Download Decidability of Monadic Theories

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
Decidability of Monadic Theories
Wolfgang Thomas
Edinburgh, October 2014
Some Undecidability Results
Wolfgang Thomas
Background: Theories of Arithmetic
The FO-theory of (N, +, ·, 0, 1) is undecidable.
Presburger arithmetic, the FO-theory of (N, +, 0, 1), is
decidable.
Skolem arithmetic, the FO-theory of (N, ·, 0, 1) is
decidable.
“Büchi’s arithmetic”, the MSO-theory of (N, +1, 0), is
decidable.
What about the MSO-theory of (N, +, 0, 1)?
We show that even the special case of x 7→ x + x gives
undecidability.
Wolfgang Thomas
Recall
Adding relation quantifiers to the FO-language of (N, +1, 0)
gives undecidability:
x + y = z iff each binary relation that contains (0, x) and is
closed under (m, n) 7→ (m + 1, n + 1) contains also (y, z)
x · y = z iff each binary relation that contains (0, 0) and is
closed under (m, n) 7→ (m + 1, n + x) contains also (y, z)
Wolfgang Thomas
Double Function
R.M. Robinson 1958:
The MSO-theory of (N, +1, 2x, 0) is undecidable.
We follow a proof idea of Elgot and Rabin (1966).
Idea: Code binary relations by sets and simulate relation
quantifiers by set quantifiers.
R = {(m0 , n0 ), (m1 , n1 ), . . .} 7→
MR = {m0′ < n0′ < m1′ < n1′ . . .}.
(m, n) ∈ R iff
in the MR -list there are successive codes m′ , n′ of m, n where
m′ sits on an odd position.
Wolfgang Thomas
Defining the Code
For each x we need an infinite set Cx of code numbers.
such that from z ∈ Cx the number x can be retrieved (by an
MSO-formula).
Set Cx := {2i · (2x + 1) | i ≥ 0}.
From a number z = 2i · (2x + 1) obtain x:
iteratively divide by 2
if this is no more possible subtract 1 and divide by 2
Obtain an MSO-formula for “z is a code of x”
Wolfgang Thomas
Details
Let ϕ2 (y, y′ ) := y′ is half of y
Then
“z is a code of x”: ∃s(ϕ∗2 (z, s) ∧ s = 2x + 1)
Translation of ∃ R( R( x, y) . . .):
∃ X (∃z∃z′ (z is code of x ∧ z′ is code of y
∧OddPos( X, z) ∧ Next( X, z, z′ ))
Wolfgang Thomas
A Sharper Result
Let f : N → N be
strictly increasing,
f − idN be monotone and unbounded.
Then MTh(N, +1, 0, f ) is undecidable.
[W. Th., A note on undecidable extensions of monadic second order
arithmetic, Arch math. Logik 17 (1975)]
Wolfgang Thomas
The Infinite Grid
The infinite grid is the structure
G2 = (N × N, (0, 0), S1 , S2 )
where S1 (i, j) = (i + 1, j), S2 (i, j) = (i, j + 1)
Wolfgang Thomas
Undecidability of Monadic Grid-Theory
The monadic second-order theory of the infinite grid is
undecidable.
Wolfgang Thomas
Another (Direct) Proof
by reduction of the halting problem for Turing machines:
For any TM M construct a sentence ϕ M of the monadic
second-order language of G2 such that
M halts when started on the empty tape iff G2 |= ϕ M .
Wolfgang Thomas
Configurations of M
Assume that M works on a left-bounded tape.
A halting computation of M can be coded by a finite sequence
of configuration words
C0 , C1 , . . . , Cm .
We can arrange the configurations row by row in a
right-infinite rectangular array:
q0
a1
q0
a3
etc.
Wolfgang Thomas
a0
q1
a1
q2
a0
a0
a2
a2
a0
a0
a0
a0
a0
a0
a0
a0
a0
a0
a0
a0
a0
a0
a0
a0
...
...
...
...
Describing an M -Run
The sentence ϕ M will express over G2 the existence of such an
array of configurations.
a0 , . . . , an are the tape symbols (a0 is the blank)
q0 , . . . , qk are the states of M , special halting state qs
We use set variables X0 , . . . , Xn , Y0 , . . . , Yk
Xi collects the grid positions where ai occurs,
Yi collects the grid positions where state qi occurs.
ϕ M : ∃ X0 , . . . , Xn , Y0 , . . . , Yk (Partition( X0 , . . . , Yk )
∧ “the first row is the initial M -configuration”
∧ “a successor row is the successor configuration of the
preceding one”
∧ “at some position the halting state is reached”)
Wolfgang Thomas
Use of Interpretations
An interpretation of a structure A = ( A, RA , . . .) in a structure
B is a description of A in B
Here we use MSO for the description and speak of
MSO-interpretations.
Assume A is MSO-interpretable in B .
Then:
MTh(A) undecidable implies MTh(B) undecidable.
MTh(B) decidable implies MTh(A) decidable.
Wolfgang Thomas
Interpretations Formally
An MSO-interpretation of a structure A = ( A, RA , . . .) in a
structure B is given by
a “domain formula” ϕ( x)
for each relation RA of A, say of arity m, an MSO-formula
ψ ( x1 , . . . , x m )
such that A is isomorphic to (ϕB , ψB , . . .)
Then there is a transformation of MSO-sentences χ (in the
signature of A) to sentences χ′ (in the signature of B ) such
that
A |= χ iff B |= χ′ .
Wolfgang Thomas
Adding the Equal Level Relation
Consider the binary tree with equal level relation E
E(u, v)
:⇔
|u| = |v|. Obtain ( T2 , E).
The MSO-theory of ( T2 , E) is undecidable.
Proof by describing the grid in ( T2 , E):
Wolfgang Thomas
A Hidden Grid
Consider the expansion of the tree T2 by the two
first-letter-adding functions:
p0 (w) = 0 · w,
p1 ( w ) = 1 · w
The MSO-theory of ( T2 , p0 , p1 ) is undecidable.
Proof: Give interpretation of G2 in ( T2 , p0 , p1 )
Domain formula, using σi (z, z′ ) : zi = z′ (i = 0, 1)
ϕ( x) : ∃y(σ0∗ (ε, y) ∧ σ1∗ (y, x))
ψ1 ( x, y) : p0 ( x) = y, ψ2 ( x, y) : x1 = y
Wolfgang Thomas
Decidability Results
Wolfgang Thomas
A First Example
Show Rabin’s Tree Theorem for T3 = ({0, 1, 2}∗ , S03 , S13 , S23 ).
Idea: Describe T3 in T2 :
Consider the T2 -vertices in (10 + 110 + 1110)∗ .
Wolfgang Thomas
Another Interpretation
The MSO-theory of (Q, <) is decidable. (Rabin 1969)
Work with the tree nodes w01
With the lexicographic order they give a countable dense
linear order.
This order is isomorphic to (Q, <) (Cantor)
So we have an MSO-interpretation of (Q, <) in T2 .
Much more difficult: MTh(R, <) is undecidable. (Shelah 1975)
Wolfgang Thomas
A Pushdown Graph
q0 ⊥
q1 ⊥
Wolfgang Thomas
a
b
b
q0 Z ⊥
q1 Z ⊥
a
b
b
q0 ZZ ⊥
q1 ZZ ⊥
a ...
b
b ...
Describing a pushdown graph in a tree
ε
⊥
Z⊥
ZZ ⊥
q0 ZZ ⊥
(q0 , Z, ZZ, q1 )
Wolfgang Thomas
ZZZ ⊥
q1 ZZZ ⊥
MSO-theory of a pushdown graph
Each pushdown graph can be obtained from a tree Tk by an
MSO-interpretation.
The MSO-theory of a pushdown graph is decidable.
(Muller Schupp 1985)
Wolfgang Thomas
Unfoldings
Given a graph (V , ( E a ) a∈Σ , (Pb )b∈Σ′ )
the unfolding of G from a given vertex v0 is the following tree
TG (v0 ) = (V ′ , ( E′a ) a∈Σ , (Pb′ )b∈Σ′ ):
V ′ consists of the vertices v0 a1 v1 . . . ar vr with
(vi−1 , vi ) ∈ E ai ,
E′a contains the pairs (v0 a1 v1 . . . ar vr , v0 a1 v1 . . . ar vr av)
with (vr , v) ∈ E a ,
Pb′ the vertices v0 a1 v1 . . . ar vr with vr ∈ Pb .
Wolfgang Thomas
Example
a
•
b
•
a
/
•
c
•
Wolfgang Thomas
•
b
•
a
/
•
c
•
a
/
•
c
•
a
/
•
c
•
a
/
•
c
•
a
/
•
c
•
a /
···
···
Andrej Muchnik (1958 - 2007)
If the MSO-theory of a graph G is decidable and v0 an
MSO-definable vertex of G, then also the MSO-theory of the
unfolding of G from v0 is decidable.
Wolfgang Thomas
Proof Architecture (for Pushdown Graphs)
Given an unfolding T of a pushdown graph G.
T is finitely branching, with labels say in Σ inherited from G.
For each MSO-formula ϕ( X1 , . . . , Xn ) find a parity tree
automaton Aϕ such that
Aϕ accepts T (P1 , . . . , Pn ) iff T (P1 , . . . , Pn ) |= ϕ( X1 , . . . , Xn )
The construction of the Aϕ follows precisely the pattern of
Rabin’s equivalence theorem.
Essential: In the complementation step we use the finite
out-degree of G.
For sentences, one checks acceptance of the unlabelled
unfolding, which is done by invoking the MSO-theory of G
Wolfgang Thomas
Caucal’s Proposal
We have now two processes which preserve decidability of
MSO-theory:
interpretation (transforming a tree into a graph)
unfolding (transforming a graph into a tree)
Let us apply them in alternation!
We obtain the Caucal hierarchy or pushdown hierarchy.
Wolfgang Thomas
Definition
T0 = the class of finite trees
Gn = the class of graphs which are MSO-interpretable in a
tree of Tn
Tn+1 = the class of unfoldings of graphs in Gn
Each structure in the pushdown hierarchy has a decidable
MSO-theory.
Nontrivial fact:
The sequence G0 , G1 , G2 , . . . is strictly increasing.
Wolfgang Thomas
The First Levels
G0 is the class of finite graphs.
T1 contains the regular trees.
G1 contains the prefix-recognizable graphs.
Wolfgang Thomas
Towards Factorial Predicate
(N, Succ, Fac)
We start with the following pushdown graph,
obtained from a finite tree by an interpretation, an unfolding,
and another interpretation:
•
b
o c
•
Wolfgang Thomas
a
/
•
b
o
•
a
c
/
•
b
a
o c
•
/
•
b
•o
a /
c
···
···
Continuation: Unfolding and Interpretation
•
b
•
•
b
•
c
•
•
b
•
•
b
•
c
c
•
c
•
c
•
•
c
•
Wolfgang Thomas
···
Obtaining (N, +1, 0, Fac)
0
1
• a •
b
•
a
b
• c
•
Wolfgang Thomas
2
•
3
•
a
b
•
•
• c
a
b
• c
•
• c
• • c
• • c
•
•
• •
• •
•
• c
• c
• c
•
On Structures (N, +1, 0, P ) with P ⊆ N
1. (N, +1, 0, P ) belongs to G1 iff P is ultimately periodic.
2. (N, +1, 0, P ) belongs to G2 iff P is morphic.
What about P √2 , Pπ , P?
Wolfgang Thomas
An Example
2
Pexp := set of iterated 2-powers 1, 2, 22 , 22 , etc.
1. (N, +1, 0, Pexp ) is outside the pushdown hierarchy.
2. MSO-Th(N, +1, 0, Pexp ) is decidable.
Method for 2:
The MSO-theory of (N, +1, 0, P ) is decidable iff the following
“acceptance problem for Büchi automata” is decidable:
Given a Büchi automaton A over {0, 1}, does A accept αP ?
(Here αP is the 0-1-sequence with α(i ) = 1 iff i ∈ P.)
Wolfgang Thomas
Outlook
The pushdown hierarchy is a very rich class of structures all of
which have a decidable MSO-theory.
Some open questions:
Understand which structures belong to the hierarchy.
Compute the smallest level on which a structure occurs.
Find ways to extend the hierarchy while keeping
decidability of the MSO-theory.
How to obtain a richer class of models when restricting to
a weaker logic, e.g. monadic transitive logic?
Wolfgang Thomas
Related documents