IRIS deployment position and topology joint optimization method and system for IRIS auxiliary relay communication

Abstract

The invention relates to the technical field of communication system optimization, in particular to an IRIS deployment position and topology joint optimization method and system for IRIS auxiliary relay communication, an IRIS auxiliary relay communication system model is established, and the system comprises a base station BS, an IRIS and a plurality of user equipment UEs. Wherein the reflective elements of the IRIS are irregularly distributed in the grid points of the IRIS expansion surface; constructing a joint optimization problem with the goal of downlink weighting and rate maximization of the system, wherein the optimization problem comprises the joint optimization of BS active beam forming, IRIS deployment position, IRIS passive beam forming and IRIS topology; decomposing the joint optimization problem into two sub-problem blocks by adopting a block alternating optimization algorithm framework; and alternately optimizing the two sub-problem blocks until convergence to obtain an optimal solution. According to the method, an optimization algorithm combining active and passive beam forming, IRIS deployment position and IRIS topology is provided, so that the multiplicative fading effect in the system is weakened, the channel estimation overhead and the system energy consumption are reduced, and the algorithm complexity is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of communication system optimization, and particularly relates to a method and system for jointly optimizing the deployment location and topology of IRIS-assisted relay communication for IRIS. Background Art

[0002] In a relay communication system assisted by a Reconfigurable Intelligent Surface (RIS), there are two key challenges: one is the impact of the "multiplicative fading effect" introduced by the RIS on the system performance. The introduction of the RIS is expected to bring significant capacity gains to users. However, in practice, significant capacity gains can usually only be achieved in communication scenarios where the direct link between the base station and the user is completely blocked or very weak. In contrast, when considering that the base station-to-user channel link is not weak or unobstructed, due to the "multiplicative fading effect" introduced by the RIS, users can only obtain limited capacity gains. Therefore, the "multiplicative fading effect" makes it almost impossible for the RIS to achieve significant capacity gains in many wireless environments. How to weaken the "multiplicative fading effect" in the system is a challenge. The other is how to effectively improve the system capacity under the condition of a limited number of RIS elements. In the RIS-assisted wireless communication system, the received signal power is quadratically proportional to the total number of RIS reflection elements. To improve the system capacity, it is usually necessary to increase the number of RIS reflection elements. However, as the number of reflection elements increases, the overhead of channel estimation and the complexity of beamforming design will become too high. Therefore, how to improve the system capacity with limited RIS elements is an important challenge.

[0003] There are two common types of methods in the deployment location and topology optimization algorithms: one is the traditional optimization algorithm, and the other is deep learning. The optimization algorithms of deep learning are not very applicable due to limitations such as high training complexity, poor adaptability, and the need for large memory and high power consumption. Although the traditional optimization algorithms also have the problem of high algorithm complexity, they still have certain advantages in the actual flexible scenario applications. Therefore, in order to improve the performance of the system, it is necessary to further solve the complexity problems of these algorithms.

[0004] For traditional optimization algorithms such as the Exhaustive Search (ES) algorithm and the Tabu Search (TS) algorithm, etc., only these two methods will be briefly introduced below, and their disadvantages will be pointed out:

[0005] 1. ES algorithm:

[0006] Algorithm principle: For a problem, assuming that its solution space is finite, the ES algorithm will systematically check all possible solutions until the optimal solution or a solution that meets a certain condition is found.

[0007] Disadvantage: Although the ES algorithm can effectively obtain the global optimal solution, when the solution space is very large, its computational cost is very high.

[0008] 2. TS algorithm:

[0009] Algorithm principle: The TS algorithm is a heuristic optimization algorithm that uses local search and memory mechanisms to avoid falling into local optimal solutions, expand the search space, and thus explore better solutions.

[0010] Disadvantage: The TS algorithm depends on appropriate parameter settings and may exhibit low efficiency or fail to find the global optimal solution in large-scale or dynamic problems. In practical applications, it needs to be combined with other optimization methods (such as Genetic Algorithm (GA), Simulated Annealing (SA), etc.) to improve the search effect and computational efficiency.

[0011] In terms of the algorithm framework, the Alternating Optimization (AO) algorithm is a widely used algorithm framework. According to the variable types of multi-variable optimization problems, the AO algorithm framework usually decomposes the optimization problem into multiple sub-problems for alternating optimization. This algorithm framework alternates the optimization of sub-problems during one iteration. When dealing with multi-variable optimization problems with strong coupling, there may be a situation of more iteration times and slower convergence speed. Summary of the Invention

[0012] The present invention aims to solve the multiplicative fading effect in the RIS-assisted relay communication system and the problem of improving system capacity under a limited number of RIS elements, and proposes a method and system for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication. By proposing an optimization algorithm for joint active and passive beamforming, IRIS deployment location, and IRIS topology, it is possible to weaken the multiplicative fading effect in the system, reduce the channel estimation overhead and system energy consumption, and reduce the algorithm complexity.

[0013] To achieve the above object, the technical solution adopted is:

[0014] The present invention provides a method for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication, including the following steps:

[0015] Establish an IRIS-assisted relay communication system model, which includes a base station BS, an irregular reconfigurable intelligent surface IRIS, and multiple user equipment UEs, where the reflection elements of the IRIS are irregularly distributed in the grid points of the IRIS extended surface;

[0016] Construct a joint optimization problem aiming at maximizing the system downlink weighted sum rate. The optimization problem includes the joint optimization of BS active beamforming, IRIS deployment location, IRIS passive beamforming, and IRIS topology;

[0017] Adopt the block alternating optimization algorithm framework to decompose the joint optimization problem into two sub-problem blocks;

[0018] Alternately optimize the two sub-problem blocks until convergence to obtain the optimal solution.

[0019] According to the method for joint optimization of IRIS deployment location and topology in IRIS-assisted relay communication of the present invention, further, adopting the block alternating optimization algorithm framework to decompose the joint optimization problem into two sub-problem blocks includes: the optimization problem of the first sub-problem block is to fix IRIS passive beamforming and IRIS topology, and jointly optimize BS active beamforming and IRIS deployment location; the optimization problem of the second sub-problem block is to fix BS active beamforming and IRIS deployment location, and jointly optimize IRIS passive beamforming and IRIS topology.

[0020] According to the method for joint optimization of IRIS deployment location and topology in IRIS-assisted relay communication of the present invention, further, adopt fractional programming FP combined with convex optimization algorithm to optimize BS active beamforming; adopt fractional programming FP combined with successive convex approximation algorithm SCA to optimize IRIS deployment location.

[0021] According to the method for joint optimization of IRIS deployment location and topology in IRIS-assisted relay communication of the present invention, further, the specific steps of optimizing BS active beamforming by combining fractional programming FP with convex optimization algorithm are as follows:

[0022] Given the IRIS deployment location q, introduce auxiliary variables through fractional programming FP τ ;

[0023] Using the quadratic transformation algorithm, reconstruct the objective function into a separated form with respect to BS active beamforming W and auxiliary variable τ ;

[0024] Fix BS active beamforming W and solve the optimal τ value in a closed form;

[0025] Fix the auxiliary variable τ , transform the non-convex sum rate maximization problem into a convex optimization problem with respect to BS active beamforming W; use the Lagrange multiplier method to solve the optimal W;

[0026] Alternately update the auxiliary variable τ and BS active beamforming W, and gradually approximate the optimal solution through multiple iterations until the objective function converges or reaches the preset number of iterations, and finally output the optimal W.

[0027] According to the method for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication of the present invention, further, the fractional programming FP combined with the successive convex approximation algorithm SCA to optimize the IRIS deployment location specifically includes the following steps:

[0028] Step a: Given the BS active beamforming W, transform the non-convex IRIS deployment location problem into a mathematical form that can be processed by SCA;

[0029] Step b: Perform a first-order Taylor expansion on the non-convex objective function at the current iteration point to construct a locally convex surrogate function;

[0030] Step c: Add slack variables to transform the constraints into convex forms;

[0031] Step d: Use CVX to solve the convex optimization problem after Taylor approximation to obtain the current optimal IRIS deployment location q;

[0032] Step e: Update the Taylor expansion point with the new solution q, and repeat steps b - d until the objective function converges, and return the finally optimized IRIS deployment location.

[0033] According to the method for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication of the present invention, further, the neighborhood extraction cross-entropy algorithm NECE is used to optimize the IRIS passive beamforming; the GA-TS algorithm is used to optimize the IRIS topology.

[0034] According to the method for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication of the present invention, further, the neighborhood extraction cross-entropy algorithm NECE to optimize the IRIS passive beamforming specifically includes the following steps:

[0035] Given the IRIS topology Z, initialize the probability matrix P, which represents the probability distribution of the phase offsets of the IRIS reflection elements;

[0036] Randomly generate multiple groups of phase configurations {Θ 1 , Θ 2 ,... Θ c} according to the current probability matrix P;

[0037] Calculate the weighted sum rate WSR of each group of configurations;

[0038] Sort in descending order of WSR, and select the first C1 phase configurations as the elite group;

[0039] Generate neighborhood solutions for the top elite Θ 1 , and select C2 solutions that are better than Θ 1 from the neighborhood solutions as supplementary elites, and merge the elite groups;

[0040] Calculate the weight η of each elite, and update the probability matrix P based on the weight and the phase configuration of the elite group;

[0041] When the maximum number of iterations is reached or the probability matrix converges, select the historical optimal phase configuration as the final solution for output.

[0042] According to the method for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication of the present invention, further, the GA-TS algorithm is used to optimize the IRIS topology, which specifically includes the following steps:

[0043] Given the IRIS passive beamforming Θ; generate an initial population, where each individual represents an IRIS topology structure;

[0044] Select high-quality individuals according to fitness, and perform crossover operation and mutation operation on the selected individuals to generate new offspring individuals;

[0045] For the high-quality individuals output by the above genetic algorithm, generate new solutions through tabu search and update the tabu list;

[0046] Repeat the genetic operation and tabu search until the termination condition is met.

[0047] According to the method for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication of the present invention, further, the base station is used to perform active beamforming; the irregular reconfigurable intelligent surface is used to reflect signals and optimize passive beamforming and topology; the multiple user equipments are used to receive signals.

[0048] Further, the present invention also provides a system for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication, which is used to implement the method for jointly optimizing the IRIS deployment location and topology in the above IRIS-assisted relay communication, and includes:

[0049] A system model construction module, which is used to establish an IRIS-assisted relay communication system model. The system includes a base station BS, an irregular reconfigurable intelligent surface IRIS, and multiple user equipments UE, where the reflection elements of the IRIS are irregularly distributed at the grid points of the IRIS extended surface;

[0050] A joint optimization module, which is used to construct a joint optimization problem with the goal of maximizing the system downlink weighted sum rate. The optimization problem includes the joint optimization of BS active beamforming, IRIS deployment location, IRIS passive beamforming, and IRIS topology; use the block alternating optimization algorithm framework to decompose the joint optimization problem into two sub-problem blocks; alternately optimize the two sub-problem blocks until convergence to obtain the optimal solution.

[0051] Adopting the above technical solutions, the beneficial effects obtained are:

[0052] 1. Significantly weaken the "multiplicative fading effect" in the system

[0053] When the direct link between the base station and the user in the traditional RIS is not completely blocked, due to the "multiplicative fading effect", the capacity gain is limited. The present invention optimizes the deployment position of the IRIS through the Fractional Programming (FP) combined with the Successive Convex Approximation (SCA) algorithm, effectively suppressing the attenuation of the signal in the reflection path of the RIS; in the non-fully blocked scenario, the system capacity (WSR, weighted sum rate) is significantly improved.

[0054] 2. Reduce the channel estimation overhead and system energy consumption

[0055] Traditional RIS requires a large number of reflection elements to achieve high capacity, but the increase in the number of elements will bring high channel estimation overhead and hardware costs. The present invention adopts an irregular IRIS topology (i.e., a given number of RIS reflection elements are sparsely arranged on the extended surface). By optimizing the selection of the feasible positions of the RIS reflection elements, the spatial degrees of freedom and diversity of the IRIS are increased, thereby achieving a significant system capacity gain while reducing the number of RIS reflection elements. This design of the irregular IRIS topology not only effectively reduces the channel estimation overhead but also reduces the system energy consumption.

[0056] 3. The optimization algorithm converges efficiently and reduces the computational complexity

[0057] Traditional optimization methods (such as exhaustive search ES, alternating optimization AO) have high computational complexity and slow convergence when optimizing strongly coupled variables. The present invention adopts the BAO (block alternating optimization) framework to decompose the problem into two sub-problem blocks, reducing variable coupling. The FP (fractional programming) + SCA (successive convex approximation) is used to optimize the IRIS deployment position to ensure convergence. The GA-TS (genetic algorithm - tabu search) is used to optimize the IRIS topology, combining global search (GA) and local optimization (TS). The GA-TS algorithm is superior to the traditional TS algorithm and approximates the global optimal solution at finite grid points. The joint optimization algorithm proposed by the present invention has a lower computational complexity, and the computational complexity is significantly lower than that of the traditional ES algorithm (reduced from exponential level to polynomial level). The simulation shows that the BAO framework converges more than 30% faster than the traditional AO algorithm.

[0058] 4. Verification of performance advantages

[0059] The simulation shows that under the same number of reflection elements, the WSR of the IRIS scheme of the present invention is improved by 20% - 40% compared with the traditional RIS scheme. Description of the drawings

[0060] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly introduced below. Among them, the accompanying drawings are only used to show some embodiments of the present invention, rather than limiting all embodiments of the present invention thereto.

[0061] Figure 1 It is a schematic flow chart of the IRIS deployment location and topology joint optimization method for IRIS-assisted relay communication in the embodiments of the present invention;

[0062] Figure 2 It is a schematic structural diagram of the IRIS-assisted relay communication system in the embodiments of the present invention;

[0063] Figure 3 It is the relationship between the weighted sum rate and the transmit power in the embodiments of the present invention;

[0064] Figure 4 It is the convergence behavior of the algorithm in the embodiments of the present invention;

[0065] Figure 5 It is the relationship between the weighted sum rate and the transmit power under different IRIS deployment locations in the embodiments of the present invention (IRIS height is 20m);

[0066] Figure 6 It is the relationship between the weighted sum rate and the transmit power under different IRIS deployment locations in the embodiments of the present invention (IRIS height is 60m);

[0067] Figure 7 It is the relationship between the weighted sum rate and the number of RIS elements in the embodiments of the present invention;

[0068] Figure 8 It is the relationship between the weighted sum rate and the number of IRIS grid points in the embodiments of the present invention. Detailed implementation manners

[0069] In the following, the exemplary solutions of the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings of the specific embodiments of the present invention. Unless otherwise defined, the technical terms or scientific terms used in the present invention should have the ordinary meanings understood by those of ordinary skill in the art.

[0070] To solve the problems of the "multiplicative fading effect" in the RIS-assisted relay communication system and improving the system capacity under a limited number of RIS elements, this embodiment discloses an IRIS deployment location and topology joint optimization method for IRIS-assisted relay communication, as Figure 1 shown, including the following steps:

[0071] Step S1: Establish an IRIS-assisted relay communication system model. The system includes a base station (BS), an irregular reconfigurable intelligent surface (IRIS), and multiple user equipments (UEs). The reflecting elements of the IRIS are irregularly distributed at the grid points of the IRIS extended surface.

[0072] Step S2: Construct a joint optimization problem aiming at maximizing the system downlink weighted sum rate (WSR). The optimization problem includes the joint optimization of BS active beamforming, IRIS deployment location, IRIS passive beamforming, and IRIS topology.

[0073] Step S3: Use the Block Alternating Optimization (BAO) algorithm framework to decompose the joint optimization problem into two sub-problem blocks, aiming to effectively solve the problems of BS active beamforming, IRIS deployment location, IRIS passive beamforming, and IRIS topology.

[0074] The BAO algorithm framework first decomposes the optimization problem into two blocks: BS active beamforming and IRIS deployment location; IRIS passive beamforming and IRIS topology. The optimization problem of the first sub-problem block is: fix IRIS passive beamforming and IRIS topology, and jointly optimize BS active beamforming and IRIS deployment location. This solution proposes an FP-convex optimization algorithm to optimize BS active beamforming and uses the FP-SCA algorithm to optimize the IRIS deployment location. The optimization problem of the second sub-problem block is: fix BS active beamforming and IRIS deployment location, and jointly optimize IRIS passive beamforming and IRIS topology. This solution proposes a Neighborhood Extraction Cross-Entropy (NECE) algorithm to optimize IRIS passive beamforming and a GA-TS optimization algorithm to optimize IRIS topology.

[0075] Step S4: Alternately optimize the two sub-problem blocks until convergence to obtain the optimal solution.

[0076] Inside each block, the AO algorithm framework is used for optimization to obtain the optimal solution of the block problem. Subsequently, the algorithm transfers to the next block for optimization, and iterates until convergence, thereby obtaining the optimal solution of the original problem.

[0077] The system model and the optimization problem are described in detail below.

[0078] First, construct an IRIS-assisted multi-user downlink wireless communication system, as Figure 2As shown. A base station (BS) equipped with M antennas and an irregular intelligent reflecting surface (IRS) equipped with N reflecting elements simultaneously serve K single-antenna user equipments (UEs). Among them, the N reflecting elements are irregularly distributed in N s grid points (N s > N). The coordinates of the BS are I BS =(0, 0, 0). The UEs are uniformly distributed within a circle with a diameter of c0, and the center coordinates of this circle are (0, y0, 0). The coordinates of the k-th UE are I UE,k =(x UE,k , y UE,k , 0), and the position coordinates of the IRS are q = (l, 0, h). It is assumed that all channels experience quasi-static flat fading and the direct links from the BS to the UEs are completely blocked by obstacles. The baseband equivalent channels from the BS to the IRS and from the IRS to the k-th UE are respectively expressed as:

[0079]

[0080] where ρ represents the path loss coefficient per unit distance, and α BR and α RU are the corresponding path loss exponents. In addition, d BR = ||q - I BS || represents the distance between the BS and the IRS, and d RU,k = ||q - I RU,k ‖ represents the distance between the IRS and the k-th UE. Meanwhile, the small-scale fading and are respectively expressed as:

[0081]

[0082] where β BR and β RU are the Rice fading factors, and the non-line-of-sight (NLoS) components and follow the Rayleigh distribution. In addition, the line-of-sight (LoS) components are expressed as and where the transmit array response vector of the BS, the receive array response vector of the IRS, and the transmit array response vector of the IRS are respectively expressed as:

[0083]

[0084] where d represents the distance between antennas, λ0 is the carrier wavelength, and and φk They respectively represent the departure angle, arrival angle of the signal on the BS-to-IRIS link, and the departure angle of the signal on the IRIS-to-k-th UE link. The phase shift matrix is usually used to represent the ability of each element on the IRIS to adjust the phase of the electromagnetic wave, and it can be expressed as:

[0085]

[0086] where β n ∈[0,1] represents the reflection amplitude of the n-th reflecting element of the IRIS. For θ n ∈[0,2π] represents its phase. In practical applications, the cost of simultaneously and independently controlling the reflection amplitude and phase shift is relatively high. Therefore, it is assumed that β n =1. In addition, the phase shift of any reflecting element can only take a finite number of discrete values. Let b represent the quantization bits of the finite discrete phase shift, where L = 2 b . Therefore, the set of discrete phase shift values of each reflecting element can be expressed as:

[0087] (9) where Δθ = 2π / L. Assume that the topology matrix of the IRIS is expressed as:

[0088] Z = diag(z), (10)

[0089] where, z n ∈{1,0} is used to represent whether a reflecting element is deployed at the n-th grid point. Set z n =1 to indicate that a reflecting element is deployed at the n-th grid point, while z n =0 indicates that no reflecting element is deployed at the n-th grid point.

[0090] The complex baseband transmitted signal at the base station is expressed as:

[0091]

[0092] where, is the transmission precoding vector for user k, s k is the transmission symbol for user k, which follows independent and identically distributed with mean 0 and variance 0. The received signal y k for the k-th user is expressed as:

[0093]

[0094] where, n k is the additive Gaussian white noise for the k-th user, which follows a complex Gaussian distribution with mean 0 and variance σ 2 .

[0095] Combined with formula (12), the signal-to-interference plus noise ratio (SINR) of user k in the IRIS-assisted wireless communication system is expressed as:

[0096]

[0097] Aiming at maximizing the weighted sum rate (WSR), its maximization optimization problem is constructed as follows:

[0098]

[0099] s.t.C1: P T ≤ P(14b)

[0100]

[0101] C4: 1 T z = N, (14e)

[0102] where ω k is the weight of the weighted sum rate, C1 is the transmit power constraint, C2 is the discrete phase shift constraint, and C3 and C4 are the sparse constraints. Due to the high coupling of W, q, Θ, Z and the non-convexity of the constraint conditions, it is difficult to directly solve the complex non-convex optimization problem. To overcome this difficulty, the logarithmic sum problem is transformed into a multi-ratio problem. Specifically, based on the quadratic transformation and the Lagrangian dual transformation, and by introducing the auxiliary variable α k , the objective function formula (14a) can be equivalently transformed into:

[0103]

[0104] where α = [α1, α2, …, α K T , and the optimal solution can be obtained by setting Given α, the problem can be re-expressed as

[0105]

[0106] s.t C1, C2, C3, C4, (16b)

[0107] where At this time, formula (16a) is a multi-ratio FP problem.

[0108] The joint optimization algorithm will be explained in detail below.

[0109] ​​This solution proposes a BAO algorithm framework that decomposes the optimization problem into two blocks, aiming to effectively solve the joint optimization problem of BS active beamforming W, IRIS deployment location q, IRIS passive beamforming Θ, and IRIS topology Z. The first block of the optimization problem is to jointly optimize the BS active beamforming and IRIS deployment location while fixing the IRIS passive beamforming and IRIS topology. This problem is a non-convex optimization problem. In this case, a convex optimization algorithm for BS active beamforming based on FP and an SCA algorithm for IRIS deployment location are designed. The second block of the optimization problem is to jointly optimize the IRIS passive beamforming and IRIS topology while fixing the BS active beamforming and IRIS deployment location. In this case, the traditional NECE is used to solve the IRIS passive beamforming and a GA-TS algorithm is designed to solve the IRIS topology.

[0110] (1) Jointly optimize the BS active beamforming and IRIS deployment location

[0111] The first block of the optimization problem is to jointly optimize the BS active beamforming W and IRIS deployment location q given the IRIS passive beamforming Θ and IRIS topology Z. The problem can be transformed into:

[0112]

[0113] s.t.C1: P T ≤ P. (17b)

[0114] The AO algorithm framework is used to alternately optimize the variables W and q until the problem converges.

[0115] (1.1) Fractional programming FP combined with a convex optimization algorithm to optimize the BS active beamforming

[0116] Given the IRIS deployment location q, the quadratic transformation algorithm is used to introduce the auxiliary variable τ = [τ1, τ2,..., τ K T , and this optimization problem is:

[0117]

[0118] s.t.C1: P T ≤ P, (18b)

[0119] where, Given W, let the optimal can be obtained, and its expression is:

[0120]

[0121] Given​ The objective function formula (18a) is a concave function, and the feasible region formed by the constraint condition formula (18b) is a convex set. The problem is a convex optimization problem, and the Lagrange multiplier method is used to solve it. The corresponding Lagrangian function is expressed as:

[0122] F(W, λ) = -f4(W) + λ(P T - P), (20)

[0123] where λ is the Lagrange multiplier, which can be obtained by the bisection search method. Given λ, and letting the optimal can be obtained as:

[0124]

[0125] (1.2) Fractional programming FP combined with the successive convex approximation SCA algorithm to optimize the IRIS deployment location

[0126] Given the BS active beamforming W, the optimization variable q of the IRIS deployment location is coupled in formula (17a). Since the optimization variable q is coupled in both d BR and d RU,k the problem is non-convex. To solve this problem, the first-order Taylor expansion method is adopted, and the optimal IRIS deployment location is solved by SCA. The specific details are as follows:

[0127]

[0128] where the problem P 3-1 can be expressed as:

[0129]

[0130] By introducing the slack variables u k and μ, this optimization problem can be expressed as:

[0131]

[0132] Using the SCA algorithm, the first-order Taylor expansion of f6 at the fixed points u k,0 and μ0 is:

[0133]

[0134] where According to formulas (24) and (25), the problem can be transformed into:

[0135]

[0136] Among them, the optimization problem is convex and can be solved by CVX.

[0137] (2) Jointly optimize the IRIS passive beamforming and IRIS topology

[0138] The second optimization problem is to jointly optimize the IRIS passive beamforming Θ and the IRIS topology Z given the BS active beamforming W and the deployment location q of the IRIS. This optimization problem is:

[0139]

[0140] s.t. C2, C3, C4. (27b)

[0141] Adopt the AO algorithm framework to alternately optimize the variables Θ and Z until the problem converges.

[0142] (2.1) Optimize the IRIS passive beamforming using the Neighborhood Extraction Cross-Entropy (NECE) algorithm

[0143] Given the IRIS topology Z, the IRIS passive beamforming Θ is solved by the NECE algorithm. Let represent the probability matrix of Θ, where is the probability vector of θ n satisfying ||p n ||1 = 1, and each component of p n represents the probability that θ n takes different values in . Given the probability matrix P, the probability distribution function of the IRIS passive beamforming Θ is:

[0144]

[0145] Among them, represents the k-th element in

[0146]

[0147] In order to search for a better Θ, the probability matrix needs to be updated, and the probability transition criterion is updated as:

[0148]

[0149] where η c is the weight of the c-th elite, and its calculation is as follows:

[0150]

[0151] The specific algorithm flow is shown in Algorithm 1

[0152]

[0153] (2.2) Optimize the IRIS topology using the GA-TS algorithm

[0154] Given the IRIS passive beamforming Θ, the topology Z is solved by the GA-TS algorithm. GA has excellent global search ability, but its local search efficiency is insufficient. On the contrary, the TS algorithm shows strong local search ability, but its performance partly depends on the choice of the initial solution. Therefore, the GA-TS algorithm is adopted to improve the overall performance. The GA-TS algorithm proposed in this scheme is summarized in Algorithm 2

[0155]

[0156] The overall framework of the optimization algorithm is shown in Algorithm 3

[0157]

[0158] AO is a widely used algorithm framework. According to the variable types of multi-variable optimization problems, the AO algorithm framework usually decomposes the optimization problem into multiple sub-problems for alternating optimization. When using the AO algorithm framework to alternately iterate and optimize multi-variable optimization problems with strong coupling, there may be situations where the number of iterations is large and the convergence speed is slow. The BAO algorithm framework proposed in this scheme decomposes the optimization problem into two blocks: BS active beamforming, the deployment location of IRIS, and IRIS passive beamforming, IRIS topology. This algorithm framework has the following characteristics: (1) realizing multi-variable decoupling; (2) by concentrating the variables with high connection degree in the optimization problem in the same stage block, realizing the centralized and coordinated processing of these variables, and enhancing the effectiveness and reliability of the optimization results; (3) an efficient iterative optimization algorithm can be used within each stage block, thereby improving the convergence speed of the algorithm

[0159] Complexity analysis: In the first block: The complexity of the BS beamforming algorithm is The complexity of the IRIS deployment location algorithm is I1 and I2 are the convergence iteration times of the FP-convex optimization algorithm and the FP-SCA algorithm respectively, and the complexity of this block is I0 is the number of alternating iterations between blocks. In the second block: The complexity of the IRIS passive beamforming algorithm is The complexity of the IRIS topology algorithm is I N 、I G and I TThey are the convergence iteration times of the NECE algorithm, GA algorithm, and TS algorithm respectively. The complexity of this block is I 00 is the number of inter-block alternating iterations, which is lower than the complexity of the traditional ES algorithm

[0160] Correspondingly, this embodiment also discloses a system for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication, including:

[0161] A system model construction module for establishing an IRIS-assisted relay communication system model, which includes a base station BS, an irregular reconfigurable intelligent surface IRIS, and multiple user equipments UE, where the reflecting elements of the IRIS are irregularly distributed at the grid points of the IRIS extended surface.

[0162] A joint optimization module for constructing a joint optimization problem aiming at maximizing the system downlink weighted sum rate, where the optimization problem includes the joint optimization of BS active beamforming, IRIS deployment location, IRIS passive beamforming, and IRIS topology; using the block alternating optimization algorithm framework to decompose the joint optimization problem into two sub-problem blocks; alternately optimizing the two sub-problem blocks until convergence to obtain the optimal solution.

[0163] The following uses experimental data to better verify the effectiveness and advancement of the present invention.

[0164] The specific implementation process of the method for jointly optimizing the IRIS deployment location and topology in IRIS-assisted relay communication is as follows:

[0165] 1. Establish an IRIS-assisted relay communication system model.

[0166] 2. Use the BAO algorithm framework to alternately optimize two blocks: BS active beamforming, IRIS deployment location, and IRIS passive beamforming, topology.

[0167] 3. Use the FP-based BS active beamforming convex optimization algorithm to solve the BS active beamforming problem.

[0168] 3. Use the FP-based SCA algorithm for the IRIS deployment location to solve the IRIS deployment location problem.

[0169] 3. Use the FP-based IRIS passive beamforming NECE algorithm to solve the IRIS passive beamforming problem.

[0170] 4. Use the FP-based IRIS topology GA-TS algorithm to solve the IRIS topology problem.

[0171] 5. Verify the performance of the algorithm through simulation, including convergence speed and sum rate performance.

[0172] Unless otherwise specified, the simulation parameters are shown in Table 1. In addition, and φ k are randomly generated in the range of [0, 2π], and ω k = 1, The iteration parameters of the GA-TS algorithm and the NECE algorithm are set as O = 50, p c1 = 0.8, p c2 = 0.3, p c3 = 0.01, p c4 = 0.1, I G = 15, Q = 15, I T = 20, C = 100, C1 = 30, I N = 15. The neighborhood distance in the GA-TS algorithm changes dynamically during the iteration, and the initial value is When the iteration reaches half of the maximum iteration number, it decreases to 2.

[0173] Table 1 Simulation parameters

[0174]

[0175] Experimental results:

[0176] Figure 3 The relationship between the WSR and the reflection power is given. The optimal solution of P1 obtained by the ES method is used as the upper bound. The comparison results show that the IRIS-assisted wireless communication system significantly improves the system performance, and the IRIS-assisted wireless communication system achieves the maximum performance gain. Under the same conditions, the proposed GA-TS algorithm is superior to the traditional TS algorithm, proving its superiority. In addition, as Figure 4 shown, compared with the traditional AO algorithm framework, the proposed BAO algorithm framework shows a faster convergence speed.

[0177] Figure 5 and Figure 6 give the relationship between the WSR and the transmit power under different IRIS deployment positions. From the comparison results in the figure, it can be seen that using the SCA algorithm for optimizing the IRIS deployment position at the deployment position of the FP-based IRIS significantly improves the WSR of the system.

[0178] Figure 7 gives the relationship between the WSR and the number of RIS elements, where P T = 20 dBm, N s = 100. Compared with the conventional RIS scheme, the IRIS scheme significantly achieves a higher WSR. Figure 8The relationship between the WSR and the number of IRIS grid points is given, where P T = 20 dBm and N = 20. Based on the conventional RIS scheme with a fixed surface size (N = 20, M = 4, K = 4), by increasing the surface size to change the irregular ratio of RIS elements to grid points, the WSR performance can be effectively improved. In addition, when the number of grid points is N s = 40, N s = 60, N s = 80, the IRIS scheme with M = 4 is superior to the conventional RIS with M = 7, M = 8, and M = 9 respectively. Therefore, the IRIS scheme can improve the system performance without increasing the base station antennas and radio frequency.

[0179] Finally, it should be noted that the above-described embodiments are only specific implementation manners of the present invention, used to illustrate the technical solutions of the present invention, rather than limiting it. The protection scope of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: Any person skilled in the art within the technical scope disclosed by the present invention can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent substitution on some of the technical features; and these modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. An IRIS deployment location and topology joint optimization method for IRIS-assisted relay communication, characterized in that It includes the following steps: Establish an IRIS-aided relay communication system model, which includes a base station BS, an irregular reconfigurable intelligent surface IRIS, and multiple user equipments UE, where the reflecting elements of the IRIS are irregularly distributed at the grid points of the IRIS extended surface; Construct a joint optimization problem aiming at maximizing the system downlink weighted sum rate, and the optimization problem includes the joint optimization of BS active beamforming, IRIS deployment location, IRIS passive beamforming, and IRIS topology; Adopt a block alternating optimization algorithm framework to decompose the joint optimization problem into two sub-problem blocks; Alternately optimize the two sub-problem blocks until convergence to obtain the optimal solution.

2. The method for jointly optimizing the IRIS deployment location and topology of IRIS-assisted relay communication according to claim 1, wherein Adopting a block alternating optimization algorithm framework to decompose the joint optimization problem into two sub-problem blocks includes: the optimization problem of the first sub-problem block is to fix the IRIS passive beamforming and IRIS topology, and jointly optimize the BS active beamforming and IRIS deployment location; the optimization problem of the second sub-problem block is to fix the BS active beamforming and IRIS deployment location, and jointly optimize the IRIS passive beamforming and IRIS topology.

3. The IRIS deployment location and topology joint optimization method for IRIS-assisted relay communication according to claim 2, characterized in that Adopt fractional programming FP combined with convex optimization algorithm to optimize the BS active beamforming; adopt fractional programming FP combined with successive convex approximation algorithm SCA to optimize the IRIS deployment location.

4. The IRIS deployment location and topology joint optimization method for IRIS-assisted relay communication according to claim 3, characterized in that, The fractional programming FP combined with convex optimization algorithm to optimize the BS active beamforming specifically includes the following steps: Given the IRIS deployment location q, introduce auxiliary variables through fractional programming FP τ ; Using the quadratic transformation algorithm, the objective function is reconstructed into a separated form with respect to the BS active beamforming W and the auxiliary variable τ ; Fixed BS active beamforming W, closed-form solution for the optimal τ value; Fixed auxiliary variable τ , transform the non-convex sum rate maximization problem into a convex optimization problem regarding the BS active beamforming matrix W; use the Lagrange multiplier method to solve for the optimal W; Alternately update the auxiliary variable τ and the BS active beamforming W, and gradually approximate the optimal solution through multiple iterations until the objective function converges or reaches the preset number of iterations, and finally output the optimal W.

5. The method for jointly optimizing the IRIS deployment location and topology of IRIS-assisted relay communication according to claim 3, characterized in that The fractional programming FP combined with successive convex approximation algorithm SCA to optimize the IRIS deployment location specifically includes the following steps: Step a: Given the BS active beamforming W, transform the non-convex IRIS deployment location problem into a mathematical form that can be processed by SCA; Step b: Perform a first-order Taylor expansion on the non-convex objective function at the current iteration point to construct a local convex surrogate function; Step c: Add slack variables to transform the constraints into convex forms; Step d: Use CVX to solve the convex optimization problem after Taylor approximation to obtain the current optimal IRIS deployment location q; Step e: Update the Taylor expansion point with the new solution q, and repeat steps b-d until the objective function converges, and return the finally optimized IRIS deployment location.

6. The method for jointly optimizing the IRIS deployment location and topology of IRIS-assisted relay communication according to claim 2, characterized in that Adopt the neighborhood extraction cross-entropy algorithm NECE to optimize the IRIS passive beamforming; adopt the GA-TS algorithm to optimize the IRIS topology.

7. The method for jointly optimizing the IRIS deployment location and topology of IRIS-assisted relay communication according to claim 6, characterized in that The neighborhood extraction cross-entropy algorithm NECE to optimize the IRIS passive beamforming specifically includes the following steps: Given the IRIS topology Z, initialize the probability matrix P, which represents the probability distribution of the phase offsets of the IRIS reflecting elements; Randomly generate multiple groups of phase configurations {Θ 1 , Θ 2 ,... Θ c} according to the current probability matrix P; Calculate the weighted sum rate WSR of each group of configurations; Sort in descending order of WSR, and select the first C1 phase configurations as the elite group; For the top-level elite Θ 1 Generate neighborhood solutions, and select C2 solutions that are better than Θ 1 from the neighborhood solutions as supplementary elites, and merge the elite group; Calculate the weight η of each elite, and update the probability matrix P based on the weights and the phase configurations of the elite group; When the maximum number of iterations is reached or the probability matrix converges, select the historical optimal phase configuration as the final solution for output.

8. The method for jointly optimizing the IRIS deployment location and topology of IRIS-assisted relay communication according to claim 6, characterized in that The GA-TS algorithm to optimize the IRIS topology specifically includes the following steps: Given the IRIS passive beamforming Θ; generate an initial population, and each individual represents an IRIS topology structure; Select high-quality individuals according to the fitness, and perform crossover operations and mutation operations on the selected individuals to generate new offspring individuals; For the high-quality individuals output by the above genetic algorithm, new solutions are generated through tabu search, and the tabu list is updated; Repeat the genetic operation and tabu search until the termination condition is met.

9. The method for jointly optimizing the IRIS deployment location and topology of IRIS-assisted relay communication according to claim 1, characterized in that, The base station is used to perform active beamforming; the irregular reconfigurable intelligent surface is used to reflect signals and optimize passive beamforming and topology; the multiple user equipments are used to receive signals.

10. An IRIS deployment location and topology joint optimization system for IRIS-assisted relay communication, characterized in that, An IRIS deployment location and topology joint optimization method for implementing the IRIS-assisted relay communication according to any one of claims 1-9, comprising: A system model construction module, configured to establish an IRIS-assisted relay communication system model, the system includes a base station BS, an irregular reconfigurable intelligent surface IRIS, and multiple user equipments UE, wherein the reflecting elements of the IRIS are irregularly distributed at the grid points of the IRIS extended surface; A joint optimization module, configured to construct a joint optimization problem with the goal of maximizing the system downlink weighted sum rate, and the optimization problem includes the joint optimization of BS active beamforming, IRIS deployment location, IRIS passive beamforming, and IRIS topology; the block alternating optimization algorithm framework is used to decompose the joint optimization problem into two sub-problem blocks; the two sub-problem blocks are alternately optimized until convergence to obtain the optimal solution.

Images