Victor Zamaraev – Boundary properties of factorial classes of graphs
Répétition soutenance
1. Dynamique symbolique
des syst`mes 2D et des arbres infinis
e
Soutenance de th`se,
e
encadr´e par Marie-Pierre Beal et Mathieu Sablik
e ´
Nathalie Aubrun
LIGM, Universit´ Paris-Est
e
22 juin 2011
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 1 / 24
2. Outline
1 Symbolic dynamics
What are symbolic dynamics ?
Subshifts
Classes of subshifts
2 Motivation
1D subshifts vs. 2D subshifts
Two orthogonal approaches
3 A study of 2D-shifts
The projective subaction
Hochman’s result
Improvement
4 Tree-shifts
Tree-shift: example
Tree automata and tree-shifts
On AFT tree-shifts
Conjugacy of tree-SFT
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 1 / 24
3. Symbolic dynamics
Outline
1 Symbolic dynamics
What are symbolic dynamics ?
Subshifts
Classes of subshifts
2 Motivation
1D subshifts vs. 2D subshifts
Two orthogonal approaches
3 A study of 2D-shifts
The projective subaction
Hochman’s result
Improvement
4 Tree-shifts
Tree-shift: example
Tree automata and tree-shifts
On AFT tree-shifts
Conjugacy of tree-SFT
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 1 / 24
4. Symbolic dynamics What are symbolic dynamics ?
Discrete dynamical systems
(X , F ) is a discrete dynamical system if:
X is a topological space, called the phase space
F is a continuous map X → X
x
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 2 / 24
5. Symbolic dynamics What are symbolic dynamics ?
Discrete dynamical systems
(X , F ) is a discrete dynamical system if:
X is a topological space, called the phase space
F is a continuous map X → X
F (x)
x
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 2 / 24
6. Symbolic dynamics What are symbolic dynamics ?
Discrete dynamical systems
(X , F ) is a discrete dynamical system if:
X is a topological space, called the phase space
F is a continuous map X → X
F 2 (x)
F (x)
x
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 2 / 24
7. Symbolic dynamics What are symbolic dynamics ?
Discrete dynamical systems
(X , F ) is a discrete dynamical system if:
X is a topological space, called the phase space
F is a continuous map X → X
F 2 (x) F 4 (x)
F (x)
F 3 (x)
x F 5 (x)
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 2 / 24
8. Symbolic dynamics What are symbolic dynamics ?
Coding of the orbits
X= n
i=1 Xi a partition of the phase space X
a color ai associated with each Xi
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 3 / 24
9. Symbolic dynamics What are symbolic dynamics ?
Coding of the orbits
X= n
i=1 Xi a partition of the phase space X
a color ai associated with each Xi
orbit (F n (x))n∈N coded by a sequence y ∈ {a1 , . . . , an }N
...
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 3 / 24
10. Symbolic dynamics Subshifts
Subshifts: topological definition
2 2
A finite alphabet, AN (or AZ , AN or AZ ) the configurations space
the configurations space is endowed with the prodiscrete topology ⇒
compact space
natural action of N (or Z, N2 or Z2 ) by translation: the shift σ
σ j (x)i = xi+j for all x ∈ AN
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 4 / 24
11. Symbolic dynamics Subshifts
Subshifts: topological definition
2 2
A finite alphabet, AN (or AZ , AN or AZ ) the configurations space
the configurations space is endowed with the prodiscrete topology ⇒
compact space
natural action of N (or Z, N2 or Z2 ) by translation: the shift σ
σ j (x)i = xi+j for all x ∈ AN
(Topological) Definition: subshift
2 2
A subshift is a closed and σ-invariant subset of AN (or AZ , AN or AZ ).
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 4 / 24
12. Symbolic dynamics Subshifts
Subshifts: combinatorial definition
A finite alphabet
pattern p ∈ AS , where S ⊂ N is finite
the pattern p = appears in the configuration
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 5 / 24
13. Symbolic dynamics Subshifts
Subshifts: combinatorial definition
A finite alphabet
pattern p ∈ AS , where S ⊂ N is finite
the pattern p = appears in the configuration
(Combinatorial) Definition: subshift
Let F be a set of finite patterns. The subshift defined by the set of forbidden
patterns F is the set
TF = x ∈ AN , no pattern of F appears in x .
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 5 / 24
14. Symbolic dynamics Subshifts
Subshifts: combinatorial definition
A finite alphabet
pattern p ∈ AS , where S ⊂ N is finite
the pattern p = appears in the configuration
(Combinatorial) Definition: subshift
Let F be a set of finite patterns. The subshift defined by the set of forbidden
patterns F is the set
TF = x ∈ AN , no pattern of F appears in x .
Proposition
The toplogical and combinatorial definitions coincide.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 5 / 24
15. Symbolic dynamics Classes of subshifts
Classes of subshifts: Subshifts of finite type (SFT)
Sets of configurations that avoid a finite set of forbidden patterns:
alphabet { , } on N
one forbidden pattern
x ∈{ , }N , ∃i ∈ N ∪ {+∞}, (xj = ⇔ j ≤ i)
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 6 / 24
16. Symbolic dynamics Classes of subshifts
Classes of subshifts: Subshifts of finite type (SFT)
Sets of configurations that avoid a finite set of forbidden patterns:
alphabet { , } on N
one forbidden pattern
x ∈{ , }N , ∃i ∈ N ∪ {+∞}, (xj = ⇔ j ≤ i)
Definition: subshift of finite type (SFT)
A subshift of finite type (SFT) is a subshift that can be defined by a finite set of
forbidden patterns.
simplest class with respect to the combinatorial definition
2D-SFT ≡ Wang tilings
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 6 / 24
17. Symbolic dynamics Classes of subshifts
Classes of subshifts: Sofic subshifts
Factor map Φ AZ → B Z given by a local map φ:
2 2
Φ(x) ∈ B Z
2
x ∈ AZ
2
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 7 / 24
18. Symbolic dynamics Classes of subshifts
Classes of subshifts: Sofic subshifts
Factor map Φ AZ → B Z given by a local map φ:
2 2
Φ(x) ∈ B Z
2
x ∈ AZ
2
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 7 / 24
19. Symbolic dynamics Classes of subshifts
Classes of subshifts: Sofic subshifts
Factor map Φ AZ → B Z given by a local map φ:
2 2
Φ(x) ∈ B Z
2
x ∈ AZ
2
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 7 / 24
20. Symbolic dynamics Classes of subshifts
Classes of subshifts: Sofic subshifts
Factor map Φ AZ → B Z given by a local map φ:
2 2
Φ(x) ∈ B Z
2
x ∈ AZ
2
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 7 / 24
21. Symbolic dynamics Classes of subshifts
Classes of subshifts: Sofic subshifts
Factor map Φ AZ → B Z given by a local map φ:
2 2
Φ(x) ∈ B Z
2
x ∈ AZ
2
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 7 / 24
22. Symbolic dynamics Classes of subshifts
Classes of subshifts: Sofic subshifts
Factor map Φ AZ → B Z given by a local map φ:
2 2
Φ(x) ∈ B Z
2
x ∈ AZ
2
Definition: sofic susbhift
A sofic subshift is the factor of an SFT.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 7 / 24
23. Symbolic dynamics Classes of subshifts
Classes of subshifts: Sofic subshifts
Factor map Φ AZ → B Z given by a local map φ:
2 2
Φ(x) ∈ B Z
2
x ∈ AZ
2
Definition: sofic susbhift
A sofic subshift is the factor of an SFT.
Recodings of SFT, using local rules.
In 1D, sofic subshifts are exactly those recognized by finite automata.
In higher dimension, decide wether a subshift is sofic or not is a difficult problem.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 7 / 24
24. Symbolic dynamics Classes of subshifts
Classes of subshifts: Effective subshifts
Definition: effective susbhift
An effective subshift is a subshift that can be defined by a recursively enumerable
set of forbidden patterns.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 8 / 24
25. Symbolic dynamics Classes of subshifts
Classes of subshifts: Effective subshifts
Definition: effective susbhift
An effective subshift is a subshift that can be defined by a recursively enumerable
set of forbidden patterns.
reasonnable susbhift
this class naturally appears as projective subactions of 2D-SFT
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 8 / 24
26. Motivation
Outline
1 Symbolic dynamics
What are symbolic dynamics ?
Subshifts
Classes of subshifts
2 Motivation
1D subshifts vs. 2D subshifts
Two orthogonal approaches
3 A study of 2D-shifts
The projective subaction
Hochman’s result
Improvement
4 Tree-shifts
Tree-shift: example
Tree automata and tree-shifts
On AFT tree-shifts
Conjugacy of tree-SFT
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 8 / 24
27. Motivation 1D subshifts vs. 2D subshifts
1D subshifts vs. 2D subshifts
1D-subshifts 2D-subshifts
Emptyness of SFT ×
∀ SFT,
Periodicity in SFT ∃ aperiodic SFT
∃ periodic configuration
Decomposition theorem sofic subshifts
N
Conjugacy of SFT ×
Z ?
finite
Recognizers for SFT
local automata
Recognizers
finite automata textile systems
for sofic subshifts
: decidable problem
×: undecidable problem
?: open problem
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 9 / 24
28. Motivation Two orthogonal approaches
Two orthogonal approaches
Two attempts to understand 2D-subshifts :
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 10 / 24
29. Motivation Two orthogonal approaches
Two orthogonal approaches
Two attempts to understand 2D-subshifts :
study 2D-SFT through operations acting on them
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 10 / 24
30. Motivation Two orthogonal approaches
Two orthogonal approaches
Two attempts to understand 2D-subshifts :
study 2D-SFT through operations acting on them
study subshifts defined on a structure between dimensions 1 and 2
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 10 / 24
31. Motivation Two orthogonal approaches
Operations on 2D-SFT
factor map operation (Fact): SFT sofic subshifts
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 11 / 24
32. Motivation Two orthogonal approaches
Operations on 2D-SFT
factor map operation (Fact): SFT sofic subshifts
operations that preserves SFT: product (P), finite type (FT) and spatial
extension
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 11 / 24
33. Motivation Two orthogonal approaches
Operations on 2D-SFT
factor map operation (Fact): SFT sofic subshifts
operations that preserves SFT: product (P), finite type (FT) and spatial
extension
operation mainly studied: projective subaction (SA)
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 11 / 24
34. Motivation Two orthogonal approaches
Tree-shifts
Structure in-between N and N2 ?
N N2
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 12 / 24
35. Motivation Two orthogonal approaches
Tree-shifts
Structure in-between N and N2 : free semi-group with two generators M2 .
N M2 N2
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 12 / 24
36. A study of 2D-shifts
Outline
1 Symbolic dynamics
What are symbolic dynamics ?
Subshifts
Classes of subshifts
2 Motivation
1D subshifts vs. 2D subshifts
Two orthogonal approaches
3 A study of 2D-shifts
The projective subaction
Hochman’s result
Improvement
4 Tree-shifts
Tree-shift: example
Tree automata and tree-shifts
On AFT tree-shifts
Conjugacy of tree-SFT
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 12 / 24
37. A study of 2D-shifts The projective subaction
The projective subaction
Idea: study subdynamics of 2D-subshifts.
Principle: given a subgroup G of Z2 and a 2D-susbhift T on alphabet A, and
consider the G-subshift defined by
SAG (T) = y ∈ AG ∃x ∈ T such that y = xT .
G = {(x, y ) ∈ Z2 y = x}
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 13 / 24
38. A study of 2D-shifts The projective subaction
The projective subaction
Idea: study subdynamics of 2D-subshifts.
Principle: given a subgroup G of Z2 and a 2D-susbhift T on alphabet A, and
consider the G-subshift defined by
SAG (T) = y ∈ AG ∃x ∈ T such that y = xT .
G = {(x, y ) ∈ Z2 y = x}
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 13 / 24
39. A study of 2D-shifts The projective subaction
The projective subaction
Idea: study subdynamics of 2D-subshifts.
Principle: given a subgroup G of Z2 and a 2D-susbhift T on alphabet A, and
consider the G-subshift defined by
SAG (T) = y ∈ AG ∃x ∈ T such that y = xT .
G = {(x, y ) ∈ Z2 y = x}
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 13 / 24
40. A study of 2D-shifts The projective subaction
The projective subaction
Idea: study subdynamics of 2D-subshifts.
Principle: given a subgroup G of Z2 and a 2D-susbhift T on alphabet A, and
consider the G-subshift defined by
SAG (T) = y ∈ AG ∃x ∈ T such that y = xT .
G = {(x, y ) ∈ Z2 y = x}
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 13 / 24
41. A study of 2D-shifts The projective subaction
The projective subaction
Idea: study subdynamics of 2D-subshifts.
Principle: given a subgroup G of Z2 and a 2D-susbhift T on alphabet A, and
consider the G-subshift defined by
SAG (T) = y ∈ AG ∃x ∈ T such that y = xT .
G = {(x, y ) ∈ Z2 y = x}
Proposition (A.&Sablik)
The class of effective subshifts is stable under projective subaction.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 13 / 24
42. A study of 2D-shifts The projective subaction
Projective subactions of 2D-SFT
Pavlov and Schraudner (2010):
what can be realized as projective subactions of 2D-SFT ?
any sofic Z-subshift of positive entropy
any zero-entropy sofic Z-subshift, with some conditions on its periods
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 14 / 24
43. A study of 2D-shifts The projective subaction
Projective subactions of 2D-SFT
Pavlov and Schraudner (2010):
what can be realized as projective subactions of 2D-SFT ?
any sofic Z-subshift of positive entropy
any zero-entropy sofic Z-subshift, with some conditions on its periods
there exist classes of Z-subshifts which are not realizable as projective
subdynamics of any Zd SFT.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 14 / 24
44. A study of 2D-shifts The projective subaction
Projective subactions of 2D-SFT
Pavlov and Schraudner (2010):
what can be realized as projective subactions of 2D-SFT ?
any sofic Z-subshift of positive entropy
any zero-entropy sofic Z-subshift, with some conditions on its periods
there exist classes of Z-subshifts which are not realizable as projective
subdynamics of any Zd SFT.
⇒ no complete characterization projective subactions of 2D-SFT.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 14 / 24
45. A study of 2D-shifts Hochman’s result
Hochman’s result
. . . no complete characterization projective subactions of 2D-SFT. . .
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 15 / 24
46. A study of 2D-shifts Hochman’s result
Hochman’s result
. . . no complete characterization projective subactions of 2D-SFT. . .
Consider factor map operations in addition to projective subactions.
Theorem (Hochman 2010)
Any effective Zd -subshift may be obtained by SA and Fact operations on a
Zd+2 -SFT.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 15 / 24
47. A study of 2D-shifts Hochman’s result
Hochman’s result
. . . no complete characterization projective subactions of 2D-SFT. . .
Consider factor map operations in addition to projective subactions.
Theorem (Hochman 2010)
Any effective Zd -subshift may be obtained by SA and Fact operations on a
Zd+2 -SFT.
Natural question: is it possible to use one dimension less ?
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 15 / 24
48. A study of 2D-shifts Improvement
Improvement of Hochman’s result
Theorem
Any effective Zd -subshift may be obtained by SA and Fact operations on a
Zd+1 -SFT.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 16 / 24
49. A study of 2D-shifts Improvement
Improvement of Hochman’s result
Theorem
Any effective Zd -subshift may be obtained by SA and Fact operations on a
Zd+1 -SFT.
Two different proofs:
Durand, Romaschenko & Shen 2010, using self-similar tilings
A.& Sablik 2010, adaptation of Robinson’s construction
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 16 / 24
50. A study of 2D-shifts Improvement
Improvement of Hochman’s result
Theorem
Any effective Zd -subshift may be obtained by SA and Fact operations on a
Zd+1 -SFT.
Two different proofs:
Durand, Romaschenko & Shen 2010, using self-similar tilings
A.& Sablik 2010, adaptation of Robinson’s construction
Many applications:
correspondance between an order on subshifts and an order on languages
multidimensional effective subshifts are sofic (A.& Sablik 2011)
construction of a tiles set whose quasi-periodic tilings have a non-recursively
bounded periodicity function (Ballier& Jeandel 2010)
...
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 16 / 24
51. Tree-shifts
Outline
1 Symbolic dynamics
What are symbolic dynamics ?
Subshifts
Classes of subshifts
2 Motivation
1D subshifts vs. 2D subshifts
Two orthogonal approaches
3 A study of 2D-shifts
The projective subaction
Hochman’s result
Improvement
4 Tree-shifts
Tree-shift: example
Tree automata and tree-shifts
On AFT tree-shifts
Conjugacy of tree-SFT
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 16 / 24
52. Tree-shifts Tree-shift: example
Example of tree-shift
alphabet A = { , }
forbidden patterns: paths containing an even number of between two
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 17 / 24
53. Tree-shifts Tree automata and tree-shifts
Tree automata
alphabet A = { , }
tree automaton A with states Q = {q0 , q1 , q }
transition rules:
q1 q0 q0 q
q0 q0 q1 q1 q , q1 q , q1 q , q1 q , q1
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 18 / 24
54. Tree-shifts Tree automata and tree-shifts
Tree automata
alphabet A = { , }
tree automaton A with states Q = {q0 , q1 , q }
transition rules:
q1 q0 q0 q
q0 q0 q1 q1 q , q1 q , q1 q , q1 q , q1
Accepted trees:
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 18 / 24
55. Tree-shifts Tree automata and tree-shifts
Tree automata and tree-shifts
(finite) tree automata, all states are accepting
a tree is accepted iff there exists a computation
deterministic tree automata ≡ non deterministic tree automata
the set of trees accepted by a tree automaton is a subshift
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 19 / 24
56. Tree-shifts Tree automata and tree-shifts
Tree automata and tree-shifts
(finite) tree automata, all states are accepting
a tree is accepted iff there exists a computation
deterministic tree automata ≡ non deterministic tree automata
the set of trees accepted by a tree automaton is a subshift
Proposition (A. & B´al 2009)
e
A tree-shift is a sofic tree-shift iff it is recognized by a tree automaton.
A tree-shift is a tree-SFT iff it is recognized by a local tree automaton.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 19 / 24
57. Tree-shifts Tree automata and tree-shifts
Main difference with finite automata on words
Synchronizing block: every calculation (there exist at least one) of A on this block
ends in the same state.
Proposition
Any finite automaton on words has a synchronizing word.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 20 / 24
58. Tree-shifts Tree automata and tree-shifts
Main difference with finite automata on words
Synchronizing block: every calculation (there exist at least one) of A on this block
ends in the same state.
Proposition
Any finite automaton on words has a synchronizing word.
But...
Proposition (A.& B´al 2009)
e
There exists a deterministic, minimal and reduced tree automaton which has no
synchronizing block.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 20 / 24
59. Tree-shifts Tree automata and tree-shifts
The context automaton
The context automaton of a tree-shift T is the deterministic tree automaton
C = (V , A, ∆) where
V is the set of non-empty contexts of finite blocks appearing in T
transitions are (contT (u), contT (v )), a → contT (a, u, v ), with u, v ∈ L(T).
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 21 / 24
60. Tree-shifts Tree automata and tree-shifts
The context automaton
The context automaton of a tree-shift T is the deterministic tree automaton
C = (V , A, ∆) where
V is the set of non-empty contexts of finite blocks appearing in T
transitions are (contT (u), contT (v )), a → contT (a, u, v ), with u, v ∈ L(T).
Proposition (A.& B´al 2010)
e
The context automaton of a sofic tree-shift is synchronized.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 21 / 24
61. Tree-shifts Tree automata and tree-shifts
The context automaton
The context automaton of a tree-shift T is the deterministic tree automaton
C = (V , A, ∆) where
V is the set of non-empty contexts of finite blocks appearing in T
transitions are (contT (u), contT (v )), a → contT (a, u, v ), with u, v ∈ L(T).
Proposition (A.& B´al 2010)
e
The context automaton of a sofic tree-shift is synchronized.
Proposition (A.& B´al 2010)
e
The context automaton of a sofic tree-shift T has a unique minimal,
irreducible and synchronized component S, called the Shannon cover of T.
The Shannon cover of a sofic tree-shift is computable.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 21 / 24
62. Tree-shifts On AFT tree-shifts
On AFT tree-shifts
AFT tree-shifts: class in-between tree-SFT and sofic tree-shifts
generalization of AFT 1D-susbhifts, used for coding purposes
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 22 / 24
63. Tree-shifts On AFT tree-shifts
On AFT tree-shifts
AFT tree-shifts: class in-between tree-SFT and sofic tree-shifts
generalization of AFT 1D-susbhifts, used for coding purposes
AFT tree-shifts: factors of tree-SFT, where the factor map satisfies syntactic
properties (right-resolving, left-closing and having a resolving block)
theses properties on the factor map are computable
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 22 / 24
64. Tree-shifts On AFT tree-shifts
On AFT tree-shifts
AFT tree-shifts: class in-between tree-SFT and sofic tree-shifts
generalization of AFT 1D-susbhifts, used for coding purposes
AFT tree-shifts: factors of tree-SFT, where the factor map satisfies syntactic
properties (right-resolving, left-closing and having a resolving block)
theses properties on the factor map are computable
Proposition (A.& B´al 2010)
e
It is decidable to say wether a sofic tree-shift is AFT or not.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 22 / 24
65. Tree-shifts Conjugacy of tree-SFT
The conjuguacy problem for tree-SFT
two subshifts are conjugate if they are both factor of the other
conjugate subshifts are the same, up to a recoding
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 23 / 24
66. Tree-shifts Conjugacy of tree-SFT
The conjuguacy problem for tree-SFT
two subshifts are conjugate if they are both factor of the other
conjugate subshifts are the same, up to a recoding
Theorem (A.& B´al 2010)
e
The conjugacy problem is decidable for tree-SFT.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 23 / 24
67. Tree-shifts Conjugacy of tree-SFT
The conjuguacy problem for tree-SFT
two subshifts are conjugate if they are both factor of the other
conjugate subshifts are the same, up to a recoding
Theorem (A.& B´al 2010)
e
The conjugacy problem is decidable for tree-SFT.
every conjuguacy can be splitted into a sequence of elementary conjuguacies
elementary conjugacies ⇒ unique minimal amalgamation of a tree-SFT
two tree-SFT are conjugate iff they have the same minimal amalgamation
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 23 / 24
68. Conclusion
Conclusion
Many properties on 2D-SFT are uncomputable. . .
. . . nevertheless it is possible to partially describe them.
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 24 / 24
69. Conclusion
Conclusion
Many properties on 2D-SFT are uncomputable. . .
. . . nevertheless it is possible to partially describe them.
Tree-shifts are very similar to N-subshifts. . .
. . . how to characterize monoids with same properties as 1D-subshifts ?
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 24 / 24
70. Conclusion
Conclusion
Many properties on 2D-SFT are uncomputable. . .
. . . nevertheless it is possible to partially describe them.
Tree-shifts are very similar to N-subshifts. . .
. . . how to characterize monoids with same properties as 1D-subshifts ?
Thank you, Спасибо, Kiitos, Merci !
Nathalie Aubrun (LIGM) Dynamique symbolique 2D et des arbres 22 juin 2011 24 / 24