Professional Documents
Culture Documents
P300-Rieffel PDF
P300-Rieffel PDF
P300-Rieffel PDF
ELEANOR RIEFFEL
FX Palo Alto Laboratory
AND
WOLFGANG POLAK
Authors’ address: E. Rieffel, FX Palo Alto Laboratory, 3400 Hillview Av., Palo Alto, CA 94304; W. Polak,
Consultant.
Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted
without fee provided that copies are not made or distributed for profit or direct commercial advantage and
that copies show this notice on the first page or initial screen of a display along with the full citation.
Copyrights for components of this work owned by others than ACM must be honored. Abstracting with
credit is permitted. To copy otherwise, to republish, to post on servers, to redistribute to lists, or to use any
component of this work in other works, requires prior specific permission and/or a fee. Permissions may
be requested from Publications Dept, ACM Inc., 1515 Broadway, New York, NY 10036 USA, fax +1 (212)
869-0481, or permissions@acm.org.
°2001
c ACM 0360-0300/01/0900-0000 $5.00
ACM Computing Surveys, Vol. 32, No. 3, September 2000, pp. 300–335.
Introduction to Quantum Computing 301
Additional interest in the subject has been the quantum state so that a common
created by the invention of quantum key property of all of the output values such
distribution and, more recently, popular as the symmetry or period of a function
press accounts of experimental successes can be read off. This technique is used
in quantum teleportation and the demon- in Shor’s factorization algorithm. Another
stration of a 3-bit quantum computer. technique transforms the quantum state
The aim of this paper is to guide com- to increase the likelihood that output of
puter scientists and other nonphysicists interest will be read. Grover’s search algo-
through the conceptual and notational rithm makes use of such an amplification
barriers that separate quantum comput- technique. This paper describes quantum
ing from conventional computing and to parallelism in detail, and the techniques
acquaint them with this new and excit- currently known for harnessing its power.
ing field. It is important for the computer Section 2, following this introduction,
science community to understand these explains of the basic concepts of quan-
new developments since they may radi- tum mechanics that are important for
cally change the way we have to think quantum computation. This section can-
about computation, programming, and not give a comprehensive view of quan-
complexity. tum mechanics. Our aim is to provide the
Classically, the time it takes to do cer- reader with tools in the form of mathemat-
tain computations can be decreased by ics and notation with which to work with
using parallel processors. To achieve an the quantum mechanics involved in quan-
exponential decrease in time requires an tum computation. We hope that this paper
exponential increase in the number of will equip readers well enough that they
processors, and hence an exponential in- can freely explore the theoretical realm of
crease in the amount of physical space quantum computing.
needed. However, in quantum systems the Section 3 defines the quantum bit, or
amount of parallelism increases exponen- qubit. Unlike classical bits, a quantum bit
tially with the size of the system. Thus, can be put in a superposition state that en-
an exponential increase in parallelism re- codes both 0 and 1. There is no good classi-
quires only a linear increase in the amount cal explanation of superpositions: a quan-
of physical space needed. This effect is tum bit representing 0 and 1 can neither
called quantum parallelism [Deutsch and be viewed as “between” 0 and 1 nor can
Jozsa 1992]. it be viewed as a hidden unknown state
There is a catch, and a big catch at that represents either 0 or 1 with a certain
that. While a quantum system can per- probability. Even single quantum bits en-
form massive parallel computation, access able interesting applications. We describe
to the results of the computation is re- the use of a single quantum bit for secure
stricted. Accessing the results is equiva- key distribution.
lent to making a measurement, which dis- But the real power of quantum compu-
turbs the quantum state. This problem tation derives from the exponential state
makes the situation, on the face of it, seem spaces of multiple quantum bits: just as a
even worse than the classical situation; we single qubit can be in a superposition of
can only read the result of one parallel 0 and 1, a register of n qubits can be in
thread, and because measurement is prob- a superposition of all 2n possible values.
abilistic, we cannot even choose which one The “extra” states that have no classical
we get. analog and lead to the exponential size of
But in the past few years, various peo- the quantum state space are the entangled
ple have found clever ways of finessing states, like the state leading to the famous
the measurement problem to exploit the EPR1 paradox (see Section 3.4).
power of quantum parallelism. This sort We discuss the two types of opera-
of manipulation has no classical analog tions a quantum system can undergo:
and requires nontraditional programming
techniques. One technique manipulates 1 EPR = Einstein, Podolsky, and Rosen
measurement and quantum state trans- takes to compute the output for a single in-
formations. Most quantum algorithms in- put state on a classical computer, a quan-
volve a sequence of quantum state trans- tum computer can compute the values for
formations followed by a measurement. all input states. This process is known as
For classical computers there are sets quantum parallelism. However, measur-
of gates that are universal in the sense ing the output states will randomly yield
that any classical computation can be per- only one of the values in the superposition,
formed using a sequence of these gates. and at the same time destroy all of the
Similarly, there are sets of primitive quan- other results of the computation. Section 5
tum state transformations, called quan- describes this situation in detail. Sections
tum gates, that are universal for quantum 6 and 7 describe techniques for taking ad-
computation. Given enough quantum bits, vantage of quantum parallelism inspite of
it is possible to construct a universal quan- the severe constraints imposed by quan-
tum Turing machine. tum mechanics on what can be measured.
Quantum physics puts restrictions on Section 6 describes the details of Shor’s
the types of transformations that can be polynomial time factoring algorithm. The
done. In particular, all quantum state fastest known classical factoring algo-
transformations, and therefore all quan- rithm requires exponential time, and it is
tum gates and all quantum computations, generally believed that there is no classi-
must be reversible. Yet all classical algo- cal polynomial time factoring algorithm.
rithms can be made reversible and can Shor’s is a beautiful algorithm that takes
be computed on a quantum computer in advantage of quantum parallelism by us-
comparable time. Some common quantum ing a quantum analog of the Fourier trans-
gates are defined in Section 4. form.
Two applications combining quantum Lov Grover developed a technique for
gates and entangled states are described searching
√ an unstructured list of n items
in Section 4.2: teleportation and dense in O( n) steps on a quantum computer.
coding. Teleportation is the transfer of a Classical computers can do no better than
quantum state from one place to another O(n), so unstructured search on a quan-
through classical channels. That telepor- tum computer is provably more efficient
tation is possible is surprising, since quan- than search on a classical computer. How-
tum mechanics tells us that it is not pos- ever, the speed-up is only polynomial, not
sible to clone quantum states or even exponential, and it has been shown that
measure them without disturbing the Grover’s algorithm is optimal for quan-
state. Thus, it is not obvious what informa- tum computers. It seems likely that search
tion could be sent through classical chan- algorithms that could take advantage of
nels that could possibly enable the recon- some problem structure could do better.
struction of an unknown quantum state Tad Hogg, among others, has explored
at the other end. Dense coding, a dual to such possibilities. We describe various
teleportation, uses a single quantum bit to quantum search techniques in Section 7.
transmit two bits of classical information. It is as yet unknown whether the power
Both teleportation and dense coding rely of quantum parallelism can be harnessed
on the entangled states described in the for a wide variety of applications. One tan-
EPR experiment. talizing open question is whether quan-
It is only in Section 5 that we see tum computers can solve NP-complete
where an exponential speed-up over clas- problems in polynomial time.
sical computers might come from. The in- Perhaps the biggest open question is
put to a quantum computation can be put whether useful quantum computers can
in a superposition state that encodes all be built. There are a number of propos-
possible input values. Performing the com- als for building quantum computers us-
putation on this initial state will result in ing ion traps, nuclear magnetic resonance
superposition of all of the corresponding (NMR), and optical and solid-state tech-
output values. Thus, in the same time it niques. All of the current proposals have
experiment can be performed with mini- visible on the screen, exactly one eighth of
mal equipment: a strong light source, such the original amount of light.
as a laser pointer, and three polaroids
(polarization filters), which can be picked
up at any camera supply store. The ex-
periment demonstrates some of the prin-
ciples of quantum mechanics through pho-
tons and their polarization.
1997] and over 0.5 km through the atmo- {w1 , w2 } respectively. The cartesian prod-
sphere [Hughes et al. 1999]. uct of these two spaces can take as its
basis the union of the bases of its com-
ponent spaces {v1 , v2 , w1 , w2 }. Note that
3.2. Multiple Qubits the order of the basis was chosen arbi-
Imagine a macroscopic physical object trarily. In particular, the dimension of the
breaking apart and multiple pieces flying state space of multiple classical particles
off in different directions. The state of this grows linearly with the number of parti-
system can be described completely by de- cles, since dim(X ×Y ) = dim(X )+dim(Y ).
scribing the state of each of its compo- The tensor product of V and W has ba-
nent pieces separately. A surprising and sis {v1 ⊗ w1 , v1 ⊗ w2 , v2 ⊗ w1 , v2 ⊗ w2 }.
unintuitive aspect of the state space of Note that the order of the basis, again,
an n-particle quantum system is that the is arbitrary.5 So the state space for two
state of the system cannot always be de- qubits, each with basis {|0i, |1i}, has ba-
scribed in terms of the state of its compo- sis {|0i ⊗ |0i, |0i ⊗ |1i, |1i ⊗ |0i, |1i ⊗ |1i},
nent pieces. It is when examining systems which can be written more compactly as
of more than one qubit that one first gets a {|00i, |01i, |10i, |11i}. More generally, we
glimpse of where the computational power write |xi to mean |bn bn−1 . . . b0 i where bi
of quantum computers could come from. are the binary digits of the number x.
As we saw, the state of a qubit A basis for a 3-qubit system is
can be represented by a vector in the
{|000i, |001i, |010i, |011i,
two-dimensional complex vector space
spanned by |0i and |1i. In classical physics, |100i, |101i, |110i, |111i}
the possible states of a system of n par-
and in general an n-qubit system has 2n
ticles, whose individual states can be de-
basis vectors. We can now see the expo-
scribed by a vector in a two-dimensional
nential growth of the state space with the
vector space, form a vector space of 2n di-
number of quantum particles. The tensor
mensions. However, in a quantum system
product X ⊗ Y has dimension d im(X ) ×
the resulting state space is much larger; a
d im(Y ).
system of n qubits has a state space of 2n
The state |00i + |11i is an example of a
dimensions.4 It is this exponential growth
quantum state that cannot be described in
of the state space with the number of par-
terms of the state of each of its components
ticles that suggests a possible exponential
(qubits) separately. In other words, we
speed-up of computation on quantum com-
cannot find a1 , a2 , b1 , b2 such that (a1 |0i +
puters over classical computers.
b1 |1i) ⊗ (a2 |0i + b2 |1i) = |00i + |11i, since
Individual state spaces of n particles
combine classically through the cartesian
(a1 |0i + b1 |1i) ⊗ (a2 |0i + b2 |1i) = a1 a2 |00i
product. Quantum states, however, com-
bine through the tensor product. Details + a1 b2 |01i + b1 a2 |10i + b1 b2 |11i
on properties of tensor products and their
expression in terms of vectors and matri- and a1 b2 = 0 implies that either a1 a2 = 0
ces are given in Appendix A. Let us look or b1 b2 = 0. States that cannot be decom-
briefly at distinctions between the carte- posed in this way are called entangled
sian product and the tensor product that states. These states represent situations
will be crucial to understanding quantum that have no classical counterpart and for
computation. which we have no intuition. These are also
Let V and W be 2 two-dimensional com- the states that provide the exponential
plex vector spaces with bases {v1 , v2 } and growth of quantum state spaces with the
number of particles.
4 Actually, as we shall see, the state space is the set
of normalized vectors in this 2n dimensional space, 5 It is only when we use matrix notation to describe
just as the state a|0i + b|1i of a qubit is normalized state transformations that the order of basis vectors
so that |a|2 + |b|2 = 1. becomes relevant.
the other hand, the state √12 (|00i + |01i) There are two standard ways that peo-
is not entangled: since √12 (|00i + |01i) = ple use to describe entangled states and
|0i ⊗ √12 (|0i + |1i), any measurement of the their measurement. Both have their posi-
tive aspects, but both are incorrect and can
first bit will yield |0i regardless of whether
lead to misunderstandings. Let us exam-
the second bit was measured. Similarly,
ine both in turn.
the second bit has a fifty-fifty chance of be-
Einstein, Podolsky, and Rosen proposed
ing measured as |0i regardless of whether
that each particle has some internal state
the first bit was measured or not. Note
that completely determines what the re-
that entanglement, in the sense that mea-
sult of any given measurement will be.
surement of one particle has an effect
This state is, for the moment, hidden from
on measurements of another particle, is
us, and therefore the best we can currently
equivalent to our previous definition of en-
do is to give probabilistic predictions. Such
tangled states as states that cannot be
a theory is known as a local hidden vari-
written as a tensor product of individual
able theory. The simplest hidden variable
states.
theory for an EPR pair is that the par-
ticles are either both in state |0i or both
3.4. The EPR Paradox in state |1i, we just don’t happen to know
which. In such a theory no communication
Einstein, Podolsky, and Rosen proposed a between possibly distant particles is nec-
gedanken experiment that uses entangled essary to explain the correlated measure-
particles in a manner that seemed to vi- ments. However, this point of view can-
olate fundamental principles of relativity. not explain the results of measurements
Imagine a source that generates two maxi- with respect to a different basis. In fact,
mally entangled particles √12 |00i + √12 |11i, Bell showed that any local hidden vari-
called an EPR pair, and sends one each able theory predicts that certain measure-
to Alice and Bob. ments will satisfy an inequality, known as
Bell’s inequality. However, the result of ac-
tual experiments performing these mea-
surements show that Bell’s inequality is
violated. Thus quantum mechanics cannot
be explained by any local hidden variable
theory. See Greenstein and Zajonc [1997]
for a highly readable account of Bell’s the-
orem and related experiments.
The second standard description is in
Alice and Bob can be arbitrarily far terms of cause and effect. For example,
apart. Suppose that Alice measures her we said earlier that a measurement per-
particle and observes state |0i. This means formed by Alice affects a measurement
that the combined state will now be |00i, performed by Bob. However, this view is
and if now Bob measures his particle he incorrect also, and results, as Einstein,
will also observe |0i. Similarly, if Alice Podolsky, and Rosen recognized, in deep
measures |1i, so will Bob. Note that the inconsistencies when combined with rela-
change of the combined quantum state oc- tivity theory. It is possible to set up the
curs instantaneously even though the two EPR scenario so that one observer sees
particles may be arbitrarily far apart. It Alice measure first, then Bob, while an-
appears that this would enable Alice and other observer sees Bob measure first,
Bob to communicate faster than the speed then Alice. According to relativity, physics
of light. Further analysis, as we shall see, must equally well explain the observa-
shows that even though there is a coupling tions of the first observer as the second.
between the two particles, there is no way While our terminology of cause and ef-
for Alice or Bob to use this mechanism to fect cannot be compatible with both ob-
communicate. servers, the actual experimental values
1
U (|c0i) = √ (U (|a0i) + U (|b0i))
2
1
= √ (|aai + |bbi).
2
But if U is a cloning transformation then The key to both dense coding and tele-
portation is the use of entangled particles.
U (|c0i) = |cci = 1/2(|aai + |abi The initial set up is the same for both pro-
cesses. Alice and Bob wish to communi-
+ |bai + |bbi), cate. Each is sent one of the entangled par-
√ ticles making up an EPR pair,
which is not equal to (1/ 2)(|aai + |bbi).
Thus there is no unitary operation that 1
ψ0 = √ (|00i + |11i).
can reliably clone unknown quantum 2
states. It is clear that cloning is not possi-
ble by using measurement, since measure- Say Alice is sent the first particle, and
ment is both probabalistic and destructive Bob the second. Until a particle is trans-
of states not in the measuring device’s as- mitted, only Alice can perform transfor-
sociated subspaces. mations on her particle, and only Bob can
It is important to understand what sort perform transformations on his.
of cloning is and isn’t allowed. It is pos-
sible to clone a known quantum state.
4.2.1 Dense Coding. Alice. Alice receives
What the no cloning principle tells us
two classical bits, encoding the num-
is that it is impossible to reliably clone
bers 0 through 3. Depending on this
an unknown quantum state. Also, it is
number Alice performs one of the trans-
possible to obtain n particles in an en-
formations {I, X , Y, Z } on her qubit of the
tangled state a|00 . . . 0i + b|11 . . . 1i from
an unknown state a|0i + b|1i. Each of
these particles will behave in exactly the
same way when measured with respect to
the standard basis for quantum compu-
tation {|00 . . . 0i, |00 . . . 01i, . . . , |11 . . . 1i},
but not when measured with respect to
other bases. It is not possible to create the
n-particle state (a|0i + b|1i) ⊗ . . . ⊗ (a|0i +
b|1i) from an unknown state a|0i + b|1i.
4.2. Examples
The use of simple quantum gates can be
studied with two simple examples: dense
entangled pair ψ0 . Transforming just one
coding and teleportation.
bit of an entangled pair means performing
Dense coding uses one quantum bit to-
the identity transformation on the other
gether with an EPR pair to encode and
bit. The resulting state is shown in the
transmit two classical bits. Since EPR
table.
pairs can be distributed ahead of time,
only one qubit (particle) needs to be physi-
cally transmitted to communicate two bits Value Transformation New state
of information. This result is surprising 0 ψ0 = (I ⊗ I )ψ0 √1 (|00i + |11i)
2
since, as was discussed in Section 3, only 1 ψ1 = (X ⊗ I )ψ0 √1 (|10i + |01i)
2
one classical bit’s worth of information can 2 ψ2 = (Y ⊗ I )ψ0 √1 (−|10i + |01i)
be extracted from a qubit. Teleportation 2
is the opposite of dense coding, in that it 3 ψ3 = (Z ⊗ I )ψ0 √1 (|00i − |11i)
2
uses two classical bits to transmit a single
qubit. Teleportation is surprising in light Alice then sends her qubit to Bob.
of the no cloning principle of quantum me-
chanics, in that it enables the transmis- Bob. Bob applies a controlled-NOT to the
sion of an unknown quantum state. two qubits of the entangled pair.
Note that Bob can now measure the sec- Alice. Alice has a qubit whose state she
ond qubit without disturbing the quantum doesn’t know. She wants to send the state
state. If the measurement returns |0i then of ths qubit
the encoded value was either 0 or 3, if the
measurement returns |1i then the encoded φ = a|0i + b|1i
value was either 1 or 2.
Bob now applies H to the first bit: to Bob through classical channels. As with
Finally, Bob measures the resulting bit, dense coding, Alice and Bob each possess
which allows him to distinguish between one qubit of an entangled pair
0 and 3, and 1 and 2.
1
ψ0 = √ (|00i + |11i).
4.2.2 Teleportation. The objective is to 2
transmit the quantum state of a particle
using classical bits and reconstruct the Alice applies the decoding step of dense
exact quantum state at the receiver. coding to the qubit φ to be transmitted and
Since quantum state cannot be copied, her half of the entangled pair. The starting
the quantum state of the given particle state is quantum state
will necessarily be destroyed. Single-bit
teleportation has been realized experi- 1
mentally [Boschi et al. 1998; Bouwmeester φ ⊗ ψ0 = √ (a|0i ⊗ (|00i + |11i)
et al. 1997; Nielsen et al. 1998]. 2
+ b|1i ⊗ (|00i + |11i))
1
= √ (a|000i + a|011i + b|100i
2
+ b|111i),
Bob can reconstruct the original state of The T gate is sufficient to construct arbi-
Alice’s qubit, φ, by applying the appropri- trary combinatorial circuits.
ate decoding transformation to his part of The following quantum circuit, for ex-
the entangled pair. Note that this is the ample, implements a 1 bit full adder using
encoding step of dense coding. Toffoli and controlled-NOT gates:
Walsh–Hadamard transformation W of 1
Section 4.1.1 to get a superposition T (H|0i ⊗ H|0i ⊗ |0i) = (|000i + |010i
2
1 + |100i + |111i).
√ (|00 . . . 0i + |00 . . . 1i + · · · + |11 . . . 1i)
2n
2n −1
The resulting superposition can be viewed
1 X as a truth table for the conjunction, or
=√ |xi more generally as the graph of a function.
2n x=0
In the output the values of x, y, and x ∧ y
which should be viewed as the superposi- are entangled in such a way that measur-
tion of all integers 0 ≤ x < 2n . Add a k-bit ing the result will give one line of the truth
register |0i then by linearity table, or more generally one point of graph
of the function. Note that the bits can be
µ 2n −1 ¶ 2n −1 measured in any order: measuring the re-
1 X 1 X
Uf √ |x, 0i = √ U f (|x, 0i) sult will project the state to a superposi-
2n x=0 2n x=0 tion of the set of all input values for which
2 −1 f produces this result, and measuring the
1 X
n
computable. Shor shows that the quan- Example. Suppose a = 11 were ran-
tum Fourier transform with base 2m can be domly chosen. Since M 2 = 441 ≤ 29 <
constructed using only m(m+1) 2
gates. The 882 = 2M 2 , we find m = 9. Thus, a total
construction makes use of two types of of 14 quantum bits, 9 for x and 5 for f (x),
gates. One is a gate to perform the famil- are required to compute the superposition
iar Hadamard transformation H. We will of equation 1.
denote by H j the Hadamard transforma-
Step 2. A state whose amplitude has the same
tion applied to the j th bit. The other type
period as f . The quantum Fourier trans-
of gate performs 2-bit transformations of
form acts on the amplitude function as-
the form
sociated with the input state. In order to
1 0 0 0 use the quantum Fourier transform to ob-
0 1 0 0 tain the period of f , a state is constructed
S j,k = 0 0 1 0 ,
whose amplitude function has the same
0 0 0 eiθk− j period as f .
To construct such a state, measure the
where θk− j = π/2k− j . This transformation last dlog2 M e qubits of the state of Eq. 1
acts on the kth and j th bits of a larger that encode f (x). A random value u is ob-
register. The quantum Fourier transform tained. The value u is not of interest in it-
is given by self; only the effect the measurement has
on our set of superpositions is of inter-
H0 S0,1 . . . S0,m−1 H1 . . . Hm−3 est. This measurement projects the state
Sm−3,m−2 Sm−3,m−1 Hm−2 Sm−2,m−1 Hm−1 space onto the subspace compatible with
the measured value, so the state after
followed by a bit reversal transformation. measurement is
If FFT is followed by measurement, as in X
Shor’s algorithm, the bit reversal can be C g (x)|x, ui,
performed classically. See Shor [1997] for x
more details.
for some scale factor C where
6.2. A Detailed Outline of Shor’s algorithm ½
1 if f (x) = u
The detailed steps of Shor’s algorithm are g (x) =
0 otherwise.
illustrated with a running example where
we factor M = 21. Note that the x’s that actually appear in
Step 1. Quantum parallelism. Choose an the sum, those with g (x) 6= 0, differ from
integer a arbitrarily. If a is not relatively each other by multiples of the period; thus
prime to M , we have found a factor of M . g (x) is the function we are looking for. If
Otherwise apply the rest of the algorithm. we could measure two successive x’s in the
Let m be such that M 2 ≤ 2m < 2M 2 . sum, we would have the period. Unfortu-
[This choice is made so that the approxi- nately the laws of quantum physics permit
mation used in Step 3 for functions whose only one measurement.
period is not a power of 2 will be good Example. Suppose that random mea-
enough for the rest of the algorithm to surement of the superposition of
work.] Use quantum parallelism as de- Eq. 1 produces 8. The state after this
scribed in Section 5.2 to compute f (x) = measurement8 (Figure 2) clearly shows
a x mod M for all integers from 0 to 2m − 1. the periodicity of f .
The function is thus encoded in the quan-
tum state Step 3. Applying a quantum Fourier trans-
form. The |ui part of the state will not be
2 −1
1 X
m
√ |x, f (x)i. (1) 8 Only the 9 bits of x are shown in Figure 2; the bits
2m x=0 of f (x) are known from the measurement.
Fig. 2. Probabilities for measuring x when measuring the state C6x∈X |x, 8i
obtained in Step 2, where X = {x|211x mod 21 = 8}}.
used, so we will no longer write it. Ap- reducing the fraction 2vm (= rj ) to its lowest
ply the quantum Fourier transform to the terms will yield a fraction whose denom-
state obtained in Step 2. inator q is the period r. The fact that
P P in general the quantum Fourier trans-
U Q F T : x g (x)|xi → c G(c)|ci
form only approximately gives multiples
of the scaled frequency complicates the
Standard Fourier analysis tells us that extraction of the period from the measure-
when the period r of the function g (x) ment. When the period is not a power
defined in Step 2 is a power of 2, the re- of 2, a good guess for the period can be
sult of the quantum Fourier transform is obtained using the continued fraction
¯
X ¯ 2m À expansion of 2vm . This classical technique
c j ¯¯ j , is described in Appendix B.
j
r
Example. Say that measurement of the
where the amplitude is 0 except at multi- state returns v = 427. Since v and 2m
ples of 2m /r. When the period r does not are relatively prime, the period r will
divide 2m , the transform approximates the most likely not divide 2m and the contin-
exact case, so most of the amplitude is at- ued fraction expansion described in Ap-
m
tached to integers close to multiples of 2r . pendix B needs to be applied. The follow-
ing is a trace of the algorithm described in
Example. Figure 3 shows the result of Appendix B:
applying the quantum Fourier transform
to the state obtained in Step 2. Note
that Figure 3 is the graph of the fast i ai pi qi ²i
Fourier transform of the function shown 0 0 0 1 0.8339844
in Figure 2. In this particular example the 1 1 1 1 0.1990632
period of f does not divide 2m . 2 5 5 6 0.02352941
3 42 211 253 0.5
Step 4. Extracting the period. Measure the
state in the standard basis for quantum which terminates with 6 = q2 < M ≤ q3 .
computation, and call the result v. In the Thus, q = 6 is likely to be the period of f .
case where the period happens to be a
power of 2, so that the quantum Fourier Step 5. Finding a factor of M . When our
transform gives exactly multiples of 2m /r, guess for the period, q, is even, use the
the period is easy to extract. In this case, Euclidean algorithm to efficiently check
m
v = j 2r for some j . Most of the time j and whether either aq/2 + 1 or aq/2 − 1 has a
r will be relatively prime, in which case nontrivial common factor with M .
The reason why aq/2 + 1 or aq/2 − 1 is 6.2.1 A Comment on Step 2 of Shor’s Algorithm.
likely to have a nontrivial common fac- The measurement in Step 2 can be skipped
tor with M is as follows. If q is indeed entirely. More generally Bernstein and
the period of f (x) = a x mod M , then Vazirani [1997] show that measurements
aq = 1 mod M, since aq a x = a x mod M for in the middle of an algorithm can always
all x. If q is even, we can write be avoided. If the measurement in Step 2
is omitted, the state consists of a superpo-
(aq/2 + 1)(aq/2 − 1) = 0 mod M . sitions of several periodic functions all of
which have the same period. By the linear-
Thus, as long as neither aq/2 +1 nor aq/2 −1 ity of quantum algorithms, applying the
is a multiple of M , either aq/2 +1 or aq/2 −1 quantum Fourier transformation leads to
has a nontrivial common factor with M . a superposition of the Fourier transforms
of these functions, each of which is entan-
Example. Since 6 is even either a6/2 −1 = gled with the corresponding u and there-
113 − 1 = 1330 or a6/2 + 1 = 113 + 1 = fore do not interfere with each other. Mea-
1332 will have a common factor with M. In surement gives a value from one of these
this particular example we find two factors Fourier transforms. Seeing how this argu-
gcd(21, 1330) = 7 and gcd(21, 1332) = 3. ment can be formalized illustrates some
of the subtleties of working with quan-
Step 6. Repeating the algorithm, if necessary.
tum superpostions. Apply the quantum
Various things could have gone wrong so
Fourier transform tensored with the iden-
that this process does not yield a factor of 2n −1
tity, U Q F T ⊗ I, to C6x=0 |x, f (x)i to get
M:
1. The value v was not close enough to a 2X
n
−1 2X
m
−1 µ ¶
0 2πicx
m
multiple of 2r . C exp |c, f (x)i,
2m
x=0 c=0
2. The period r and the multiplier j could
have had a common factor so that the
which is equal to
denominator q was actually a factor of
the period, not the period itself. X X X µ ¶
2πicx
3. Step 5 yields M as M ’s factor. C0 exp |c, ui
2m
4. The period of f (x) = a x mod M is odd. u x| f (x)=u c
Shor shows that few repetitions of this al- for u in the range of f (x). What results
gorithm yields a factor of M with high is a superposition of the results of Step 3
probability. for all possible u’s. The quantum Fourier
→C 0
G u (c)|c, ui, uations of P . Thus Grover’s search algo-
rithm [Grover 1996] is provably more ef-
u∈R x=0 c=0
ficient than any algorithm that could run
where G u (c) is the discrete Fourier trans- on a classical computer.
form of g u (x) and R is the range of f (x). While Grover’s algorithm is optimal
Measure c and run Steps 4 and 5 as [Bennett et al. 1997; Boyer et al. 1996;
before. Zalka 1997] for completely unstructured
searches, most search problems involve
searching a structured solution space.
7. SEARCH PROBLEMS
Just as there are classical heuristic al-
A large class of problems can be speci- gorithms that exploit problem structure,
fied as search problems of the form “find one would expect that there are more
some x in a set of possible solutions such efficient quantum algorithms for cer-
that statement P (x) is true.” Such prob- tain structured problem instances. Cerf
lems range from database search to sort- et al. [1998] use Grover’s search al-
ing to graph coloring. For example, the gorithm in place of classical searches
graph coloring problem can be viewed as a within a heuristic algorithm to show that
search for an assignment of colors to ver- a quadratic speed-up is possible over
tices so that the statement “all adjacent a particularly simple classical heuristic
vertices have different colors” is true. Sim- for solving NP-hard problems. Brassard
ilarly, a sorting problem can be viewed as et al. [1998], using the techniques of
a search for a permutation for which the Grover’s search algorithm in a less obvious
statement “the permutation x takes the way, show that general heuristic searches
initial state to the desired sorted state” is have quantum analogs with quadratic
true. speed-up.
An unstructured search problem is one There is hope that for certain struc-
where nothing is know (or no assumption tured problems a speed-up greater than
are used) about the structure of the solu- quadratic is possible. Such algorithms will
tion space and the statement P . For exam- likely require new approaches that are not
ple, determining P (x0 ) provides no infor- merely quantum implementations of clas-
mation about the possible value of P (x1 ) sical algorithms. Shor’s algorithm, when
for x0 6= x1 . A structured search problem viewed as a search for factors, is an ex-
is one where information about the search ample of an algorithm that achieves expo-
space and statement P can be exploited. nential speed-up by using problem struc-
For instance, searching an alphabet- ture (number theory) in new ways unique
ized list is a structured search problem to quantum computation.
2 −1
1 X
n
Tad Hogg has developed heuristic quan-
tum search algorithms that exploit prob- √ |x, P (x)i. (2)
lem structure. His approach is distincly 2n x=0
nonclassical and uses unique properties of The difficult step is to obtain a useful re-
quantum computation. One problem with sult from this superposition.
this approach is that, like most heuristic For any x0 such that P (x0 ) is true, |x0 , 1i
algorithms, the use of problem structure will be part of the superposition of Eq. 2.
is complicated enough that it is hard to Since the amplitude of such a state is √12n ,
determine the probability that a single
the probability that a random measure-
execution of an algorithm will give a cor-
ment of the superposition produces x0 is
rect answer. Therefore, it is unknown how
only 2−n . The trick is to change the quan-
efficient Hogg’s algorithms are. Classi-
tum state in Eq. 2 so as to greatly increase
cally the efficiency of heuristic algorithms
the amplitude of vectors |x0 , 1i for which
is estimated by empirically testing the
P is true and decrease the amplitude of
algorithm. But as there is an exponential
vectors |x, 0i for which P is false.
slowdown when simulating a quantum
Once such a transformation of the quan-
computer on a classical one, empirical test-
tum state has been performed, one can
ing of quantum algorithms is currently in-
simply measure the last qubit of the quan-
feasible except in small cases. Small cases
tum state which represents P (x). Because
indicate that Hogg’s algorithms are more
of the amplitude change, there is a high
efficient than Grover’s algorithm applied
probability that the result will be 1. If
to structured search problems, but that
this is the case, the measurement has pro-
the speed-up is likely to be only polyno-
jected the state of Eq. 2 onto the subspace
mial. While less interesting theoretically,
√1 6 k |xi , 1i where k is the number of so-
even a small polynomial speed-up on av- k i=1
erage for these computationally difficult lutions. Further measurement of the re-
problems is of significant practical inter- maining bits will provide one of these so-
est. Until sufficiently large quantum com- lutions. If the measurement of qubit P (x)
puters are built, or better techniques for yields 0, then the whole process is started
analyzing such algorithms are found, the over and the superposition of Eq. 2 must
efficiency cannot be determined for sure. be computed again.
Grover’s algorithm then consists of the
following steps:
7.1. Grover’s Search Algorithm
1. Prepare a register containing a super-
Grover’s algorithm searches an unstruc- position of all possible values xi ∈
tured list of size N for an x that makes a [0 . . . 2n − 1].
statement true. Let n be such that 2n ≥ N , 2. Compute P (xi ) on this register.
and let U p be the quantum gate that im- 3. Change amplitude a j to −a j for x j such
plements the classical function P (x) that that P (x j ) = 1. An efficient algorithm
tests the truth of the statement, where for changing selected signs is described
true is encoded as 1. in Section 7.1.2. A plot of the ampli-
tudes after this step is shown here.
U P : |x, 0i → |x, P (x)i
given in Section 7.1.1. The resulting not increase the overall complexity of the
amplitudes are shown, where the am- algorithm.
plitude of all the xi ’s with P (xi ) = 0 have Grover has extended his algorithm to
been diminished imperceptibly. achieve quadratic speed-up for other non-
search problems such as computing the
mean and median of a function [Grover
1998]. Using similar techniques Grover
has also shown that certain search prob-
lems that classically run in O(log N ) can
be solved in O(1) on a quantum computer.
Grover’s search can used as a subrou-
tine in other quantum computations, since
Biron et al. [1998] show how the technique
π
√ can be used with arbitrary initial ampli-
5. Repeat steps 2 through 4 2n times. tude
4 √ distributions while still maintaining
6. Read the result. O( N ) complexity.
Boyer et al. [1996] provide a detailed
analysis of the performance of Grover’s al- 7.1.1 Inversion about the Average. To perform
gorithm. They prove that Grover’s algo- inversion about the average on a quantum
rithm is optimal up to a constant factor; computer the inversion must be a unitary
no quantum algorithm can perform an un- transformation. Furthermore, in order for
structured search faster. They also show the algorithm
√ as a whole to solve the prob-
lem in O( N ) time, the inversion must be
that if there is only a single√ x0 such that able to be performed efficiently. As will be
P (x0 ) is true, then after π8 2n iterations
of steps 2 through shown shortly, the inversion can be accom-
√ 4 the failure rate is 0.5. plished with O(n) = O(log(N )) quantum
After iterating π4 2n times the failure rate
gates.
drops to 2−n . Interestingly, additional it-
It is easy to see that the transformation
erations will increase
√ the failure rate. For
example, after π2 2n iterations the failure
rate is close to 1. X
N −1 X
N −1
=√ (−1)sn−1 rn−1 |sn−1 i ⊗ · · · dence with the sets. For example, Figure 5
2n s=0 shows the lattice of Figure 4 rewritten in
ket notation, where the elements v1 = 0,
⊗ (−1)s0 r0 |s0 i v1 = 1, v2 = 0 and v2 = 1 have been enu-
merated in that order.
2 −1
1 X
n
If a state violates a constraint, then so
=√ (−1)s·r |si. do all states above it in the lattice. The
2n s=0
approach Hogg takes in designing quan-
tum algorithms for constraint satisfaction
7.2.2 Overview of Hogg’s Algorithms. A con- problems is to begin with all the amplitude
straint satisfaction problem (CSP) has n concentrated in the |0 . . . 0i state and to
variables V = {v1 , . . . , vn } which can take iteratively move amplitude up the lattice
m different values X = {x1 , . . . , xm } sub- from sets to supersets and away from sets
ject to certain constraints C1 , . . . , Cl . So- that violate the constraints. Note that this
lutions to a constraint satisfaction prob- algorithm begins differently than Shor’s
lem lie in the space of assignments of xi ’s algorithm and Grover’s algorithm, which
to v j ’s, V × X . There is a natural lattice both begin by computing a function on a
structure on this space given by set con- superposition of all the input values at
tainment. Figure 4 shows the assignment once.
space and its lattice structure for n = 2, Hogg gives two ways [Hogg 1996; 1998]
m = 2, x1 = 0, and x2 = 1. Note that of constructing a unitary matrix for
the lattice includes both incomplete and moving amplitude up the lattice. We will
inconsistent assignments. describe both methods, and then describe
Grover’s algorithm has been shown to be and Clearwater [1998] contains an in-
optimal, there is hope that faster algo- troduction to early results in quantum
rithms can be found by exploiting prop- complexity.
erties of the problem structure. We have Of course, there are daunting physical
described one such approach taken by problems that must be overcome if any-
Hogg. one is ever to build a useful quantum
There are a few other known quantum computer. Decoherence, the distortion of
algorithms that we did not discuss. Jones the quantum state due to interaction with
and Mosca [1998] describe the implemen- the environment, is a key problem. A big
tation on a 2-bit quantum computer of breakthrough for dealing with decoher-
a constant time algorithm [Deutsch and ence came from the algorithmic, rather
Jozsa 1992] that can distinguish whether than the physical, side of the field with the
a function is balanced or constant. Grover development of quantum error correction
[1998] describes an efficient algorithm for techniques. We have described some of the
estimating the median of a set of values principles involved. Further advances in
and both Grover [1998] and Terhal and quantum error correction and the devel-
Smolin [1997], using different methods, opment of robust algorithms will be as im-
can solve the coin weighing problem in a portant for the development of practical
single step. quantum computers as advances in the
Beyond these algorithms not much more hardware side.
is known about what could be done with
a practical quantum computer. It is an
9.1. Further Reading
open question whether or not we can
find quantum algorithms that provide ex- Andrew Steane’s survey article “Quan-
ponential speed-up for problems other tum Computing” [Steane 1998] is aimed at
than factoring. There is some speculation physicists. We recommend reading his pa-
among physicists that quantum transfor- per for his viewpoint on this subject, par-
mations might be slightly nonlinear. So ticularly for his description of connections
far all experiments that have been done between information theory and quantum
are consistent with the standard linear computing and for his discussion of er-
quantum mechanics, but a slight nonlin- ror correction, of which he was one of the
earity is still possible. Abrams and Lloyd main developers. He also has an overview
[1998] show that even a very slight nonlin- of the physics involved in actually build-
earity could be exploited to solve all NP- ing quantum computers and a survey of
hard problems on a quantum computer in what had been done up to July 1997. His
polynomial time. This result further high- article contains a more detailed history
lights the fact that computation is funda- of the ideas related to quantum comput-
mentally a physical process, and that what ing than the present paper, and has more
can be computed efficiently may depend on references as well. Another shorter and
subtle issues in physics. very readable tutorial can be found in
The unique properties of quantum com- Berthiaume [1997].
puters give rise to new kinds of complex- Richard Feynman’s Lectures on Compu-
ity classes. For instance, BQP is the set tation [Feynman 1996] contains a reprint
of all languages accepted by a quantum of the lecture “Quantum Mechanical Com-
Turing machine in polynomial time with puters” [Feynman 1985], which began the
bounded probability of error. Details of whole field. It also discusses the thermo-
the extensive research done in the field dynamics of computations, which is closely
of quantum complexity theory is beyond tied with reversible computing and infor-
the scope of this paper. The interested mation theory.
reader may start by consulting Bennett Colin Williams and Scott Clearwater’s
et al. [1997] and Watrous [1998] respec- book Explorations in Quantum Comput-
tively for analyses of time and space com- ing [Williams and Clearwater 1998] comes
plexity of quantum computation. Williams with software, in the form of Mathematica
BERTHIAUME. 1997. Quantum computation. In ALAN FEYNMAN, R. 1985. Quantum mechanical comput-
L. SELMAN, ED., Complexity Theory Retrospective, ers. Optics News 11, 11–20. Also in Foundations
In Honor of Juris Hartmanis on the Occasion of of Physics, 16(6), 507–531, 1986.
His Sixtieth Birthday, (July 5, 1988), Vol. 2, 23– FEYNMAN, R. 1996. In Feynman Lectures on Compu-
51. tation. A. J. HEY AND R. W. ALLEN EDS., Addison-
BIRON, D., BIHAM, O., BIHAM, E., GRASSEL, M., AND LI- Wesley.
DAR, D. A. 1998. Generalized grover search al-
FEYNMAN, R. P., LEIGHTON, R. B., AND SANDS, M.
gorithm for arbitrary initial amplitude distri- 1965. Lectures on Physics, Vol. III. Addison-
bution. Los Alamos Physics Preprint Archive, Wesley.
http://xxx.lanl.gov/abs/quant-ph/9801066.
GERSHENFELD, N. A. AND CHUANG, I. L. 1997.
BOSCHI, D., BRANCA, S., MARTINI, F. D., HARDY, L., AND Bulk spin resonance quantum computing. Sci-
POPESCU, S. 1998. Experimental realization of ence 275, 350–356.
teleporting an unknown pure quantum state via
dual classical and einstein-podolski-rosen chan- GREENSTEIN, G. AND ZAJONC, A. G. 1997. The Quan-
nels. Physical Review Letters 80, 1121–1125. tum Challenge. Jones and Bartlett Publishers,
Sudbury, Mass.
BOUWMEESTER, D., PAN, J.-W., MATTLE, K., EIBL,
M., WEINFURTER, H., AND ZEILINGER, A. GROVER, L. K. 1996. A fast quantum mechanical al-
1997. Experimental quantum teleportation. gorithm for database search. In Proceedings of
Nature 390, 575. the 28th Annual ACM Symposium on the The-
ory of Computing (Philadelphia, Pennsylvania,
BOYER, M., BRASSARD, G., HøYER, P., AND TAPP, A. 1996.
22–24 May 1996), pp. 212–219.
Tight bounds on quantum search. In Proceed-
ings of the Workshop on Physics of Computation: GROVER, L. K. 1998. A framework for fast quan-
PhysComp ’96 (Los Alamitos, CA, 1996), 36–43. tum mechanical algorithms. Proceedings of the
Institute of Electrical and Electronic Engineers 30th Annual ACM Symposium on the The-
Computer Society Press. Preprint at Los Alamos ory of Computing, 53–62. Preprint at Los
Physics Preprint Archive, http://xxx.lanl. Alamos Physics Preprint Archive, http://xxx.
gov/abs/quant-ph/9605034. lanl.gov/abs/quant-ph/9711043.
BRASSARD, G., HøYER, P., AND TAPP, A. 1998. HARDY, G. H. AND WRIGHT, E. M. 1979. An Introduc-
Quantum counting. Preprint at Los Alamos tion to the Theory of Numbers. Oxford University
Physics Preprint Archive, http://xxx.lanl. Press.
gov/abs/quant-ph/9805082. HOGG, T. 1996. Quantum computing and phase
CERF, N. J., GROVER, L. K., AND WILLIAMS, C. P. 1998. transitions in combinatorial search. Journal
Nested quantum search and np-complete prob- of Artificial Intelligence Research 4, 91–
lems. Preprint at Los Alamos Physics Preprint 128. Preprint at Los Alamos Physics Preprint
Archive, http://xxx.lanl.gov/abs/ Archive, http://xxx.lanl.gov/abs/quant-ph/
quant-ph/9806078. 9508012.
CIRAC, J. I. AND ZOLLER, P. 1995. Quantum compu- HOGG, T. 1998. Highly structured searches with
tations with cold trapped ions. Physical Review quantum computers. Physical Review Letters 80,
Letters 74, 4091–4094. 2473–2473.
CORY, D. G., MASS, W., PRICE, M., KNILL, E., HUGHES, R. J., BUTTLER, W. T., KWIAT, P. G.,
LAFLAMME, R., ZUREK, W. H., HAVEL, T. F., AND LAMOREAUX, S. K., MORGAN, G. L., NORDHOLT,
SOMAROO, S. S. 1998. Experimental quantum J. E., AND PETERSON, C. G. 1999. Practical Quan-
error correction. Preprint at Los Alamos Physics tum Cryptography for Secure Free-Space Com-
Preprint Archive, http://xxx.lanl.gov/abs/ munications. Preprint at Los Alamos Physics
quant-ph/9802018. Preprint Archive, http://xxx.lanl.gov/abs/
DEUTSCH, D. 1985. Quantum theory, the Church- quant-ph/9905009.
Turing principle and the universal quantum HUGHES, R. J., BUTTLER, W. T., KWIAT, P. G., LUTHER,
computer. Proceedings of the Royal Society of G. G., MORGAN, G. L., NORDHOLT, J. E., PETERSON,
London Series A A400, 97–117. C. G., AND SIMMONS, C. M. 1997. Secure commu-
DEUTSCH, D. AND JOZSA, R. 1992. Rapid solution of nications using quantum cryptography. In S. P.
problems by quantum computation. Proceedings HOTALING AND A. R. PIRICH EDS., Photonic Quan-
of the Royal Society of London Series A A439, tum Computing, Vol. 3076, 2–11.
553–558. HUNGERFORD, T. A. 1974. Algebra. Springer Verlag,
DIRAC, P. 1958. The Principles of Quantum Mechan- New York, Heidelberg, Berlin.
ics (4th ed.). Oxford University Press. JONES, J. A. AND MOSCA, M. 1998. Implementation
EKERT, A. K., RARITY, J., TAPSTER, P., AND PALMA, G. of a quantum algorithm on a nuclear mag-
1992. Practical quantum cryptography based netic resonance quantum computer. Journal of
on two-photon interferometry. Physical Review Chemical Physics 109, 5, 1648–1653. Preprint
Letters 69, 1293–1295. at Los Alamos Physics Preprint Archive,
FEYNMAN, R. 1982. Simulating physics with com- http://xxx.lanl.gov/abs/quant-ph/9801027.
puters. International Journal of Theoretical LAFLAMME, R., KNILL, E., ZUREK, W., CATASTI, P., AND
Physics 21, 6&7, 467–488. MARIAPPAN, S. 1997. NMR GHZ. Los Alamos
Physics Preprint Archive, http://xxx.lanl. peared in the Proceedings of the 35th Annual
gov/abs/quant-ph/9709025. Symposium on Foundations of Computer Sci-
LENSTRA, A. AND LENSTRA, H. EDS. 1993. The Devel- ence.
opment of the Number Field Sieve, Vol. 1554 of STEANE, A. 1996. The Ion Trap Quantum Informa-
Lecture Notes in Mathematics. Springer Verlag. tion Processor. Los Alamos Physics Preprint
LIBOFF, R. L. 1997. Introductory Quantum Mechan- Archive, http://xxx.lanl.gov/abs/quant-ph/
ics (3rd ed.). Addison-Wesley, Reading, Mass. 9608011.
LO, H.-K. AND CHAU, H. F. 1999. Unconditional secu- STEANE, A. 1998. Quantum computing. Reports on
rity of quantum key distribution over arbitrarily Progress in Physics 61, 2, 117–173. Preprint at
long distances. Science 283, 2050–2056. Los Alamos Physics Preprint Archive, http://
MAYERS, D. 1998. Unconditional Security in Quan- xxx.lanl.gov/abs/quant-ph/9708022.
tum Cryptography. Preprint at Los Alamos TERHAL, B. M. AND SMOLIN, J. A. 1997. Single Quan-
Physics Preprint Archive, http://xxx.lanl. tum Querying of a Database. Los Alamos Physics
gov/abs/quant-ph/9802025. Preprint Archive, http://xxx.lanl.gov/abs/
NIELSEN, M. A., KNILL, E., AND LAFLAMME, R. 1998. quant-ph/9705041.
Complete Quantum Teleportation using Nuclear VANDERSYPEN, L. M. K., YANNONI, C. Y., SHERWOOD,
Magnetic Resonance. Preprint at Los Alamos M. H., AND CHUANG, I. L. 1999. Realization
Physics Preprint Archive, http://xxx.lanl. of Effective Pure States for Bulk Quantum
gov/abs/quant-ph/9811020. Computation. Preprint at Los Alamos Physics
SCHULMAN, L. J. AND VAZIRANI, U. 1998. Scalable Preprint Archive, http://xxx.lanl.gov/abs/
NMR Quantum Computation. Los Alamos quant-ph/9905041.
Physics Preprint Archive, http://xxx.lanl. VEDRAL, V., BARENCO, A., AND EKERT, A. K. 1996.
gov/abs/quant-ph/9804060. Quantum Networks for Elementary Arithmetic
SHOR, P. W. 1994. Algorithms for quantum compu- Operations. Physical Review A. Preprint at Los
tation: Discrete log and factoring. In Proceedings Alamos Physics Preprint Archive, http://xxx.
of the 35th Annual Symposium on Foundations lanl.gov/abs/quant-ph/9511018.
of Computer Science (Nov.), 124–134. Institute WATROUS, J. 1998. Relationships between quan-
of Electrical and Electronic Engineers Computer tum and classical space-bounded complexity
Society Press. ftp://netlib.att.com/netlib/ classes. In 13th Annual IEEE Conference on
att/math/shor/quantum.algorithms.ps.Z. Computational Complexity (June 1998), 210–
SHOR, P. W. 1997. Polynomial-time algorithms for 227.
prime factorization and discrete logarithms on WILLIAMS, C. P. AND CLEARWATER, S. H. 1998. Explo-
a quantum computer. Society for Industrial rations in Quantum Computing. Telos, Springer-
and Applied Mathematics Journal on Comput- Verlag.
ing 26, 5, 1484–1509. Expanded version of Shor WOOTTERS, W. K. AND ZUREK, W. H. 1982. A single
[1994]. quantum cannot be cloned. Nature 299, 802.
SIMON, D. R. 1997. On the power of quantum compu- ZALKA, C. 1997. Grover’s Quantum Searching
tation. Society for Industrial and Applied Math- Algorithm is Optimal. Los Alamos Physics
ematics Journal on Computing 26, 5, 1474– Preprint Archive, http://xxx.lanl.gov/abs/
1483. A preliminary version of this paper ap- quant-ph/9711070.