isospectral graphs) ? Is it possible to have two different graphs with the same spectrum (isospectral graphs) ? ‘Can one hear the shape of a graph ?’ Marc Kac ( )

Metric Graphs - Introduction A graph Γ consists of a finite set of vertices V={v i }and E={e j }. A graph Γ consists of a finite set of vertices V={v i } and a finite set of undirected edges E={e j }. A metric graph has a finite length (L e >0) assigned to each edge. E v Let E v be the set of all edges connected to a vertex v. The degree of v is A function on the graph is a vector of functions on the edges: L 13 L 23 L 34 L 45 L 46 L 15 L 45 L 34 L 23 L 12 L 25 L 35 L 14

Quantum Graphs - Introduction A quantum graph is a metric graph equipped with an operator, such as the negative Laplacian: For each vertex v, define: We impose boundary conditions of the form: Where A v and B v are complex d v x d v matrices. To ensure the self-adjointness of the Laplacian, we require that the matrix (A v |B v ) has rank d v, and that the matrix A v B v † is self-adjoint (Kostrykin and Schrader, 1999).

Quantum Graphs - Introduction A quantum graph is a metric graph equipped with an operator, such as the negative Laplacian: For each vertex v, define: We impose boundary conditions of the form: Where A v and B v are complex d v x d v matrices. A common boundary condition is the Kirchhoff condition, namely all functions agree at each vertex, and the sum of the derivatives vanishes. This corresponds to the matrices: For d v =1 vertices, there are two special cases of boundary conditions, denoted Dirichlet: A v = (1), B v =(0) (so that ) Neumann: A v = (0), B v =(1) (so that )

Quantum Graphs - Introduction A quantum graph is a metric graph equipped with an operator, such as the negative Laplacian: For each vertex v, define: We impose boundary conditions of the form: Where A v and B v are complex d v x d v matrices. A common boundary condition is the Kirchhoff condition, namely all functions agree at each vertex, and the sum of the derivatives vanishes. For d v =1 vertices, there are two special cases of boundary conditions, denoted Dirichlet: A v = (1), B v =(0) (so that ) Neumann: A v = (0), B v =(1) (so that )

The spectrum is With the set of corresponding eigenfunctions: Use the notation: The Spectrum of Quantum Graphs Examples of several functions of the graph: λ 8 ≈9.0 λ 13 ≈24.0 λ 16 ≈ √2 D D N N √3 √2

One can hear the shape of a simple graph if the lengths are incommensurate (Gutkin, Smilansky 2001) One can hear the shape of a simple graph if the lengths are incommensurate (Gutkin, Smilansky 2001) Otherwise, we do have isospectral graphs: Otherwise, we do have isospectral graphs: Von Below (2001) Von Below (2001) Band, Shapira, Smilansky (2006) Band, Shapira, Smilansky (2006) There are several methods for construction of isospectrality – the main is due to Sunada (1985). There are several methods for construction of isospectrality – the main is due to Sunada (1985). We present a method based on representation theory arguments. We present a method based on representation theory arguments. a 2b a 2c DN 2a b b c c D D N N ‘Can one hear the shape of a graph ?’

rxrxrxrx e Groups & Graphs Example: The Dihedral group – the symmetry group of the square D 4 = { e, a, a 2, a 3, r x, r y, r u, r v } Example: The Dihedral group – the symmetry group of the square D 4 = { e, a, a 2, a 3, r x, r y, r u, r v } y x u v a How does the Dihedral group act on a square ? A few subgroups of the Dihedral group: H 1 = { e, a 2, r x, r y } H 2 = { e, a 2, r u, r v } H 3 = { e, a, a 2, a 3 } A few subgroups of the Dihedral group: H 1 = { e, a 2, r x, r y } H 2 = { e, a 2, r u, r v } H 3 = { e, a, a 2, a 3 }

Representation – Given a group G, a representation R is an assignment of a matrix R(g) to each group element g  G, such that:  g 1,g 2  G R(g 1 )R(g 2 )=R(g 1 g 2 ). Representation – Given a group G, a representation R is an assignment of a matrix R(g) to each group element g  G, such that:  g 1,g 2  G R(g 1 )R(g 2 )=R(g 1 g 2 ). Example 1 - D 4 has the following 1-dimensional rep. S 1 : Example 1 - D 4 has the following 1-dimensional rep. S 1 : Example 2 - D 4 has the following 2-dimensional rep. S 2 : Example 2 - D 4 has the following 2-dimensional rep. S 2 : Restriction - is the following rep. of H 1 : Restriction - is the following rep. of H 1 : Induction - is the following rep. of G: Induction - is the following rep. of G: Groups - Representations

F = Knowing the matrix representation gives us information on the functions. Knowing the matrix representation gives us information on the functions. The group can act on a graph and … the group can act on a function which is defined on the graph and may give new functions: The group can act on a graph and … the group can act on a function which is defined on the graph and may give new functions: We have So that, form a representation of the group. We have So that, form a representation of the group. Groups & Graphs How does the group act on the function ? a Example: The Dihedral group – the symmetry group of the square D 4 = { e, a, a 2, a 3, r x, r y, r u, r v } Example: The Dihedral group – the symmetry group of the square D 4 = { e, a, a 2, a 3, r x, r y, r u, r v }

Consider the following rep. R 1 of the subgroup H 1 : Consider the following rep. R 1 of the subgroup H 1 : H 1 = { e, a 2, r x, r y } R 1 : This is a 1d rep. – what do we know about the function F ? Consider the following rep. R 2 of the subgroup H 2 : Consider the following rep. R 2 of the subgroup H 2 : H 2 = { e, a 2, r u, r v } R 2 : This is a 1d rep. – what do we know about the function F ? D N N D Groups & Graphs Examine the graph Γ: D N D D D D N N N N N N D D

Theorem – Let Γ be a graph which obeys a symmetry group G. Let H 1, H 2 be two subgroups of G with representations R 1, R 2 that obey then the graphs, are isospectral. Theorem – Let Γ be a graph which obeys a symmetry group G. Let H 1, H 2 be two subgroups of G with representations R 1, R 2 that obey then the graphs, are isospectral. In our example: In our example: Γ = H 1 = {e, a 2, r x, r y } R 1 : H 2 = {e, a 2, r u, r v } R 2 : And it can be checked that Groups, Graphs & Isospectrality D N N D D N G = D 4

Extending our example: Γ = H 1 = { e, a 2, r x, r y } R 1 : H 2 = { e, a 2, r u, r v } R 2 : H 3 = { e, a, a 2, a 3 } R 3 : Extending the Isospectral pair D N N D D N ×i×i×i×i ×i×i×i×i ×i×i×i×i G = D 4

Extending our example: Γ = H 1 = { e, a 2, r x, r y } R 1 : H 2 = { e, a 2, r u, r v } R 2 : H 3 = { e, a, a 2, a 3 } R 3 : Extending the Isospectral pair D N N D D N ×i×i×i×i ×i×i×i×i ×i×i×i×i G = D 4

Theorem – Let Γ be a graph which obeys a symmetry group G. Let H 1, H 2 be two subgroups of G with representations R 1, R 2 that obey then the graphs, are isospectral. Theorem – Let Γ be a graph which obeys a symmetry group G. Let H 1, H 2 be two subgroups of G with representations R 1, R 2 that obey then the graphs, are isospectral. Proof: Lemma : There exists a quantum graph such that: Using the Lemma and Frobenius reciprocity theorem gives: Hence, are isospectral. Applying the same for the rep. R 2 and using finishes the proof. Proof: Lemma : There exists a quantum graph such that: Using the Lemma and Frobenius reciprocity theorem gives: Hence, are isospectral. Applying the same for the rep. R 2 and using finishes the proof. Groups, Graphs & Isospectrality

Theorem – Let Γ be a graph which obeys a symmetry group G. Let H 1, H 2 be two subgroups of G with representations R 1, R 2 that obey then the graphs, are isospectral. Theorem – Let Γ be a graph which obeys a symmetry group G. Let H 1, H 2 be two subgroups of G with representations R 1, R 2 that obey then the graphs, are isospectral. Proof: Lemma : There exists a quantum graph such that: Interesting issues in the proof of the lemma: Proof: Lemma : There exists a quantum graph such that: Interesting issues in the proof of the lemma: A group which does not act freely on the edges. A group which does not act freely on the edges. Representations which are not 1-d. Representations which are not 1-d. The dependence of in the choice of basis for the representation. The dependence of in the choice of basis for the representation. Groups, Graphs & Isospectrality

Γ is the Cayley graph of D 4 (with respect to the generators a, r x ): Take again G=D 4 and the same subgroups: H 1 = { e, a 2, r x, r y } with the rep. R 1 H 2 = { e, a 2, r u, r v } with the rep. R 2 H 3 = { e, a, a 2, a 3 } with the rep. R 3 Arsenal of isospectral examples e a a3a3 a2a2 rxrx ryry rvrv ruru L1L1 L2L2 L2L2 L1L1 The resulting quotient graphs are: L1L1 L1L1 L1L1 L1L1 L2L2 L2L2 L2L2 L2L2 L1L1 L1L1 L2L2 L2L2

G = D 6 = {e, a, a 2, a 3, a 4, a 5, r x, r y, r z, r u, r v, r w } with the subgroups: H 1 = { e, a 2, a 4, r x, r y, r z } with the rep. R 1 H 2 = { e, a 2, a 4, r u, r v, r w } with the rep. R 2 H 3 = { e, a, a 2, a 3, a 4, a 5 } with the rep. R 3 Arsenal of isospectral examples The resulting quotient graphs are: L2L2 L1L1 2L 2 2L 3 2L 1 L2L2 L3L3 L3L3 2L 2 2L 3 2L 1 2L 3 2L 2 2L 1 L1L1 L1L1 2L 2 2L 3 2L 1

G = S 4 acts on the tetrahedron. with the subgroups: H 1 = S 3 with the rep. R 1 H 2 = S 4 with the rep. R 2 |S 4 |=24, |S 3 |=6 Arsenal of isospectral examples The resulting quotient graphs are: L D L L/2 D N D D N D

Γ is the a graph which obeys the O h (octahedral group with reflections). |O h |=48. Take G= O h and the subgroups: H 1 = O, the octahedral group. H 2 = T d, the tetrahedral group with reflections. with reflections. |H 1 |= |H 2 |= 24 |H 1 |= |H 2 |= 24 Arsenal of isospectral examples

A puzzle : construct an isospectral pair out of the following familiar graph:

G = S 3 (D 3 ) acts on Γ with no fixed points. To construct the quotient graph, we take the same rep. of G, but use two different bases for the matrix representation. Arsenal of isospectral examples The resulting quotient graphs are: L1L1 L2L2 L3L3 L3L3 L2L2 L1L1 L3L3 L3L3 L3L3 L2L2 L2L2 L2L2 L1L1 L3L3 L3L3 L3L3 L2L2 L2L2 L2L2 L1L1 L2L2 L2L2 L3L3 L3L3 L1L1 L2L2 L3L3 L1L1 L2L2 L3L3 L1L1 L2L2 L3L3 L3L3 L2L2

Arsenal of isospectral examples ‘One cannot hear the shape of a drum’ ‘One cannot hear the shape of a drum’ Gordon, Webb and Wolpert (1992) Isospectral drums D N D N G 0 =*444 G 0 =*444 (using Conway’s orbifold notation) acts on the hyperbolic plane. G 0 G=PSL(3,2) H 1 H 2 R 1 R 2 Considering a homomorphism of G 0 onto G=PSL(3,2) and taking two subgroups H 1, H 2 such that: and R 1, R 2 are the sign representations, we obtain the known isospectral drums of Gordon et al. but with new boundary conditions:

Arsenal of isospectral examples ‘Spectral problems with mixed Dirichlet-Neumann boundary conditions: isospectrality and beyond’ D. Jacobson, M. Levitin, N. Nadirashvili, I. Polterovich (2004) ‘Isospectral domains with mixed boundary conditions’ M. Levitin, L. Parnovski, I. Polterovich (2005) Isospectral drums This isospectral quartet can be obtained when acting with the group D 4 xD 4 on the following torus and considering the subgroups H 1 X H 1, H 1 X H 2, H 2 X H 1, H 2 X H 2 with the reps. R 1 X R 1, R 1 X R 2, R 2 X R 1, R 2 X R 2 (using the notation presented before for the main dihedral example).

The relation to Sunada’s construction H 1, H 2 be two subgroups of G Let G be a group. Let H 1, H 2 be two subgroups of G. Then the triple (G, H 1, H 2 ) satisfies Sunada’s condition if: where [g] is the cunjugacy class of g in G. For such a triple (G, H 1, H 2 ) we get that, are isospectral. Pesce (94) proved Sunada’s theorem using the observation that Sunada’s condition is equivalent to the following: The relation to the construction method presented so far is via the identification:

Further on … What is the strength of this method ? What is the strength of this method ? Having two isospectral graphs – how to construct the ‘parent’ graph from which they were born ? Having two isospectral graphs – how to construct the ‘parent’ graph from which they were born ? Having such a ‘parent’ graph, can it be shown that it obeys a symmetry group such that the conditions of the theorem are fulfilled ? Having such a ‘parent’ graph, can it be shown that it obeys a symmetry group such that the conditions of the theorem are fulfilled ? What are the conditions which guarantee that the quotient graphs are not isometric ? What are the conditions which guarantee that the quotient graphs are not isometric ? A graph with a self-adjoint operator might be isospectral to a graph with a non self-adjoint one. A graph with a self-adjoint operator might be isospectral to a graph with a non self-adjoint one. What other properties of the functions can be used to resolve isospectrality ? What other properties of the functions can be used to resolve isospectrality ?

Groups, Graphs & Isospectrality Rami Band Ori Parzanchevski Gilad Ben-Shach Uzy Smilansky Acknowlegments: M. Sieber, I. Yaakov.