Zero-forcing transmitting and receiving beam forming joint optimization method, device and equipment based on original decomposition method and medium

Abstract

The invention discloses a zero-forcing transmitting and receiving beam forming joint optimization method, device and equipment based on an original decomposition method and a medium, and belongs to the field of wireless communication networks, and the method comprises the following steps: constructing a zero-forcing transmitting and receiving beam forming joint optimization model of all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network, the zero-forcing transmit-receive beamforming joint optimization model comprises a network utility objective function, transmit-receive beamforming power constraint and zero interference constraint; and decomposing the zero-forcing transmit-receive beamforming joint optimization model based on an original decomposition method, designing an original decomposition algorithm based on a parallel continuous convex approximation algorithm, and carrying out parallel iteration to update transmit-receive beamforming until a preset iteration termination condition is satisfied, thereby obtaining feasible transmit-receive beamforming of all links after optimization, the transmitting and receiving beamforming obtained by each iteration of the original decomposition algorithm is feasible.

Description

Technical Field

[0001] This invention relates to the field of wireless communication networks, and in particular to a method, apparatus, device, and medium for joint optimization of zero-forcing transmit and receive beamforming based on the original decomposition method. Background Technology

[0002] To eliminate interference in multi-user multiple-input multiple-output (MIMO) wireless communication networks, some studies have considered zero-forcing transmit beamforming optimization design. This involves optimizing transmit beamforming under zero-interference constraints and total transmit power constraints to maximize network throughput. This problem is non-convex, and the classic algorithm for solving it is block diagonalization (BD). However, zero-forcing transmit beamforming optimization design does not consider the contribution of receive beamforming to zero-forcing, which can lead to a loss of network utility. Furthermore, some studies focus on joint transmit and receive beamforming design to achieve interference alignment and thus achieve zero-forcing, but these methods are generally not optimization algorithms and cannot maximize network utility. Summary of the Invention

[0003] This invention provides a method, apparatus, device, and medium for joint optimization of zero-forcing transmit / receive beamforming based on the original decomposition method. It can jointly optimize transmit / receive beamforming under zero interference constraints and transmit / receive beamforming power constraints to maximize the effectiveness of multi-user multiple-input multiple-output wireless communication networks.

[0004] This invention provides a joint optimization method for zero-forcing transmit / receive beamforming based on the original decomposition method, comprising: A joint optimization model for zero-forcing transmit and receive beamforming of all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network is constructed. The joint optimization model for zero-forcing transmit and receive beamforming includes a network utility objective function, transmit and receive beamforming power constraints, and zero interference constraints. The zero-forcing transmit / receive beamforming joint optimization model is decomposed based on the original decomposition method. An original decomposition algorithm is designed based on a parallel continuous convex approximation algorithm to perform parallel iterative updates of transmit / receive beamforming until a preset iteration termination condition is met, resulting in optimized feasible transmit / receive beamforming for all links. Specifically, the zero-forcing transmit / receive beamforming joint optimization model is decomposed into a main model for all transmit beamforming and multiple sub-models for receive beamforming. In each iteration, each sub-model is solved in parallel based on the transmit beamforming of the previous iteration to obtain the receive beamforming for the current iteration. Based on the receive beamforming of the current iteration, the main model is approximated into multiple sub-main models for transmit beamforming using a parallel continuous convex approximation algorithm. Each sub-main model is solved in parallel, and then each transmit beamforming is updated in parallel. The transmit / receive beamforming obtained in each iteration of the original decomposition algorithm is feasible.

[0005] This invention constructs a joint optimization model for zero-forcing transmit / receive beamforming that includes a network utility objective function, transmit / receive beamforming power constraints, and zero-interference constraints. This provides a mathematical model foundation for maximizing network performance in zero-forcing scenarios. By employing a primitive decomposition method to decompose the joint optimization model into a master model and sub-models and performing parallel iterations, large-scale non-convex optimization models can be solved efficiently, resulting in optimized transmit / receive beamforming. Compared to existing technologies that do not consider the contribution of receive beamforming to zero-forcing and cannot maximize network utility, this application can jointly optimize transmit / receive beamforming under zero-interference and transmit / receive beamforming power constraints to maximize the utility of multi-user multiple-input multiple-output wireless communication networks.

[0006] Furthermore, the construction of the joint optimization model for zero-forcing transmit / receive beamforming of all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network includes: A joint optimization model for zero-forcing transmit / receive beamforming is constructed for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network to maximize network utility under transmit / receive beamforming power constraints and zero-interference constraints. Specifically, the joint optimization model for zero-forcing transmit / receive beamforming is as follows: ; in, For network utility; These are the optimization variables that imply joint optimization of transmit and receive beamforming; Channel state information; For all transmitter sets; for all , For transmitter The corresponding set of receivers; For the set of all subcarriers; for all , and ,transmitter In subcarrier Upward receiver The maximum number of data streams that can be transmitted is The data stream set is denoted as For all , For transmitter In subcarrier Upward receiver Transmitted data stream The corresponding transmit beamforming vector, For receiver In subcarrier upper receiver transmitter Transmitted data stream The corresponding receiving beamforming vector; for Norm; for all , For transmitter Maximum transmission power.

[0007] This invention provides a solution object and objective for the subsequent original decomposition algorithm by establishing a joint optimization model for zero-forcing transmit and receive beamforming that includes an objective function and three constraints.

[0008] Further, the zero-forcing transmit / receive beamforming joint optimization model is decomposed based on the original decomposition method, and the original decomposition algorithm is designed based on the parallel continuous convex approximation algorithm. Parallel iterative updates of the transmit / receive beamforming are performed until a preset iteration termination condition is met, resulting in optimized feasible transmit / receive beamforming for all links, including: The zero-forcing transmit / receive beamforming joint optimization model is decomposed into a master model for all transmit beamformings and multiple sub-models for receive beamformings using the primal decomposition method. The master model and each sub-model are iteratively solved using the primal decomposition algorithm based on parallel continuous convex approximation until a preset iteration termination condition is met, outputting the receive and transmit beamformings for the current iteration. In each iteration, the sub-models are solved in parallel based on the transmit beamformings from the previous iteration to obtain the receive beamformings for the current iteration. Based on the receive beamformings for the current iteration, the master model is approximated into multiple sub-master models for transmit beamformings using the parallel continuous convex approximation algorithm. Each sub-master model is solved in parallel, and the transmit beamformings are then updated in parallel. The transmit / receive beamformings obtained in each iteration of the primal decomposition algorithm are all feasible.

[0009] The embodiments of the present invention design an iterative framework that solves the sub-model in parallel before solving the main model in each iteration. This enables the algorithm to update the points of the original model in each iteration and eventually converge to the stationary point of the model.

[0010] Furthermore, the original decomposition method decomposes the joint optimization model of zero-forcing transmit and receive beamforming into a master model for all transmit beamforming and multiple sub-models for receive beamforming, including: Regarding any and The specific sub-model for receiving beamforming is as follows: ; The main model for all transmitted beamforming is as follows: ; in, , , For the sub-model in A post in the area.

[0011] The embodiments of the present invention decompose the original model into a master model and a sub-model, which lays the foundation for subsequent parallel computing and closed-form solving, thereby transforming a large-scale non-convex optimization model into a sub-task that is easy to process.

[0012] Furthermore, the parallel solution of each sub-model based on the previous iteration's transmit beamforming includes: The stationary points of each sub-model are computed in parallel; wherein, the zero-disturbance constraints of each sub-model are eliminated by variable substitution, and the equivalent transformation sub-models are solved in parallel based on the Cauchy-Schwarz inequality; Construct a set of equations for the Lagrangian function of each sub-model with zero gradient with respect to the corresponding receiving beamforming vector, and solve each set of equations in parallel to obtain the Lagrangian multipliers with respect to zero interference constraints corresponding to the stationary points of each sub-model.

[0013] This invention employs variable substitution and Cauchy-Schwarz inequality to solve the sub-model in parallel, and simultaneously solves the Lagrange multipliers corresponding to its stationary points. This enables the acquisition of the optimal solution for receiving beamforming while providing the necessary key information for the subsequent gradient calculation of the main model.

[0014] Furthermore, based on each received beamforming iteration in the current iteration, the main model is approximated into multiple sub-main models with respect to the transmitted beamforming using a parallel continuous convex approximation algorithm, and each sub-main model is solved in parallel, including: Multiple sub-master models are constructed based on the master model; wherein, any transmitter corresponds to one sub-master model; the first... In the next iteration, regarding any transmitter The sub-main model is specifically as follows: ; in, These are algorithm parameters; For Frobenius norm; gradient It is calculated based on the stationary points of all the sub-models and their corresponding Lagrange multipliers; By analyzing the KKT conditions, the optimal solutions of each of the sub-principal models are calculated in parallel.

[0015] The embodiments of the present invention can efficiently handle the updates of transmit beamforming in the main model by constructing an independent sub-master model for each transmitter and solving it in parallel using KKT conditions.

[0016] Furthermore, after analyzing the KKT conditions and calculating the optimal solutions of each of the sub-principal models in parallel, the method further includes: All transmit beamformings are updated in parallel based on the optimal solutions of each of the sub-master models.

[0017] The embodiments of the present invention ensure that the iterative update direction of the transmit beamforming is an effective direction that has been optimized and calculated by parallel updating all transmit beamforming based on the optimal solution of each of the sub-master models, thereby driving the algorithm to converge toward a solution with better performance.

[0018] Another embodiment of the present invention provides a joint optimization device for zero-forcing transmit and receive beamforming based on the original decomposition method, comprising: a joint optimization model module and an original decomposition algorithm module; The joint optimization model module is used to construct a zero-forcing transmit / receive beamforming joint optimization model for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network. The zero-forcing transmit / receive beamforming joint optimization model includes a network utility objective function, transmit / receive beamforming power constraints, and zero-interference constraints. The original decomposition algorithm module is used to decompose the zero-forcing transmit / receive beamforming joint optimization model based on the original decomposition method, and to design the original decomposition algorithm based on the parallel continuous convex approximation algorithm to perform parallel iterative updates of transmit / receive beamforming until a preset iteration termination condition is met, thereby obtaining the feasible transmit / receive beamformings for all optimized links. The zero-forcing transmit / receive beamforming joint optimization model is decomposed into a main model concerning all transmit beamformings and multiple sub-models concerning receive beamformings. In each iteration, each sub-model is solved in parallel based on the transmit beamformings of the previous iteration to obtain the receive beamformings of the current iteration. Based on the receive beamformings of the current iteration, the main model is approximated into multiple sub-main models concerning transmit beamformings using the parallel continuous convex approximation algorithm. Each sub-main model is solved in parallel, and then each transmit beamforming is updated in parallel. The transmit / receive beamformings obtained in each iteration of the original decomposition algorithm are all feasible.

[0019] Another embodiment of the present invention provides a terminal device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the steps of a zero-forcing transmit / receive beamforming joint optimization method based on the original decomposition method of the present invention.

[0020] Another embodiment of the present invention also provides a computer-readable storage medium item, including: a stored computer program, which, when the computer program is running, controls the device where the computer-readable storage medium is located to perform the steps of a zero-forcing transmit / receive beamforming joint optimization method based on the original decomposition method of the present invention. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating an embodiment of the zero-forcing transmit / receive beamforming joint optimization method based on the original decomposition method provided by the present invention. Figure 2 A flowchart illustrating an embodiment of the original decomposition method provided by the present invention; Figure 3 This is a schematic diagram of an embodiment of the zero-forcing transmit / receive beamforming joint optimization device based on the original decomposition method provided by the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0024] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0025] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0026] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0027] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0028] See Figure 1 To address the problem that existing technologies do not consider the contribution of receive beamforming to zero-forcing and cannot maximize network utility, an embodiment of the present invention provides a joint optimization method for zero-forcing transmit and receive beamforming based on the original decomposition method, including steps S101 to S102: Step S101: Construct a joint optimization model for zero-forcing transmit / receive beamforming for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network. The joint optimization model for zero-forcing transmit / receive beamforming includes a network utility objective function, transmit / receive beamforming power constraints, and zero-interference constraints.

[0029] It should be noted that the multi-user multiple-input multiple-output wireless communication network described in this invention contains... One transmitter and One receiver ( and (Cannot all be 1), the set of transmitters is The set of receivers is Each transmitter Towards The set of receivers that transmit signals is defined as follows: ,and Each receiver Receive from The set of transmitters corresponding to the signals from each transmitter is defined as follows: ,and For all ,transmitter Equipped The transmitting antenna has a maximum transmitting power of [number] units. (Unit: Watt). For all receiver Equipped Root receiving antenna. All transmitters and receivers are in bandwidth of It operates on the frequency band, and the number of subcarriers is The set of subcarriers is denoted as In particular, when and , Time (can be inferred) , This wireless communication network can represent a multi-user downlink cellular network, in which Represents a single community. Representing multiple communities; when and , Time (can be inferred) , This wireless communication network can represent a multi-user uplink cellular network, where Represents a single community. Representing multiple communities; when , , and , Time (in combination) and It can be inferred This wireless communication network can represent a multi-user D2D network; when And for all that satisfy of All , In this case, the wireless communication network can represent a hybrid network of cellular networks and D2D networks.

[0030] Furthermore, large-scale fading and small-scale fading are used to model the wireless channel, and block fading and frequency-selective fading models are used to model small-scale fading. Considering any time, for all , and ,make and They represent the receivers respectively. and transmitter Between subcarriers Large-scale fading power and small-scale fading coefficient of the up-channel transmitter and receiver Between subcarriers The total fading coefficient of the channel can be expressed as When designing a resource management method based on instantaneous channel state information (CSI), it is assumed that the channel state information... It is known (and can be obtained by traditional channel estimation methods).

[0031] Furthermore, each transceiver employs linear beamforming to enhance the transmitted signal and reduce interference. For all , and ,transmitter In subcarrier Upward receiver The maximum number of data streams that can be transmitted is The data stream set is denoted as For all ,make Indicates transmitter In subcarrier Upward receiver Transmitted data stream The corresponding transmit beamforming vector, let Indicates receiver In subcarrier upper receiver transmitter Transmitted data stream The corresponding receiving beamforming vector. Let Indicates transmitter In subcarrier Upward receiver The transmit beamforming matrix corresponding to all transmitted data streams, let Indicates receiver In subcarrier upper receiver transmitter The receive beamforming matrix corresponding to all transmitted data streams. The corresponding transmit and receive beamforming power constraints are as follows: , ; in, express Norm. To eliminate interference between all links and data streams, a zero-interference constraint needs to be further introduced: ; To ensure that the set of feasible points defined by the zero-interference constraint is always non-empty, this invention assumes... .

[0032] Furthermore, for all , , and , Represents the transmitter In subcarrier Upward receiver Transmitted data stream , signal power Represents the transmitter In subcarrier Upward receiver Transmitted data stream beam direction, Represents the transmitter In subcarrier Upward receiver Transmit data stream , Represents the transmitter In subcarrier There was no signal to the receiver. Transmit data stream ,thus, Represents the transmitter In subcarrier Upward receiver The actual number of data streams transmitted, of which This represents the indicator function. Therefore, the beam vectors of all transceivers... , , , , , It can characterize transmit beamforming, power control, frequency selection, and data stream allocation.

[0033] Furthermore, for all and ,transmitter In subcarrier On the transmission signal as follows: ; in, It is a transmitter In subcarrier Upward receiver Launched The symbols of a data stream are independent and identically distributed, with a mean of zero and a variance of 1. For all and receiver In subcarrier The signal received above as follows: ; in, It is a receiver In subcarrier The vector of additive white Gaussian noise (AWGN) on the vector. yes An identity matrix of order 1. Indicates receiver In subcarrier The noise power on. For all , and After applying receive beamforming, the receiver In subcarrier Received from the transmitter signal as follows: .

[0034] Furthermore, for all , , and Under zero interference constraints, the receiver In subcarrier From the transmitter data stream The signal-to-noise ratio (SNR) is: ; Among them, the transmitter In subcarrier Upward receiver Transmitted data stream The achievable rate is denoted as It is given by the following formula: .

[0035] Furthermore, for all and ,transmitter to the receiver The achievable launch rate is: ; in, .

[0036] Furthermore, the performance indicators of wireless communication networks can be expressed as: ; in, , , For from the transmitter to receiver A monotonically increasing, non-negative, and differentiable utility function. For receiver The utility function is monotonically increasing, non-negative, and differentiable. The above-mentioned performance indicators of wireless communication networks... Simply put, it is network utility, usually... The non-convex function.

[0037] Preferably, the construction of the joint optimization model for zero-forcing transmit / receive beamforming of all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network includes: A joint optimization model for zero-forcing transmit / receive beamforming is constructed for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network to maximize network utility under transmit / receive beamforming power constraints and zero-interference constraints. Specifically, the joint optimization model for zero-forcing transmit / receive beamforming is as follows: ; in, For network utility; These are the optimization variables that imply joint optimization of transmit and receive beamforming; Channel state information; For all transmitter sets; for all , For transmitter The corresponding set of receivers; For the set of all subcarriers; for all , and ,transmitter In subcarrier Upward receiver The maximum number of data streams that can be transmitted is The data stream set is denoted as For all , For transmitter In subcarrier Upward receiver Transmitted data stream The corresponding transmit beamforming vector, For receiver In subcarrier upper receiver transmitter Transmitted data stream The corresponding receiving beamforming vector; for Norm; for all , For transmitter Maximum transmission power.

[0038] Specifically, the objective function Receive beamforming power constraint and zero-interference constraints Since both are non-convex, the zero-forcing transmit / receive beamforming joint optimization model is a large-scale non-convex optimization model.

[0039] Furthermore, due to Equivalent to , and when When, the objective function is due to The operator is non-differentiable. Therefore, the joint optimization model for zero-forcing transmit / receive beamforming can be transformed to obtain the following model: ; in, The value can be determined according to The differences can be categorized into three situations: In the first case, when hour, .

[0040] In the second scenario, when When the objective function is not differentiable, use the differentiable log-sum-exp function. To approximate the non-differentiable max function ,in .at this time, .in, and The upper bound of the approximate error between them, i.e. , It decreases as it increases, and can be arbitrarily small.

[0041] The third scenario, when When the objective function is not differentiable, similar to the second case, a differentiable log-sum-exp function should be used. To approximate the non-differentiable max function .at this time, The transformed objective function It is a differentiable nonconvex function, receiving beamforming power constraint and zero-interference constraints Both are non-convex. Therefore, the transformed zero-forcing transmit / receive beamforming joint optimization model is still a large-scale non-convex optimization model.

[0042] Step S102: The zero-forcing transmit / receive beamforming joint optimization model is decomposed based on the original decomposition method. An original decomposition algorithm is designed based on a parallel continuous convex approximation algorithm, and the transmit / receive beamforming is updated in parallel iteratively until a preset iteration termination condition is met, resulting in feasible transmit / receive beamformings for all optimized links. The zero-forcing transmit / receive beamforming joint optimization model is decomposed into a main model for all transmit beamformings and multiple sub-models for receive beamformings. In each iteration, each sub-model is solved in parallel based on the transmit beamformings of the previous iteration to obtain the receive beamformings for the current iteration. Based on the receive beamformings of the current iteration, the main model is approximated into multiple sub-main models for transmit beamformings using a parallel continuous convex approximation algorithm. Each sub-main model is solved in parallel, and then each transmit beamforming is updated in parallel. The transmit / receive beamformings obtained in each iteration of the original decomposition algorithm are feasible.

[0043] Preferably, the decomposition of the zero-forcing transmit / receive beamforming joint optimization model based on the original decomposition method, and the design of the original decomposition algorithm based on the parallel continuous convex approximation algorithm, are performed to update the transmit / receive beamforming in parallel iteratively until a preset iteration termination condition is met, resulting in the optimized feasible transmit / receive beamforming for all links, including: The zero-forcing transmit / receive beamforming joint optimization model is decomposed into a master model for all transmit beamformings and multiple sub-models for receive beamformings using the primal decomposition method. The master model and each sub-model are iteratively solved using the primal decomposition algorithm based on parallel continuous convex approximation until a preset iteration termination condition is met, outputting the receive and transmit beamformings for the current iteration. In each iteration, the sub-models are solved in parallel based on the transmit beamformings from the previous iteration to obtain the receive beamformings for the current iteration. Based on the receive beamformings for the current iteration, the master model is approximated into multiple sub-master models for transmit beamformings using the parallel continuous convex approximation algorithm. Each sub-master model is solved in parallel, and the transmit beamformings are then updated in parallel. The transmit / receive beamformings obtained in each iteration of the primal decomposition algorithm are all feasible.

[0044] Specifically, fix all transmitted beamforming Later, due to And utility function and and functions , and All are monotonically increasing, model The objective function and constraints can be about Complete decoupling. Therefore, It can be regarded as a model Coupled variables, It can be regarded as a model Local variables. Therefore, this invention will model Decomposed into all received beamforming Sub-models and about all transmitted beamforming The main models are all non-convex models.

[0045] Preferably, the joint optimization model for zero-forcing transmit / receive beamforming based on the original decomposition method is decomposed into a main model for all transmit beamforming and multiple sub-models for receive beamforming, including: Regarding any and The specific sub-model for receiving beamforming is as follows: ; The main model for all transmitted beamforming is as follows: ; in, , , For the sub-model in A post in the area.

[0046] Specifically, under certain conditions, the model The stationary points can be determined by the aforementioned main model. Stationary points and sub-models , The stationary points are constructed. Therefore, this invention solves the model by designing a primitive decomposition algorithm to solve the above-mentioned main model and each sub-model. The steps of the original decomposition algorithm are as follows: Figure 2 As shown. Due to the main model objective function Explicit expressions of are difficult to obtain; only local information about them at any point can be obtained (e.g., their position at 0°). function value at and gradient value Therefore, iterative solutions to each sub-model are required. and based on the acquired Local information updates coupling variables Based on this, this invention proposes an algorithm based on Parallel Successive Convex Approximation (PSCA) to solve the master model. The original decomposition algorithm is suitable for degraded system models, such as single / multi-cell downlink / uplink cellular networks, device-to-device (D2D) networks, and hybrid networks.

[0047] Preferably, the parallel solution of each sub-model based on the previous iteration's transmit beamforming includes: The stationary points of each sub-model are computed in parallel; wherein, the zero-disturbance constraints of each sub-model are eliminated by variable substitution, and the equivalent transformation sub-models are solved in parallel based on the Cauchy-Schwarz inequality; Construct a set of equations for the Lagrangian function of each sub-model with zero gradient with respect to the corresponding receiving beamforming vector, and solve each set of equations in parallel to obtain the Lagrangian multipliers with respect to zero interference constraints corresponding to the stationary points of each sub-model.

[0048] Specifically, in the first In the next iteration, fix (in the (obtained during the next iteration), for all information regarding receive beamforming sub-model The following steps are executed in parallel: First, calculate each sub-model. The optimal point (i.e., the stationary point) First, because in a fixed Sub-model Zero-disturbance constraint For about Linear equality constraints can be eliminated in sub-models through variable substitution. Zero disturbance constraint in Specifically, define sub-models. All zero-disturbance constraints coefficient matrix ,in It is the corresponding index in the matrix. of rows, by It can be inferred rank The matrix is ​​decomposed using the standard Singular Value Decomposition (SVD) method. Decomposed into , where the matrix It is a left singular matrix, a matrix It is a singular value matrix whose diagonal elements are arranged in descending order. Singular values ​​of the matrix It is a right singular matrix. , The columns are matrices A set of orthonormal bases in null space, (Depend on It can be inferred ), and satisfy for An identity matrix of order 1. Therefore, an auxiliary variable is defined. By variable substitution Eliminate sub-model All zero-disturbance constraints and sub-model Equivalently transformed into the following form: ; Based on the Cauchy-Schwarz inequality, the optimal solution for this model can be obtained as follows: Therefore, it can be seen that the non-convex submodel... The optimal point (i.e., the stationary point) is: .

[0049] Second, calculate each sub-model. Best Corresponding to zero-disturbance constraints Lagrange multipliers Solving the sub-model The Lagrange function with respect to The system of equations whose gradient is zero: ; in, The solution to this system of linear equations It can be obtained by conventional methods (such as LU decomposition, Cholesky decomposition, QER decomposition, or SVD).

[0050] Preferably, the receiving beamforming based on the current iteration approximates the main model into multiple sub-main models with respect to the transmitting beamforming using a parallel continuous convex approximation algorithm, and solves each sub-main model in parallel, including: Multiple sub-master models are constructed based on the master model; wherein, any transmitter corresponds to one sub-master model; the first... In the next iteration, regarding any transmitter The sub-main model is specifically as follows: ; in, These are algorithm parameters; For Frobenius norm; gradient It is calculated based on the stationary points of all the sub-models and their corresponding Lagrange multipliers; By analyzing the KKT conditions, the optimal solutions of each of the sub-principal models are calculated in parallel.

[0051] Specifically, according to the function exist The second-order Taylor approximation at the point is used to construct the principal model. exist Regarding the transmission beamforming Sub-main model .in, The calculation can be based on The different expressions can be categorized into four cases: In the first case, when hour, .

[0052] In the second scenario, when hour, .

[0053] The third scenario, when hour, .

[0054] The fourth scenario, when hour, .

[0055] Furthermore, all sub-master models are computed in parallel. The best advantage According to the sub-master model The optimal solution can be obtained by applying the KKT (Karush-Kuhn-Tucker) conditions: ; in, .

[0056] Preferably, after analyzing the KKT conditions and calculating the optimal solutions of each of the sub-principal models in parallel, the method further includes: All transmit beamformings are updated in parallel based on the optimal solutions of each of the sub-master models.

[0057] Specifically, all transmit beamformings are updated in parallel. : ; in, It is iteration The step size.

[0058] In summary, the flowchart of the original decomposition algorithm is shown in Table 1: Table 1 - Flowchart of the Primitive Decomposition Algorithm

[0059] Specifically, in steps 4, 5, and 7-9, parallel computing can be applied to perform matrix (vector) multiplication and addition. The updates obtained in each iteration of this original decomposition algorithm represent feasible points. The algorithm parameters of the original decomposition algorithm proposed in this invention are parameters. and each iteration Step size Algorithm parameters can be specifically chosen, or they can be obtained through deep unrolling methods and optimization using neural networks and data samples. When the model... The algorithm iterative update satisfies certain conditions, and the iteration step size satisfies the following conditions: ; It can be proven that the original decomposition algorithm converges to the model. The outpost.

[0060] It should be noted that the algorithm parameters of the original decomposition algorithm proposed in this application can be specifically selected, or obtained through deep expansion methods and optimization using neural networks and data samples. Furthermore, besides the original decomposition algorithm described in this application, other approximate implementation methods can be designed to solve the zero-forcing transmit / receive beamforming joint optimization model described in this application. Simultaneously, the complex operations in the original decomposition algorithm described in this application, such as solving linear equations and calculating gradients, can also be approximated by designing neural network modules to reduce computation time within the allowable range of computational accuracy errors.

[0061] This invention constructs a joint optimization model for zero-forcing transmit / receive beamforming that includes a network utility objective function, transmit / receive beamforming power constraints, and zero-interference constraints. This provides a mathematical model foundation for maximizing network performance in zero-forcing scenarios. By employing a primitive decomposition method to decompose the joint optimization model into a master model and sub-models and performing parallel iterations, large-scale non-convex optimization models can be solved efficiently, resulting in optimized transmit / receive beamforming. Compared to existing technologies that do not consider the contribution of receive beamforming to zero-forcing and cannot maximize network utility, this application can jointly optimize transmit / receive beamforming under zero-interference and transmit / receive beamforming power constraints to maximize the utility of multi-user multiple-input multiple-output wireless communication networks.

[0062] Optionally, in this embodiment of the invention, the construction of a joint optimization model for zero-forcing transmit / receive beamforming for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network includes: A joint optimization model for zero-forcing transmit / receive beamforming is constructed for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network to maximize network utility under transmit / receive beamforming power constraints and zero-interference constraints. Specifically, the joint optimization model for zero-forcing transmit / receive beamforming is as follows: ; in, For network utility; These are the optimization variables that imply joint optimization of transmit and receive beamforming; Channel state information; For all transmitter sets; for all , For transmitter The corresponding set of receivers; For the set of all subcarriers; for all , and ,transmitter In subcarrier Upward receiver The maximum number of data streams that can be transmitted is The data stream set is denoted as For all , For transmitter In subcarrier Upward receiver Transmitted data stream The corresponding transmit beamforming vector, For receiver In subcarrier upper receiver transmitter Transmitted data stream The corresponding receiving beamforming vector; for Norm; for all , For transmitter Maximum transmission power.

[0063] This invention provides a solution object and objective for the subsequent original decomposition algorithm by establishing a joint optimization model for zero-forcing transmit and receive beamforming that includes an objective function and three constraints.

[0064] Optionally, in this embodiment of the invention, the decomposition of the zero-forcing transmit / receive beamforming joint optimization model based on the original decomposition method, and the design of the original decomposition algorithm based on the parallel continuous convex approximation algorithm, and the parallel iterative updating of transmit / receive beamforming until the preset iteration termination condition is met, to obtain the optimized feasible transmit / receive beamforming for all links, includes: The zero-forcing transmit / receive beamforming joint optimization model is decomposed into a master model for all transmit beamformings and multiple sub-models for receive beamformings using the primal decomposition method. The master model and each sub-model are iteratively solved using the primal decomposition algorithm based on parallel continuous convex approximation until a preset iteration termination condition is met, outputting the receive and transmit beamformings for the current iteration. In each iteration, the sub-models are solved in parallel based on the transmit beamformings from the previous iteration to obtain the receive beamformings for the current iteration. Based on the receive beamformings for the current iteration, the master model is approximated into multiple sub-master models for transmit beamformings using the parallel continuous convex approximation algorithm. Each sub-master model is solved in parallel, and the transmit beamformings are then updated in parallel. The transmit / receive beamformings obtained in each iteration of the primal decomposition algorithm are all feasible.

[0065] The embodiments of the present invention design an iterative framework that solves the sub-model in parallel before solving the main model in each iteration. This enables the algorithm to update the points of the original model in each iteration and eventually converge to the stationary point of the model.

[0066] Optionally, in this embodiment of the invention, the decomposition of the zero-forcing transmit / receive beamforming joint optimization model based on the original decomposition method into a main model for all transmit beamformings and multiple sub-models for receive beamformings includes: Regarding any and The specific sub-model for receiving beamforming is as follows: ; The main model for all transmitted beamforming is as follows: ; in, , , For the sub-model in A post in the area.

[0067] The embodiments of the present invention decompose the original model into a master model and a sub-model, which lays the foundation for subsequent parallel computing and closed-form solving, thereby transforming a large-scale non-convex optimization model into a sub-task that is easy to process.

[0068] Optionally, in this embodiment of the invention, the parallel solution of each sub-model based on the previous iteration of each transmitted beamforming includes: The stationary points of each sub-model are computed in parallel; wherein, the zero-disturbance constraints of each sub-model are eliminated by variable substitution, and the equivalent transformation sub-models are solved in parallel based on the Cauchy-Schwarz inequality; Construct a set of equations for the Lagrangian function of each sub-model with zero gradient with respect to the corresponding receiving beamforming vector, and solve each set of equations in parallel to obtain the Lagrangian multipliers with respect to zero interference constraints corresponding to the stationary points of each sub-model.

[0069] This invention employs variable substitution and Cauchy-Schwarz inequality to solve the sub-model in parallel, and simultaneously solves the Lagrange multipliers corresponding to its stationary points. This enables the acquisition of the optimal solution for receiving beamforming while providing the necessary key information for the subsequent gradient calculation of the main model.

[0070] Optionally, in this embodiment of the invention, the step of approximating the main model with respect to the transmit beamforming based on each received beamforming in the current iteration, and solving each of the sub-main models in parallel, includes: Multiple sub-master models are constructed based on the master model; wherein, any transmitter corresponds to one sub-master model; the first... In the next iteration, regarding any transmitter The sub-main model is specifically as follows: ; in, These are algorithm parameters; For Frobenius norm; gradient It is calculated based on the stationary points of all the sub-models and their corresponding Lagrange multipliers; By analyzing the KKT conditions, the optimal solutions of each of the sub-principal models are calculated in parallel.

[0071] The embodiments of the present invention can efficiently handle the updates of transmit beamforming in the main model by constructing an independent sub-master model for each transmitter and solving it in parallel using KKT conditions.

[0072] Optionally, in this embodiment of the invention, after analyzing the KKT conditions and calculating the optimal solutions of each of the sub-principal models in parallel, the method further includes: All transmit beamformings are updated in parallel based on the optimal solutions of each of the sub-master models.

[0073] The embodiments of the present invention update the transmit beamforming in parallel based on the optimal solutions of each of the sub-master models, which can ensure that the iterative update direction of the transmit beamforming is an effective direction after optimization calculation, thereby driving the algorithm to converge toward a solution with better performance.

[0074] like Figure 3 As shown, based on the above method embodiments, corresponding apparatus embodiments are provided; One embodiment of the present invention provides a joint optimization device for zero-forcing transmit and receive beamforming based on the original decomposition method, comprising: a joint optimization model module 301 and an original decomposition algorithm module 302; The joint optimization model module 301 is used to construct a zero-forcing transmit / receive beamforming joint optimization model for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network. The zero-forcing transmit / receive beamforming joint optimization model includes a network utility objective function, transmit / receive beamforming power constraints, and zero-interference constraints. The original decomposition algorithm module 302 is used to decompose the zero-forcing transmit / receive beamforming joint optimization model based on the original decomposition method, and to design the original decomposition algorithm based on the parallel continuous convex approximation algorithm to perform parallel iterative updates of transmit / receive beamforming until a preset iteration termination condition is met, thereby obtaining the feasible transmit / receive beamformings for all optimized links. The zero-forcing transmit / receive beamforming joint optimization model is decomposed into a main model concerning all transmit beamformings and multiple sub-models concerning receive beamformings. In each iteration, each sub-model is solved in parallel based on the transmit beamformings of the previous iteration to obtain the receive beamformings of the current iteration. Based on the receive beamformings of the current iteration, the main model is approximated into multiple sub-main models concerning transmit beamformings using the parallel continuous convex approximation algorithm. Each sub-main model is solved in parallel, and then each transmit beamforming is updated in parallel. The transmit / receive beamformings obtained in each iteration of the original decomposition algorithm are all feasible.

[0075] Optionally, in this embodiment of the invention, the joint optimization model module 301 includes: a joint optimization model submodule; The joint optimization model submodule is used to construct a zero-forcing transmit / receive beamforming joint optimization model for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network, so as to maximize the network utility under transmit / receive beamforming power constraints and zero interference constraints; wherein, the zero-forcing transmit / receive beamforming joint optimization model is specifically as follows: ; in, For network utility; These are the optimization variables that imply joint optimization of transmit and receive beamforming; Channel state information; For all transmitter sets; for all , For transmitter The corresponding set of receivers; For the set of all subcarriers; for all , and ,transmitter In subcarrier Upward receiver The maximum number of data streams that can be transmitted is The data stream set is denoted as For all , For transmitter In subcarrier Upward receiver Transmitted data stream The corresponding transmit beamforming vector, For receiver In subcarrier upper receiver transmitter Transmitted data stream The corresponding receiving beamforming vector; for Norm; for all , For transmitter Maximum transmission power.

[0076] This invention provides a solution object and objective for the subsequent original decomposition algorithm by establishing a joint optimization model for zero-forcing transmit and receive beamforming that includes an objective function and three constraints.

[0077] Optionally, in this embodiment of the invention, the original decomposition algorithm module 302 includes: a parallel iteration submodule; The parallel iterative submodule is used to decompose the zero-forcing transmit / receive beamforming joint optimization model into a master model for all transmit beamformings and multiple sub-models for receive beamformings based on the primal decomposition method. It then uses the primal decomposition algorithm based on parallel continuous convex approximation to iteratively solve the master model and each of the sub-models until a preset iteration termination condition is met, outputting the receive beamformings and transmit beamformings for the current iteration. In each iteration, based on the transmit beamformings of the previous iteration, each of the sub-models is solved in parallel to obtain the receive beamformings for the current iteration. Based on the receive beamformings of the current iteration, the master model is approximated into multiple sub-master models for transmit beamformings using the parallel continuous convex approximation algorithm. Each sub-master model is solved in parallel, and then each transmit beamforming is updated in parallel. The transmit / receive beamformings obtained in each iteration of the primal decomposition algorithm are all feasible.

[0078] The embodiments of the present invention design an iterative framework that solves the sub-model in parallel before solving the main model in each iteration. This enables the algorithm to update the points of the original model in each iteration and eventually converge to the stationary point of the model.

[0079] Optionally, in this embodiment of the invention, the parallel iterative submodule includes: a sub-model unit and a main model unit; The sub-model unit is used for any... and The specific sub-model for receiving beamforming is as follows: ; The main model unit, specifically the main model for shaping all transmitted beams, is as follows: ; in, , , For the sub-model in A post in the area.

[0080] The embodiments of the present invention decompose the original model into a master model and a sub-model, which lays the foundation for subsequent parallel computing and closed-form solving, thereby transforming a large-scale non-convex optimization model into a sub-task that is easy to process.

[0081] Optionally, in this embodiment of the invention, the parallel iterative submodule further includes: a first sub-model solving unit and a second sub-model solving unit; The first sub-model solving unit is used to calculate the stationary points of each sub-model in parallel; wherein, the zero-disturbance constraints of each sub-model are eliminated by variable substitution, and the equivalent transformed sub-models are solved in parallel based on the Cauchy-Schwarz inequality; The second sub-model solving unit is used to construct a set of equations for the Lagrangian function of each sub-model with the gradient of the corresponding receiving beamforming vector being zero, and solve each set of equations in parallel to obtain the Lagrangian multipliers with respect to the zero-interference constraint corresponding to the stationary point of each sub-model.

[0082] This invention employs variable substitution and Cauchy-Schwarz inequality to solve the sub-model in parallel, and simultaneously solves the Lagrange multipliers corresponding to its stationary points. This enables the acquisition of the optimal solution for receiving beamforming while providing the necessary key information for the subsequent gradient calculation of the main model.

[0083] Optionally, in this embodiment of the invention, the parallel iterative submodule further includes: a sub-master model unit and a sub-master model solving unit; The sub-master model unit is used to construct multiple sub-master models based on the master model; wherein, any transmitter corresponds to one sub-master model; the first... In the next iteration, regarding any transmitter The sub-main model is specifically as follows: ; in, These are algorithm parameters; For Frobenius norm; gradient It is calculated based on the stationary points of all the sub-models and their corresponding Lagrange multipliers; The sub-principal model solving unit is used to calculate the optimal solution of each sub-principal model in parallel by analyzing the KKT conditions.

[0084] The embodiments of the present invention can efficiently handle the updates of transmit beamforming in the main model by constructing an independent sub-master model for each transmitter and solving it in parallel using KKT conditions.

[0085] Optionally, in this embodiment of the invention, a transmit beamforming update unit is further included after the parallel iterative submodule; The transmit beamforming update unit is used to update all transmit beamformings in parallel based on the optimal solutions of each of the sub-master models.

[0086] The embodiments of the present invention update all transmit beamformings in parallel based on the optimal solutions of each of the sub-master models, which ensures that the iterative update direction of transmit beamforming is an effective direction after optimization calculation, thereby driving the algorithm to converge toward a solution with better performance.

[0087] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the zero-forcing transmit / receive beamforming joint optimization method based on the original decomposition method provided by any of the above-described method embodiments of the present invention.

[0088] This invention constructs a joint optimization model for zero-forcing transmit / receive beamforming, including a network utility objective function, transmit / receive beamforming power constraints, and zero-interference constraints, through a joint optimization model module 301. This provides a mathematical model foundation for maximizing network performance in zero-forcing scenarios. The primal decomposition algorithm module 302 decomposes the joint optimization model into a master model and sub-models using a primal decomposition method, performing parallel iterations to efficiently solve large-scale non-convex optimization models, thereby obtaining optimized transmit / receive beamforming. Compared to existing technologies that do not consider the contribution of receive beamforming to zero-forcing and cannot maximize network utility, this application can jointly optimize transmit / receive beamforming under zero-interference and transmit / receive beamforming power constraints to maximize the utility of multi-user multiple-input multiple-output wireless communication networks.

[0089] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0090] Based on the above embodiment of the zero-forcing transmit / receive beamforming joint optimization method based on the primal decomposition method, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the zero-forcing transmit / receive beamforming joint optimization method based on the primal decomposition method of any embodiment of the present invention.

[0091] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0092] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0093] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0094] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute a zero-forcing transmit / receive beamforming joint optimization method based on the original decomposition method described in any of the above-described method embodiments of the present invention.

[0095] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0096] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A joint optimization method for zero-forcing transmit / receive beamforming based on the original decomposition method, characterized in that, include: A joint optimization model for zero-forcing transmit and receive beamforming of all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network is constructed. The joint optimization model for zero-forcing transmit and receive beamforming includes a network utility objective function, transmit and receive beamforming power constraints, and zero interference constraints. The zero-forcing transmit / receive beamforming joint optimization model is decomposed based on the original decomposition method. An original decomposition algorithm is designed based on a parallel continuous convex approximation algorithm to perform parallel iterative updates of transmit / receive beamforming until a preset iteration termination condition is met, resulting in optimized feasible transmit / receive beamforming for all links. Specifically, the zero-forcing transmit / receive beamforming joint optimization model is decomposed into a main model for all transmit beamforming and multiple sub-models for receive beamforming. In each iteration, each sub-model is solved in parallel based on the transmit beamforming of the previous iteration to obtain the receive beamforming for the current iteration. Based on the receive beamforming of the current iteration, the main model is approximated into multiple sub-main models for transmit beamforming using a parallel continuous convex approximation algorithm. Each sub-main model is solved in parallel, and then each transmit beamforming is updated in parallel. The transmit / receive beamforming obtained in each iteration of the original decomposition algorithm is feasible.

2. The joint optimization method for zero-forcing transmit / receive beamforming based on the original decomposition method as described in claim 1, characterized in that, The construction of the joint optimization model for zero-forcing transmit / receive beamforming of all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network includes: A joint optimization model for zero-forcing transmit / receive beamforming is constructed for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network to maximize network utility under transmit / receive beamforming power constraints and zero-interference constraints. Specifically, the joint optimization model for zero-forcing transmit / receive beamforming is as follows: ； in, For network utility; These are the optimization variables that imply joint optimization of transmit and receive beamforming; Channel state information; For all transmitter sets; for all , For transmitter The corresponding set of receivers; For the set of all subcarriers; for all , and ,transmitter In subcarrier Upward receiver The maximum number of data streams that can be transmitted is The data stream set is denoted as For all , For transmitter In subcarrier Upward receiver Transmitted data stream The corresponding transmit beamforming vector, For receiver In subcarrier upper receiver transmitter Transmitted data stream The corresponding receiving beamforming vector; for Norm; for all , For transmitter Maximum transmission power.

3. The joint optimization method for zero-forcing transmit / receive beamforming based on the original decomposition method as described in claim 2, characterized in that, The original decomposition method is used to decompose the joint optimization model of zero-forcing transmit / receive beamforming. An original decomposition algorithm is designed based on a parallel continuous convex approximation algorithm, and the transmit / receive beamforming is updated in parallel iteratively until a preset iteration termination condition is met, resulting in the optimized feasible transmit / receive beamforming for all links. This includes: The zero-forcing transmit / receive beamforming joint optimization model is decomposed into a master model for all transmit beamformings and multiple sub-models for receive beamformings using the primal decomposition method. The master model and each sub-model are iteratively solved using the primal decomposition algorithm based on parallel continuous convex approximation until a preset iteration termination condition is met, outputting the receive and transmit beamformings for the current iteration. In each iteration, the sub-models are solved in parallel based on the transmit beamformings from the previous iteration to obtain the receive beamformings for the current iteration. Based on the receive beamformings for the current iteration, the master model is approximated into multiple sub-master models for transmit beamformings using the parallel continuous convex approximation algorithm. Each sub-master model is solved in parallel, and the transmit beamformings are then updated in parallel. The transmit / receive beamformings obtained in each iteration of the primal decomposition algorithm are all feasible.

4. The joint optimization method for zero-forcing transmit / receive beamforming based on the original decomposition method as described in claim 3, characterized in that, The original decomposition method decomposes the zero-forcing transmit / receive beamforming joint optimization model into a master model for all transmit beamformings and multiple sub-models for receive beamformings, including: Regarding any and The specific sub-model for receiving beamforming is as follows: ； The main model for all transmitted beamforming is as follows: ； in, , , For the sub-model in A post in the area.

5. The joint optimization method for zero-forcing transmit / receive beamforming based on the original decomposition method as described in claim 3, characterized in that, The parallel solution of each sub-model based on the previous iteration's transmit beamforming includes: The stationary points of each sub-model are computed in parallel; wherein, the zero-disturbance constraints of each sub-model are eliminated by variable substitution, and the equivalent transformation sub-models are solved in parallel based on the Cauchy-Schwarz inequality; Construct a set of equations for the Lagrangian function of each sub-model with zero gradient with respect to the corresponding receiving beamforming vector, and solve each set of equations in parallel to obtain the Lagrangian multipliers with respect to zero interference constraints corresponding to the stationary points of each sub-model.

6. The joint optimization method for zero-forcing transmit / receive beamforming based on the original decomposition method as described in claim 5, characterized in that, The receiving beamforming based on the current iteration approximates the main model into multiple sub-main models with respect to the transmitting beamforming using a parallel continuous convex approximation algorithm, and solves each sub-main model in parallel, including: Multiple sub-master models are constructed based on the master model; wherein, any transmitter corresponds to one sub-master model; the first... In the next iteration, regarding any transmitter The sub-main model is specifically as follows: ； in, These are algorithm parameters; For Frobenius norm; gradient It is calculated based on the stationary points of all the sub-models and their corresponding Lagrange multipliers; By analyzing the KKT conditions, the optimal solutions of each of the sub-principal models are calculated in parallel.

7. The joint optimization method for zero-forcing transmit / receive beamforming based on the original decomposition method as described in claim 6, characterized in that, After analyzing the KKT conditions and calculating the optimal solutions of each of the sub-principal models in parallel, the method further includes: All transmit beamformings are updated in parallel based on the optimal solutions of each of the sub-master models.

8. A joint optimization device for zero-forcing transmit / receive beamforming based on the original decomposition method, characterized in that, include: Joint optimization model module and original decomposition algorithm module; The joint optimization model module is used to construct a zero-forcing transmit / receive beamforming joint optimization model for all transmitters and receivers in a multi-user multiple-input multiple-output wireless communication network. The zero-forcing transmit / receive beamforming joint optimization model includes a network utility objective function, transmit / receive beamforming power constraints, and zero-interference constraints. The original decomposition algorithm module is used to decompose the zero-forcing transmit / receive beamforming joint optimization model based on the original decomposition method, and to design the original decomposition algorithm based on the parallel continuous convex approximation algorithm to perform parallel iterative updates of transmit / receive beamforming until a preset iteration termination condition is met, thereby obtaining the feasible transmit / receive beamformings for all optimized links. The zero-forcing transmit / receive beamforming joint optimization model is decomposed into a main model concerning all transmit beamformings and multiple sub-models concerning receive beamformings. In each iteration, each sub-model is solved in parallel based on the transmit beamformings of the previous iteration to obtain the receive beamformings of the current iteration. Based on the receive beamformings of the current iteration, the main model is approximated into multiple sub-main models concerning transmit beamformings using the parallel continuous convex approximation algorithm. Each sub-main model is solved in parallel, and then each transmit beamforming is updated in parallel. The transmit / receive beamformings obtained in each iteration of the original decomposition algorithm are all feasible.

9. A terminal device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a joint optimization method for zero-forcing transmit / receive beamforming based on the primal decomposition method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, the device containing the computer-readable storage medium is controlled to perform a joint optimization method for zero-forcing transmit / receive beamforming based on the primal decomposition method as described in any one of claims 1-7.

Images