Scribd
Upload
24 views

Tutorial10 Solution

Grover's algorithm can be used as a database search algorithm to find an item s in an unstructured database of N=2^n items. (1) Classically, on average N/2 queries are needed, and worst case is N queries. (2) With a quantum processing unit, a conditional phase flip is applied that introduces a phase if the register matches s, allowing Grover's algorithm to need only O(√N) operations. (3) A quantum addressing scheme is proposed using O(N logN) switches for quantum superposition, but a fully quantum memory is not practical currently due to noise and cost.

Uploaded by

veena panini
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
24 views

Tutorial10 Solution

Grover's algorithm can be used as a database search algorithm to find an item s in an unstructured database of N=2^n items. (1) Classically, on average N/2 queries are needed, and worst case is N queries. (2) With a quantum processing unit, a conditional phase flip is applied that introduces a phase if the register matches s, allowing Grover's algorithm to need only O(√N) operations. (3) A quantum addressing scheme is proposed using O(N logN) switches for quantum superposition, but a fully quantum memory is not practical currently due to noise and cost.

Uploaded by

veena panini
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Introduction to uantum Computing (IN2381) – W2022/2023 Tutorial 10 solution

Christian B. Mendl, Pedro Hack, Keefe Huang, Irene López Gutiérrez

Tutorial 10 (Grover as a database search algorithm1 )


Grover’s algorithm is sometimes referred to as a database search algorithm. In this tutorial we will examine how
the algorithm could in principle be used to search in an unstructured database, and discuss the feasibility of this
approach.
Assume we have a database containing N = 2n items, each of length l bits: {d1 , d2 , . . . , dN }. We want to
determine where a particular item, s, is in this database.

(a) Discuss how a classical computer (with a CPU and a memory) would approach this problem. How many queries
to the memory are required on average? What is the worst-case scenario?

Now imagine we are given a “quantum processing unit” (QPU) containing four registers:
• An n qubit register for the database index.

• An l qubit register for our query |si.


• An l qubit register for items loaded from the database, initialized as |0i.
• A 1 qubit register initialized as |−i = √1 (|0i
2
− |1i).

With this QPU, we can perform the following load operation: for an index x
LOAD
|xi |si |ti |−i 7→ |xi |si |t ⊕ dx i |−i .

In particular, for |ti = |0i the third register will contain |dx i. Then, the second and the third register are compared
and, if they are the same, a bit flip is applied to the forth register.

(b) What is the effect of this operation? What is its connection to Grover’s algorithm?

Recall, however, that in Grover’s algorithm we make use of


superposition states. Thus, it may seem that in order to
implement this we need a quantum memory besides our QPU.
But in fact, we only need a classical memory which can be
addressed by a quantum scheme.

The figure to the right shows a conceptual diagram of a 32 cell


memory with a five qubit addressing scheme. The tree diagram
represents the possible paths taken by an input qubit. At each
node, the input qubit is sent left or right depending on the value
of the qubit inside the circle. A superposition of paths is possible.

(c) What is the advantage of this scheme as opposed to a fully


quantum memory?
(d) Discuss in general the practicality of this algorithm. Is the
unstructured database problem common? Is the required
hardware achievable?

Solution
(a) The CPU must have enough memory to store a bit string of length n. One can then iterate through the indexes
of the database entries. Each entry, dx , is loaded from the memory into the CPU and the condition dx = s is
checked until satisfied. On average, this would require N/2 queries to the memory. The worst case scenario is
when s is the very last item checked, i.e., one performs N queries.
1 M. A. Nielsen, I. L. Chuang: Quantum Computation and Quantum Information. Cambridge University Press (2010), section 6.5

1
(b) A conditional bit flip on the forth register results in
(
|xi |si |dx i √12 (|1i − |0i) = − |xi |si |dx i |−i , if s = dx
|xi |si |dx i |−i 7→
|xi |si |dx i |−i otherwise

One can then perform the load operation again to reset the third qubit to zero, since |dx ⊕ dx i = |0i. In
summary, when the second and third register are the same, a phase of √ −1 is introduced. This is precisely the
oracle in Grover’s algorithm. Grover’s algorithm will then only need O( N ) such load and flip operations.
(c) Given that quantum states are very sensitive to noise, for long term storage it is preferable to use classical
hardware.

(d) Unstructured databases are not common. Typically, the database is designed with some kind of structure to
optimize the number of queries to memory needed (e.g., by storing items in alphabetical order). In terms of
hardware, the quantum addressing scheme depicted above uses O(N log(N )) “switches” to send the input
qubit left or right. At this point in time, such a setup would be too expensive, and not advantageous compared
to distributed classical computing approaches.

You might also like