Uplink energy efficiency optimization method for RIS-assisted cellular-removal large-scale MIMO system

Abstract

The invention provides a method for optimizing the energy efficiency of an uplink of an RIS-assisted cellular-removed large-scale MIMO (Multiple-Input Multiple-Output) system. Based on a Rayleigh channel fading model, the method firstly derives an expression of the energy efficiency of an uplink under perfect channel state information, and provides a method for jointly optimizing the user transmitting power and the RIS phase shift matrix with the purpose of maximizing the energy efficiency. In view of the non-convexity of the optimization problem, the user transmitting power is optimized by utilizing a water injection method, and the RIS phase shift matrix is optimized by adopting a genetic algorithm, so that the energy efficiency is quickly improved. According to the invention, the total power consumption of the system is effectively reduced, the wireless resource utilization rate of the system is improved, and green communication requirements are met.

Description

Technical Field

[0001] The present invention belongs to the field of wireless communication technologies, and particularly relates to an uplink energy efficiency optimization method for RIS-assisted cell-free massive MIMO systems. Background Art

[0002] With the rapid development of wireless communication technologies, future wireless networks are required to support higher data transmission rates, lower latency, and stronger energy efficiency management capabilities. As an emerging network architecture, the cell-free massive MIMO system has received extensive attention because it can eliminate cell-edge effects, provide a uniform user experience, and improve spectrum utilization. In a cell-free massive MIMO system, multiple access points cooperate to serve all users, avoiding the inter-cell interference problem in traditional cellular networks and achieving higher diversity gain and spatial multiplexing capabilities through joint processing and cooperative signal transmission.

[0003] However, although the cell-free massive MIMO system can improve network capacity and reliability, its energy efficiency still faces severe challenges. On the one hand, the cell-free architecture requires a large number of access points to provide full coverage services, which leads to a significant increase in the overall system power consumption; on the other hand, during multi-user uplink transmission, the access point side needs to jointly detect multiple signals, increasing the complexity of signal processing. As an intelligent reflecting surface technology, RIS can enhance the signal propagation path by dynamically adjusting the phase shift of the reflecting elements, thereby improving the signal-to-interference-plus-noise ratio of the system. In an RIS-assisted cell-free massive MIMO system, RIS can optimize the propagation path of user signals, enhance useful signals, suppress interference, and at the same time reduce the detection complexity at the access point side. Therefore, reasonably optimizing the joint allocation of the RIS phase shift matrix and user transmit power is crucial for improving the energy efficiency of the system.

[0004] Currently, most studies only focus on the spectrum efficiency optimization or power allocation strategy of cell-free massive MIMO systems, and there is less research on the joint optimization of RIS phase shift optimization and power control. Moreover, most existing methods use greedy algorithms or heuristic methods, which are difficult to obtain the global optimal solution in the energy efficiency optimization problem. To address the above problems, the present invention proposes an uplink energy efficiency optimization method for RIS-assisted cell-free massive MIMO systems, which maximizes the energy efficiency of the system by jointly optimizing the user transmit power and the RIS phase shift matrix. The present invention adopts an alternating optimization strategy, in which the water-filling method is used to optimize the user transmit power, and the genetic algorithm is used to optimize the RIS phase shift matrix, thereby improving the energy efficiency of the system while ensuring the communication quality. The present invention can effectively reduce the total power consumption of the system, achieve the co-optimization of power and phase shift, thereby improving the wireless resource utilization rate of the cell-free massive MIMO system and providing a new technical solution for future high-energy efficiency wireless communication systems. Summary of the Invention

[0005] To solve the above problems, the present invention discloses an uplink energy efficiency optimization method for a RIS-assisted cell-free massive MIMO system. By alternately optimizing the transmit power of users and the phase shift matrix of the RIS, the uplink energy efficiency of the system is improved, and the system energy consumption is reduced.

[0006] Technical Solution:

[0007] An uplink energy efficiency optimization method for a RIS-assisted cell-free massive MIMO system, the method comprising the following steps:

[0008] Step (1) Based on the RIS-assisted cell-free massive MIMO system, establish an uplink channel model, where the channel model is based on the Rayleigh fading assumption;

[0009] Step (2) Based on the channel model, derive the received signal expression of the access point during the uplink data transmission process, and calculate the analytical expression of the uplink energy efficiency;

[0010] Step (3) With the maximization of the system energy efficiency as the optimization goal, construct a non-convex optimization problem including power constraints and phase shift constraints;

[0011] Step (4) Adopt an alternating optimization method, use the water-filling method to optimize the transmit power of users and the genetic algorithm to optimize the phase shift matrix of the RIS respectively, and calculate the system energy efficiency based on the optimized parameters.

[0012] Furthermore, in the uplink of the RIS-assisted cell-free MIMO system, the access point can obtain perfect channel state information; the number of access points in the system is M, the number of users is K, and the number of RISs is N.

[0013] Furthermore, based on the Rayleigh fading assumption, the channel model can be expressed as:

[0014]

[0015] where, f mk ~CN(0,β mk ), g m ~CN(0,ν m I N ), h k ~CN(0,ρ k I N ) represent the direct link from the access point to the user, the reflected link from the RIS to the access point, and the reflected link from the user to the RIS respectively. Among them, β mk , ν m and ρ kThey are the large-scale fading coefficients between the m-th access point and the k-th user, between the m-th access point and the RIS, and between the RIS and the k-th user, respectively. denotes the phase shift matrix of the RIS. Where G = diag(g1, g2,..., g N ) and denote the amplitude amplification matrix and the phase shift matrix of the RIS, respectively.

[0016] Furthermore, based on the channel model, in the uplink data transmission phase, the signal expression received by the access point m is:

[0017]

[0018] where, p k , s k ~CN(0,1), denote the transmit power, data symbol of the k-th user, the noise at the access point, and the noise at the RIS, respectively. The m-th access point applies the receive filter to process the received signal and then forwards it to the central processing unit for further processing. The present invention considers using a matched filter for signal reception at the access point because of its simple calculation and distributed application ability.

[0019] Therefore, the observed signal of the k-th user is expressed as:

[0020]

[0021] The receive filter w mk is designed based on the local channel state information obtained by the m-th access point. The m-th access point obtains the channel information by estimating the cascaded channel defined by the user-RIS-access point and user-access point channels, i.e., e mk The present invention assumes that the access point has perfect channel state information about this cascaded channel.

[0022] Furthermore, based on the channel model, the signal-to-interference-plus-noise ratio (SINR) expression at the user k in the uplink of the system is as follows:

[0023]

[0024] where I kj , Z A,k and Z m,k are the inter-user interference generated by the j-th user to the k-th user, the amplified noise at the active RIS, and the noise at the access point, respectively, and their definitions are as follows:

[0025]

[0026] Furthermore, with the goal of maximizing the system energy efficiency, a non-convex optimization problem including user transmit power constraints and RIS phase shift constraints is constructed, and its mathematical expression is as follows:

[0027]

[0028] where, is the spectral efficiency of the k-th user, B is the bandwidth, is the signal-to-interference-plus-noise ratio of the k-th user, G max is the gain constraint of a single RIS element, P R is the maximum transmission power of the RIS.

[0029] P total is the total system power consumption, which consists of user transmit power, RIS power consumption, and access point power consumption, and the expression is as follows:

[0030] P total = P UE + P RIS + P AP ,

[0031] where, is the sum of user transmit powers;

[0032] P RIS = P amp + P control + P noise is the power consumption at the RIS. Among them, is the signal amplification power consumption of the RIS, P in,n is the input power of this element; P control = NP unit is the control circuit power consumption of the RIS, P unit is the fixed control power consumption of each RIS element; is the amplified noise power consumption of the RIS.

[0033] P AP = P AP,receive + P AP,processing is the power consumption at the access point, including the reception power P AP,receive and the processing power P AP,processing in two parts, which are respectively defined as: P AP,processing = M * P proc , P proc is the processing power consumption of each access point.

[0034] Furthermore, for the non-convex optimization problem (P1), an alternating optimization method is adopted, the water-filling method is used to optimize the user transmit power, the genetic algorithm is used to optimize the RIS phase shift matrix, and the system energy efficiency is calculated based on the optimized parameters.

[0035] Furthermore, with the fixed RIS phase shift Θ, the water-filling method is used to optimize the user transmit power to ensure maximizing the energy efficiency while meeting the system constraints. The optimization problem is reformulated as:

[0036]

[0037] The user transmit power optimization method in the present invention includes the following steps:

[0038] (1) Channel gain calculation and sorting

[0039] Calculate the equivalent channel gain of each user:

[0040] Sort the users in descending order of channel gain, such that λ (1) ≥λ (2) ≥...≥λ (K) .

[0041] (2) Use the bisection method to calculate the water level

[0042] Set the water level search interval [μ min ,μ max :

[0043]

[0044] Use bisection search to solve for the optimal water level μ * that satisfies the total power constraint:

[0045]

[0046] (3) User transmit power allocation

[0047] Based on the calculated optimal water level μ * , allocate the user transmit power:

[0048]

[0049] (4) Power constraint check

[0050] Calculate the optimized total power: If P UE_opt >P max , then adjust the water level and repeat the steps until the constraint is satisfied.

[0051] (5) Output the optimization result

[0052] Calculate the optimized system energy efficiency:

[0053]

[0054] If the energy efficiency converges or reaches the set number of iterations, the optimal user transmit power allocation scheme is output and enter the RIS phase shift optimization stage.

[0055] Furthermore, with the user transmit power P fixed, the genetic algorithm is used to optimize the RIS phase shift matrix Θ. This method gradually optimizes the phase shift matrix of the RIS through genetic operations to maximize the energy efficiency of the system while satisfying the system constraints. The optimization problem is reformulated as:

[0056]

[0057] The RIS phase shift optimization method in the invention includes the following steps:

[0058] (1) Individual initialization:

[0059] Set the number N of RIS reflection units and the population size P pop , for subsequent iterative optimization;

[0060] Randomly generate an initial population within the defined phase shift constraints. Each individual Θ j is composed of the phase shift vector θ j = [θ1, θ2,..., θ N ;

[0061] (2) Fitness evaluation:

[0062] Calculate the system energy efficiency corresponding to each individual Θ j in the population:

[0063]

[0064] Sort the individuals according to the fitness values, and retain the individuals with higher fitness values for the next generation of evolution.

[0065] (3) Selection:

[0066] Adopt the roulette wheel selection method to select individuals with higher fitness values from the current population as the parents. The selection probability is weighted according to the fitness values to ensure that excellent individuals have a higher survival rate.

[0067] (4) Crossover:

[0068] Adopt uniform crossover for genetic operations.

[0069] For two parent individuals Θ a and Θ b , according to the crossover probability P cross exchange some RIS phase shift values:

[0070]

[0071] (5) Mutation:

[0072] Random mutation is used to perturb the phase shift values of some individuals to enhance the diversity of the population.

[0073] For the mutation probability P m , if r < P m , then update the phase shift of the nth RIS reflection unit to:

[0074]

[0075] where δ is the maximum mutation amplitude to ensure that the phase shift remains within the feasible range.

[0076] (6) Convergence judgment:

[0077] Calculate the energy efficiency of the optimal individual in the current generation population and judge the convergence condition:

[0078]

[0079] If the convergence condition is satisfied or the maximum number of iterations I max is reached, then output the optimal RIS phase shift matrix Θ * ;

[0080] Otherwise, update the population and return to step (2) for the next round of optimization.

[0081] (7) Output the optimal solution:

[0082] After the iteration ends, output the optimal RIS phase shift matrix Θ * , and calculate the final system energy efficiency:

[0083]

[0084] Use the optimal phase shift matrix for the next stage of the alternating optimization process to ensure that the system achieves maximum energy efficiency under the optimal transmit power and RIS phase shift configuration.

[0085] Furthermore, summarize the algorithm steps of the alternating optimization as follows:

[0086] (1) Algorithm initialization

[0087] Set system parameters, including the number of users K, the number of access points M, the number of RIS reflection units N, the maximum power constraint Pmax, the convergence threshold ∈, and the maximum number of iterations Imax; initialize the user transmit power and the RIS phase shift matrix Θ (0) ; calculate the initial energy efficiency

[0088] (2) Optimize the user transmit power p k (Fix Θ)

[0089] With the fixed RIS phase shift matrix Θ (t) the water-filling method is adopted to optimize the user transmit power.

[0090] (3) Optimize the RIS phase shift matrix Θ (fix p k )

[0091] With the fixed user transmit power the genetic algorithm is adopted to optimize the RIS phase shift.

[0092] (4) Convergence determination

[0093] Calculate the energy efficiency increment: If Δη EE < ε or the maximum number of iterations I max is reached, stop the iteration; otherwise, let t = t + 1 and return to step (2) to continue the optimization.

[0094] (5) Output the optimal solution

[0095] After the iteration ends, output the optimal user transmit power and the optimal phase shift matrix Θ * and calculate the final system energy efficiency

[0096] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0097] (1) Different from the existing independent methods that only optimize the user power or the RIS phase shift, the present invention adopts an alternating optimization method to optimize the RIS phase shift matrix while optimizing the user transmit power, ensuring the balance between the system energy efficiency and the spectral efficiency. By maximizing the system energy efficiency, the de-cellular massive MIMO system reduces energy consumption while improving communication quality and resource utilization rate, providing effective technical support for the development of green communication;

[0098] (2) Most of the prior art adopts greedy algorithms or convex optimization methods to solve the power control and phase shift optimization problems. However, the present invention calculates the optimal user power allocation by the water-filling method and combines the genetic algorithm to optimize the RIS phase shift, thus avoiding the local optimal problem caused by non-convexity in the traditional methods. Experimental results show that this method can quickly converge to the optimal solution, and has low computational complexity, improving the practicability of the algorithm;

[0099] (3) Through simulation experiments, it is verified that the optimization method proposed in the present invention can converge quickly under different numbers of users, RIS scales, and system power consumption models, and can significantly improve the system energy efficiency. At the same time, compared with the traditional method, this method achieves a better energy efficiency optimization effect while maintaining a low computational complexity, and is suitable for the deployment of actual wireless communication systems. Description of the Drawings

[0100] Figure 1 It is a graph showing the relationship between uplink energy efficiency and the number of access points M;

[0101] Figure 2 It is a graph showing the relationship between uplink energy efficiency and the number of iterations. Detailed Embodiment

[0102] The present invention will be further clarified below in conjunction with the drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. It should be noted that the words "front", "rear", "left", "right", "up" and "down" used in the following description refer to the directions in the drawings, and the words "inner" and "outer" refer to the directions towards or away from the geometric center of a specific component respectively.

[0103] This embodiment provides an uplink energy efficiency optimization method for a RIS-assisted cell-free massive MIMO system. By alternately optimizing the transmit power of users and the phase shift matrix of the RIS, the uplink energy efficiency of the system is improved, and the system energy consumption is reduced.

[0104] The main steps of this method include:

[0105] 1. Based on the uplink channel model of a RIS-assisted cell-free massive MIMO system, where the channel model is based on the Rayleigh fading assumption;

[0106] 2. Based on the channel model, derive the received signal expression of the access point during the uplink data transmission process, and calculate the analytical expression of the uplink energy efficiency;

[0107] 3. With the maximization of the system energy efficiency as the optimization goal, construct a non-convex optimization problem including power constraints and phase shift constraints;

[0108] 4. Adopt an alternating optimization method, use the water filling method to optimize the transmit power of users and the genetic algorithm to optimize the phase shift matrix of the RIS respectively, and calculate the system energy efficiency based on the optimized parameters.

[0109] Furthermore, based on the Rayleigh fading assumption, the channel model can be expressed as:

[0110]

[0111] Among them, f mk ~CN(0, β mk ), g m ~CN(0, ν m I N ), h k ~CN(0, ρ k I N ) respectively represent the direct link between the access point and the user, the reflected link from the RIS to the access point, and the reflected link from the user to the RIS. Among them, β mk , ν m and ρ k are respectively the large-scale fading coefficients between the m-th access point and the k-th user, between the m-th access point and the RIS, and between the RIS and the k-th user. represents the phase shift matrix of the RIS. Among them, G = diag(g1, g2,..., g N ) and respectively represent the amplitude amplification matrix and the phase shift matrix of the RIS.

[0112] Further, based on the channel model, in the uplink data transmission phase, the signal expression received by the access point m is:

[0113]

[0114] Among them, p k , s k ~CN(0, 1), respectively represent the transmission power of the k-th user, the data symbol, the noise at the access point, and the noise at the RIS. The m-th access point applies the receive filter to process the received signal, and then forwards it to the central processing unit for further processing. The present invention considers using a matched filter at the access point because its calculation is simple and it can be applied distributively.

[0115] Therefore, the signal of the observed k-th user is expressed as:

[0116]

[0117] The receive filter w mk is designed based on the local channel state information obtained by the m-th access point. The m-th access point obtains the channel information by estimating the cascaded channel defined by the user-RIS-access point and user-access point channels, that is, e mk . The present invention assumes that the access point has perfect channel state information about this cascaded channel.

[0118] Furthermore, based on the channel model, the signal-to-interference-plus-noise ratio (SINR) expression at the uplink of the system at user k is as follows:

[0119]

[0120] where I kj , Z A,k and Z m,k are the inter-user interference generated by the j-th user to the k-th user, the amplified noise at the active RIS, and the noise at the access point, respectively, and their definitions are as follows:

[0121]

[0122] Furthermore, with the goal of maximizing the system energy efficiency, a non-convex optimization problem including user transmit power constraints and RIS phase shift constraints is constructed, and its mathematical expression is as follows:

[0123]

[0124] where, is the spectral efficiency of the k-th user, B is the bandwidth, is the signal-to-interference-plus-noise ratio of the k-th user, G max is the gain constraint of a single RIS unit, P R is the maximum transmission power of the RIS,

[0125] P total is the total system power consumption, which consists of user transmit power, RIS power consumption, and access point power consumption, and the expression is as follows:

[0126] P total = P UE + P RIS + P AP

[0127] where, is the sum of user transmit powers;

[0128] P RIS = P amp + P control + P noise is the power consumption at the RIS. Among them, is the signal amplification power consumption of the RIS, P in,n is the input power of this unit; P control = NP unit is the control circuit power consumption of the RIS, P unit is the fixed control power consumption of each RIS unit; is the amplified noise power consumption of the RIS.

[0129] P AP = PAP,receive +P AP,processing is the power consumption at the access point, including the receiving power consumption P AP,receive and the processing power consumption P AP,processing These two parts are respectively defined as: P AP,processing = M * P proc , P proc is the processing power consumption of each access point.

[0130] Furthermore, for the non-convex optimization problem (P1), an alternating optimization method is adopted. The water-filling method is used to optimize the user transmit power, the genetic algorithm is used to optimize the RIS phase shift matrix, and the system energy efficiency is calculated based on the optimized parameters.

[0131] The algorithm steps of the alternating optimization are summarized as follows:

[0132] (1) Algorithm initialization

[0133] Set the system parameters, including the number of users K, the number of access points M, the number of RIS reflection units N, the maximum power constraint Pmax, the convergence threshold ∈, and the maximum number of iterations Imax; initialize the user transmit power and the RIS phase shift matrix Θ (0) ; calculate the initial energy efficiency

[0134] (2) Optimize the user transmit power p k (Fix Θ)

[0135] With the RIS phase shift matrix Θ (t) fixed, the water-filling method is used to optimize the user transmit power.

[0136] (3) Optimize the RIS phase shift matrix Θ (fix p k )

[0137] With the user transmit power fixed, the genetic algorithm is used to optimize the RIS phase shift.

[0138] (4) Convergence determination

[0139] Calculate the energy efficiency increment: If Δη EE < ε or the maximum number of iterations I max is reached, stop the iteration; otherwise, let t = t + 1 and return to step (2) to continue the optimization.

[0140] (5) Output the optimal solution

[0141] After the iteration ends, output the optimal user transmit power and the optimal phase shift matrix Θ *, and calculate the final system energy efficiency

[0142] Furthermore, referring to Figure 1 , when the number of users K = 100, the model of the present invention significantly improves the uplink energy efficiency. Under the same number of access points, its energy efficiency is always higher than that of the traditional MIMO scheme without RIS assistance and the unoptimized RIS scheme, verifying the effectiveness of the optimized RIS design in improving system performance. In addition, the energy efficiency of all schemes increases with the increase in the number of access points, but the growth rate of the model of the present invention is faster, indicating that it has more obvious performance advantages in the scenario of large-scale AP deployment, thus verifying the innovation and superiority of this technical solution in the optimization of wireless communication energy efficiency.

[0143] Furthermore, referring to Figure 2 When the number of users K = 100 and the number of access points M = 50, with the increase in the number of iterations, the uplink energy efficiency is significantly improved and converges after approaching 25 iterations, indicating that the adopted optimization algorithm can achieve significant energy efficiency improvement within a small number of iterations and has a fast convergence speed.

[0144] The technical means disclosed in the solution of the present invention are not limited to the technical means disclosed in the above embodiments, but also include technical solutions composed of any combination of the above technical features.

Claims

1. A method for optimizing uplink energy efficiency for a RIS-assisted decellularized massive MIMO system, characterized in that: The following steps are involved: Step (1.1) establishing an uplink channel model based on the RIS-assisted decellularized massive MIMO system, wherein the uplink channel model is based on the Rayleigh fading assumption; Step (1.2) derives the received signal expression of the access point during uplink data transmission based on the uplink channel model, and calculates the analytical expression of the uplink energy efficiency; Step (1.3) takes maximizing the system energy efficiency as the optimization goal and constructs a non-convex optimization problem including power constraints and phase shift constraints; Step (1.4) adopts an alternating optimization method, using the water injection method to optimize the user transmission power, the genetic algorithm to optimize the RIS phase shift matrix, and calculates the system energy efficiency based on the optimized parameters.

2. The uplink energy efficiency optimization method for a RIS-assisted decellularized massive MIMO system according to claim 1, characterized in that: The channel model assumes that the uplink channel obeys Rayleigh fading, including a direct link from the user to the access point and an indirect link from the user to the access point via RIS reflection, where the access point has perfect channel state information and uses a matched filter for signal reception; the channel model in step (1.1) is expressed as follows: Where: e mk represents the equivalent uplink channel between the kth user and the mth access point; f mk ~CN(0,β mk ) represents the direct link between the mth access point and the kth user, β mk is the large-scale fading coefficient between the mth access point and the kth user; g m ~CN(0,v m I N ) represents the reflection link between RIS and the mth access point, v m is the large-scale fading coefficient between RIS and the mth access point; h k ~CN(0,ρ k I N ) represents the reflection link between the kth user and RIS, ρ k is the large-scale fading coefficient from the kth user to the RIS; Represents the phase shift matrix of RIS. Where G = diag(g1,g2,...,g N )and They represent the amplitude amplification matrix and phase shift matrix of RIS respectively.

3. The uplink energy efficiency optimization method for a RIS-assisted decellularized massive MIMO system according to claim 1, characterized in that: Based on the uplink channel model, the received signal expression of the access point during the uplink data transmission process in step (1.2) is derived as follows: Where: y m represents the signal received by the mth access point; e mk represents the equivalent uplink channel between the kth user and the mth access point; p k is the transmission power of the kth user; s k The symbol sent by the kth user; is the receiving noise of the access point; is the noise at RIS; The access point uses a matched filter for signal detection, and the combined signal expression is: Where: receiving filter w mk It is designed based on the local channel state information obtained at the mth AP.

4. The uplink energy efficiency optimization method for a RIS-assisted decellularized massive MIMO system according to claim 1, characterized in that: Taking maximizing the system energy efficiency as the optimization goal, the non-convex optimization problem containing the user transmission power constraint and the RIS phase shift constraint in step (1.3) is constructed, and its mathematical expression is as follows: in: is the spectrum efficiency of the kth user; is the signal-to-interference-noise ratio of the kth user; P total is the total system power consumption, including user transmit power, RIS power consumption, and access point power consumption; G max is the gain constraint of a single RIS unit; P R is the maximum transmission power of RIS.

5. The uplink energy efficiency optimization method for a RIS-assisted decellularized massive MIMO system according to claim 1, characterized in that: The optimization problem described in step (1.4) is solved using an alternating optimization algorithm, which specifically includes the following steps: Step 4.1: Initialize the initial value of user transmission power RIS phase shift matrix initial value Θ (0) , maximum number of iterations I max and convergence threshold ε; Step 4.2: Optimize user transmission power p k ,,fix Θ; (421) Calculate channel gain And sort the channel gains of all users from large to small, denoted as λ (1) ≥λ (2) ≥...≥λ (K) ; (422) Use the binary method to calculate the water level μ: Set the water level search interval [μ min ,μ max ], and binary search is used to solve the optimal water level μ that satisfies the total power constraint. * ; (423) Power allocation based on water level (424) calculating the updated system energy efficiency; Step 4.3, optimize the RIS phase shift matrix Θ, fix p k ; (431) Individual initialization: Randomly initialize the RIS phase shift matrix population; (432) Fitness evaluation: Calculate the energy efficiency η corresponding to each individual EE,j ; (433) Selection: individuals with higher fitness are selected to enter the next generation; (434) Crossover: Randomly exchange some RIS phase shift values; (435) Mutation: fine-tune individual RIS phase shift values ​​to ensure search space coverage; (436) Convergence judgment: If the convergence condition is met or the maximum number of iterations is reached, the optimal phase shift matrix Θ is output * And calculate the final system energy efficiency; Step 4.4: Convergence judgment; Calculating Energy Efficiency Gains If Δη EE <ε or reaches the maximum number of iterations I max , then stop the iteration; otherwise, set t = t + 1 and return to step 4.2 to continue the optimization; Step 4.5: Output the optimal solution After the iteration, the optimal user transmit power is output and the optimal RIS phase shift matrix Θ * , and calculate the final system energy efficiency

Images