Marked Systems and Circular Splicing

Transcript

Marked Systems and Circular Splicing
Marked Systems and Circular Splicing
Clelia De Felice
Gabriele Fici
Rosalba Zizza
Dipartimento di Informatica ed Applicazioni
Università di Salerno
FCT 2007 – August 27-30, 2007
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
COMPUTING
STANDARD
NATURAL
ALPHABET {0,1}
ALPHABET {A,C,G,T}
CONCATENATION
DNA SPLICING
TURING MACHINES
SPLICING SYSTEMS
CHOMSKY HIERARCHY
???
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Splicing Systems
Splicing Systems (Head 87, Paun 96, Pixton 96):
Generate strings on an alphabet starting from an initial set
through rules:
S = (A, I, R)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Splicing Systems
Splicing Systems (Head 87, Paun 96, Pixton 96):
Generate strings on an alphabet starting from an initial set
through rules:
S = (A, I, R)
Strings in the initial set can be linear, circular or both.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Splicing Systems
Splicing Systems (Head 87, Paun 96, Pixton 96):
Generate strings on an alphabet starting from an initial set
through rules:
S = (A, I, R)
Strings in the initial set can be linear, circular or both.
We deal with finite (i.e. I and R both finite) circular Paun
splicing systems.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Circular words and languages
Conjugacy equivalence on A∗ :
w ∼ w0
⇔
w = xy , w 0 = yx
(x, y ∈ A∗ )
Example: abbc ∼ bcab
A circular word ∼w ∈ ∼A∗ is a conjugacy class.
A circular language is C ⊆ ∼A∗ .
Lin(C) ⊆ A∗ is the set of all linearizations of circular words
in C.
C is circular regular ⇔ Lin(C) is regular.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Circular splicing system
Paun Circular Splicing System:
SC = (A, I, R)
A is the alphabet
I ⊆ ∼A∗ is the initial set
R is the set of rules
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Circular splicing system
Paun Circular Splicing System:
SC = (A, I, R)
A is the alphabet
I ⊆ ∼A∗ is the initial set
R is the set of rules
A rule in R is of the form r = u1 #u2 $u3 #u4 :
∼u hu , ∼u ku
2
1
4
3
generate
Clelia De Felice, Gabriele Fici, Rosalba Zizza
∼u hu u ku
2
1 4
3
(ui , h, k ∈ A∗ )
Marked Systems and Circular Splicing
Circular splicing system
Paun Circular Splicing System:
SC = (A, I, R)
A is the alphabet
I ⊆ ∼A∗ is the initial set
R is the set of rules
A rule in R is of the form r = u1 #u2 $u3 #u4 :
∼u hu , ∼u ku
2
1
4
3
generate
∼u hu u ku
2
1 4
3
(ui , h, k ∈ A∗ )
The words u1 u2 and u3 u4 are called the SITES of the rule r
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Circular splicing system
Paun Circular Splicing System:
SC = (A, I, R)
A is the alphabet
I ⊆ ∼A∗ is the initial set
R is the set of rules
A rule in R is of the form r = u1 #u2 $u3 #u4 :
∼u hu , ∼u ku
2
1
4
3
generate
∼u hu u ku
2
1 4
3
(ui , h, k ∈ A∗ )
The words u1 u2 and u3 u4 are called the SITES of the rule r
Example
r = a#1$cb#b
∼ba, ∼bacb
Clelia De Felice, Gabriele Fici, Rosalba Zizza
`r
∼babacb
Marked Systems and Circular Splicing
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
CIRCULAR SPLICING
Additional hypotheses
R is reflexive:
u1 #u2 $u3 #u4 ∈ R ⇒ u1 #u2 $u1 #u2 , u3 #u4 $u3 #u4 ∈ R
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Additional hypotheses
R is reflexive:
u1 #u2 $u3 #u4 ∈ R ⇒ u1 #u2 $u1 #u2 , u3 #u4 $u3 #u4 ∈ R
R is symmetric:
u1 #u2 $u3 #u4 ∈ R
Clelia De Felice, Gabriele Fici, Rosalba Zizza
⇒
u3 #u4 $u1 #u2 ∈ R
Marked Systems and Circular Splicing
Additional hypotheses
R is reflexive:
u1 #u2 $u3 #u4 ∈ R ⇒ u1 #u2 $u1 #u2 , u3 #u4 $u3 #u4 ∈ R
R is symmetric:
u1 #u2 $u3 #u4 ∈ R
⇒
u3 #u4 $u1 #u2 ∈ R
Self-splicing:
∼
hu1 u2 ku3 u4
` u1 #u2 $u3 #u4
Clelia De Felice, Gabriele Fici, Rosalba Zizza
∼
hu1 u2 , ∼ku3 u4
Marked Systems and Circular Splicing
Additional hypotheses
R is reflexive:
u1 #u2 $u3 #u4 ∈ R ⇒ u1 #u2 $u1 #u2 , u3 #u4 $u3 #u4 ∈ R
R is symmetric:
u1 #u2 $u3 #u4 ∈ R
⇒
u3 #u4 $u1 #u2 ∈ R
Self-splicing:
∼
hu1 u2 ku3 u4
` u1 #u2 $u3 #u4
∼
hu1 u2 , ∼ku3 u4
Remark
We can assume that R is symmetric (see the definition of
splicing)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
The language generated by a Splicing System
Definition
The language generated by a circular splicing system
S = (A, I, R) is the smallest circular language on A containing I
and closed under application of the rules in R.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
The language generated by a Splicing System
Definition
The language generated by a circular splicing system
S = (A, I, R) is the smallest circular language on A containing I
and closed under application of the rules in R.
The class of languages generated by finite circular Paun
splicing systems is denoted by C(Fin, Fin).
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Computational power
Theorem (Head, Paun, Pixton – 96)
I ∈ Reg ∼, R finite reflexive, self-splicing ⇒ L(I, R) ∈ Reg ∼
(Thus: using additional hypotheses C(Fin, Fin) ⊆ Reg ∼)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Computational power
Theorem (Head, Paun, Pixton – 96)
I ∈ Reg ∼, R finite reflexive, self-splicing ⇒ L(I, R) ∈ Reg ∼
(Thus: using additional hypotheses C(Fin, Fin) ⊆ Reg ∼)
Without additional hypotheses:
∼an b n
∈ C(Fin, Fin)
(Siromoney, Subramanian, Dare – 92)
∼((aa)∗ b)
∈
/ C(Fin, Fin)
(Bonizzoni, De Felice, Mauri, Zizza – 03)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Computational power
Theorem (Head, Paun, Pixton – 96)
I ∈ Reg ∼, R finite reflexive, self-splicing ⇒ L(I, R) ∈ Reg ∼
(Thus: using additional hypotheses C(Fin, Fin) ⊆ Reg ∼)
Without additional hypotheses:
∼an b n
∈ C(Fin, Fin)
(Siromoney, Subramanian, Dare – 92)
∼((aa)∗ b)
∈
/ C(Fin, Fin)
(Bonizzoni, De Felice, Mauri, Zizza – 03)
C(Fin, Fin) ⊆ CS ∼
(Fagnot – 04)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Our problem
Problem
Characterize Reg ∼ ∩ C(Fin, Fin)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Our problem
Problem
Characterize Reg ∼ ∩ C(Fin, Fin)
Solved if |A| = 1. Moreover Reg ∼ ∩ C(Fin, Fin) = C(Fin, Fin)
(Bonizzoni, De Felice, Mauri, Zizza – 04,05)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Our problem
Problem
Characterize Reg ∼ ∩ C(Fin, Fin)
Solved if |A| = 1. Moreover Reg ∼ ∩ C(Fin, Fin) = C(Fin, Fin)
(Bonizzoni, De Felice, Mauri, Zizza – 04,05)
Partial results if |A| > 1:
Theorem (Bonizzoni, De Felice, Mauri, Zizza – 04)
If X ∗ is a cycle closed star language (ex. X regular group code
or X finite with X ∗ closed under conjugacy) then
∼X ∗ ∈ Reg ∼ ∩ C(Fin, Fin)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
CSSH
Definition (Ceterchi, Martin-Vide, Subramanian – 04)
A (1, 3)-Circular Semi-simple Splicing System is a finite Paun
circular splicing system in which the rules have the form
(a#1$b#1)
a, b ∈ A
To shorten notation we write the rule above
Clelia De Felice, Gabriele Fici, Rosalba Zizza
(a, b)
Marked Systems and Circular Splicing
CSSH
Definition (Ceterchi, Martin-Vide, Subramanian – 04)
A (1, 3)-Circular Semi-simple Splicing System is a finite Paun
circular splicing system in which the rules have the form
(a#1$b#1)
a, b ∈ A
To shorten notation we write the rule above
(a, b)
So:
∼ha, ∼kb
`(a,b)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
∼hakb
(h, k ∈ A∗ )
Marked Systems and Circular Splicing
Marked Systems
A Marked System is a (1, 3)-CSSH system with
I = SITES(R) = A.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Marked Systems
A Marked System is a (1, 3)-CSSH system with
I = SITES(R) = A.
Example
I = {a, b, c}, R = {(a, b), (b, c), (c, c)}
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Marked Systems
A Marked System is a (1, 3)-CSSH system with
I = SITES(R) = A.
Example
I = {a, b, c}, R = {(a, b), (b, c), (c, c)}
Example
I = {a, b, c}, R = {(a, b), (c, c)}
The first one is transitive (all letters are "linked" by rules).
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Marked Systems
A Marked System is a (1, 3)-CSSH system with
I = SITES(R) = A.
Example
I = {a, b, c}, R = {(a, b), (b, c), (c, c)}
Example
I = {a, b, c}, R = {(a, b), (c, c)}
The first one is transitive (all letters are "linked" by rules).
Proposition
Every marked system admits a canonical decomposition in
transitive marked subsystems.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Distance and Diameter
The distance between two letters ai , aj is 1+ the length of the
shortest path in R linking ai and aj .
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Distance and Diameter
The distance between two letters ai , aj is 1+ the length of the
shortest path in R linking ai and aj .
The diameter of a Marked System is the maximum value of the
distance between two different letters (or 2 if |I| = 1).
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Distance and Diameter
The distance between two letters ai , aj is 1+ the length of the
shortest path in R linking ai and aj .
The diameter of a Marked System is the maximum value of the
distance between two different letters (or 2 if |I| = 1).
Example
I = {a, b, c}, R = {(a, b), (b, c), (c, c)}
d(a, c) = 3 d(a, b) = 2 d(a, a) = 3 d(c, c) = 2
Clelia De Felice, Gabriele Fici, Rosalba Zizza
d(S) = 3
Marked Systems and Circular Splicing
Distance and Diameter
The distance between two letters ai , aj is 1+ the length of the
shortest path in R linking ai and aj .
The diameter of a Marked System is the maximum value of the
distance between two different letters (or 2 if |I| = 1).
Example
I = {a, b, c}, R = {(a, b), (b, c), (c, c)}
d(a, c) = 3 d(a, b) = 2 d(a, a) = 3 d(c, c) = 2
d(S) = 3
Theorem
If d(S) < 3 then L(S) ∈ Reg ∼. If d(S) > 3 then L(S) ∈
/ Reg ∼.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity when d(S) = 3
Regularity Condition
Let S = (I, R) be a marked system. S satisfies the Regularity
Condition if ∀ J = {a1 , a2 , a3 , a4 } ⊆ I one has
R ∩ (J × J) 6= {(a1 , a2 ), (a2 , a3 ), (a3 , a4 )}
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity when d(S) = 3
Regularity Condition
Let S = (I, R) be a marked system. S satisfies the Regularity
Condition if ∀ J = {a1 , a2 , a3 , a4 } ⊆ I one has
R ∩ (J × J) 6= {(a1 , a2 ), (a2 , a3 ), (a3 , a4 )}
Theorem
L(S) is regular
⇔
S satisfies the Regularity Condition.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity when d(S) = 3
Regularity Condition
Let S = (I, R) be a marked system. S satisfies the Regularity
Condition if ∀ J = {a1 , a2 , a3 , a4 } ⊆ I one has
R ∩ (J × J) 6= {(a1 , a2 ), (a2 , a3 ), (a3 , a4 )}
Theorem
L(S) is regular
⇔
S satisfies the Regularity Condition.
Moreover, if L(S) is regular we can characterize it:
[
L(S) = I ∪
J⊆I, J
∼
(∩ai ∈J J ∗ ai J ∗ )
transitive
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity when d(S) = 3
Regularity Condition
Let S = (I, R) be a marked system. S satisfies the Regularity
Condition if ∀ J = {a1 , a2 , a3 , a4 } ⊆ I one has
R ∩ (J × J) 6= {(a1 , a2 ), (a2 , a3 ), (a3 , a4 )}
Theorem
L(S) is regular
⇔
S satisfies the Regularity Condition.
Moreover, if L(S) is regular we can characterize it:
L(S) = {w ∈ I + : alph(w) ⊆ I is transitive}
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity
Let S = (I, R) be a marked system.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity
Let S = (I, R) be a marked system.
We compute the canonical decomposition: S = ∪i Si
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity
Let S = (I, R) be a marked system.
We compute the canonical decomposition: S = ∪i Si
For each Si we compute its diameter d(Si )
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity
Let S = (I, R) be a marked system.
We compute the canonical decomposition: S = ∪i Si
For each Si we compute its diameter d(Si )
If for each i one has:
d(Si ) < 3
or
d(Si ) = 3 + Reg.Cond.
then L(S) = ∪i L(Si ) is regular
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity
Let S = (I, R) be a marked system.
We compute the canonical decomposition: S = ∪i Si
For each Si we compute its diameter d(Si )
If for each i one has:
d(Si ) < 3
or
d(Si ) = 3 + Reg.Cond.
then L(S) = ∪i L(Si ) is regular
Else L(S) is not regular
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity
Let S = (I, R) be a marked system.
We compute the canonical decomposition: S = ∪i Si
For each Si we compute its diameter d(Si )
If for each i one has:
d(Si ) < 3
or
d(Si ) = 3 + Reg.Cond.
then L(S) = ∪i L(Si ) is regular
Else L(S) is not regular
If L(S) is regular we can give an algebraic characterization of
its structure
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Remark
Remark
Let A be an alphabet. There exists a finite number of possible
Marked Systems over A (each one coming with its canonical
decomposition).
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Remark
Remark
Let A be an alphabet. There exists a finite number of possible
Marked Systems over A (each one coming with its canonical
decomposition).
So, given a regular circular language C over an alphabet A we
can test if a Marked System S exists such that C = L(S).
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Self Splicing
The self-splicing operation:
∼hu
1 u2 ku3 u4
`u1 #u2 $u3 #u4
Clelia De Felice, Gabriele Fici, Rosalba Zizza
∼hu
1 u2 ,
∼ku
Marked Systems and Circular Splicing
3 u4
Self Splicing
The self-splicing operation:
∼hakb
`(a,b)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
∼ha,
∼kb
Marked Systems and Circular Splicing
Self Splicing
The self-splicing operation:
∼hakb
`(a,b)
∼ha,
∼kb
Theorem
Let S = (I, R) be a transitive Marked System with self-splicing.
Then
L(S) = ∼I +
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Self Splicing
The self-splicing operation:
∼hakb
∼ha,
`(a,b)
Theorem
Let S = (I, R) be a
Then
∼kb
Marked System with self-splicing.
[
L(S) =
∼ +
J
J⊆I, J transitive
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity with self-splicing
Let S = (I, R) be a Marked System with self-splicing.
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity with self-splicing
Let S = (I, R) be a Marked System with self-splicing.
The language L(S) generated by S is always regular
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Regularity with self-splicing
Let S = (I, R) be a Marked System with self-splicing.
The language L(S) generated by S is always regular
We can give an algebraic characterization of L(S)
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing
Thank you for your attention
Clelia De Felice, Gabriele Fici, Rosalba Zizza
Marked Systems and Circular Splicing