Handout Ohya Velovich P NP

P=NP
Solving SAT with a Chaotic Quantum Computer

Wilson Rosa de Oliveira Jr.
(Based on Ohya&Volovichs Work)
DEInfo-UFRPE
November 19, 2012
Apresentacao no Grupo de Transporte Quantico e
Caos Ondulatorio DF-UFPE
Wilson Rosa (DEInfo-UFRPE) P=NP November 19, 2012 DF-UFPE 1 / 37
Outline:
1
Introduction
2
SAT Problem
3
Chaotic Dynamics
4
Main Results
5
Proofs
Proof Theorem 1
Proof Theorem 2
Comments
6
Conclusion
7
Perspective for further research
Introduction
NP-complete problems: the napsack problem, the traveling salesman
problem, the integer programming problem, the subgraph
isomorphism problem, the satisability problem;
have been studied for decades and for which all known algorithms
have a running time that is exponential in the length of the input.
Any problem that can be solved in polynomial time on a
nondeterministic Turing machine is polynomially transformed to an
NP-complete problem [1].
it seems that such problems are very dicult and probably
exponential.
If so, solutions are still needed, and in this talk we consider an
approach to these problems based on quantum computing and
chaotic dynamics.
P versus NP
Introduction (cont.)
It is widely believed that quantum computers are more ecient than
classical computers.
In particular Shor [2] gave a remarkable quantum polynomial-time
algorithm for the factoring problem.
However, it is unknown whether this problem is NP-complete.
The computational power of quantum computers has been explored in
a number of papers. Bernstein and Vasirani [3] proved that
BPPBQP PSPACE.
The question whether NPBQP, is open.
BPP : class of problems eciently solvable that can be solved in
polynomial time by probabilistic Turing machines with error
probability bounded by 1/3 for all inputs.
BQP : The quantum analogue of BPP, is the class problems that
can be solved in polynomial time by quantum Turing machines with
error probability bounded by 1/3 for all inputs.
BPP, BQP
Introduction (cont.)
For a recent discussion of computational complexity in quantum
computing see [5, 6, 7].
Mathematical features of quantum computing and quantum
information theory are summarized in [8].
A possibility to exploit nonlinear quantum mechanics so that the class
of problems NP may be solved in polynomial time has been
considered by Abrams and Lloyd in [9].
It is mentioned in [9] that such nonlinearity is purely hypotetical; all
known experiments conrm the linearity of quantum mechanics.
Introduction
In this talk we argue that chaotic dynamics plays a constructive role
in computations.
Chaos and quantum decoherence are considered normally as the
degrading eects which lead to an unwelcome increase of the error
rate with the input size.
However, in this talk we argue that under some circumstances chaos
can play a constructive role in computer science.
In particular we show how to combine quantum computer with the
chaotic dynamics amplier.
We will argue that such a chaotic quantum computer can solve the
SAT problem in polynomial time.
As a possible specic implementation of chaotic quantum
computations it has been recently proposed atomic quantum
computer [11].
It has been argued [11] that in the atomic quantum computer one can
build also nonlinear quantum gates
SAT Problem
Let X = x
1
, , x
n
be a set.
Then x
k
and its negation x
k
(k = 1, 2, , n) are called literals
the set of all such literals is denoted by X
= x
1
, x
1
, , x
n
, x
n
.
The set of all subsets of X
is denoted by T(X
) and an element
C T(X
) is called a clause.
We take a truth assignment to all variables x
k
.
If at least one element of C is asssigned true, then C is called
satisable.
When C is satisable, the truth value t(C) of C is regarded as true,
otherwise, that of C is false.
Then
C is satisable i t(C) = 1.
SAT Problem (cont.)
Let L = 0, 1 be a Boolean lattice with usual join and meet ,
and t(x) be the truth value of a literal x in X.
Then the truth value of a clause C can be written as
t(C)
xC
t(x).
A set ( of all clauses C
j
(j = 1, 2, m) is called satisable i the
meet of all truth values of C
j
is 1;
t(()
m
j=1
t(C
j
) = 1.
SAT Problem (cont.)
Denition
SAT Problem: Given a set X x
1
, , x
n
and a set
( =C
1
, C
2
, , C
m
of clauses, determine whether ( is satisable or not.
That is, this problem is to ask whether there exsits a truth assignment
to make ( satisable.
it is polynomial time to check the satisability when a specic truth
assignment is given, but it is not known how to determine the
satisability in polynomial time when an assignment is not specied.
Note that a formula made by the product (AND ) of the disjunction
(OR ) of literals is said to be in the product of sums (POS) form.
For example:
(x
1
x
2
) (x
1
) (x
2
x
3
)
SAT Problem (cont.)
SAT problem can be regarded as determining whether or not a
formula in POS form is satisable.
The following analytical formulation of SAT problem is useful: dene
a family of Boolean polynomials f
A
, indexed by /
/ = S
1
, ..., S
N
, T
1
, ..., T
N
,
where S
i
, T
i
1, ..., n , and f
A
is dened as
f
A
(x
1
, , x
n
) =
N
i=1
1 +
aS
i
(1 x
a
)
bT
i
x
b
.
We assume here the addition modulo 2.
The SAT problem now is to determine whether or not there exists a
value of x = (x
1
, , x
n
) such that f
A
(x) = 1.
Quantum Algorithm
we will work in the (n + 1)-tuple tensor product Hilbert space 1
n+1
1
C
2
with the computational basis
[x
1
, ..., x
n
, y =
n
i=1
[x
i
[y
where x
1
, ..., x
n
, y = 0 or 1.
We denote [x
1
, ..., x
n
, y = [x, y .
The quantum version of the function f(x) := f
A
(x) is given by the
unitary operator U
f
[x, y = [x, y + f(x) .
We assume that the unitary matrix U
f
can be build in the polynomial
time, see [10].
The usual quantum algorithm:
(i) From [0,0 get the superposition
[v :=
1
2
n
x
[x, 0 .
(ii) Use the unitary matrix U
f
to calculate f(x) :
[v
f
= U
f
[v =
1
2
n
x
[x, f(x)
If we measure the last qubit, i.e., apply the projector
P = I [1 1[
to [v
f
, the probability to nd the result f(x) = 1 is
|P [v
f
|
2
= r/2
n
where r is the number of roots of the equation f(x) = 1.
Quantum Algorithm (cont.)
If r is suitably large to detect it, then the SAT problem is solved in
polynominal time.
However, for small r, the probability is very small and this means we
in fact dont get information about the existence of the solution of
f(x) = 1, so we need further work.
Let us simplify our notations. After the step (ii) the quantum
computer will be in the state
[v
f
=
1 q
2
[
0
[0 + q [
1
[1
where [
1
and [
0
are normalized n qubit states and q =
r/2
n
.
So, our problem is reduced to the following 1 qubit problem. We have
the state
[ =
1 q
2
[0 + q [1
and we want to distinguish between the cases q = 0 and q > 0(small
positive number).
Quantum Algorithm (cont.)
The amplication could not be possible in the standard model of
quantum computations with a unitary evolution.
The idea here is to combine quantum computer with a chaotic
dynamics amplier.
Such a quantum chaos computer is a new model of computations
going beyond usual scheme of quantum computation and we
demonstrate that the amplication is possible in the polynomial time.
One could object that we dont suggest a practical realization of the
new model of computations. But at the moment nobody knows of
how to make a practically useful implementation of the standard
model of quantum computing ever.
Quantum circuit or quantum Turing machine is a mathematical model
though convincing one. It seems to us that the quantum chaos
computer deserves an investigation and has a potential to be
realizable.
Chaotic Dynamics
Chaotic behaviour in a classical system usually is considered as an
exponential sensitivity to initial conditions.
It is this sensitivity we would like to use to distinquish between the
cases q = 0 and q > 0.
Consider the logistic map:
x
n+1
= ax
n
(1 x
n
) g(x), x
n
[0, 1] .
The properties of the map depend on the parameter a.
If we take, for example, a = 3.71, then the Lyapunov exponent is
positive, the trajectory is very sensitive to the initial value and one
has the chaotic behaviour.
It is important to notice that if the initial value x
0
= 0, then x
n
= 0
for all n.
Bifurcation Diagram for the iterates of the Logistic Map, x
0
= 0.1
and x
n+1
= rx
n
(1 x
n
), r [0, 4]. With r = 3.71 in red.
Bifurcation Diagram for the Logistic Map: zoomed in r [3.60, 4.00].
With r = 3.71 in red.
Plot of the Lyapunov exponent versus r. With r = 3.71 in red.
Chaotic Dynamics (cont.)
Any classical algorithm can be implemented on quantum computer.
Our quantum chaos computer will be consisting of two blocks. One
block is the ordinary quantum computer performing computations
with the output
[ =
1 q
2
[0 + q [1
.
The second block is a computer performing computations of the
classical logistic map.
This two blocks should be connected in such a way that the state [
is transformed into the density matrix of the form
= q
2
P
1
+
1 q
2
P
0
where P
1
and P
0
are projectors to the state vectors [1 and [0 .
This connection is in fact nontrivial and actually it should be
considered as the third block.
Notice that P
1
and P
0
generate an Abelian algebra which can be
considered as a classical system.
In the second block the density matrix above is interpreted as the
initial data
0
, and we apply the logistic map as
m
=
(I + g
m
(
0
)
3
)
2
where I is the identity matrix and
3
is the z-component of Pauli
matrix on C
2
.
To nd a proper value m we nally measure the value of
3
in the
state
m
such that
M
m
tr
m
3
.
After simple computation we obtain
m
=
(I + g
m
(q
2
)
3
)
2
, and M
m
= g
m
(q
2
).
The question is whether we can nd such a m in polynomial steps of
n satisfying the inequality M
m

1
2
for very small but non-zero q
2
.
Here we have to remark that if one has q = 0 then
0
= P
0
and we
obtain M
m
= 0 for all m. If q ,= 0, the stochastic dynamics leads to
the amplication of the small magnitude q in such a way that it can
be detected as is explained below.
The transition from
0
to
m
is nonlinear and can be considered as a
classical evolution because our algebra generated by P
0
and P
1
is
abelian.
Since g
m
(q
2
) is x
m
of the logistic map x
m+1
= g(x
m
) with x
0
= q
2
,
we use the notation x
m
in the logistic map for simplicity.
Main Results
The amplication can be done within atmost 2n steps due to the
following theorems.
Theorem (1)
For the logistic map x
n+1
= ax
n
(1 x
n
) with a [0, 4] and x
0
[0, 1] ,
let x
0
be
1
2
n
and a set J be 0, 1, 2, , n, 2n . If a is 3.71, then there
exists an integer m in J satisfying x
m
>
1
2
.
Theorem (2)
Let a and n be the same in the above proposition. If there exists m
0
in J
such that x
m
0
>
1
2
, then m
0
>
n1
log
2
3.71
.
Proof Theorem 1
Proof.
Suppose that there does not exist such m in J. Then x
m

1
2
for any
m J. The inequality x
m

1
2
implies
x
m
= 3.71(1 x
m1
)x
m1

3.71
2
x
m1
.
Thus we have
1
2
x
m

3.71
2
x
m1

3.71
2
m
x
0
=
3.71
2
m
1
2
n
,
from which we get
2
n+m1
(3.71)
m
.
Proof Theorem 1
Proof. (cont)
According to the above inequality, we obtain
m
n 1
log
2
3.71 1
.
Since log
2
3.71 1.8912, we have
m
n 1
log
2
3.71 1
<
5
4
(n 1) ,
which is denitely less than 2n 1 and it is contradictory to the statement
x
m

1
2
for any m J. Thus there exists m in J satisfying x
m
>
1
2
.
Proof Theorem 2
Proof.
Since 0 x
m
1, we have
x
m
= 3.71(1 x
m1
)x
m1
3.71x
m1
,
which reduces to
x
m
(3.71)
m
x
0
.
For m
0
in J satisfying x
m
0
>
1
2
, it holds
x
0

1
(3.71)
m
0
x
m
0
>
1
2 (3.71)
m
0
.
Proof Theorem 2
Proof. (cont)
It follows that from x
0
=
1
2
n
log
2
2 (3.71)
m
0
> n,
which implies
m
0
>
n 1
log
2
3.71
.
Comments
According to these propositions, it is enough to check the value x
m
(M
m
) around the above m
0
when q is
1
2
n
for a large n.
More generally, when q=
k
2
n
with some integer k, it is easily checked
that the above two propositions are held and the value x
m
(M
m
)
becomes over
1
2
around the m
0
above.
One can think about various possible implementations of the idea of
using chaotic dynamics for computations and/or realization of
nonlinear quantum gates, which is open problem
In Fig.1 we can see how to amplify the small q in several steps.
n
x
n
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20 25 30 35 40 45 50
Fig.1. Change of x
n
w.r.t. time n
...
Conclusion
The complexity of the quantum algorithm for SAT was shown to be
polynomial time.
It is the probabilistic part of the construction that requires one to
compute several times to be able to distingish the cases q = 0 and
q > 0.
the quantum chaos algorithm is proposed combines the ordinary
quantum algorithm with quantum chaotic dynamics amplier.
The physical implementation of this algorithm is another question ...
Perspective for further research
What is so special about the logistic map?
Vanishes at 0 and amplies 0 < q << 1.
Others chaotic maps?
Physically realizable chaotic maps?
Nonlinear Optics, Random Matrices?
Lorentz maps?
Hyperturing?
What if we replace the quantum circuit part with say
neural networks?
cellular automata?
DiscreteContnuous?
References I
M. Garey and D. Johnson, Computers and Intractability - a guide to
the theory of NP-completeness, Freeman, 1979.
P.W. Shor, Algorithm for quantum computation : Discrete logarithm
and factoring algorithm, Proceedings of the 35th Annual IEEE
Symposium on Foundation of Computer Science, pp.124-134, 1994.
E. Bernstein and U. Vazirani, Quantum Complexity Theory, in: Proc.
of the 25th Annual ACM Symposium on Theory of Comuting, (ACM
Press, New York,1993), pp.11-20.
C. H. Bennett, E. Bernstein, G. Brassard, U. Vazirani,
Strengths and Weaknesses of Quantum Computing, quant-ph/9701001
L. Fortnow and J. Rogers, Complexity Limitations on Quantum
Computation, cs.CC/9811023.
References II
R. Cleve, An Introduction to Quantum Complexity Theory,
quant-ph/9906111.
E. Hemaspaandra, L.A. Hemaspaandra and M. Zimand,
Almost-Everywhere Superiority for Quantum Polynomial Time,
quant-ph/9910033.
M. Ohya, Mathematical Foundation of Quantum Computer, Maruzen
Publ. Company, 1998
D. S. Abrams and S. Lloyd, Nonlinear quantum mechanics implies
polynomial-time solution for NP-complete and #P problems,
quant-ph/9801041.
M. Ohya and N. Masuda, NP problem in Quantum Algorithm,
quant-ph/9809075.
I.V. Volovich, Atomic Quantum Computer, quant-ph/9911062.
References III
M. Ohya, Complexities and Their Applications to Characterization of
Chaos, Int. Journ. of Theoret. Physics, 37 (1998) 495.
S.A. Gardiner, J.I. Cirac and P. Zoller, Phys. Rev. Lett. 79(1997) 4790.
R. Schack, Phys. Rev. A57 (1998) 1634; T. Brun and R. Schack,
quant-ph/9807050.
I. Kim and G. Mahler, Quantum Chaos in Quantum Turing Machine,
quant-ph/9910068.
D. Deutsch, Quantum theory, the Church-Turing principle and the
universal quantum computer, Proc. of Royal Society of London series
A, 400, pp.97-117, 1985.
J.I. Cirac and P. Zoller, Phys. Rev. Lett., 74 (1995) 74.
N.A. Gershenfeld and I.L. Chuang, Science, 275 (1997) 350.
References IV
G. Burkard, D. Loss and D.P. DiVincenzo, cond-mat/9808026.
A. Ekert and R. Jozsa, Quantum computation and Shors factoring
algorithm, Reviews of Modern Physics, 68 No.3,pp.733-753, 1996.
I.I. Sobelman, Atomic Spectra and Radiative Transitions,
Springer-Verlag, 1991.
Accardi, L. and Ohya, M.: Teleportation of general quantum states,
quant-ph/9912087.
Fichtner, K.-H. and Ohya, M.:Quantum Teleportation with Entangled
States given by Beam Splittings, quant-ph/9912083.
Thanks for your attention!

Handout Ohya Velovich P NP

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Handout Ohya Velovich P NP

Uploaded by

Copyright:

Available Formats

P=NP

Solving SAT with a Chaotic Quantum Computer

Vanishes at 0 and amplies 0 < q << 1.

Others chaotic maps?

Physically realizable chaotic maps?

Nonlinear Optics, Random Matrices?

You might also like