+ All Categories
Home > Documents > Tribonacci tiling is sofic. - IRISA · Tribonacci tiling is so c ... Speci cities of Tribonacci...

Tribonacci tiling is sofic. - IRISA · Tribonacci tiling is so c ... Speci cities of Tribonacci...

Date post: 08-Jun-2018
Category:
Upload: phamdieu
View: 216 times
Download: 0 times
Share this document with a friend
44
Introduction Metaresult Tribonacci Tribonacci tiling is sofic. Xavier Bressaud Universit´ e Paul Sabatier Institut de Math´ ematiques de Toulouse Substitutive Tiling and Fractal Geometry, Guangzhou July 2010 Xavier Bressaud Tribonacci / Sofic
Transcript

IntroductionMetaresultTribonacci

Tribonacci tiling is sofic.

Xavier Bressaud

Universite Paul SabatierInstitut de Mathematiques de Toulouse

Substitutive Tiling and Fractal Geometry, Guangzhou

July 2010

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

1 Introduction

2 Metaresult

3 Tribonacci

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

A discrete plane

Let Π be the contracting plane of the matrix A =0@ 1 1 0

1 0 11 0 0

1AApproximation by unit cubes : discrete plane.

Projection (parallel to the PF direction) of the faces of thecubes on a transversal plane.

Tiling by a family of three rhombi T (Tribonacci).

Observe that this tiling has some self-similar stucture.

Space Ω generated by translations (R2) and closure.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci Tiling

Fig.: A piece of the discrete plane

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Sofic

Tiling by a family of rhombi T , labelled, C.

Local rules L : some glueing are forbiden.

Underlying tiling : forget the labels.

Space ΩT ,C,L of tilings by labelled tiles (T , C) satisfying thelocal rules L (SFT)

Space U(ΩT ,C,L) of the underlying (non labelled) tilings(Sofic).

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling is sofic

Theorem

There exist a set of colors C and a set of local rules L such that

U(ΩT ,C,L) = Ω.

Last computation : 9776 labelled tiles

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Contrast with dimension 1

Discrete line of slope α : Sturmian sequences and tilings.

For α rational : the tiling is periodic (an SFT).

For α quadratic : the tiling is self-similar.

Such tiling is (certainly) not of finite type (no perodic orbits).

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Motivations

The result is not surprising ONCE you know the metaresult.

Existing proofs did not yield the result : it had to be checked.

Unerstand existing proofs.

Following a general proof yields a huge set of tiles.

Specificities of Tribonacci tiling (symetries, geometry of tiles)

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tilings

Tilings and spaces of tilings. Topology. Action of R2.

Patch, Language, Finite local complexity.

Substitutive tilings and space of tilings.

SFT and sofic spaces of tilings.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Substitutive tilings are sofic

Preliminaries : Does a set of tiles tile ? Can we decide ?

Berger, Robinson and aperiodic set of tiles.

Moses for squared substitutive tilings.

Goodman-Strauss for polygonal substitutive tilings.

Solomyak to see pseudo-self-similar tilings as self-similar ones.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Hierarchical structure (informal)

A tiling T by a set of tiles T .

A sequence of tilings T (n) by sets of tiles T (n).

Tiles in T (n) are unions of tiles in T (n−1).

Decomposition of all tiles of T (n) yields T (n−1).

The tiling is self-similar if the families T (n) are similar.

Pseudo self-similar if the decomposition is the ”same” at all n.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Hierarchical structure

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

BB

B

BB

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

BB

B

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Metatile

Consider a self similar tiling and its hierarchy.

Skeleton and boundary of a metatile.

Skeleton intersects the boundary (dim > 1 !)

In the tiling : union of skeletons = union of boundaries

Partitions of the boundaries (scale changes).

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Betaproof

A priori the (initial) set of tiles may have a huge universe.

Put more informations on tiles to specify the roles they canplay in the hierarchy.

Put local constraints that force them play the specified role.

So that the tilings obtained following these rules respect thehierarchical structure.

It remains to forget about the additional information.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

First idea

Specify the position of the tile in the upper level of thehierarchy (for instance say which part of its boundary it mustglue to which tile to produce a metatile).

Consider the SFT where the rules allow only these glueings.

Automatically, a given tile belongs to a patch forming ametatile.

Hence we arrive at level 2. But how to iterate ?

New tiles may not have the good colors on their boundaries.And we need to put only a finite amount of information oneach tile.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Second idea

Send information from one level to the next. Through theskeletons.

Recall the skeletons are connected to the boundary of theirmetatile.

Let the skeleton ”know” the global configuration of itsmetatile.

Check at the connection that the information is coherent.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Conclusion

On the one hand, this is enough to force the SFT to respectthe hierarchical structure.

On the other hand, it must be checked that the self-similartiling is itself in the SFT.

In practice, we produce a decoration of the self-similar tilingto determine the alphabet that is to be used.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Case of Tribonacci

Understand the combinatorics of the tiles as the scale grows.

Then, use the geometry of tiles to recover information.

Lot of symmetries : the three tiles play similar similar roles.

Main goal : reduce the number of tiles needed.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci

Discrete plane.

Generalized substitution (dual). and on Z2.

IFS.

”Combinatorial” substitution.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling : substitution rule

B

Fig.: τ (12,1)Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling : inflation

T (n+1) = T (n) ∪(T (n−1) + ω

(n)1

)∪(T (n−2) + ω

(n)2

).

Fig.: Tiles T (3) ⊂ · · · ⊂ T (11).

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling

τ (n,n−3)

τ (n,k).

τ (∞,k). Combinatorics appears in the geometry from k = 5.

µ−kτ (∞,k) → τ

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling

Fig.: τ (12,1)

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling

Fig.: τ (15,1)

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling

Fig.: τ (∞,∞)

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling : scale 5

1

2

3

4

5

6

7

1

2

3

4

5

6

1

2

3

4 5

6

7

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4 5

6

7

W1 W2

W3

W4

W5W6

W7

W1

W2W3

W4

W5

W6

W7

W1

W2

W3W4

W5

W6

W7

1

2

3

4

5

6

7

-

-

W1

W2

W3

W4

W5

W6

W7

Fig.: T (5)

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

BB

B

BB

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

B

BB

B

Fig.: Skeletons

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling

GG

PA

MAGA

AB

PP

MP

GP

PM

MM

GM

GG

MG

PG

PA

MA

GA

AB

BC

AB

PG

GG

MG

GP

MP

PP

GM

PM

MM

GA

PA

MA

BC

On veut comprendre ce qu'il faut transmettre sur les aretes pour controler ce qui se passe sur les bords "exterieurs"1. Le long de chaque frontiere on dit le nom de la cellule (de gauche) et celui à laquelle elle veut se coller. La ou c'est clair. 2. La ou c'est indeterminé, on laisse de a liberté. Mais bien sur ce "sera" fixé. 3. Il faut qd meme transmettre des contraintes aux bords exterieurs. Il faut pouvoir demander une uniformité de la couleur.

Fig.: SkeletonsXavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling

Fig.: Skeletons

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Tribonacci tiling : geometry of the boundary

(W

(1)1 , . . . ,W

(1)7 ) = (∅, c , ∅, a, c−1, ∅, a−1),

(W(2)1 , . . . ,W

(2)7 ) = (∅, a, ∅, b, a−1, ∅, b−1),

(W(3)1 , . . . ,W

(3)7 ) = (b, ∅, ∅, c , b−1, ∅, c−1).

W(n+1)1 = W

(n−2)1 W

(n−2)2 W

(n−1)6

W(n+1)2 = W

(n−1)7

W(n+1)3 = W

(n−1)1

W(n+1)4 = W

(n−1)2 W

(n)6

W(n+1)5 = W

(n)7

W(n+1)6 = W

(n)1

W(n+1)7 = W

(n)2 W

(n)3 W

(n−2)7 .

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Topological metatile

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4 5

6

7

W1 W2

W3

W4

W5W6

W7

W1

W2W3

W4

W5

W6

W7

W1

W2

W3W4

W5

W6

W7

1

2

3

4

5

6

7

-

-

W1

W2

W3

W4

W5

W6

W7

Fig.: Meta-tile T (8) with informations on marked vertices.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

v -configuration

State of a vertex : 0, ±1 or 2.

2

-10

12

2

Seven states of the vertices of a tile T : v -configuration V (T ).

V = VP ∪ VM ∪ VG set of possible v -configurations.

Describes how the boundary is cut in pieces of skeletons ofdifferent scales.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Automaton

Let T be a metatile : T = TP ∪ TM ∪ TG .

If V = V (T ), V P = V (TP),V M = V (TM),V G = V (TG ),

then,V = V P

1 V M7 V M

1 V M2 V G

7 V G1 V G

2 .

and, V P = V1 0 1 0 2 0 −1V M = V3 V4 1 0 2 −1 V2

V G = V6 V7 0 1 2 −1 V5

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

V

VP 0 0 1 0 2 0 −1−1 0 1 0 2 0 −1

1 0 1 0 2 0 −1

VM 1 0 1 0 2 −1 11 0 1 0 2 −1 00 1 1 0 2 −1 20 1 1 0 2 −1 −10 1 1 0 2 −1 10 1 1 0 2 −1 0

VG −1 2 0 1 2 −1 20 −1 0 1 2 −1 2−1 1 0 1 2 −1 2−1 0 0 1 2 −1 2−1 −1 0 1 2 −1 2

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Labelled tiles

Tn : T (n),T (n+1),T (n+2) × VP (3× 3 = 9)

T (n+1),T (n+2) × VM (2× 6 = 12)

T (n+2) × VG (1× 5 = 5)

]Tn = 26.

Fig.: Labelled tiles

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Labelled tiles

Fig.: The labelled tiles (partial)

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

First step

Set T5 of labelled tiles.

Local rule : at all triple point : (0, 0), (2, 2, 2) ou (0,−1, 1).

Thanks to the geometry/combinatorics this constraint isenough to ”climb one scale”, i.e. all tile belong to a metatile.

To control the configuration of the metatile, need to transmit”information”

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Coloring edges

Each edge E belong to the skeleton of a metatile T (of scaleK ) ; T = TP ∪ TM ∪ TG .

Type : S(E) = P,M ou G according to whether the edgebelongs to TP ∩ TM , TM ∩ TG ou TG ∩ TP .

The edge E is colored by :

c(E) = (S(E),V (T ))

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Topological metatile

1

2

3

4

5

6

7

1

2

3

4

5

6

7

1

2

3

4 5

6

7

W1 W2

W3

W4

W5W6

W7

W1

W2W3

W4

W5

W6

W7

W1

W2

W3W4

W5

W6

W7

1

2

3

4

5

6

7

-

-

W1

W2

W3

W4

W5

W6

W7

Fig.: Meta-tile T (8) with informations on marked vertices.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Automaton (2)

Again we use the automatic rule :

C = CP1 CM

7 CM1 CM

2 CG7 CG

1 CG2 .

and, CP = C1 C1 (P,V ) (P,V ) (G ,V ) (G ,V ) C7

CM = C3 C4 (M,V ) (M,V ) (P,V ) C2 C2

CG = C6 C7 C7 (G ,V ) (G ,V ) C6 C5

to check the colorings that really appear in Tribonacci tiling.

Denote V ⊂ (P,M,G × V)7 these colorings. ]V = 4893.

Let Dn be the set of decorated tiles (label + color) at scale n.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Rules

It remains to fix a set rules L.

Constraint on the vertices as before.

Tiles have same colors on common boundaries.

At vertices with states (−1, 1, 0), with ”incoming” edge oftype M : check that the v -configuration on the incoming edgeis coherent with that on the passing edge.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

SFT

Consider the SFT space of tilings with tiles in Dn satisfying localrules L. We claim that the space of underlying tilings is exactly theTribonacci tiling space.

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Labelled tiles

]D5 = 9776.

Fig.: Labelled tiles. Put colors

Xavier Bressaud Tribonacci / Sofic

IntroductionMetaresultTribonacci

Further

Reduce the number of tiles (to something that can be drawn).

Tilings arising from other substitutions. What are theimportant combinatorial/topological properties of the tiles ?

Non stationary hierarchical structure (among a finite set andin an effective order).

Xavier Bressaud Tribonacci / Sofic


Recommended