Next Article in Journal
Some New Applications of the Faber Polynomial Expansion Method for Generalized Bi-Subordinate Functions of Complex Order γ Defined by q-Calculus
Previous Article in Journal
Dynamic Event-Triggered Consensus for Fractional-Order Multi-Agent Systems without Intergroup Balance Condition
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Quantum Weighted Fractional-Order Transform

1
Information Science Teaching and Research Section, Northeastern University at Qinhuangdao, Qinhuangdao 066004, China
2
College of Marxism, Northeastern University at Qinhuangdao, Qinhuangdao 066004, China
*
Author to whom correspondence should be addressed.
Fractal Fract. 2023, 7(3), 269; https://doi.org/10.3390/fractalfract7030269
Submission received: 14 February 2023 / Revised: 11 March 2023 / Accepted: 17 March 2023 / Published: 18 March 2023
(This article belongs to the Special Issue Recent Developments in Fractional Quantum Mechanics)

Abstract

:
Quantum Fourier transform (QFT) transformation plays a very important role in the design of many quantum algorithms. Fractional Fourier transform (FRFT), as an extension of the Fourier transform, is particularly important due to the design of its quantum algorithm. In this paper, a new reformulation of the weighted fractional Fourier transform (WFRFT) is proposed in order to realize quantum FRFT; however, we found that this reformulation can be applied to other transformations, and therefore, this paper presents the weighted fractional Hartley transform (WFRHT). For the universality of application, we further propose a general weighted fractional-order transform (WFRT). When designing the quantum circuits, we realized the quantum WFRFT via QFT and quantum phase estimation (QPE). Moreover, after extending our design to the WFRHT, we were able to formulate the quantum WFRHT. Finally, in accordance with the research results, we designed the quantum circuit of the general WFRT, and subsequently proposed the quantum WFRT. The research in this paper has great value as a reference for the design and application of quantum algorithms.

1. Introduction

The human desire for computing power is endless, and improvements in computing power are closely related to the progress of civilization. Quantum computing comprises parallel computations; this shows that quantum computing quantum algorithms have great advantages over traditional, classical algorithms. Computers are composed of the following four elements: electron tubes, transistors, integrated circuits, and large-scale integrated circuits. As electronic computers continue to develop, chip integration is becoming more common, and a single transistor will eventually be measured on the nanoscopic scale; however, as transistors continue to shrink in size, the physical and chemical properties of the material will qualitatively change, which will lead to the gradual disappearance of the Molar law [1], and the emergence of the quantum effect. Therefore, quantum computing will be important for the next generation of computer research. In 1980, Benioff proposed the quantum Turing machine model, which opened new avenues for quantum computing research [2]. Furthermore, Feynman pointed out that quantum computing has the ability to complete tasks that classical computers cannot [3]. Moreover, the Deutsch–Jozsa algorithm was one of the first to prove the superiority of quantum computing; this algorithm also conceived the physical implementation of the quantum Turing machine and the general quantum computer [4,5]. In 1994, Shor’s algorithm was proposed. Shor’s algorithm can effectively solve the problem of order finding and factorization using quantum Fourier transform, quantum phase estimation, and other basic operations that involve quantum computing. Compared with classical algorithms, Shor’s algorithm can achieve exponential acceleration; undoubtedly, this is a great challenge for public-key cryptography [6]. In 1996, Grover proposed a square root accelerated quantum search algorithm for N unstructured databases with disordered records [7]. The Grover algorithm aims to use quantum computing technology to solve search problems, and it can achieve square-level acceleration for classical search algorithms. With the introduction of two efficient quantum algorithms, researchers have paid an increasing level of attention to quantum computing; as a result, many improved Grover algorithms appeared [8,9,10,11,12], which has further promoted the development of quantum computing. In addition, applied quantum algorithms have emerged in large numbers; these algorithms are also based on the Grover algorithm [13,14,15,16,17,18]. In 2009, the emergence of the HHL algorithm once again proved the superiority of the quantum algorithm [19], and improvements to the HHL algorithm have also been proposed [20,21,22]. In 2012, Preskill proposed the term “quantum supremacy”; this term highlights that quantum computers may easily answer a class of problems that are complex for classical computers [23]. Moreover, quantum supremacy can be demonstrated on quantum computers using random quantum circuits of more than 50 qubits [24]. A series of quantum computing techniques, such as the quantum Fourier transform (QFT) [25], quantum phase estimation (QPE) [26], and HHL algorithm [19], play a key role in quantum algorithm design; hence, they are called quantum basic linear algebra assemblies [27]. Moreover, these basic operations may be applied to quantum singular value decomposition problems [28] and quantum gradient solving problems [29]. In addition, quantum algorithms play an important role in solving linear equations and differential equations [30,31,32,33,34]. At present, quantum computing has been widely used and it shows great potential.
Out of the abovementioned quantum algorithms, QFT and QPE play an important role in quantum computing; hence, they are called quantum toolbox algorithms. Similar quantum algorithms include the quantum wavelet transform [35], quantum cosine transform [36], quantum Hartley transform [37], quantum black-box [38,39], and quantum random walk [40,41] algorithms, among others. As an extension of the Fourier transform, fractional Fourier transform (FRFT) has a wide range of applications; therefore, the proposal of the quantum fractional Fourier transform (QFRFT) is particularly important. Recently, some studies concerning QFRFT have been conducted. Unfortunately, these studies do not present a complete quantum circuit [42,43,44,45]. Parasa et al. proposed a quantum pseudo-fractional Fourier transform (QPFRFT) using multi-valued logic [42]; however, it is not a proper QFRFT because it is not unitary. Recently, we presented a QFRFT using a quantum artificial neural network, which required using more qubits [46]. There are several diverse definitions of FRFT [47,48]. In this paper, the weighted fractional Fourier transform (WFRFT) was studied because it is unitary. First, we propose a new reformulation of the WFRFT, and we designed its quantum circuit using QFT and QPE; secondly, we propose a definition for the weighted fractional Hartley transform (WFRHT), and the design for the quantum circuit is presented; and finally, the general weighted fractional-order transform (WFRT) is defined and the design for its quantum circuit is proposed.
The remainder of this paper is organized as follows. The theories behind WFRFT, WFRHT, and WFRT are analyzed in Section 2. The quantum WFRFT is proposed in Section 3. The quantum WFRHT is proposed in Section 4. Section 5 discusses the quantum circuit of the WFRT. Finally, the conclusions are presented in Section 6.

2. Weighted Fractional-Order Transform

In 1995, the WFRFT was proposed [49]. We have studied this algorithm in great detail, and in this paper, we propose a new reformulation of this algorithm; we found that this new reformulation can be used with many new algorithms, such as the WFRHT, weighted fractional sine transform (WFRST), weighted fractional cosine transform (WFRCT), and weighted fractional Hadamard transform (WFRHaT) algorithms, among others. Furthermore, we found that when a periodic matrix is given, we can define a new algorithm; thus, we present the definition of the general WFRT algorithm in Section 2.3.

2.1. Weighted Fractional Fourier Transform

In 1995, Shih proposed the WFRFT using the integer power operation of the Fourier transform [49]. As a generalized version of the Fourier transform, the algorithm can be widely used in signal processing and image encryption, among other fields [50,51]; it is defined as follows:
( F α f ) ( t ) = l = 0 3 A l α f l ( t ) .
Here, f 0 ( t ) = f ( t ) , f 1 ( t ) = ( F f 0 ) ( t ) , f 2 ( t ) = ( F f 1 ) ( t ) and f 3 ( t ) = ( F f 2 ) ( t ) (F denotes the Fourier transform). The weighting coefficient A l α can be expressed as:
A l α = cos ( α l ) π 4 cos 2 ( α l ) π 4 exp 3 ( α l ) i π 4 .
For Equation (1), we can also write
F α f t = l = 0 3 A l α f l t = A 0 α f 0 t + A 1 α f 1 t + A 2 α f 2 t + A 3 α f 3 t = A 0 α I f t + A 1 α F f t + A 2 α F 2 f t + A 3 α F 3 f t = A 0 α I + A 1 α F + A 2 α F 2 + A 3 α F 3 f t = I , F , F 2 , F 3 A 0 α A 1 α A 2 α A 3 α f t .
From [49], we can obtain Equation (4),
B 0 α B 1 α B 2 α B 3 α = 1 1 1 1 1 i 1 i 1 1 1 1 1 i 1 i A 0 α A 1 α A 2 α A 3 α ,
where B k α = exp 2 π i k α 4 ; k = 0 , 1 , 2 , 3 . Via inverse transformation, we can obtain Equation (5),
A 0 α A 1 α A 2 α A 3 α = 1 4 1 1 1 1 1 i 1 i 1 1 1 1 1 i 1 i B 0 α B 1 α B 2 α B 3 α .
Next, Equation (5) is substituted into Equation (3), so that Equation (6) is obtained.
F α f t = I , F , F 2 , F 3 A 0 α A 1 α A 2 α A 3 α f t = 1 4 I , F , F 2 , F 3 1 1 1 1 1 i 1 i 1 1 1 1 1 i 1 i B 0 α B 1 α B 2 α B 3 α f t .
Let
Y 0 = I + F + F 2 + F 3 Y 1 = I i F F 2 + i F 3 Y 2 = I F + F 2 F 3 Y 3 = I + i F F 2 i F 3
Thus, a new reformulation of WFRFT is obtained, as shown in Equation (8).
F α f t = 1 4 Y 0 , Y 1 , Y 2 , Y 3 B 0 α B 1 α B 2 α B 3 α f t .
Here B k α = exp 2 π i k α 4 ; k = 0 , 1 , 2 , 3 .
Unitarity is a prerequisite for quantum algorithm design. We can easily prove that the WFRFT is unitary using Equation (8) [52]. Moreover, in Section 3, Equation (8) was used to design quantum circuits. The new reformulation that we propose can easily be used with other new algorithms.

2.2. Weighted Fractional Hartley Transform

The abovementioned WFRFT uses the Fourier transform as the base function. We know that the discrete Fourier transform (DFT) is a periodic matrix with eigenvalues (1, i, −1, −i). Similar to the WFRFT, we can also try to define new fractional-order transforms for other periodic matrices; therefore, we used the Hartley transform as an example. The discrete Hartley transform (DHT) has been proposed [53], and its definition is as follows:
H = 1 N cos 2 π m n N + sin 2 π m n N .
The DHT comprises period 2; hence, its eigenvalues are 1 and −1. Therefore, we present the definition of WFRHT as:
( H α f ) ( t ) = l = 0 1 A l α f l ( t ) .
Here, f 0 ( t ) = f ( t ) and f 1 ( t ) = ( H f 0 ) ( t ) (H denotes the Hartley transform). The weighting coefficient A l α can be obtained using Equation (11),
B 0 α B 1 α = 1 1 1 1 A 0 α A 1 α ,
where B k α = exp π i k α ; k = 0 , 1 . Thus, Equation (12) is obtained,
A 0 α A 1 α = 1 2 1 1 1 1 B 0 α B 1 α .
Then, Equation (10) can be re-expressed as:
F α f t = I , H A 0 α A 1 α f t = 1 2 I , H 1 1 1 1 B 0 α B 1 α f t .
Let
Y 0 = I + H Y 1 = I H
Therefore, the WFRHT can also be expressed as:
H α f t = 1 2 Y 0 , Y 1 B 0 α B 1 α f t .
Here B k α = exp π i k α ; k = 0 , 1 .
The discrete sine transform, discrete cosine transform, and discrete Hadamard transform, among other discrete matrices, have period 2, and their eigenvalues are 1 and −1; therefore, a new algorithm can be defined by replacing the Hartley transform in Equation (14).

2.3. Weighted Fractional-Order Transform

In the analysis of Section 2.1 and Section 2.2, it is evident that for periodic transforms (that is, when the matrix of a transform is periodic), we can present the definition of its weighted fractional-order. Next, we present the definition of a general WFRT.
Let T denote a periodic matrix (strictly speaking, the matrix should be symmetrical) and T M = I (M is the period of the matrix T, and I denotes the identity matrix); therefore, the WFRT can be defined as:
( T α f ) ( t ) = l = 0 M 1 A l α f l ( t ) .
Here, f l t = T l f t ; l = 0 , 1 , , M 1 . The weighting coefficient A l α can be obtained by Equation (17),
B 0 α B 1 α B M 1 α = w 0 × 0 w 0 × 1 w 0 × ( M 1 ) w 1 × 0 w 1 × 1 w 1 × ( M 1 ) w ( M 1 ) × 0 w ( M 1 ) × 1 w ( M 1 ) × ( M 1 ) A 0 α A 1 α A M 1 α ,
where w = e 2 π i / M and B k α = e 2 π i α k / M ; k = 0 , 1 , , M 1 . Furthermore, we can obtain Equation (18) as follows:
A 0 α A 1 α A M 1 α = 1 M u 0 × 0 u 0 × 1 u 0 × ( M 1 ) u 1 × 0 u 1 × 1 u 1 × ( M 1 ) u ( M 1 ) × 0 u ( M 1 ) × 1 u ( M 1 ) × ( M 1 ) B 0 α B 1 α B M 1 α ,
where u = e 2 π i / M . Next, Equation (16) can be re-expressed as:
( T α f ) ( t ) = l = 0 M 1 A l α f l ( t ) = I , T , , T M 1 A 0 α A 1 α A M 1 α f t .
Substituting Equation (18) into Equation (19), we can obtain
( T α f ) ( t ) = l = 0 M 1 A l α f l ( t ) = 1 M I , T , , T M 1 u 0 × 0 u 0 × 1 u 0 × ( M 1 ) u 1 × 0 u 1 × 1 u 1 × ( M 1 ) u ( M 1 ) × 0 u ( M 1 ) × 1 u ( M 1 ) × ( M 1 ) B 0 α B 1 α B M 1 α f t .
Let
Y 0 = u 0 × 0 I + u 1 × 0 T + + u M 1 × 0 T M 1 Y 1 = u 0 × 1 I + u 1 × 1 T + + u M 1 × 1 T M 1 Y 2 = u 0 × 2 I + u 1 × 2 T + + u M 1 × 2 T M 1           Y M 1 = u 0 × M 1 I + u 1 × M 1 T + + u M 1 × M 1 T M 1
Therefore, the WFRT can also be expressed as:
( T α f ) ( t ) = l = 0 M 1 A l α f l ( t ) = 1 M Y 0 , Y 1 , , Y M 1 B 0 α B 1 α B M 1 α f t = 1 M k = 0 M 1 Y k B k α f t .
Here B k α = e 2 π i α k / M and the WFRT is proposed; thus, the WFRFT and the WFRHT presented above are its special cases. For example, M = 4, and T denotes the Fourier transform in Equation (21), which is defined as the WFRFT. If M = 2, T denotes the Hartley transform in Equation (21), which is the WFRHT. This new reformulation can be acceptably applied to the fractional-order definition of periodic functions. Our findings have important value as they can be used as a reference for future information processing and quantum algorithm designs. With increasing demand for greater computing power, we will eventually enter the quantum age. Next, we designed quantum circuits for these algorithms.

3. Quantum Weighted Fractional Fourier Transform

Recently, studies concerning QFRFT have been conducted [42,43,44,45]; however, these research methods describe the QFRFT from the perspective of quantum mechanics, and they do not mention quantum circuits. FRFT definitions are diverse, and some definitions are not unitary [47,48]; this causes issues when designing the QFRFT. As a result of such difficulties, Parasa et al. showed that QFRFT cannot be realized [42]. Recently, we proposed a method to help realize the QFRFT using a quantum artificial neural network; however, this method uses more qubits, and thus, it requires more resources [46]. In this section, we will use QFT and QPE to redesign the QFRFT quantum circuit. Studies by Gidney et al. assisted with our research, as follows [54,55].
First, we used QPE to obtain the eigenvalues of Equation (7), and the quantum circuit of Equation (7) is shown in Figure 1. The upper part represents the control register, the lower part represents the target register, and u represents the feature vector that corresponds with the operator in the target register. Here, we used QFT instead of the inverse quantum Fourier transform (IQFT) in the control register, as the purpose of this design is to obtain the same results as Equation (7).
From Figure 1, we know
ψ 0 = 0 0 u .
Via the H gate, we obtain
ψ 1 = 1 2 0 + 1 1 2 0 + 1 u = 1 2 00 + 01 + 10 + 11 u .
Then, the QFT operation of the target register is,
ψ 2 = 1 2 00 u + 01 F u + 10 u + 11 F u ,
and
ψ 3 = 1 2 00 I u + 01 F u + 10 F 2 u + 11 F 3 u .
Here, u is the eigenvector of Fourier transform; therefore, F u = D u (D is the eigenvalue of Fourier transform). The eigenvalue D can be expressed as:
D = λ 0 0 λ 1 0 λ n 1 ,
where λ j 1 , i , 1 , i ; j = 0 , 1 , , n 1 . Therefore, Equation (26) can be expressed as:
ψ 3 = 1 2 D 0 00 + D 1 01 + D 2 10 + D 3 11 u .
The eigenvector u can be written in accordance with standard orthogonal basis conditions, as shown in Equation (29),
u = j b j u j ,
where b j is the projection length. Thus, we can obtain
D u = j λ j b j u j .
Equation (28) can be expressed as:
ψ 3 = 1 2 j λ j 0 b j u j 00 + j λ j 1 b j u j 01 + j λ j 2 b j u j 10 + j λ j 3 b j u j 11 .
The two-qubit QFT is used in the control register, and its matrix can be expressed as:
Q F T = 1 4 1 1 1 1 1 i 1 i 1 1 1 1 1 i 1 i .
Therefore, we can obtain the following quantum state ψ 4 ,
ψ 4 = j ϕ j b j u j ,
where ϕ j 00 , 01 , 10 , 11 ; the results are shown in Figure 2.
Here, we analyzed the evolutionary process of ϕ 3 to ϕ 4 in detail. Equations (28) and (31) are equivalent; therefore, we can continue to use Equation (28). Then, after the application of the two-qubit QFT, we can obtain
ψ 4 = 1 4 W 0 00 + W 1 01 + W 2 10 + W 3 11 u ,
where
W 0 = D 0 + D 1 + D 2 + D 3 W 1 = D 0 i D 1 D 2 + i D 3 W 2 = D 0 D 1 + D 2 D 3 W 3 = D 0 + i D 1 D 2 i D 3
Equation (35) is consistent with Equation (7); however, we used the description of the eigenvalues in Equation (7). Furthermore, we know that
W 0 = D 0 + D 1 + D 2 + D 3 = λ 0 0 + λ 0 1 + λ 0 2 + λ 0 3 λ 1 0 + λ 1 1 + λ 1 2 + λ 1 3 λ n 1 0 + λ n 1 1 + λ n 1 2 + λ n 1 3 ,
where λ j 1 , i , 1 , i ; j = 0 , 1 , , n 1 . When λ j = 1 ,
λ j 0 + λ j 1 + λ j 2 + λ j 3 = 4 .
When λ j = i ,
λ j 0 + λ j 1 + λ j 2 + λ j 3 = 0 .
When λ j = 1 ,
λ j 0 + λ j 1 + λ j 2 + λ j 3 = 0 .
When λ j = i ,
λ j 0 + λ j 1 + λ j 2 + λ j 3 = 0 .
Regarding W 0 , only in the following circumstance, when λ j = 1 , is the corresponding element not zero; notably, the elements at the other positions are zero. For W 1 ,
W 1 = D 0 i D 1 D 2 + i D 3 = λ 0 0 i λ 0 1 λ 0 2 + i λ 0 3 λ 1 0 i λ 1 1 λ 1 2 + i λ 1 3 λ n 1 0 i λ n 1 1 λ n 1 2 + i λ n 1 3 .
When λ j = i ,
λ j 0 i λ j 1 λ j 2 + i λ j 3 = 4 .
When λ j = 1 , 1 , i ,
λ j 0 i λ j 1 λ j 2 + i λ j 3 = 0 .
For W 2 ,
W 2 = D 0 D 1 + D 2 D 3 = λ 0 0 λ 0 1 + λ 0 2 λ 0 3 λ 1 0 λ 1 1 + λ 1 2 λ 1 3 λ n 1 0 λ n 1 1 + λ n 1 2 λ n 1 3 .
When λ j = 1 ,
λ j 0 λ j 1 + λ j 2 λ j 3 = 4 .
When λ j = 1 , i , i ,
λ j 0 λ j 1 + λ j 2 λ j 3 = 0 .
For W 3 ,
W 3 = D 0 + i D 1 D 2 i D 3 = λ 0 0 + i λ 0 1 λ 0 2 i λ 0 3 λ 1 0 + i λ 1 1 λ 1 2 i λ 1 3 λ n 1 0 + i λ n 1 1 λ n 1 2 i λ n 1 3 .
When λ j = i ,
λ j 0 + i λ j 1 λ j 2 i λ j 3 = 4 .
When λ j = 1 , i , 1 ,
λ j 0 + i λ j 1 λ j 2 i λ j 3 = 0 .
Therefore, in Equation (34), we know that for W 0 , only in the following circumstance, when the eigenvalue is 1, is the corresponding element not zero. For W 1 , W 2 , and W 3 , the corresponding elements are not zero when the eigenvalues are i, −1, and −i, respectively. The number of eigenvalues is shown in Table 1 [56,57]. Then, the combined eigenvectors of W 0 , W 1 , W 2 , and W 3 comprise the exact eigenvectors of the DFT. It is evident that Equations (33) and (34) are equivalent.
Next, we introduced two phase gates (Equations (50) and (51)) and added them to the control register, as shown in Figure 3.
The phase gates R Z π α and R Z π α / 2 are expressed as:
R Z π α = 1 0 0 e π i α ,
and
R Z π α / 2 = 1 0 0 e π i α / 2 .
Thus, we can obtain
1 0 0 e π i α 1 0 0 e π i α / 2 = 1 0 e π i α / 2 e 2 π i α / 2 0 e 3 π i α / 2 .
The diagonal element of the matrix is exactly B k α of Equation (8). Then, ψ 5 is obtained
ψ 5 = j e π i α ϕ j / 2 ϕ j b j u j ,
where ϕ j 0 , 1 , 2 , 3 and ϕ j 00 , 01 , 10 , 11 . Next, we used inverse quantum phase estimation (IQPE), as shown in Figure 4.
The two-qubit IQFT of the control register can be expressed as:
I Q F T = 1 4 1 1 1 1 1 i 1 i 1 1 1 1 1 i 1 i .
After the IQFT of ϕ j is calculated, the eigenvalues are restored,
ψ 6 = 1 2 j e π i α ϕ j / 2 λ j 0 b j u j 00 + j e π i α ϕ j / 2 λ j 1 b j u j 01 + j e π i α ϕ j / 2 λ j 2 b j u j 10 + j e π i α ϕ j / 2 λ j 3 b j u j 11 = 1 2 j e π i α ϕ j / 2 b j u j 00 + j e π i α ϕ j / 2 F b j u j 01 + j e π i α ϕ j / 2 F 2 b j u j 10 + j e π i α ϕ j / 2 F 3 b j u j 11 .
Next,
ψ 7 = 1 2 j e π i α ϕ j / 2 λ j 0 a j u j 00 + j e π i α ϕ j / 2 F b j u j 01 + j e π i α ϕ j / 2 I F 2 F 2 b j u j 10 + j e π i α ϕ j / 2 I F 2 F 3 b j u j 11 ,
where IF denotes the inverse Fourier transform.
ψ 8 = 1 2 j e π i α ϕ j / 2 λ j 0 b j u j 00 + j e π i α ϕ j / 2 I F F b j u j 01 + j e π i α ϕ j / 2 I F 2 F 2 b j u j 10 + j e π i α ϕ j / 2 I F 3 F 3 b j u j 11 = 1 2 j e π i α ϕ j / 2 b j u j 00 + j e π i α ϕ j / 2 b j u j 01 + j e π i α ϕ j / 2 b j u j 10 + j e π i α ϕ j / 2 b j u j 11 = j e π i α ϕ j / 2 b j u j 1 2 0 + 1 1 2 0 + 1 .
Finally, through the H gate, we obtain ψ 9
ψ 9 = 0 0 j e π i α ϕ j / 2 b j u j ,
where ϕ j 0 , 1 , 2 , 3 . We know
j e π i ϕ j / 2 b j u j = F u ,
so
j e π i α ϕ j / 2 b j u j = F α u .
Then, Equation (58) is expressed as:
ψ 9 = 0 0 F α u .
Therefore, the quantum WFRFT (QWFRFT) is obtained.
The QWFRFT was designed as shown in Figure 4. This design provides the foundation for subsequent quantum WFRHT designs (QWFRHT).

4. Quantum Weighted Fractional Hartley Transform

The quantum Hartley transform (QHT) has been proposed [37]. Next, we use QPE and QHT to design the circuit of the QWFRHT, as shown in Figure 5. The eigenvalues of the Hartley transform are 1 and −1; therefore, the control register only needs one qubit to complete the circuit.
Here, u is the eigenvector of the Hartley transform; thus, we can obtain
ψ 0 = 0 u ,
ψ 1 = 1 2 0 + 1 u ,
and
ψ 2 = 1 2 I 0 u + Q H T 1 u = 1 2 D 0 0 + D 1 u .
We know Q H T u = D u . Here, D is the eigenvalue of the Hartley transform and it can be expressed as:
D = λ 0 0 λ 1 0 λ n 1 ,
where λ j 1 , 1 ; j = 0 , 1 , , n 1 . This is because
u = j r j u j ,
therefore,
D u = j λ j r j u j .
Equation (64) can be expressed as:
ψ 2 = 1 2 j λ j 0 r j u j 0 + j λ j 1 r j u j 1 .
Similarly to the abovementioned QWFRFT, after passing through the H gate, we obtain
ψ 3 = j ϕ j r j u j ,
where ϕ j 0 , 1 ; thus, the phase gate of Equation (50) is used,
ψ 4 = j e π i α ϕ j ϕ j r j u j .
Next, the IQPE is used, so that the eigenvalues can be recovered:
ψ 5 = 1 2 j e π i α ϕ j λ j 0 r j u j 0 + j e π i α ϕ j λ j 1 r j u j 1 ,
and
ψ 6 = 1 2 j e π i α ϕ j r j u j 0 + j e π i α ϕ j r j u j 1 = j e π i α ϕ j r j u j 1 2 0 + 1 .
Finally, we obtain
ψ 7 = 0 j e π i α ϕ j r j u j .
This is because
j e π i ϕ j r j u j = H u ,
where H denotes the Hartley transform; therefore,
j e π i α ϕ j r j u j = H α u .
Then, Equation (73) is expressed as
ψ 7 = 0 H α u .
Thus, the QWFRHT is obtained.
Similarly to the QWFRHT, the quantum sine transform, quantum Hadamard transform, quantum cosine transform, and so on, are periodic functions, and their eigenvalues are 1 and −1; therefore, the designs of these functions are consistent with the findings in this section.

5. Discussion

With the help of the designs in Section 3 and Section 4, in this section, we present the quantum circuit of the WFRT (QWFRT); that is, the quantum circuit design of the WFRT in Section 2.3. Moreover, in this section, the unitary matrix T M = I (M is the period and I is the identity matrix) is determined. In order to fulfil the design requirements, the period should satisfy M = 2 q . We also used QPE to design quantum circuits, as shown in Figure 6.
Here, u is the eigenvector of matrix T; thus, we can obtain
ψ 0 = 0 q u ,
where q = log 2 M . Thus
ψ 1 = 1 2 q 0 + 1 q u ,
and
ψ 2 = 1 2 q 0 + T 2 q 1 1 0 + T 2 q 2 1 0 + T 1 u .
This is because M = 2 q and u = j c j u j ; therefore, Equation (79) can be expressed as:
ψ 2 = 1 2 q k = 0 M 1 k T k u = 1 2 q k = 0 M 1 k j λ j k c j u j .
Then, after the q-qubit IQFT of k = 0 M 1 λ j k k , we obtain
ψ 3 = j ϕ j c j u j .
Via the phase gates, we can obtain
ψ 4 = j e 2 π i α ϕ j / M ϕ j c j u j .
Next, the IQPE is used. First, for the q-qubit QFT of ϕ j , the eigenvalues are recovered,
ψ 5 = 1 2 q j e 2 π i α ϕ j / M k = 0 M 1 λ j k k c j u j .
Hence,
ψ 6 = 1 2 q j e 2 π i α ϕ j / M k = 0 M 1 k c j u j .
where ϕ j 0 , 1 , , M 1 . Finally, we obtain
ψ 7 = 0 q j e 2 π i α ϕ j / M c j u j .
Thus
ψ 7 = 0 q T α u .
Ergo, the QWFRT is obtained.
In the quantum circuit of Figure 6, the upper part represents the control register, and the lower part represents the target register. The quantum bits required in the control register are closely related to the period of the unitary matrix T, which is q = log 2 M . In the quantum circuit of Figure 4, the period of the Fourier transform is 4; that is, M = 4; therefore, in the circuit design, the control register needs two qubits. Similarly, in the quantum circuit of Figure 5, the period of the Hartley transform is 2; that is, M = 2; therefore, we used a qubit in the control register. Therefore, it is evident that the schemes proposed in Section 3 and Section 4 are special cases of the scheme proposed in Section 5.
Due to the complexity of the designed quantum algorithm, in space, in order to store eigenvalues, we use an additional q qubits (control registers). The qubits required for the control registers of different algorithms may also be different. For example, for the QWFRFT in Figure 4, we use two auxiliary qubits to store eigenvalues; for the QWFRHT in Figure 5, we use only one auxiliary qubit to store the eigenvalues. That is to say, the qubit of the control register depends on the number of eigenvalues of the algorithm. In time, when the target register is large, we only consider the complexity of the quantum gate in the circuit. Taking the QWFRFT in Figure 4 as an example, we use three QFTs and three IQFTs, and the complexity of each QFT and IQFT is O n 2 . Therefore, the time complexity of the QWFRT depends on the complexity of quantum algorithm T.

6. Conclusions

As a generalized Fourier transform algorithm, the quantum circuit design of FRFT is particularly important. In order to design a QFRFT, we first analyzed the definitions for WFRFT; then, we proposed a new reformulation. Based on the new reformulation, we presented the WFRHT, and this fractional-order definition may also be applicable to the sine transform, cosine transform, and Hadamard transform algorithms; therefore, we proposed an additional weighted fractional-order definition based on the periodic matrix, which has wide applicability. Moreover, the proposed new reformulation contributes to the design of quantum algorithms. We first used QFT and QPE to design the quantum circuit of WFRFT; then, we designed the QWFRHT, a design which has wide applicability. Moreover, by designing the quantum circuit of the WFRT, we were able to realize the QWFRT. The quantum algorithms proposed in this paper will have great value in terms of their ability to be used as a reference for future quantum information processing.

Author Contributions

Methodology, T.Z.; validation, T.Z. and Y.C.; investigation, T.Z.; writing—original draft preparation, T.Z.; writing—review and editing, T.Z.; supervision, T.Z.; project administration, T.Z.; funding acquisition, T.Z. All authors have read and agreed to the published version of the manuscript.

Funding

This study was supported by the Fundamental Research Funds for the Central Universities (N2123016); and the Scientific Research Projects of Hebei colleges and universities (QN2020511).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Moore, G.E. Cramming More Components onto Integrated Circuits; McGraw-Hill: New York, NY, USA, 1965; Volume 38, pp. 114–117. [Google Scholar]
  2. Benioff, P. The computer as a physical system: A microscopic quantum mechanical Hamiltonian model of computers as represented by Turing machines. J. Stat. Phys. 1980, 22, 563–591. [Google Scholar] [CrossRef]
  3. Feynman, R.P. Simulating physics with computers. Int. J. Theor. Phys. 1982, 21, 467–488. [Google Scholar] [CrossRef]
  4. Deutsch, D.; Jozsa, R. Rapid solution of problems by quantum computation. Proc. R. Soc. Lond. Ser. A Math. Phys. Sci. 1992, 439, 553–558. [Google Scholar] [CrossRef]
  5. Deutsch, D. Quantum theory, the Church–Turing principle and the universal quantum computer. Proc. R. Soc. Lond. Ser. A Math. Phys. Sci. 1985, 400, 97–117. [Google Scholar]
  6. Shor, P.W. Algorithms for Quantum Computation: Discrete Logarithms and Factoring. In Proceedings of the 35th Annual Symposium on Foundation of Computer Science, Washington, DC, USA, 20–22 November 1994; pp. 124–134. [Google Scholar]
  7. Grover, L.K. A fast quantum mechanical algorithm for database search. In Proceedings of the Twenty-Eighth Annual ACM Symposium on Theory of Computing, Philadelphia, PA, USA, 1 July 1996; pp. 212–219. [Google Scholar]
  8. Boyer, M.; Brassard, G.; Høyer, P.; Tapp, A. Tight bounds on quantum searching. Fortschr. Phys. Prog. Phys. 1998, 46, 493–505. [Google Scholar] [CrossRef]
  9. Grover, L.K. Quantum Computers Can Search Rapidly by Using Almost Any Transformation. Phys. Rev. Lett. 1998, 80, 4329–4332. [Google Scholar] [CrossRef] [Green Version]
  10. Biham, E.; Biham, O.; Biron, D.; Grassl, M.; Lidar, D.A. Grover’s quantum search algorithm for an arbitrary initial amplitude distribution. Phys. Rev. A 1999, 60, 2742–2745. [Google Scholar] [CrossRef] [Green Version]
  11. Long, G.L. Grover algorithm with zero theoretical failure rate. Phys. Rev. A 2001, 64, 022307. [Google Scholar] [CrossRef] [Green Version]
  12. Li, P.; Li, S. Phase matching in Grover’s algorithm. Phys. Lett. A 2007, 366, 42–46. [Google Scholar] [CrossRef]
  13. Brassard, G.; Høyer, P.; Tapp, A. Quantum counting, Automata, Languages and Programming. In Proceedings of the 25th International Colloquium, ICALP’98, Aalborg, Denmark, 13–17 July 1998; Springer: Berlin/Heidelberg, Germany, 1998; pp. 820–831. [Google Scholar]
  14. Grover, L.K. Synthesis of Quantum Superpositions by Quantum Computation. Phys. Rev. Lett. 2000, 85, 1334–1337. [Google Scholar] [CrossRef]
  15. Hao, G.; Gui-Lu, L.; Feng, L. Quantum Algorithms for Some Well-Known NP Problems. Commun. Theor. Phys. 2002, 37, 424–426. [Google Scholar] [CrossRef]
  16. Pang, C.-Y.; Zhou, R.-G.; Ding, C.-B.; Hu, B.-Q. Quantum search algorithm for set operation. Quantum Inf. Process. 2012, 12, 481–492. [Google Scholar] [CrossRef] [Green Version]
  17. Furrow, B. A panoply of quantum algorithms. arXiv 2006, arXiv:quant-ph/0606127. [Google Scholar] [CrossRef]
  18. Baritompa, W.P.; Bulger, D.W.; Wood, G.R. Grover’s quantum algorithm applied to global optimization. SIAM J. Optim. 2005, 15, 1170–1184. [Google Scholar] [CrossRef]
  19. Harrow, A.W.; Hassidim, A.; Lloyd, S. Quantum Algorithm for Linear Systems of Equations. Phys. Rev. Lett. 2009, 103, 150502. [Google Scholar] [CrossRef]
  20. Clader, B.D.; Jacobs, B.C.; Sprouse, C.R. Preconditioned Quantum Linear System Algorithm. Phys. Rev. Lett. 2013, 110, 250504. [Google Scholar] [CrossRef] [Green Version]
  21. Childs, R.K.A.M.; Kothari, R.; Somma, R.D. Quantum Algorithm for Systems of Linear Equations with Exponentially Improved Dependence on Precision. SIAM J. Comput. 2017, 46, 1920–1950. [Google Scholar] [CrossRef] [Green Version]
  22. Wossnig, L.; Zhao, Z.; Prakash, A. Quantum Linear System Algorithm for Dense Matrices. Phys. Rev. Lett. 2018, 120, 050502. [Google Scholar] [CrossRef] [Green Version]
  23. Preskill, J. Quantum computing and the entanglement frontier. arXiv 2012, arXiv:1203.5813. [Google Scholar]
  24. Boixo, S.; Isakov, S.V.; Smelyanskiy, V.N.; Babbush, R.; Ding, N.; Jiang, Z.; Bremner, M.J.; Martinis, J.M.; Neven, H. Characterizing quantum supremacy in near-term devices. Nat. Phys. 2018, 14, 595–600. [Google Scholar] [CrossRef] [Green Version]
  25. Coppersmith, D. An approximate Fourier transform useful in quantum factoring. arXiv 2002, arXiv:quant-ph/0201067. [Google Scholar]
  26. Nielsen, M.A.; Chuang, I.L. Quantum computation and quantum information. Phys. Today 2001, 54, 60. [Google Scholar] [CrossRef] [Green Version]
  27. Biamonte, J.; Wittek, P.; Pancotti, N.; Rebentrost, P.; Wiebe, N.; Lloyd, S. Quantum machine learning. Nature 2017, 549, 195–202. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  28. Rebentrost, P.; Steffens, A.; Marvian, I.; Lloyd, S. Quantum singular-value decomposition of nonsparse low-rank matrices. Phys. Rev. A 2018, 97, 012327. [Google Scholar] [CrossRef] [Green Version]
  29. Rebentrost, P.; Schuld, M.; Wossnig, L.; Petruccione, F.; Lloyd, S. Quantum gradient descent and Newton’s method for constrained polynomial optimization. New J. Phys. 2019, 21, 073023. [Google Scholar] [CrossRef]
  30. Higgott, O.; Wang, D.; Brierley, S. Variational Quantum Computation of Excited States. Quantum 2019, 3, 156. [Google Scholar] [CrossRef] [Green Version]
  31. Bravo-Prieto, C.; Larose, R.; Cerezo, M.; Subasi, Y.; Cincio, L.; Coles, P. Variational Quantum Linear Solver: A Hybrid Algorithm for Linear Systems. Am. Phys. Soc. 2020, 65, 1. [Google Scholar]
  32. LaRose, R.; Tikku, A.; O’Neel-Judy, É.; Cincio, L.; Coles, P.J. Variational quantum state diagonalization. NPJ Quantum Inf. 2019, 5, 57. [Google Scholar] [CrossRef] [Green Version]
  33. Arrazola, J.M.; Kalajdzievski, T.; Weedbrook, C.; Lloyd, S. Quantum algorithm for nonhomogeneous linear partial differential equations. Phys. Rev. A 2019, 100, 032306. [Google Scholar] [CrossRef] [Green Version]
  34. Childs, A.M.; Liu, J.-P. Quantum Spectral Methods for Differential Equations. Commun. Math. Phys. 2020, 375, 1427–1457. [Google Scholar] [CrossRef] [Green Version]
  35. Fijany, A.; Williams, C.P. Quantum wavelet transforms: Fast algorithms and complete circuits. In Proceedings of the Quantum Computing and Quantum Communications: First NASA International Conference, QCQC’98, Palm Springs, CA, USA, 17–20 February 1998; Springer: Berlin/Heidelberg, Germany, 1999; pp. 10–33. [Google Scholar]
  36. Klappenecker, A.; Rotteler, M. Discrete cosine transforms on quantum computers. In Proceedings of the ISPA 2001. Proceedings of the 2nd International Symposium on Image and Signal Processing and Analysis. In Conjunction with 23rd International Conference on Information Technology Interfaces, Pula, Croatia, 19–21 June 2001; pp. 464–468. [Google Scholar]
  37. Tseng, C.-C.; Hwang, T.-M. Quantum Circuit Design of Discrete Hartley Transform using Recursive Decomposition Formula. In Proceedings of the 2005 IEEE International Symposium on Circuits and Systems (ISCAS), Kobe, Japan, 23–26 May 2005; Volume 1, pp. 824–827. [Google Scholar] [CrossRef]
  38. Mermin, N.D. Quantum mysteries refined. Am. J. Phys. 1994, 62, 880–887. [Google Scholar] [CrossRef]
  39. Berthiaume, A.; Brassard, G. Oracle Quantum Computing. J. Mod. Opt. 1994, 41, 2521–2535. [Google Scholar] [CrossRef] [Green Version]
  40. Aharonov, Y.; Davidovich, L.; Zagury, N. Quantum random walks. Phys. Rev. A 1993, 48, 1687. [Google Scholar] [CrossRef]
  41. Venegas-Andraca, S.E.; Ball, J.L.; Burnett, K.; Bose, S. Quantum walks with entangled coins. New J. Phys. 2005, 7, 221. [Google Scholar] [CrossRef]
  42. Parasa, V.; Perkowski, M. Quantum Pseudo-Fractional Fourier Transform Using Multiple-Valued Logic. In Proceedings of the 2012 IEEE 42nd International Symposium on Multiple-Valued Logic, Victoria, BC, Canada, 14–16 May 2012; pp. 311–314. [Google Scholar] [CrossRef]
  43. Lv, C.-H.; Fan, H.-Y.; Li, D. From fractional Fourier transformation to quantum mechanical fractional squeezing transformation. Chin. Phys. B 2015, 24, 020301. [Google Scholar] [CrossRef]
  44. Weimann, S.; Perez-Leija, A.; Lebugle, M.; Keil, R.; Tichy, M.; Gräfe, M.; Heilmann, R.; Nolte, S.; Moya-Cessa, H.; Weihs, G.; et al. Implementation of quantum and classical discrete fractional Fourier transforms. Nat. Commun. 2016, 7, 11027. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  45. Madrid, Y.; Molina, M.; Torres, R. Quantum Fractional Fourier Transform. In Frontiers in Optics; Optica Publishing Group: Washington, DC, USA, 2018. [Google Scholar]
  46. Zhao, T.; Yang, T.; Chi, Y. Quantum Weighted Fractional Fourier Transform. Mathematics 2022, 10, 1896. [Google Scholar] [CrossRef]
  47. Tao, R.; Zhang, F.; Wang, Y. Research progress on discretization of fractional Fourier transform. Sci. China Inf. Sci. 2008, 51, 859–880. [Google Scholar] [CrossRef]
  48. Gómez-Echavarría, A.; Ugarte, J.P.; Tobón, C. The fractional Fourier transform as a biomedical signal and image processing tool: A review. Biocybern. Biomed. Eng. 2020, 40, 1081–1093. [Google Scholar] [CrossRef]
  49. Shih, C.-C. Fractionalization of Fourier transform. Opt. Commun. 1995, 118, 495–498. [Google Scholar] [CrossRef]
  50. Liang, Y.; Da, X.; Xu, R.; Ni, L.; Zhai, D.; Pan, Y. Research on Constellation-Splitting Criterion in Multiple Parameters WFRFT Modulations. IEEE Access 2018, 6, 34354–34364. [Google Scholar] [CrossRef]
  51. Zhao, T.; Ran, Q. The Weighted Fractional Fourier Transform and Its Application in Image Encryption. Math. Probl. Eng. 2019, 2019, 4789194. [Google Scholar] [CrossRef]
  52. Zhao, T.; Chi, Y. Multiweighted-Type Fractional Fourier Transform: Unitarity. Fractal Fract. 2021, 5, 205. [Google Scholar] [CrossRef]
  53. Bracewell, R.N. Discrete hartley transform. JOSA 1983, 73, 1832–1835. [Google Scholar] [CrossRef]
  54. Gidney, C. Computing Fractional Fourier Transforms. 2017. Available online: https://algassert.com/post/1710 (accessed on 11 June 2017).
  55. Pang, H. Circuit Construction of Quantum Fractional Fourier Transform. Zhihu. 2022. Available online: https://zhuanlan.zhihu.com/p/489812535 (accessed on 31 March 2022).
  56. McClellan, J.; Parks, T. Eigenvalue and eigenvector decomposition of the discrete Fourier transform. IEEE Trans. Audio Electroacoust. 1972, 20, 66–74. [Google Scholar] [CrossRef]
  57. Dickinson, B.; Steiglitz, K. Eigenvectors and functions of the discrete Fourier transform. IEEE Trans. Acoust. Speech Signal Process. 1982, 30, 25–31. [Google Scholar] [CrossRef] [Green Version]
Figure 1. The QPE for Equation (7).
Figure 1. The QPE for Equation (7).
Fractalfract 07 00269 g001
Figure 2. Quantum states corresponding with eigenvalues.
Figure 2. Quantum states corresponding with eigenvalues.
Fractalfract 07 00269 g002
Figure 3. Phase gates are added.
Figure 3. Phase gates are added.
Fractalfract 07 00269 g003
Figure 4. Circuit of the QWFRFT.
Figure 4. Circuit of the QWFRFT.
Fractalfract 07 00269 g004
Figure 5. The QWFRHT circuit.
Figure 5. The QWFRHT circuit.
Fractalfract 07 00269 g005
Figure 6. The circuit of QWFRT.
Figure 6. The circuit of QWFRT.
Fractalfract 07 00269 g006
Table 1. Multiplicities of the DFT eigenvalues.
Table 1. Multiplicities of the DFT eigenvalues.
N1−1ii
4nn + 1nnn − 1
4n + 1n + 1nnn
4n + 2n + 1n + 1nn
4n + 3n + 1n + 1n + 1n
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhao, T.; Chi, Y. Quantum Weighted Fractional-Order Transform. Fractal Fract. 2023, 7, 269. https://doi.org/10.3390/fractalfract7030269

AMA Style

Zhao T, Chi Y. Quantum Weighted Fractional-Order Transform. Fractal and Fractional. 2023; 7(3):269. https://doi.org/10.3390/fractalfract7030269

Chicago/Turabian Style

Zhao, Tieyu, and Yingying Chi. 2023. "Quantum Weighted Fractional-Order Transform" Fractal and Fractional 7, no. 3: 269. https://doi.org/10.3390/fractalfract7030269

Article Metrics

Back to TopTop