A multi-antenna wireless communication system beamforming method and system based on quantum approximate optimization

Abstract

The present application relates to a kind of quantum approximation optimization-based multi-antenna wireless communication system beam forming method, system, belong to multi-antenna wireless communication field, comprising the following steps: S1, according to the channel state information and the wireless resource weight, construct the weighted channel correlation matrix reflecting the characteristics of multi-user integrated channel;S2, the continuous phase of each transmitting antenna is discretized, and a gradually doubled binary variable is sequentially assigned to the discrete phase, the phase combination of all antennas is jointly constructed to determine the state space corresponding to quantum approximation optimization algorithm corresponding to quantum state representation of quantum approximation optimization algorithm;S3, by quantum approximation optimization algorithm, quantum circuit evolution and observation are executed, and the power allocation quantum optimization initial parameter is updated and optimized by the gradient descent of classical algorithm, to obtain power allocation quantum optimization optimal parameter;S4, power allocation quantum optimization optimal parameter is converted into the phase combination of all antennas, and the optimal power allocation beam is obtained.

Description

A beamforming method and system for multi-antenna wireless communication systems based on quantum approximation optimization Technical Field

[0001] This invention relates to the field of multi-antenna wireless communication, specifically to a beamforming method and system for multi-antenna wireless communication systems based on quantum approximation optimization. Background Technology

[0002] Multi-antenna technology significantly increases the number of transmitting and receiving antennas, and its transmission modes are also more diverse. Its basic principle is to use multiple transmitting and receiving antennas at both the transmitting and receiving ends, enabling the differentiation of signals sent to or originating from different spatial directions. It can also improve system capacity, coverage, and signal-to-noise ratio, and enhance the transmission quality of wireless signals without increasing bandwidth or transmission power. Its difference from traditional signal processing methods lies in its simultaneous study of signal processing from both temporal and spatial perspectives.

[0003] Wireless power transfer using multi-antenna technology enables technologies such as wireless charging. To improve the energy transfer efficiency of wireless charging, the energy transmitter is equipped with multiple antennas. By optimizing the beamforming matrix, channel fading can be effectively combated. However, for the combinatorial optimization problem introduced by phase quantization, traditional methods based on exhaustive search, alternating optimization, or heuristic algorithms experience exponential growth in computational complexity as antenna size and quantization accuracy increase. Furthermore, these methods are prone to getting trapped in local optima, making it difficult to achieve stable and efficient optimization control in practical systems. Existing beamforming methods often focus on a single communication performance indicator as the optimization objective, failing to fully consider the service priority differences among users in multi-user scenarios, as well as new communication requirements such as wireless energy harvesting. This makes it difficult to achieve an effective trade-off between communication rate and energy transfer, thus limiting their applicability in applications such as smart energy-carrying networks.

[0004] The purpose of this invention is to design a beamforming method and system for a multi-antenna wireless communication system based on quantum approximation optimization to address the problems existing in the prior art. Summary of the Invention

[0005] To address the problems existing in the prior art, the present invention provides a beamforming method and system for a multi-antenna wireless communication system based on quantum approximation optimization, which can effectively solve at least one of the problems existing in the prior art.

[0006] The technical solution of this invention is:

[0007] A beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization includes the following steps:

[0008] S1, obtain channel state information of multiple users, allocate radio resource weights to multiple users, and construct a weighted channel correlation matrix reflecting the comprehensive channel characteristics of multiple users based on the channel state information and the radio resource weights;

[0009] S2, Discretize the continuous phase of each transmitting antenna, and assign a binary variable that gradually doubles to the discrete phase. Based on the binary variable, the discrete phase, and the weighted channel correlation matrix, combine the phases of all antennas to jointly construct the quantum state representation corresponding to the quantum approximation optimization algorithm, determine the state space corresponding to the quantum approximation optimization algorithm, and initialize the quantum circuit corresponding to the state space.

[0010] S3, the quantum circuit evolution and observation are performed through a quantum approximation optimization algorithm to obtain the initial parameters of the power allocation quantum optimization and the corresponding mathematical expectation. Then, the initial parameters of the power allocation quantum optimization are updated and optimized through the gradient descent algorithm of the classical algorithm to obtain the optimal parameters of the power allocation quantum optimization.

[0011] S4. The optimal parameters for quantum optimization of power allocation are converted into a phase combination of all antennas based on the binary variables and the discrete phases to obtain the optimal power allocation beam.

[0012] Further, step S1 includes:

[0013] S1.1, Obtain the total number of users K that need to be served simultaneously, the number of antennas N_T of the base station, and the channel gain h from each antenna to the k-th user. k This forms an N_T×1 complex vector, which is used to obtain the channel vector {h} of the k-th user. k};

[0014] S1.2, calculate the service priority w from each antenna to the k-th user using a network scheduling algorithm or a quality of service algorithm. k This forms an N_T×1 complex vector, from which the user weight {w} of the k-th user is obtained. k};

[0015] S1.3, for all users' channel vectors {h k} and user weight {w k} via Q = Σ k w k h k h k The weighted summation is performed to obtain the weighted channel correlation matrix Q.

[0016] Further, in step S2, the continuous phase of each transmitting antenna is discretized, and a binary variable that gradually doubles is assigned to each discrete phase, including:

[0017] The number of bits b of the binary variable is quantized according to θ. j = (2π / 2 b )·2 j Phase discretization is performed to obtain the discrete phase θ. j Where j = 0, 1, ..., b-1;

[0018] Through 2 j The binary representation of the corresponding number of bits b is used as the binary variable, and a mapping relationship is established between the binary variable and the discrete phase.

[0019] Further, in step S2, based on the binary variables, the discrete phases, and the weighted channel correlation matrix, the phase combinations of all antennas are used to jointly construct the quantum state representation corresponding to the quantum approximation optimization algorithm, and the state space corresponding to the quantum approximation optimization algorithm is determined to include:

[0020] The state space corresponding to the quantum approximation optimization algorithm includes H-gate, Rz-gate, and Rx-gate, which initialize the quantum state |+ ^{ N} is used as an H gate to separate the discrete phase θ j The weighted channel correlation matrix is ​​combined with the rotation angle to construct the quantum state, and the rotation angle is used as the quantum state representation, based on each discrete phase θ of all antennas. j The rotation angles are used to determine the rotation amounts of the Rz gate and the Rx gate, thus obtaining the state space corresponding to the quantum approximation optimization algorithm.

[0021] Furthermore, the discrete phase θ j The rotation angle, constructed by combining the weighted channel correlation matrix with the above, includes:

[0022] With discrete phase θ j The larger the corresponding binary variable, the greater the weight provided by the weighted channel correlation matrix for the discrete phase θ. j The weighted channel correlation matrix is ​​jointly weighted, and then combined with the first quantum approximation optimization algorithm. The optimizable parameter γ of the layer The rotation angle is established using the following formula. ,

[0023] = 2γ ·[Q] nn ·θ j Where n represents the nth antenna, and J represents θ j The j-th bit of the corresponding binary variable, [Q] nn This represents the nth diagonal element of the weighted channel correlation matrix.

[0024] Further, in step S2, initializing the quantum circuit corresponding to the state space includes:

[0025] Set the initial phase of all antennas to 0.

[0026] Further, in step S3, quantum circuit evolution and observation are performed using a quantum approximation optimization algorithm to obtain the initial parameters and corresponding mathematical expectations for power allocation quantum optimization. Then, the initial parameters for power allocation quantum optimization are updated and optimized using gradient descent using a classical algorithm to obtain the optimal parameters for power allocation quantum optimization, including:

[0027] For each computational layer of the quantum approximation optimization algorithm, the following is executed sequentially: The rotation angle is... The data to be processed, as the Rz rotating gate, is introduced into the quantum circuit. The quantum circuit evolution and observation are performed through a quantum approximation optimization algorithm, resulting in the first quantum approximation optimization algorithm. The optimizable parameter γ of the layer The rotation angle The data to be processed, used as the Rx rotating gate, is introduced into the quantum circuit. The quantum circuit evolution and observation are then performed using a quantum approximation optimization algorithm, yielding the first result of the quantum approximation optimization algorithm. The optimizable parameter β of the layer The optimizable parameter γ and the optimizable parameter β As initial parameters for quantum optimization of power allocation;

[0028] According to the optimizable parameter γ and the optimizable parameter β Calculate the mathematical expectation C of the final state on the quantum processor in the current state to obtain the optimizable parameter γ. and the optimizable parameter β The solution for quantum approximation optimization under the current weights;

[0029] The optimizable parameter γ is calculated using the classic gradient descent algorithm. The optimizable parameter β Within a preset range, the parameter displacement change is compensated for by learning and iterated. When the classical optimizer deems the parameter convergent or has reached a preset number of iterations, the resulting optimizable parameter γ is obtained. The optimizable parameter β As the optimal parameter for power allocation quantum optimization.

[0030] Further, step S4 includes:

[0031] The final state on the quantum processor in the current state is calculated based on the optimal parameters of the power allocation quantum optimization. The final state is then projected onto the computational basis to obtain the binary optimal parameters.

[0032] Based on the mapping relationship between the binary variables and the discrete phases, the binary optimal parameters are decoded into the phase of each antenna.

[0033] Furthermore, a beamforming system for a multi-antenna wireless communication system based on quantum approximation optimization is provided to implement the aforementioned beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization, comprising the following modules:

[0034] The weighted channel correlation matrix construction module is used to obtain channel state information of multiple users, assign radio resource weights to multiple users, and construct a weighted channel correlation matrix that reflects the comprehensive channel characteristics of multiple users based on the channel state information and the radio resource weights.

[0035] The state space establishment module is used to discretize the continuous phase of each transmitting antenna and assign a binary variable that gradually doubles to the discrete phase. Based on the binary variable, the discrete phase, and the weighted channel correlation matrix, the phase combination of all antennas is used to jointly construct the quantum state representation corresponding to the quantum approximation optimization algorithm, determine the state space corresponding to the quantum approximation optimization algorithm, and initialize the quantum circuit corresponding to the state space.

[0036] The optimal parameter calculation module is used to perform quantum circuit evolution and observation through a quantum approximation optimization algorithm to obtain the initial parameters of power allocation quantum optimization and the corresponding mathematical expectation. Then, the initial parameters of power allocation quantum optimization are updated and optimized through gradient descent of the classical algorithm to obtain the optimal parameters of power allocation quantum optimization.

[0037] The optimal power allocation beam conversion module is used to convert the optimal parameters of the quantum optimization of power allocation into a phase combination of all antennas based on the binary variables and the discrete phases, so as to obtain the optimal power allocation beam.

[0038] Therefore, the present invention provides the following effects and / or advantages:

[0039] This application proposes a beamforming design method for multi-antenna wireless communication systems. The phase quantization problem of beamforming antennas in large-scale antenna systems, caused by hardware limitations in traditional communication engineering, is transformed into a discrete combinatorial optimization problem. A quantum-classical hybrid optimization framework is constructed using the quantum approximation optimization algorithm (QAOA). By executing parameterized quantum circuits on a quantum processor and optimizing the circuit parameters on a classical processor, the optimal quantized phase configuration is efficiently searched, thereby optimizing the overall system performance, including weighted sum rate and wireless energy harvesting, under limited phase quantization accuracy (b-bits).

[0040] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0041] It should be understood that the above summary and the following detailed description of the invention are exemplary and explanatory, and are intended to provide further explanation of the invention as claimed. Attached Figure Description

[0042] Figure 1 is a flowchart illustrating one embodiment of the present invention.

[0043] Figure 2 shows a partial quantum circuit diagram obtained by the present invention.

[0044] Figure 3 is a schematic diagram of the rate and energy curves of the user as the number of antennas changes.

[0045] Figure 4 is a schematic diagram of the superimposed power field.

[0046] Figure 5 is a schematic diagram of the user information rate-energy power curve. Detailed Implementation

[0047] To facilitate understanding by those skilled in the art, the present invention will now be described in further detail with reference to the embodiments:

[0048] Referring to Figure 1, a beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization includes the following steps:

[0049] S1, obtain channel state information of multiple users, allocate radio resource weights to multiple users, and construct a weighted channel correlation matrix reflecting the comprehensive channel characteristics of multiple users based on the channel state information and the radio resource weights;

[0050] S2, Discretize the continuous phase of each transmitting antenna, and assign a binary variable that gradually doubles to the discrete phase. Based on the binary variable, the discrete phase, and the weighted channel correlation matrix, combine the phases of all antennas to jointly construct the quantum state representation corresponding to the quantum approximation optimization algorithm, determine the state space corresponding to the quantum approximation optimization algorithm, and initialize the quantum circuit corresponding to the state space.

[0051] S3, the quantum circuit evolution and observation are performed through a quantum approximation optimization algorithm to obtain the initial parameters of the power allocation quantum optimization and the corresponding mathematical expectation. Then, the initial parameters of the power allocation quantum optimization are updated and optimized through the gradient descent algorithm of the classical algorithm to obtain the optimal parameters of the power allocation quantum optimization.

[0052] S4. The optimal parameters for quantum optimization of power allocation are converted into a phase combination of all antennas based on the binary variables and the discrete phases to obtain the optimal power allocation beam.

[0053] This embodiment proposes a beamforming design method for multi-antenna wireless communication systems. The phase quantization problem of beamforming antennas in large-scale antenna systems, caused by hardware limitations in traditional communication engineering, is transformed into a discrete combinatorial optimization problem. A quantum-classical hybrid optimization framework is constructed using the quantum approximation optimization algorithm (QAOA). By executing parameterized quantum circuits on a quantum processor and optimizing the circuit parameters on a classical processor, the optimal quantized phase configuration is efficiently searched, thereby optimizing the overall system performance, including weighted sum rate and wireless energy harvesting, under limited phase quantization accuracy (b-bits).

[0054] This embodiment explicitly introduces the quantum approximation optimization algorithm (QAOA) into multi-antenna beamforming. By establishing a weighted channel correlation matrix and discretizing the continuous phase of each transmit antenna, and assigning a binary variable that gradually doubles to the discrete phase, the physical state of the phase can be converted into a quantum state. This transforms the continuous, non-convex beamforming problem into a discrete optimization problem that can be solved quantumly. The solution is then achieved using quantum circuits and the quantum approximation optimization algorithm, resulting in the optimal power allocation beam for the best antenna.

[0055] The following is a detailed implementation of each step.

[0056] Further, step S1 includes:

[0057] S1.1, Obtain the total number of users K that need to be served simultaneously, the number of antennas N_T of the base station, and the channel gain h from each antenna to the k-th user. k This forms an N_T×1 complex vector, which is used to obtain the channel vector {h} of the k-th user. k};

[0058] S1.2, calculate the service priority w from each antenna to the k-th user using a network scheduling algorithm or a quality of service algorithm. k This forms an N_T×1 complex vector, from which the user weight {w} of the k-th user is obtained. k};

[0059] S1.3, for all users' channel vectors {h k} and user weight {w k} via Q = Σ k w k h k h k The weighted summation is performed to obtain the weighted channel correlation matrix Q.

[0060] In this step, the channel gain h k The service priority w can be obtained by channel estimation calculation using uplink pilot signals. k The priority can be determined by network scheduling algorithms or Quality of Service (QoS) requirements, or by the different user devices. A higher weight corresponds to a higher priority. For example, it can be set based on the power requirements of the devices, with the device with the highest power among multiple devices having the highest service priority. k Set to 5, service priority of the device with the lowest power. k Set to 1, and allocate service priority to the remaining devices according to this ratio. k .

[0061] Through Q = Σ k w k h k h k The weighted summation is performed to simultaneously characterize the spatial directional characteristics and service priorities of a multi-user channel within a unified matrix representation, thereby transforming the multi-user beamforming optimization problem into a quadratic objective that can be used for quantum approximation optimization. k h k It represents the overall contribution of the beam direction to the received power, indicating the direction in which the k-th user wants the antenna energy to be aligned in the spatial domain, expressed by w. k By superimposing the spatial requirements of multiple users, a weighted channel correlation matrix is ​​constructed by weighted summation of the channel vectors and user weights of all users. This facilitates the subsequent mapping of the multi-user beamforming problem into a combinatorial optimization problem suitable for solving by quantum approximation optimization algorithms.

[0062] Further, in step S2, the continuous phase of each transmitting antenna is discretized, and a binary variable that gradually doubles is assigned to each discrete phase, including:

[0063] The number of bits b of the binary variable is quantized according to θ. j = (2π / 2 b )·2 j Phase discretization is performed to obtain the discrete phase θ. j Where j = 0, 1, ..., b-1;

[0064] Through 2 j The binary representation of the corresponding number of bits b is used as the binary variable, and a mapping relationship is established between the binary variable and the discrete phase.

[0065] In this embodiment, the continuous phase of the antenna is discretized into 2... b For example, b=3 means that the phase of each antenna can only be in 2... 3 =Selected from 8 discrete values, b is determined by the resolution of the phase shifter in the hardware RF link; the larger b is, the higher the resolution. This results in a value consisting of b bits q0, q1, ..., q b The binary number composed of 1s corresponds to the corresponding antenna angle. Taking b=3 as an example, the corresponding binary variable is a three-digit binary value, which is obtained by using 2... j We can obtain three binary values: 100, 010, and 001. This allows us to discretize the continuous phase of the antenna into the corresponding θ values ​​calculated from these three binary values. j These discrete phases are doubled. Then, the binary value is assigned to each phase to obtain the mapping relationship between the binary variable and the discrete phase.

[0066] Further, in step S2, based on the binary variables, the discrete phases, and the weighted channel correlation matrix, the phase combinations of all antennas are used to jointly construct the quantum state representation corresponding to the quantum approximation optimization algorithm, and the state space corresponding to the quantum approximation optimization algorithm is determined to include:

[0067] The state space corresponding to the quantum approximation optimization algorithm includes H-gate, Rz-gate, and Rx-gate, which initialize the quantum state |+ ^{ N} is used as an H gate to separate the discrete phase θ j The weighted channel correlation matrix is ​​combined with the rotation angle to construct the quantum state, and the rotation angle is used as the quantum state representation, based on each discrete phase θ of all antennas. j The rotation angles are used to determine the rotation amounts of the Rz gate and the Rx gate, thus obtaining the state space corresponding to the quantum approximation optimization algorithm.

[0068] In this step, all qubits are initialized to a searchable superposition state using an H-gate; then, the antenna phase signal and channel weights are encoded into the phase structure of the quantum state using an Rz-gate; finally, an Rx-gate is used to mix these phase structures, thereby constructing the entire quantum state space that QAOA can search. During this process, the initialized quantum state |+ ^{ N} is used as an H gate, where |+ =(|0 +|1 ) / It is a uniform superposition state of a single quantum bit. N represents the tensor product of N qubits, all of which are fabricated on |+ The state is the starting point for optimization, where N = N_T × b, and N_T is the number of transmit antennas.

[0069] Discrete phase θ j The weighted channel correlation matrix is ​​combined with the rotation angle to construct the antenna phase from the physical state to the quantum state, and the rotation angle is used as data that can be recognized by the quantum circuit.

[0070] Furthermore, the discrete phase θ j The rotation angle, constructed by combining the weighted channel correlation matrix with the above, includes:

[0071] With discrete phase θ j The larger the corresponding binary variable, the greater the weight provided by the weighted channel correlation matrix for the discrete phase θ. j The weighted channel correlation matrix is ​​jointly weighted, and then combined with the first quantum approximation optimization algorithm. The optimizable parameter γ of the layer The rotation angle is established using the following formula. ,

[0072] = 2γ ·[Q] nn ·θ j Where n represents the nth antenna, and J represents θ j The j-th bit of the corresponding binary variable, [Q] nn This represents the nth diagonal element of the weighted channel correlation matrix. [Q] nn This is the nth diagonal element of the channel correlation matrix Q, reflecting the contribution of the nth antenna to the overall channel gain for all users.

[0073] This step describes the specific method for converting the antenna phase from a physical state to a quantum state. This conversion combines the antenna's discrete phase with the user's weighted channel correlation matrix. Specifically, this is achieved by using the discrete phase θ... jThe larger the corresponding binary variable, the larger the corresponding θ. j The larger the phase amplitude, the larger the rotation angle each time, and therefore the greater its contribution to the result. Furthermore, the more important the nth antenna is in a multi-user system, the greater its contribution to the result; therefore, it is necessary to combine the phase magnitude information and the antenna channel importance information into a single weighting factor. Finally, this is combined with the optimizable parameter γ used in the quantum approximation optimization algorithm to adjust the cost evolution intensity. By encoding each phase bit of each antenna to construct the corresponding quantum rotation angle, the optimization objective of the communication system is mapped to the phase evolution process in the quantum circuit.

[0074] Further, in step S2, initializing the quantum circuit corresponding to the state space includes:

[0075] Set the initial phase of all antennas to 0.

[0076] Further, in step S3, quantum circuit evolution and observation are performed using a quantum approximation optimization algorithm to obtain the initial parameters and corresponding mathematical expectations for power allocation quantum optimization. Then, the initial parameters for power allocation quantum optimization are updated and optimized using gradient descent using a classical algorithm to obtain the optimal parameters for power allocation quantum optimization, including:

[0077] For each computational layer of the quantum approximation optimization algorithm, the following is executed sequentially: The rotation angle is... The data to be processed, as the Rz rotating gate, is introduced into the quantum circuit. The quantum circuit evolution and observation are performed through a quantum approximation optimization algorithm, resulting in the first quantum approximation optimization algorithm. The optimizable parameter γ of the layer The rotation angle The data to be processed, used as the Rx rotating gate, is introduced into the quantum circuit. The quantum circuit evolution and observation are then performed using a quantum approximation optimization algorithm, yielding the first result of the quantum approximation optimization algorithm. The optimizable parameter β of the layer The optimizable parameter γ and the optimizable parameter β As initial parameters for quantum optimization of power allocation;

[0078] In this step, the Rz rotation gate is a single-qubit operation in quantum circuits. The entire operation involves rotating around the Z-axis, which changes the relative phase of the qubit's state but does not change its state at |0. state and |1 The probability of the state. In this treatment, γ This is the [number]. The optimizable parameters of the layered QAOA circuit, adjusted by a classical computer, control the overall intensity of the rotation. This step primarily involves introducing the antenna phase as data to be processed into the quantum circuit to obtain the initial optimizable parameter γ. .

[0079] Then, Rx rotates the door around the X-axis the entire way, and the Rx door will be at |0 and |1. The oscillations between states are introduced to create a quantum tunneling effect, helping the algorithm escape local optima and explore the entire solution space. The goal is to find the quantum state corresponding to the optimal phase from all phases of the antenna's quantum states, thus obtaining the initial optimizable parameter β. .

[0080] Specifically, no. Layer calculation of the optimizable parameter γ and the optimizable parameter β The complete operation is as follows:

[0081]

[0082] in, It is the unitary evolution operator of the QAOA circuit, representing the unitary evolution operator of the QAOA circuit. Layered quantum circuit, where p represents the total number of layers in the QAOA circuit. This represents applying an Rx rotation gate to all N-layer quantum circuits. This represents applying an Rz rotation gate to a specific qubit (n,j).

[0083] The evolutionary operation of the overall p-layer quantum circuit is as follows:

[0084]

[0085] Where p represents the total number of layers in the quantum circuit. This represents the formula for obtaining the entire circuit after the p-layer processing.

[0086] The initial quantum state evolves through this circuit to become:

[0087]

[0088] in, This represents the final quantum state processed by the quantum circuit. Processed initial state ,in This represents the initial quantum state introduced by the antenna phase. γ and β It was determined during the application of the revolving door that this step was for use with γ. and β To obtain the final quantum state, and then to use the final quantum state to determine γ and β These two parameters will be further optimized.

[0089] According to the optimizable parameter γ and the optimizable parameter β Calculate the mathematical expectation C of the final state on the quantum processor in the current state to obtain the optimizable parameter γ. and the optimizable parameter β The solution for quantum approximation optimization under the current weights;

[0090] In this step, C is calculated using the following formula:

[0091]

[0092] Where t represents the number of iterations. This indicates the t-th iteration.

[0093] H C H represents the Hamiltonian of the problem. C This is the key to mapping classical optimization problems to quantum systems.

[0094] H C The construction is as follows:

[0095]

[0096] in, This indicates that the qubit is in the |0 state. and |1 The probability difference.

[0097] Different constant terms are given based on the baseline value of Hc, and the distance is equal to the constant term.

[0098] ;

[0099] Therefore, the Hamiltonian of the problem is...

[0100]

[0101] This ensures that the problem Hamiltonian Hc is greater than 0.

[0102] The optimizable parameter γ is calculated using the classic gradient descent algorithm. The optimizable parameter β Within a preset range, the parameter displacement change is compensated for by learning and iterated. When the classical optimizer deems the parameter convergent or has reached a preset number of iterations, the resulting optimizable parameter γ is obtained. The optimizable parameter β As the optimal parameter for power allocation quantum optimization.

[0103] In this step, the preset range can be set to... At this point, the optimizable parameter γ is calculated using the following formulas. The optimizable parameter β The change in parameter displacement:

[0104]

[0105]

[0106] During this process, all other parameters remain constant, except for γ. Replace with γ +π / 2, run the quantum circuit once, and measure the target function value C( +π / 2). Then add γ Replace with γ π / 2, run the circuit again, and get C(γ) (π / 2), subtract the two values ​​and divide by 2 to obtain the gradient component. For β The operation is exactly the same.

[0107] Degree obtained through parameter shifting rules C=( C / , C / β After that, classical computers use gradient descent to update the optimization parameters.

[0108] Next, the optimization parameters are updated using the following formula:

[0109]

[0110] Where α>0 is the learning compensation, used to control the magnitude of each update.

[0111] Once the classical optimizer determines that the parameters have converged or the preset number of iterations has been reached, it proceeds to this step to obtain the final beamforming scheme.

[0112] Further, step S4 includes:

[0113] The final state on the quantum processor in the current state is calculated based on the optimal parameters of the power allocation quantum optimization. The final state is then projected onto the computational basis to obtain the binary optimal parameters.

[0114] Based on the mapping relationship between the binary variables and the discrete phases, the binary optimal parameters are decoded into the phase of each antenna.

[0115] In this step, the optimal parameters are optimized using power allocation quantum optimization. Run the QAOA circuit one last time and check the final state. Projective measurements are performed under a computational basis. The measurement result is an N-bit binary string, denoted as {q}. n,j}, where each q n,j ∈{0,1}.

[0116] Based on the defined mapping relationship, the phase of each antenna n is decoded. :

[0117]

[0118] The generated beamforming vector, the final quantized beamforming vector f is:

[0119]

[0120] in It is the power normalization factor, ensuring that the total transmit power is 1, e is the natural logarithm, and i is the imaginary number. This represents the exponential form of each antenna phase after normalization.

[0121] In summary, the antenna phase (n,j) is encoded as a coefficient of Z_{n,j} in H_C. This coefficient determines the rotation angle Θ_{n,j} of the R_z gate required to simulate the evolution of H_C. By adjusting the classical parameter γ, the intensity of the R_z gate rotation is controlled, thereby guiding the phase distribution of the quantum state to evolve towards the ground state (i.e., the optimal solution) of H_C. Finally, by measuring the qubits and decoding, the obtained {q_{n,j}} is the antenna phase control word that optimizes the system performance.

[0122] Furthermore, a beamforming system for a multi-antenna wireless communication system based on quantum approximation optimization is provided to implement the aforementioned beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization, comprising the following modules:

[0123] The weighted channel correlation matrix construction module is used to obtain channel state information of multiple users, assign radio resource weights to multiple users, and construct a weighted channel correlation matrix that reflects the comprehensive channel characteristics of multiple users based on the channel state information and the radio resource weights.

[0124] The state space establishment module is used to discretize the continuous phase of each transmitting antenna and assign a binary variable that gradually doubles to the discrete phase. Based on the binary variable, the discrete phase, and the weighted channel correlation matrix, the phase combination of all antennas is used to jointly construct the quantum state representation corresponding to the quantum approximation optimization algorithm, determine the state space corresponding to the quantum approximation optimization algorithm, and initialize the quantum circuit corresponding to the state space.

[0125] The optimal parameter calculation module is used to perform quantum circuit evolution and observation through a quantum approximation optimization algorithm to obtain the initial parameters of power allocation quantum optimization and the corresponding mathematical expectation. Then, the initial parameters of power allocation quantum optimization are updated and optimized through gradient descent of the classical algorithm to obtain the optimal parameters of power allocation quantum optimization.

[0126] The optimal power allocation beam conversion module is used to convert the optimal parameters of the quantum optimization of power allocation into a phase combination of all antennas based on the binary variables and the discrete phases, so as to obtain the optimal power allocation beam.

[0127] Experimental data

[0128] The partial quantum circuit established through this embodiment can be seen in Figure 2.

[0129] As shown in Figure 3, the orange line represents the transmission rate and harvested energy corresponding to the antenna phase calculated by the existing SDR semidefinite programming method, while the blue line represents the transmission rate and harvested energy corresponding to the antenna phase calculated in this embodiment. As the antenna array size increases, the QAOA method shows a stable upward trend in both the rate and energy indicators and is superior to the existing methods, resulting in higher antenna utilization in multi-antenna scenarios.

[0130] As shown in Figure 4, the normalized received power distribution of the existing algorithm and the quantum approximation optimization algorithm in the same coordinate system of superimposed power shows that, in terms of overall shape and peak distribution, the quantum approximation optimization algorithm forms a more concentrated and sharper main lobe in the neighborhood of the target user compared to the baseline, effectively focusing energy in the desired direction while maintaining the overall energy constraint.

[0131] As can be seen from Figure 5, under the same total power conditions, the QAOA curve is significantly higher than that of existing methods for all users in terms of the trade-off between user rate, information and energy, and the slope is steeper.

[0132] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0133] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0134] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0135] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0136] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms should not be construed as necessarily referring to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

Claims

1. A beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization, characterized in that: Includes the following steps: S1. Obtain channel state information for multiple users, assign radio resource weights to multiple users, and construct a weighted channel correlation matrix reflecting the comprehensive channel characteristics of multiple users based on the channel state information and the radio resource weights; S2. Discretize the continuous phase of each transmit antenna, and assign a binary variable that gradually doubles to the discrete phase. Based on the binary variable, the discrete phase, and the weighted channel correlation matrix, combine the phases of all antennas to jointly construct the quantum state representation corresponding to the quantum approximation optimization algorithm, determine the state space corresponding to the quantum approximation optimization algorithm, and initialize the quantum circuit corresponding to the state space; Based on the binary variable, the discrete phase, and the weighted channel correlation matrix, combine the phases of all antennas to jointly construct the quantum state representation corresponding to the quantum approximation optimization algorithm, and determine the state space corresponding to the quantum approximation optimization algorithm, including: the state space corresponding to the quantum approximation optimization algorithm includes H gate, Rz gate, and Rx gate, and initialize the quantum state |+ ^{ N} as an H gate N represents the tensor product of N qubits, where the discrete phase θ j The weighted channel correlation matrix is ​​combined with the rotation angle to construct the quantum state, and the rotation angle is used as the quantum state representation, based on each discrete phase θ of all antennas. j The corresponding rotation angle determines the rotation amount of the Rz gate and the Rx gate, obtaining the state space corresponding to the quantum approximation optimization algorithm; the discrete phase θ j The rotation angle, constructed by combining the weighted channel correlation matrix with the aforementioned matrix, includes: using the discrete phase θ j The larger the corresponding binary variable, the greater the weight provided by the weighted channel correlation matrix for the discrete phase θ. j The weighted channel correlation matrix is ​​jointly weighted, and then combined with the first quantum approximation optimization algorithm. The optimizable parameter γ of the layer The rotation angle is established using the following formula. ， = 2c ·[Q] nn ·θ j Where n represents the nth antenna, and J represents θ j The j-th bit of the corresponding binary variable, [Q] nn S3, The initial parameters of the power allocation quantum optimization and the corresponding mathematical expectation are obtained by performing quantum circuit evolution and observation through the quantum approximation optimization algorithm. Then, the initial parameters of the power allocation quantum optimization are updated and optimized by the gradient descent algorithm of the classical algorithm to obtain the optimal parameters of the power allocation quantum optimization. S4, The optimal parameters of the power allocation quantum optimization are converted into the phase combination of all antennas according to the binary variable and the discrete phase to obtain the optimal power allocation beam.

2. The beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization according to claim 1, characterized in that: Step S1 includes: S1.1, obtaining the total number of users K that need to be served simultaneously, the number of antennas N_T of the base station, and the channel gain h from each antenna to the k-th user. k This forms an N_T×1 complex vector, which is used to obtain the channel vector {h} of the k-th user. k S1.2, calculate the service priority w from each antenna to the k-th user using a network scheduling algorithm or a quality of service algorithm. k This forms an N_T×1 complex vector, from which the user weight {w} of the k-th user is obtained. k };S1.3, for all users' channel vectors {h k } and user weight {w k } via Q = Σ k w k h k h k The weighted summation is performed to obtain the weighted channel correlation matrix Q.

3. The beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization according to claim 1, characterized in that: In step S2, discretizing the continuous phase of each transmitting antenna and assigning a binary variable that gradually doubles to the discrete phase includes: quantizing the number of bits b of the binary variable, based on θ. j = (2π / 2 b )·2 j Phase discretization is performed to obtain the discrete phase θ. j Where j = 0, 1, ..., b-1; through 2 j The binary representation of the corresponding number of bits b is used as the binary variable, and a mapping relationship is established between the binary variable and the discrete phase.

4. The beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization according to claim 1, characterized in that: In step S2, initializing the quantum circuit corresponding to the state space includes setting the initial phase of all antennas to 0.

5. The beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization according to claim 1, characterized in that: In step S3, quantum circuit evolution and observation are performed using a quantum approximation optimization algorithm to obtain the initial parameters and corresponding mathematical expectations for power allocation quantum optimization. Then, the initial parameters for power allocation quantum optimization are updated and optimized using gradient descent of a classical algorithm to obtain the optimal parameters for power allocation quantum optimization. This includes sequentially executing the following steps for each computational layer of the quantum approximation optimization algorithm: adjusting the rotation angle... The data to be processed, as the Rz rotating gate, is introduced into the quantum circuit. The quantum circuit evolution and observation are performed through a quantum approximation optimization algorithm, resulting in the first quantum approximation optimization algorithm. The optimizable parameter γ of the layer The rotation angle The data to be processed, used as the Rx rotating gate, is introduced into the quantum circuit. The quantum circuit evolution and observation are then performed using a quantum approximation optimization algorithm, yielding the first result of the quantum approximation optimization algorithm. The optimizable parameter β of the layer The optimizable parameter γ and the optimizable parameter β As initial parameters for power allocation quantum optimization; according to the optimizable parameter γ and the optimizable parameter β Calculate the mathematical expectation C of the final state on the quantum processor in the current state to obtain the optimizable parameter γ. and the optimizable parameter β The solution for quantum approximation optimization under the current weights; the optimizable parameter γ is calculated using the gradient descent method of the classical algorithm. The optimizable parameter β Within a preset range, the parameter displacement change is compensated for by learning and iterated. When the classical optimizer deems the parameter convergent or has reached a preset number of iterations, the resulting optimizable parameter γ is obtained. The optimizable parameter β As the optimal parameter for power allocation quantum optimization.

6. The beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization according to claim 1, characterized in that: Step S4 includes: calculating the final state on the quantum processor in the current state based on the optimal parameters of the power allocation quantum optimization; performing projection measurement on the final state under the computational basis to obtain the binary optimal parameters; and decoding the binary optimal parameters into the phase of each antenna based on the mapping relationship between the binary variables and the discrete phase.

7. A beamforming system for a multi-antenna wireless communication system based on quantum approximation optimization, characterized in that: A beamforming method for a multi-antenna wireless communication system based on quantum approximation optimization as described in any one of claims 1-6, comprising the following modules: a weighted channel correlation matrix construction module, used to obtain channel state information of multiple users, allocate radio resource weights to multiple users, and construct a weighted channel correlation matrix reflecting the comprehensive channel characteristics of multiple users based on the channel state information and the radio resource weights; The state space establishment module is used to discretize the continuous phase of each transmitting antenna and assign a binary variable that gradually doubles to the discrete phase. Based on the binary variable, the discrete phase, and the weighted channel correlation matrix, the phase combination of all antennas is used to jointly construct the quantum state representation corresponding to the quantum approximation optimization algorithm, determine the state space corresponding to the quantum approximation optimization algorithm, and initialize the quantum circuit corresponding to the state space. The optimal parameter calculation module is used to perform quantum circuit evolution and observation through a quantum approximation optimization algorithm to obtain the initial parameters of power allocation quantum optimization and the corresponding mathematical expectation. Then, the initial parameters of power allocation quantum optimization are updated and optimized through gradient descent of the classical algorithm to obtain the optimal parameters of power allocation quantum optimization. The optimal power allocation beam conversion module is used to convert the optimal parameters of power allocation quantum optimization into the phase combination of all antennas according to the binary variables and the discrete phase to obtain the optimal power allocation beam.

Images