This action might not be possible to undo. Are you sure you want to continue?
r
X
i
v
:
q
u
a
n
t

p
h
/
0
1
0
8
0
0
4
v
2
1
9
M
a
r
2
0
0
2
Quantum walks in higher dimensions
T. D. Mackay, S. D. Bartlett, L. T. Stephenson, and B. C. Sanders
Department of Physics, Macquarie University, Sydney, New South Wales 2109, Australia
(Dated: March 19, 2002)
We analyze the quantum walk in higher spatial dimensions and compare classical and quantum
spreading as a function of time. Tensor products of Hadamard transformations and the discrete
Fourier transform arise as natural extensions of the “quantum coin toss” in the one–dimensional walk
simulation, and other illustrative transformations are also investigated. We ﬁnd that entanglement
between the dimensions serves to reduce the rate of spread of the quantum walk. The classical limit
is obtained by introducing a random phase variable.
I. INTRODUCTION
Classical random walks (also known as ‘drunken
walks’) have found practical applications in mathematics,
physics and computational science, for example in stud
ies of diﬀusion, Wiener processes and search algorithms,
respectively. Quantum physics introduces new perspec
tives, such as quantum diﬀusion [1], quantum stochas
tics [2], and quantum walks [3, 4, 5, 6]. The quantum
walk is particularly appealing as an intuitively accessi
ble model underpinning quantum diﬀusion and quantum
stochastics. Remarkable properties of these quantum
walks (QWs) have been discovered; of particular inter
est is that the spread (standard deviation) for the quan
tum walk is proportional to elapsed time t, as opposed
to
√
t for the classical random walk; thus, the QW oﬀers
a quadratic gain over its classical counterpart.
Physical implementations of the quantum walk have
been proposed [3, 7], and possess the attractive property
that they are inherently local in the sense that the spa
tial state shifts by one step along the lattice at each time
step. One potential use of the QW is as a benchmark
for assessing the non–classical performance of a quantum
computer [7]. It is critically important in quantum infor
mation to develop algorithms and processes that behave
in a distinct, observably diﬀerent way than any classical
one; the quantum walk is one such example.
We extend studies of QWs to a higher number of spa
tial dimensions and examine the time dependence of the
standard deviation, which reveals the universal feature
of a quadratic gain over the classical random walk. We
analyze and discuss the eﬀects of entanglement between
the diﬀerent spatial degrees of freedom. We also compare
with the equivalent classical random walk, and obtain the
classical limit from the quantum model via the introduc
tion of a random phase variable at each time step and
performing an ensemble average.
II. THE ONE–DIMENSIONAL QW
The classical random walk in one dimension describes
a particle that moves in the positive or negative direc
tion according to the random outcome of some unbiased
binary variable (e.g., a fair coin). The one–dimensional
lattice on which the particle moves could be inﬁnite or
bounded (as in a circle). We may extend this to a QW
by giving the particle an internal degree of freedom; for
example, the particle may be a spin–1/2 system with
internal Hilbert space H
2
and basis states ±. The spa
tial state of the particle is given by a state in a Hilbert
space H
spatial
of a one–dimensional regular lattice. Let
i, with i an integer, denote the state of a particle lo
cated at position i; the set {i} forms an orthonormal
basis for H
spatial
. The total state of the particle is given
by a state in the tensor product space
H
T
= H
spatial
⊗H
2
. (1)
Let the particle initially be in the spatial state 0 (i.e.,
localized at the origin) with internal state +. To re
alize the 1–D QW [4], this particle is subjected to two
alternating unitary transformations. The ﬁrst step is the
Hadamard transformation [8],
H =
1
√
2
_
1 1
1 −1
_
, (2)
which acts only on the internal state of the particle (i.e.,
on H
2
), and transforms the initial state + into the su
perposition
1
√
2
(++−). Following this transformation,
we apply a unitary operator F that translates the posi
tion of the particle conditionally on the internal state: if
the particle has internal state +, it is moved one unit
to the right, and if the internal state is −, it is moved
to the left, i.e.,
F
_
i ⊗+
_
= i + 1 ⊗+ ,
F
_
i ⊗−
_
= i −1 ⊗− . (3)
The translation does not alter the internal state, i.e., the
states ± are internal translation eigenstates. Since the
transformation is linear, it will transform the superposi
tion state
1
√
2
(+ +−) into a superposition state of the
particle having moved left and right. Thus, the internal
and spatial degrees of freedom become entangled. The
Hadamard transformation is applied again, followed by
F, and these transformations are repeated alternately.
After n iterations, the particle is in an entangled state
Ψ
n
∈ H
T
. The probability P
i
that the particle will be
found at the i
th
location is given by
P
i
=
¸
¸
_
i ⊗+
_
Ψ
n
¸
¸
2
+
¸
¸
_
i ⊗−
_
Ψ
n
¸
¸
2
. (4)
2
−100 −50 0 50 100
0
0.05
0.1
0.15
i
P
i
FIG. 1: The probability distribution of the 1–D quantum walk
after 100 iterations. The internal state transformation used is
the Hadamard transformation, and the initial internal state
is −.
In Fig. 1, we plot the probability distribution of this 1–
D QW as a function of i [4, 5]. Analytical results are
possible for the 1–D QW, and the n → ∞ asymptotic
behaviour has been investigated [4]. A key feature of the
quantum walk is quantum interference, whereby two sep
arate paths between two nodes can interfere according to
the phase diﬀerence. In contrast, the classical model has
additive probabilities for alternate paths. Perhaps most
interesting is the relative uniformity of the central por
tion of the distribution (−25 < i < 25) and the standard
deviation of the distribution increases linearly with the
number of steps t; this result is in contrast to the square
root dependence of the classical random walk. Another
peculiar feature is the asymmetry of the spatial probabil
ity distribution; this asymmetry is a consequence of the
choice of initial state. The distribution resulting from the
initial internal state ψ
s
=
1
√
2
(+ +i−) is symmetric.
III. QW IN HIGHER DIMENSIONS
The analysis of the one–dimensional walk can be ex
tended to higher dimensions. We deﬁne generalizations of
the Hadamard gate, which place the internal state of the
particle in superpositions of internal translation eigen
states, plus a generalization of F, which moves the parti
cle in the d–dimensional space conditional on the internal
state of the particle.
For a QW in d–dimensions, we require the particle to
have an internal state in a 2
d
–dimensional Hilbert space.
This internal state is simply described as the state of
d coupled qubits [8]; thus, we can express the internal
Hilbert space H
int
as
H
int
= H
2
⊗H
2
⊗· · · ⊗ H
2
= ⊗
d
H
2
, (5)
and give a basis for internal states in binary notation as
ǫ
1
ǫ
2
. . . ǫ
d
= ǫ
1
⊗ǫ
2
⊗. . . ⊗ǫ
d
, (6)
where ǫ
i
= ±. The state of the i
th
qubit (with basis
±) will determine the direction (positive or negative)
that the particle moves in the i
th
dimension. That is,
we deﬁne a translation operator F which translates the
state of the particle by one unit in every dimension: the
direction in the i
th
dimension is conditional on the state
of the i
th
qubit. The internal translation eigenstates are
those given in Eq. (6).
For the 1–D QW, the quantum analogue of the classical
“coin–ﬂip” was the application of the Hadamard transfor
mation of Eq. (2). This transformation maps an internal
translation eigenstate of the translation operator F (ei
ther + or −) into an equally weighted superposition of
the two. The choice of phases in this transformation was
to some extent arbitrary; the Hadamard transformation
represents a choice with real entries.
For the d–dimensional QW, there exists a wide vari
ety of unitary transformations on the internal state that
could be used as a generalization of the Hadamard trans
formation for the 1–D case. One obvious generalization
would be to apply a Hadamard transformation Hto each
qubit in the decomposition of Eq. (5); i.e., the transfor
mation
H
d
= H⊗H⊗. . . ⊗H. (7)
This internal transformation is separable, in the sense
that it does not produce entanglement between the spa
tial degrees of freedom. This choice could be viewed as
the quantum analogue of using d independent coin tosses,
one for each spatial dimension.
Another obvious generalization, which is not separa
ble and does produce entanglement between spatial de
grees of freedom, is the 2
d
–dimensional discrete Fourier
transform (DFT) D
d
, deﬁned as follows. Expressing
the basis of Eq. (6) as labelled by its numerical value
{µ, µ = 0, 1, . . . , 2
d
−1}, the DFT acts on this basis as
D
d
µ =
1
√
2
d
2
d
−1
ν=0
e
2πiµν/2
d
ν . (8)
Note that the Hadamard transformation is the d = 1 dis
crete Fourier transform D
1
. As the Hadamard transfor
mation does for the 1–D case, this DFT transforms any
internal translation eigenstate into an equally weighted
superposition of all the eigenstates. Unlike the ten
sor product of Hadamard transformations, it is non–
separable and highly entangles the diﬀerent internal
qubits. Although this internal transformation can also
be viewed as a quantum analogue of d independent coin
tosses, this entanglement between the spatial degrees of
freedom is a genuinely quantum eﬀect.
3
The DFT transformation represents a natural choice
for the phase relationship between the translation eigen
states of the superpositions. However, this choice of
phases is arbitrary, and we may consider other choices,
which will have a diﬀerent eﬀect on the QW. We
also investigate another internal state transformation
(the Grover operator [8]) that also produces an equally
weighted superposition is the transformation (deﬁned on
the same basis as used above)
G
d
µ =
1
√
2
d
_
−2µ +
2
d
−1
ν=0
ν
_
. (9)
This choice, like the Hadamard transformation, possesses
only real entries.
There are, of course, an inﬁnite variety of other non–
separable choices for the internal transformation by em
ploying diﬀerent phase relationships. Also, a bias could
be introduced into the transformation, which would give
an unequally weighted superposition of translation eigen
states; however, we consider only unbiased transforma
tions here.
One of the remarkable properties of the 1–D QW is
that, unlike its classical counterpart, it can produce an
asymmetric distribution. Note, however, that with ap
propriate initial conditions (such as the state ψ
s
=
1
√
2
(+ + i−)) a symmetric distribution is obtained. It
is of interest to question what eﬀect the initial condi
tions will have on the higher–dimensional QWs. (Note
that a symmetric distribution can always be obtained by
averaging over initial conditions.)
IV. CALCULATIONS OF QWS
We begin our analysis with the straightforward gener
alization to higher dimensions of using the Hadamard
transformation on each qubit. Fig. 1 and Fig. 2
show simulation results for the Hadamard walk both in
one–dimension and a tensor product H ⊗ H for two–
dimensions respectively. The initial condition for the
internal state was chosen to be the separable state com
posed of all qubits in the − state, which leads to an
asymmetric probability distribution.
For the case of separable transformations with separa
ble initial conditions, the diﬀerent spatial dimensions be
have independently; thus, the variance can be expressed
in terms of the one–dimensional case. For example, con
sider the family
H, H⊗H, H⊗H⊗H, . . . ; (10)
the time dependence of the standard deviation for these
walks is plotted in Fig. 3, and the corresponding slopes
∆σ/∆t are presented in Table I. We observe that
(
∆σ
1
∆t
,
∆σ
2
∆t
,
∆σ
3
∆t
, . . . ) = (
∆σ
1
∆t
,
√
2
∆σ
1
∆t
,
√
3
∆σ
1
∆t
, . . . ) ,
(11)
−100 −50 0 50 100
−100
−50
0
50
100
X
Y
FIG. 2: Probability distribution of the 2–D quantum walk
using the separable internal transformation H⊗ H over 100
iterations, with initial condition given by − ⊗−.
0 10 20 30 40 50
0
10
20
30
40
Time
S
t
a
n
d
a
r
d
D
e
v
i
a
t
i
o
n
(
σ
)
1−D
2−D
3−D
FIG. 3: Time dependence of standard deviation for the series
H, H⊗H and H⊗H⊗H with initial state given by ⊗
d
−.
where σ
d
is the standard deviation for the d–dimensional
QW, as expected for independent distributions. Also,
by calculating a d–dimensional QW using this separa
ble internal transformation and projecting the state onto
the Hilbert space for any one dimension, the state of
the 1–D QW is recovered. Again, this property illus
trates that the diﬀerent spatial dimensions are indepen
dent. Analytical results for these QWs follow from the
1–D case in a straightforward manner.
We now consider the behaviour of higher–dimensional
QWs that possess entanglement between the spatial de
grees of freedom, i.e., QWs that have non–separable in
4
Transformation ∆σ/∆t (
√
d)∆σ1/∆t
H 0.4544 ±0.0012 0.4544 ±0.0012
H⊗H 0.6427 ±0.0017 0.6427 ±0.0017
H⊗H⊗H 0.7871 ±0.0021 0.7871 ±0.0021
TABLE I: The slope of the standard deviation as a function
of time for the family (H, H ⊗ H, H ⊗ H ⊗ H, . . . ). The
slope ∆σ/∆t is found by linear regression of data points where
t ≥ 10 (such as to allow stabilization of irregularities caused
by initial condition). σ1 refers to the 1–D case.
−100 −50 0 50 100
−100
−50
0
50
100
X
Y
FIG. 4: Probability distribution for the quantum walk using
the d = 2 DFT (D2) over 100 iterations, with initial condition
given by − ⊗−.
ternal transformations, such as the DFT of Eq. (8). Fig. 4
shows the spatial probability distribution of the QW
with internal transformation given by the d = 2 discrete
Fourier transform D
2
; note that this distribution is dis
tinct from that of the H⊗ H QW. In particular, it has
the feature that the density of the distribution is signiﬁ
cant near the origin, in constrast to the separable H⊗H
QW which possesses only average density at the origin.
Note also that it is asymmetric for the initial condition
− ⊗−; the asymmetry appears to be a general prop
erty of the higher–dimensional QWs as it is for the 1–D
case.
The time dependence of the standard deviations for
the d = 1, 2, 3 DFT walks are plotted in Fig. 5 and
summarized in Table II. In contrast to the separable
case, the trend observed in the DFT family is ∆σ
d
/∆t =
_
(d + 1)/2∆σ
1
/∆t. This trend is in agreement with the
three calculations (d = 1, 2, 3). For the family of DFT
walks, the standard deviation grows linearly with time,
but the slope is less than that for the separable case (the
tensor products of Hadamard transformations); see Ta
ble II. This suggests that the entanglement between the
spatial degrees of freedom serves to reduce the rate of
spread.
0 10 20 30 40 50
0
10
20
30
40
Time
S
t
a
n
d
a
r
d
D
e
v
i
a
t
i
o
n
(
σ
)
1−D
2−D
3−D
FIG. 5: Time dependence of the standard deviation for the
D
d
DFT series with initial state given by ⊗
d
−. Details are
given in Table II.
Transformation ∆σ/∆t
_
(d + 1)/2∆σ1/∆t
D1 (H) 0.4544 ±0.0012 0.4544 ±0.0012
D2 0.5569 ±0.0006 0.5565 ±0.0015
D3 0.6449 ±0.0007 0.6426 ±0.0017
TABLE II: Slope of the standard deviation as a function of
time, and comparison to the suggested pattern. ∆σ/∆t is the
slope found by linear regression of data points where t ≥ 10
(such as to allow stabilization of irregularities caused by initial
condition).
Choosing diﬀerent relative phases in the internal state
transformation can lead to vastly diﬀerent distribu
tions. Fig. 6 shows the results of using the internal
transformation G of Eq. (9). This distribution is chara
terized by its marked localization at the centre, as well as
possessing peaks at the “maximum distance” attainable
in the number of iterations (100 units from the origin).
Note that the time dependence of variance for the non–
separable 2–D transformations (D
2
, G
2
) are quite simi
lar, although the probability density functions are quite
diﬀerent in appearance. (See Table III.) The choice of
initial condition does not appear to have a signiﬁcant ef
fect on the time dependence of the standard deviation.
V. OBTAINING THE CLASSICAL RANDOM
WALK FROM THE QUANTUM MODEL
A classical distribution can be obtained from the quan
tum model by introducing a random element into the
5
−100 −50 0 50 100
−100
−50
0
50
100
X
Y
FIG. 6: Probability distribution of the 2–D quantum walk
with internal transformation given by G (see Eq. 9) over 100
iterations with initial condition − ⊗−.
Transformation Initial State ∆σ/∆t
(H⊗H) − ⊗− 0.6427 ±0.0017
D2 − ⊗− 0.5569 ±0.0006
D2 + ⊗+ 0.5569 ±0.0006
D2 ψs ⊗ψs 0.6234 ±0.0005
D2 ψ− 0.6009 ±0.0006
G (see Eq. 9) − ⊗− 0.5418 ±0.0020
G + ⊗+ 0.5418 ±0.0020
G ψs ⊗ψs 0.5988 ±0.0006
G ψ− 0.5440 ±0.0008
TABLE III: Slope of the standard deviation as a function of
time for various 2–D transformations and initial conditions.
∆σ/∆t is the slope found by linear regression of data points
where t ≥ 10 (to allow stabilization of irregularities caused
by initial condition). Here ψs =
1
√
2
(+ + i−) is the state
that produces the 1–D symmetric distribution, and ψ− =
1
√
2
(+ ⊗− −− ⊗+) is the entangled singlet state.
transformation at each time step. As shown previously,
the “quantum” behaviour of the QW is due to the phase
relationship (interference) between the separate paths of
the walk. By adding a random element to the phase and
averaging over many trials, we show that the quantum
inteference can be made to disappear and that the dis
tribution of the classical random walk is regained. The
introduction of this random phase is an example of de
coherence.
Let us ﬁrst investigate the one–dimensional case. In
the internal translation eigenstate basis ±, the unitary
operator that transforms the relative phase between these
states is
R(β) = e
i
2
βˆ σz
=
_
e
iβ/2
0
0 e
−iβ/2
_
, (12)
where ˆ σ
z
=
_
1 0
0 −1
_
is the Pauli spin matrix, and β ∈
[0, 2π). We then consider a QW where the phase between
the + and − states is randomly selected at each inter
val from a uniform prior distribution over [0, 2π). Rather
than applying the Hadamard transformation as the inter
nal transformation, we apply
H(β) = R(β) HR(β)
−1
=
_
1 e
iβ
e
−iβ
−1
_
, (13)
with a phase β chosen randomly from the set [0, 2π) at
each time step. The resulting distribution has a stardard
deviation comparable to that of the corresponding bino
mial distribution, but exhibits strong interference eﬀects.
By averaging oven many trials, the distribution rapidly
converges to the binomial distribution.
These results for the one–dimensional case can easily
be generalized to higher dimensions. For the separable
d–dimensional QW, the generalization is straightforward:
one simply replaces each Hadamard transformation in the
tensor product with a random H(β) at each step. The
separability ensures that the resulting walk is equivalent
to the 1–D walk in each dimension.
For non–separable internal transformations such as the
DFT, a straightforward extension is to apply
(R
1
(β
1
) ⊗· · · ⊗R
d
(β
d
)) D
d
(R
d
(β
d
)
−1
⊗· · · ⊗R
1
(β
1
)
−1
) ,
(14)
where β
1
, . . . , β
d
are random phases, each from the set
[0, 2π). That is, an independent random phase is added
for each dimension (qubit). Again, by averaging 400
walks of 50 iterations each, we obtained the 2–D bino
mial distribution to a high degree of conﬁdence.
VI. CONCLUSIONS AND DISCUSSION
We present here a framework for calculating and an
alyzing quantum walks in higher dimensions. The gen
eralization of these walks beyond one dimension gives
a wide variety of choice for the phases involved in the
“quantum coin toss”. We discuss the role of entangle
ment between the diﬀerent spatial degrees of freedom as
a possible non–classical property of the higher dimen
sional QWs. As diﬀerent choices lead to diﬀerent spatial
probability distributions, it may be that speciﬁc unitary
transformations of the internal Hilbert space are partic
ularly well suited for certain computational tasks.
As with the one–dimensional QW, the increased rate
of spread (given by the linear dependence of the standard
deviation on time) is present in the higher dimensional
walks. This property may be particularly valuable for
6
classical random walk based algorithms, such as quantum
searches. We show that entanglement between the spatial
degrees of freedom reduces the slope of this linear growth
but not the linear dependence on t. These results are
shown to be independent of the initial internal state in
the cases investigated.
We show that the classical distribution can be obtained
from the QW by introducing an internal transformation
with a random phase and then averaging over many tri
als. This result is expected; the quantum behaviour of
the QW is due to interference eﬀects between the phases
of diﬀerent paths. For higher dimensional QW, more
random parameters (one for each spatial dimension) are
needed.
Acknowledgments
This project has been supported by an Australian Re
search Council Large Grant and a Macquarie University
Research Grant. SDB acknowledges the support of a
Macquarie University Research Fellowship. We acknowl
edge helpful discussions with G. J. Milburn, T. E. Free
man, T. Rudolph and B. C. Travaglione.
[1] G. J. Milburn, Phys. Rev. A 33, 674 (1986).
[2] C. W. Gardiner, “Handbook of Stochastic Methods: For
Physics, Chemistry and the Natural Sciences,” Springer
Series in Synergetics, Vol. 13, Springer, Berlin (1996).
[3] Y. Aharonov, L. Davidovich, and N. Zagury, Phys. Rev.
A 48, 1687 (1992).
[4] A. Ambainis, E. Bach, A. Nayak, A. Vishwanath, and J.
Watrous, “One–dimensional quantum walks,” in Proceed
ings of the 33rd Annual Symposium on Theory of Com
puting, ACM Press, New York (2001), p. 37.
[5] D. Aharonov, A. Ambainis, J. Kempe, and U. Vazirani,
“Quantum walks on graphs,” in Proceedings of the 33rd
Annual Symposium on Theory of Computing, ACM Press,
New York (2001), p. 50.
[6] C. Moore and A. Russell, “Quantum Walks on the Hyper
cube,” quantph/0104137, (2001).
[7] B. C. Travaglione and G. J. Milburn, “Implementing the
quantum random walk,” quantph/0109076, (2001).
[8] M. A. Nielsen and I. L. Chuang, “Quantum Computation
and Quantum Information,” Cambridge University Press,
Cambridge (2000).
the Hadamard transformation represents a choice with real entries. the transformation Hd = H ⊗ H ⊗ . Another obvious generalization.05 0 −100 −50 0 i 50 100 FIG. the classical model has additive probabilities for alternate paths.. one for each spatial dimension. QW IN HIGHER DIMENSIONS where ǫi = ±. is the 2d –dimensional discrete Fourier transform (DFT) Dd . (6) as labelled by its numerical value {µ . which is not separable and does produce entanglement between spatial degrees of freedom. this asymmetry is a consequence of the choice of initial state. 2d − 1}. deﬁned as follows. . . in the sense that it does not produce entanglement between the spatial degrees of freedom. . ǫd = ǫ1 ⊗ ǫ2 ⊗ . (5). The internal translation eigenstates are those given in Eq. this DFT transforms any internal translation eigenstate into an equally weighted superposition of all the eigenstates. 1. . d (8) The analysis of the one–dimensional walk can be extended to higher dimensions. The distribution resulting from the 1 initial internal state ψs = √2 (+ + i− ) is symmetric. which moves the particle in the d–dimensional space conditional on the internal state of the particle. Unlike the tensor product of Hadamard transformations. Although this internal transformation can also be viewed as a quantum analogue of d independent coin tosses. . it is non– separable and highly entangles the diﬀerent internal qubits. we require the particle to have an internal state in a 2d –dimensional Hilbert space. ǫ1 ǫ2 . 1: The probability distribution of the 1–D quantum walk after 100 iterations. In Fig. Perhaps most interesting is the relative uniformity of the central portion of the distribution (−25 < i < 25) and the standard deviation of the distribution increases linearly with the number of steps t. . the quantum analogue of the classical “coin–ﬂip” was the application of the Hadamard transformation of Eq.15 Hilbert space Hint as Hint = H2 ⊗ H2 ⊗ · · · ⊗ H2 = ⊗d H2 . (2). For a QW in d–dimensions. The internal state transformation used is the Hadamard transformation. . ⊗ H . we plot the probability distribution of this 1– D QW as a function of i [4. For the 1–D QW. Analytical results are possible for the 1–D QW. . we can express the internal Note that the Hadamard transformation is the d = 1 discrete Fourier transform D1 . the DFT acts on this basis as 1 Dd µ = √ 2d 2d −1 ν=0 e2πiµν/2 ν . This internal state is simply described as the state of d coupled qubits [8].1 Pi (6) 0. 1. thus. This transformation maps an internal translation eigenstate of the translation operator F (either + or − ) into an equally weighted superposition of the two. this entanglement between the spatial degrees of freedom is a genuinely quantum eﬀect. whereby two separate paths between two nodes can interfere according to the phase diﬀerence.e. plus a generalization of F. For the d–dimensional QW. there exists a wide variety of unitary transformations on the internal state that could be used as a generalization of the Hadamard transformation for the 1–D case. ⊗ ǫd . III. One obvious generalization would be to apply a Hadamard transformation H to each qubit in the decomposition of Eq. (6). 5].2 0. i. this result is in contrast to the square root dependence of the classical random walk. As the Hadamard transformation does for the 1–D case. This choice could be viewed as the quantum analogue of using d independent coin tosses. That is. we deﬁne a translation operator F which translates the state of the particle by one unit in every dimension: the direction in the ith dimension is conditional on the state of the ith qubit. We deﬁne generalizations of the Hadamard gate. and the n → ∞ asymptotic behaviour has been investigated [4]. The state of the ith qubit (with basis ± ) will determine the direction (positive or negative) that the particle moves in the ith dimension. . (7) This internal transformation is separable. (5) and give a basis for internal states in binary notation as 0. The choice of phases in this transformation was to some extent arbitrary. which place the internal state of the particle in superpositions of internal translation eigenstates. and the initial internal state is − . Expressing the basis of Eq. µ = 0. . A key feature of the quantum walk is quantum interference. . In contrast. Another peculiar feature is the asymmetry of the spatial probability distribution.
by calculating a d–dimensional QW using this separable internal transformation and projecting the state onto the Hilbert space for any one dimension. . Also. However. . 40 30 20 10 0 0 1−D 2−D 3−D 10 20 30 Time 40 50 FIG. H ⊗ H and H ⊗ H ⊗ H with initial state given by ⊗d − ..e.). and the corresponding slopes ∆σ/∆t are presented in Table I. 2 . 1 and Fig. We observe that ( ∆σ1 ∆σ2 ∆σ3 ∆σ1 √ ∆σ1 √ ∆σ1 .. 2 show simulation results for the Hadamard walk both in one–dimension and a tensor product H ⊗ H for two– dimensions respectively. . One of the remarkable properties of the 1–D QW is that. it can produce an asymmetric distribution.3 The DFT transformation represents a natural choice for the phase relationship between the translation eigenstates of the superpositions. and we may consider other choices. the time dependence of the standard deviation for these walks is plotted in Fig..) = ( . H ⊗ H ⊗ H. Note. For the case of separable transformations with separable initial conditions. the variance can be expressed in terms of the one–dimensional case. i. . For example. the state of the 1–D QW is recovered. 3: Time dependence of standard deviation for the series H. Again. Analytical results for these QWs follow from the 1–D case in a straightforward manner. which would give an unequally weighted superposition of translation eigenstates. (10) Standard Deviation (σ) This choice. however. a bias could be introduced into the transformation. The initial condition for the internal state was chosen to be the separable state composed of all qubits in the − state.) −100 −100 −50 0 X 50 100 FIG. (9) Y −50 IV. Also. (Note that a symmetric distribution can always be obtained by averaging over initial conditions. as expected for independent distributions. with initial condition given by − ⊗ − . ∆t ∆t ∆t ∆t ∆t ∆t (11) where σd is the standard deviation for the d–dimensional QW. which will have a diﬀerent eﬀect on the QW. however. that with appropriate initial conditions (such as the state ψs = 1 √ (+ + i− )) a symmetric distribution is obtained. the diﬀerent spatial dimensions behave independently. unlike its classical counterpart. an inﬁnite variety of other non– separable choices for the internal transformation by employing diﬀerent phase relationships.. consider the family H. H ⊗ H. possesses only real entries. QWs that have non–separable in . It 2 is of interest to question what eﬀect the initial conditions will have on the higher–dimensional QWs. . Fig. We also investigate another internal state transformation (the Grover operator [8]) that also produces an equally weighted superposition is the transformation (deﬁned on the same basis as used above) 1 −2µ + Gd µ = √ 2d 2d −1 ν=0 100 50 0 ν . 2: Probability distribution of the 2–D quantum walk using the separable internal transformation H ⊗ H over 100 iterations.. this property illustrates that the diﬀerent spatial dimensions are independent. . we consider only unbiased transformations here. We now consider the behaviour of higher–dimensional QWs that possess entanglement between the spatial degrees of freedom. 3.. which leads to an asymmetric probability distribution. 3 . like the Hadamard transformation.. CALCULATIONS OF QWS We begin our analysis with the straightforward generalization to higher dimensions of using the Hadamard transformation on each qubit. thus. of course. There are. this choice of phases is arbitrary.
7871 ± 0.0017 −100 −100 −50 0 X 50 100 FIG. (8).0012 0. ternal transformations.4544 ± 0. with initial condition given by − ⊗ − . 30 20 100 10 50 0 Y 0 0 10 20 30 Time 40 50 −50 FIG.0012 0. OBTAINING THE CLASSICAL RANDOM WALK FROM THE QUANTUM MODEL A classical distribution can be obtained from the quantum model by introducing a random element into the .0012 0. V. Note that the time dependence of variance for the non– separable 2–D transformations (D2 . such as the DFT of Eq. and comparison to the suggested pattern.6427 ± 0. . the asymmetry appears to be a general property of the higher–dimensional QWs as it is for the 1–D case. the standard deviation grows linearly with time. 3). Choosing diﬀerent relative phases in the internal state transformation can lead to vastly diﬀerent distributions.7871 ± 0. Details are given in Table II. In particular. 2. This distribution is charaterized by its marked localization at the centre. it has the feature that the density of the distribution is signiﬁcant near the origin. The time dependence of the standard deviations for the d = 1. This suggests that the entanglement between the spatial degrees of freedom serves to reduce the rate of spread. ∆σ/∆t is the slope found by linear regression of data points where t ≥ 10 (such as to allow stabilization of irregularities caused by initial condition). 2.4544 ± 0.4 Transformation H H⊗H H⊗H⊗H ∆σ/∆t 0. but the slope is less than that for the separable case (the tensor products of Hadamard transformations). (9). although the probability density functions are quite diﬀerent in appearance. 4: Probability distribution for the quantum walk using the d = 2 DFT (D2 ) over 100 iterations. Fig. H ⊗ H ⊗ H. (See Table III.0007 0. For the family of DFT walks. 5 and summarized in Table II.5569 ± 0.0006 0.6426 ± 0. see Table II. Fig.4544 ± 0. σ1 refers to the 1–D case. In contrast to the separable case. .0017 0.0012 0.5565 ± 0.6427 ± 0. 4 shows the spatial probability distribution of the QW with internal transformation given by the d = 2 discrete Fourier transform D2 . the trend observed in the DFT family is ∆σd /∆t = (d + 1)/2∆σ1 /∆t. The slope ∆σ/∆t is found by linear regression of data points where t ≥ 10 (such as to allow stabilization of irregularities caused by initial condition). ). 6 shows the results of using the internal transformation G of Eq.) The choice of initial condition does not appear to have a signiﬁcant effect on the time dependence of the standard deviation.0021 40 Standard Deviation (σ) 1−D 2−D 3−D TABLE I: The slope of the standard deviation as a function of time for the family (H. H ⊗ H. G2 ) are quite similar. in constrast to the separable H ⊗ H QW which possesses only average density at the origin.0015 0. Transformation D1 (H) D2 D3 (d + 1)/2∆σ1 /∆t ∆σ/∆t 0. TABLE II: Slope of the standard deviation as a function of time.0017 0. 3 DFT walks are plotted in Fig. This trend is in agreement with the three calculations (d = 1. . 5: Time dependence of the standard deviation for the Dd DFT series with initial state given by ⊗d − . Note also that it is asymmetric for the initial condition − ⊗ − . note that this distribution is distinct from that of the H ⊗ H QW. as well as possessing peaks at the “maximum distance” attainable in the number of iterations (100 units from the origin).0021 √ ( d)∆σ1 /∆t 0.4544 ± 0.6449 ± 0.
. Again. the “quantum” behaviour of the QW is due to the phase relationship (interference) between the separate paths of the walk. .0017 0. the unitary operator that transforms the relative phase between these We present here a framework for calculating and analyzing quantum walks in higher dimensions. For non–separable internal transformations such as the DFT. 6: Probability distribution of the 2–D quantum walk with internal transformation given by G (see Eq. but exhibits strong interference eﬀects. The generalization of these walks beyond one dimension gives a wide variety of choice for the phases involved in the “quantum coin toss”. As diﬀerent choices lead to diﬀerent spatial probability distributions.0020 0.6009 ± 0.0006 0.5 100 states is ˆ R(β) = e 2 β σz = i 50 eiβ/2 0 0 e−iβ/2 . Let us ﬁrst investigate the one–dimensional case.5418 ± 0.5988 ± 0. Transformation (H ⊗ H) D2 D2 D2 D2 G (see Eq. ∆σ/∆t is the slope found by linear regression of data points where t ≥ 10 (to allow stabilization of irregularities caused 1 by initial condition).0020 0. 9) G G G Initial State − ⊗ − − ⊗ − + ⊗ + ψs ⊗ ψs ψ− − ⊗ − + ⊗ + ψs ⊗ ψs ψ− ∆σ/∆t 0. As with the one–dimensional QW. VI. it may be that speciﬁc unitary transformations of the internal Hilbert space are particularly well suited for certain computational tasks. we obtained the 2–D binomial distribution to a high degree of conﬁdence. 2π). CONCLUSIONS AND DISCUSSION TABLE III: Slope of the standard deviation as a function of time for various 2–D transformations and initial conditions.0006 0. The separability ensures that the resulting walk is equivalent to the 1–D walk in each dimension. . . (12) 0 −50 0 where σz = 1 −1 is the Pauli spin matrix. 2π) at each time step. Rather than applying the Hadamard transformation as the internal transformation.6427 ± 0.6234 ± 0. the distribution rapidly converges to the binomial distribution.0005 0. We discuss the role of entanglement between the diﬀerent spatial degrees of freedom as a possible non–classical property of the higher dimensional QWs. 2π). The introduction of this random phase is an example of decoherence. This property may be particularly valuable for . by averaging 400 walks of 50 iterations each. (13) −50 0 X 50 100 FIG. 9) over 100 iterations with initial condition − ⊗ − . (14) where β1 . We then consider a QW where the phase between the + and − states is randomly selected at each interval from a uniform prior distribution over [0. a straightforward extension is to apply (R1 (β1 ) ⊗ · · · ⊗ Rd (βd )) Dd (Rd (βd )−1 ⊗ · · · ⊗ R1 (β1 )−1 ) . βd are random phases. In the internal translation eigenstate basis ± . and β ∈ ˆ 0 [0.0008 with a phase β chosen randomly from the set [0. 2 transformation at each time step. Here ψs = √2 (+ + i− ) is the state that produces the 1–D symmetric distribution. and ψ− = 1 √ (+ ⊗ − − − ⊗ + ) is the entangled singlet state. As shown previously. That is.0006 0. For the separable d–dimensional QW. we show that the quantum inteference can be made to disappear and that the distribution of the classical random walk is regained.5569 ± 0. the increased rate of spread (given by the linear dependence of the standard deviation on time) is present in the higher dimensional walks. an independent random phase is added for each dimension (qubit). By averaging oven many trials.5440 ± 0. The resulting distribution has a stardard deviation comparable to that of the corresponding binomial distribution. the generalization is straightforward: one simply replaces each Hadamard transformation in the tensor product with a random H(β) at each step. These results for the one–dimensional case can easily be generalized to higher dimensions.0006 0. 2π). we apply Y H(β) = R(β) H R(β)−1 = 1 e −iβ −100 −100 eiβ −1 .5569 ± 0.5418 ± 0. each from the set [0. By adding a random element to the phase and averaging over many trials.
L. Rudolph and B.” quantph/0104137. C. T. Nielsen and I. 37. 50.6 classical random walk based algorithms. A. Aharonov. Acknowledgments This project has been supported by an Australian Research Council Large Grant and a Macquarie University Research Grant. Rev. Chemistry and the Natural Sciences. “One–dimensional quantum walks. E.” in Proceedings of the 33rd Annual Symposium on Theory of Computing. Rev. ACM Press. E. Cambridge (2000).” Cambridge University Press. “Quantum walks on graphs. [1] G. [2] C. Freeman. p. A. Gardiner. Vol. Kempe. For higher dimensional QW. . Moore and A. p. We acknowledge helpful discussions with G. Berlin (1996). A. Bach. A 48. “Quantum Walks on the Hypercube. J. Ambainis. Travaglione and G. and U. Milburn. New York (2001). L. Chuang. [5] D. Russell. These results are shown to be independent of the initial internal state in the cases investigated. [8] M.” quantph/0109076. J. ACM Press. Davidovich. Aharonov. New York (2001). the quantum behaviour of the QW is due to interference eﬀects between the phases of diﬀerent paths. such as quantum searches. 674 (1986). Springer. J. This result is expected. A 33. “Implementing the quantum random walk. and N. [4] A. Zagury. Vazirani. We show that entanglement between the spatial degrees of freedom reduces the slope of this linear growth but not the linear dependence on t. 1687 (1992). Nayak. T. (2001). Ambainis. W. (2001). [7] B. “Handbook of Stochastic Methods: For Physics. and J. more random parameters (one for each spatial dimension) are needed. C. Watrous. [3] Y. J.” Springer Series in Synergetics. Milburn. [6] C.” in Proceedings of the 33rd Annual Symposium on Theory of Computing. Phys. We show that the classical distribution can be obtained from the QW by introducing an internal transformation with a random phase and then averaging over many trials. Milburn. 13. A. SDB acknowledges the support of a Macquarie University Research Fellowship. Vishwanath. “Quantum Computation and Quantum Information. Travaglione. Phys.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.