A method, system, device and medium for wireless resource management in a hybrid mobile communication network based on convex optimization

Abstract

A kind of mixed mobile communication network wireless resource management method, system, equipment and medium based on convex optimization belong to digital information transmission technical field, method includes: the parameter initialization of mobile communication network system is carried out, and its downlink transmission scene model is established;Equivalent delay-Doppler domain channel matrix and the spectral efficiency of high-speed mobile user and low-speed mobile user are calculated, and joint optimization model about user scheduling, beam width and transmit power is constructed, and it is decomposed into user scheduling submodel and beam width and transmit power joint optimization submodel;User scheduling submodel is convex and iteratively solved, and the optimal user matching scheme of mixed mobile user and channel is obtained;Beam width and transmit power joint optimization submodel is convex and iteratively solved, and the optimal beam width and transmit power allocation scheme is obtained;Transmission efficiency evaluation and feedback are completed;The transmission rate and communication reliability of mixed mobile communication network are effectively improved by the application.

Description

Technical Field

[0001] This invention belongs to the field of digital information transmission technology, and specifically relates to a method, system, device and medium for wireless resource management of hybrid mobile communication networks based on convex optimization. Background Technology

[0002] Achieving highly reliable and high-speed wireless resource management is a core challenge for high-speed mobile communication networks. In hybrid mobile communication networks, the severe Doppler frequency shift caused by high-speed users severely restricts transmission performance. Traditional solutions mainly rely on compensation techniques such as measurement calculation and pilot estimation, which are widely used in civilian fields such as high-speed railways and low-Earth orbit satellite communications. The US Starlink system, based on precise satellite orbits and user Global Positioning System (GPS) information, adopts a "prediction + compensation" strategy to pre-correct Doppler frequency shift before signal transmission. However, it relies on predictable user motion models and is difficult to adapt to the dynamic and ever-changing communication scenarios of ground mobile users. To reduce its dependence on GPS, European telecom operator Sateliot developed an IoT system based on low-Earth orbit satellites. This system directly analyzes positioning and timing information through communication signals and uses complex algorithms to estimate and compensate for channel changes in real time. However, this solution requires high computing power from terminals, limiting its practical deployment. my country has made progress in the field of integrated sensing and communication. By fusing radar sensing and communication functions and leveraging radar's ability to predict Doppler interference, a compensation carrier has been designed to counteract frequency offset caused by high-speed movement, significantly improving the system's anti-interference performance. This has led to its application in high-speed rail networks. However, this system requires high real-time signal processing capabilities in complex ground environments, and with the expansion of mixed user scale, inter-device interference problems are becoming increasingly prominent. In summary, although frequency offset compensation technology has been extensively studied, the industry still urgently needs to develop an efficient wireless resource management method with inherent Doppler resistance for mixed mobile communication scenarios.

[0003] In high-speed mobile communication environments, while existing frequency offset compensation schemes can partially suppress the Doppler effect, they still have significant limitations. These schemes heavily rely on accurate real-time channel sensing, and the dense pilot deployments required for complex channel estimation consume scarce radio spectrum resources, increasing system overhead and signaling latency. Furthermore, passive compensation-based resistance mechanisms struggle to adapt to the rapidly changing characteristics of the channel, failing to provide stable, reliable, and high-quality communication services for high-speed mobile users.

[0004] Preliminary research indicates that there is currently a lack of mature wireless resource management solutions in China capable of achieving high-reliability, high-speed transmission in hybrid mobile communication network environments. Existing methods largely rely on direct estimation of complex time-varying channels, leading to high signal detection complexity and large system overhead. Taking Orthogonal Frequency Division Multiplexing (OFDM) systems as an example, their performance depends on accurate time-frequency domain channel state information. However, the rapid time-varying nature of the channel makes accurate estimation difficult, and inaccurate channel information further causes inter-carrier interference and equalization failure, resulting in a significant decrease in the system's signal-to-noise ratio and making it difficult to meet the high-reliability, high-throughput transmission requirements of high-mobility services.

[0005] Patent application CN102905356B discloses a power control method in heterogeneous mobile communication networks. This method involves a macro base station determining an acceptable interference value based on the average interference noise level and informing neighboring low-power nodes via the X2 interface or air interface. The low-power nodes then dynamically determine their uplink transmission power based on this value, thereby achieving interference coordination in the heterogeneous network. This scheme focuses on managing static or quasi-static network interference through signaling interaction between base stations. However, because it relies on preset, periodic interference threshold information exchange and a responsive power adjustment mechanism, its dynamic adaptability is insufficient when facing rapidly changing channel conditions caused by high-speed user movement. It struggles to provide timely and accurate compensation for instantaneously changing channel states, thus failing to effectively address the fundamental problem of decreased transmission reliability caused by drastic channel changes in high-speed mobile scenarios. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, the present invention aims to provide a method, system, device, and medium for radio resource management in hybrid mobile communication networks based on convex optimization. This involves initializing parameters of the mobile communication network system, utilizing sector partitioning technology to implement non-orthogonal multiple access (NOMA) user packet and resource block partitioning, and establishing a downlink transmission scenario model for the hybrid mobile communication network. Furthermore, it introduces orthogonal time-frequency-space (OTFS) modulation technology to convert the time-varying time-frequency (TF) domain channel into a sparse quasi-static delay. This invention utilizes a Doppler (DD) domain channel to fundamentally suppress Doppler frequency shift and ensure high-reliability transmission. By calculating the equivalent DD domain channel matrix and the spectral efficiency of high-speed and low-speed mobile users, a joint optimization model for user scheduling, beamwidth, and transmit power is constructed. Based on this, the original joint optimization model is decomposed into a user scheduling sub-model and a beamwidth and transmit power joint optimization sub-model, which are then solved using convex optimization methods to obtain the optimal user matching scheme and the optimal beamwidth and transmit power allocation scheme. This improves the transmission performance of high-speed mobile users while ensuring the minimum average throughput requirement for low-speed mobile users within the NOMA group. Therefore, this invention effectively improves the transmission rate and communication reliability of hybrid mobile communication networks.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for radio resource management in hybrid mobile communication networks based on convex optimization includes the following steps: Step 1: Initialize the parameters of the mobile communication network system and establish its downlink transmission scenario model; Step 2: Based on the downlink transmission scenario model, calculate the equivalent delay-Doppler DD domain channel matrix and the spectral efficiency of high-speed mobile users and low-speed mobile users, construct a joint optimization model for user scheduling, beamwidth and transmit power, and decompose it into a user scheduling sub-model and a beamwidth and transmit power joint optimization sub-model. Step 3: Perform convexification and iterative solution on the user scheduling sub-model to obtain the optimal user matching scheme for hybrid mobile users and channels; Step 4: Based on the optimal user matching scheme, perform convexification and iterative solution on the joint optimization sub-model of beamwidth and transmit power to obtain the optimal beamwidth and transmit power allocation scheme; Step 5: Based on the optimal user matching scheme and the optimal beamwidth and transmit power allocation scheme, perform transmission efficiency evaluation and feedback.

[0008] The specific process of step 1 includes: Step 1.1: Obtain the system parameters of the mobile communication network, including network configuration parameters: the number of time intervals N and the number of subcarriers M in orthogonal time-frequency space-time (OTFS) modulation, and the symbol duration. Subcarrier spacing Preset sector angle, base station location, pilot transmission time Total working hours User parameters: number of users, initial user location, user speed, multipath delay. With Doppler frequency shift Channel parameters: Number of available frequency bands Q The system parameters are initialized using path loss factor and Nakagami distribution parameters; where the subscript... i Indicates the first i A high-speed mobile user non-orthogonal multiple access (NOMA) group, subscript k Indicates the number of NOMA groups contained in a set. k A high-speed or low-speed mobile user, subscript p Indicates the first p One transmission path; Step 1.2: Based on the system parameters obtained in Step 1.1, implement NOMA access user grouping and resource block allocation, and establish a downlink transmission scenario model for the hybrid mobile communication network. The specific process is as follows: The downlink transmission scenario model is divided into sectors based on the number of users, initial user locations, base station locations, and preset sector angles. The number of NOMA access groups within each sector is determined based on the available frequency bands. Based on user speed, a corresponding number of high-speed mobile users and low-speed mobile users are configured for each NOMA user group according to user distribution. Simultaneously, the base station transmits narrow beams to users within the group for alignment. Resource blocks in the DD domain are allocated to the high-speed mobile user group according to network configuration parameters, and resource blocks in the TF domain are allocated to the low-speed mobile user group. User information is overlaid using inverse sine Fourier transform and power domain NOMA access technology, completing the NOMA access user grouping and resource block division.

[0009] The specific process of step 2 includes: Step 2.1: Calculate the equivalent DD domain channel matrix for each user in the downlink transmission scenario model obtained in Step 1. The specific process is as follows: Calculate the large-scale fading of the channel using the path loss factor obtained in step 1.1, and calculate the small-scale fading of the channel using the Nakagami distribution parameters. Multiply the large-scale fading and the small-scale fading by the total fading coefficient of the channel to obtain the total fading coefficient. subscript q Indicates the first q Available frequency bands; using the multipath delay obtained in step 1.1 With Doppler frequency shift It is then converted into a DD-domain tap index through discretization sampling, including Doppler taps and time-delay taps, wherein the Doppler taps include an integer part. and fractional part , represented as The delay tap is ; Based on the NOMA access user packets and resource block partitioning obtained in step 1.2, the equivalent DD domain channel matrix for each path is constructed using the DD domain tap index through block cyclic matrix operations, and then all... The channel matrix of the i-th path is calculated. i In the NOMA group, the first k The high-speed user in the first q Equivalent DD domain channel matrix on each frequency band , represented as: (1) in, For the fractional Doppler case, the approximate constant is... d This indicates the selection tap acting on the Doppler approximation. Represent a A cyclic matrix is ​​created by cyclically shifting each row of the identity matrix. get, This represents modulo N operation. Indicates the integer part of the Doppler tap. This represents the fractional part of the Doppler tap. Indicates a delayed tap. Indicates the Kronecker product; Step 2.2: Based on the downlink transmission scenario model and the equivalent DD domain channel matrix, construct the user signal reception model and calculate the spectral efficiency expression. The specific process is as follows: Based on the downlink transmission scenario model parameters obtained in step 1, the beam alignment time is calculated as follows: (2) in, I Indicates the number of NOMA groups for high-speed mobile users. J This indicates the number of NOMA groups for low-speed mobile users and defines... I=J=Q , K This indicates the number of high-speed or low-speed mobile users included in a NOMA group. For pilot transmission time, To fix the sector angle value, and These are the beamwidth variables related to high-speed mobile users and low-speed mobile users, respectively, where the subscripts... j Indicates the first j For each low-speed mobile user NOMA group, the effective transmission time is then calculated. The effective antenna gain in the main lobe direction for high-speed and low-speed mobile users was calculated respectively. and ,in Indicates sidelobe gain; For high-speed mobile users, zero-forcing equalization detection in the DD domain is employed. The equivalent DD domain channel matrix is ​​diagonalized using the discrete Fourier transform matrix. ,use dimensional discrete Fourier transform matrix The diagonal matrix of the equivalent DD domain channel matrix is ​​obtained. N and M are the dimensions of the OTFS frame, combined with the power of high-speed mobile users. Low-speed mobile user power and Boolean scheduling variables and Receiver antenna gain Noise power and diagonal matrix The One main diagonal element A spectrum efficiency expression for high-speed mobile users is constructed using Shannon's formula. , means as follows: (3) For low-speed mobile users, frequency-domain zero-forcing equalization detection combined with continuous interference cancellation technology is adopted, based on the power of the low-speed mobile user. Receiver antenna gain Noise power and the TF domain channel matrix for low-speed mobile users The expression for the spectral efficiency of low-speed mobile users is constructed as follows; (4) in, This indicates the number of subcarriers allocated to each user within the NOMA group; Step 2.3: Based on the spectral efficiency expressions for high-speed mobile users and low-speed mobile users, construct a set of user scheduling variables. Beamwidth variable set With the set of transmit power variables The joint optimization model includes: setting the objective function as the sum of the average spectral efficiencies of all high-speed mobile users, expressed as: (5) The constraints include: low-speed user throughput requirements, expressed as: (6) in, Indicates the subcarrier width. The minimum throughput rate limit is indicated by: The system power limit is indicated by: , Indicates the base station's maximum transmit power; NOMA decoding constraints: Total time constraint: Beamwidth constraints: , Indicates the minimum beamwidth. Indicates the maximum beamwidth; Boolean variable constraints between user groups: and ,in ; Step 2.4: Decompose the joint optimization model constructed in Step 2.3 into two sub-models: a user scheduling sub-model obtained by fixing the beamwidth and transmit power, and a joint optimization sub-model of beamwidth and transmit power obtained by fixing the user scheduling scheme.

[0010] In step 2.4, when fixing the beamwidth and transmit power, the beamwidth is initialized to the arithmetic mean of the adjustable maximum beamwidth and minimum beamwidth of the base station, and the transmit power is initialized to the system's preset average transmit power. After solving the user scheduling sub-model, the optimal user matching scheme output by the user scheduling sub-model is used as the fixed user scheduling variable of the joint optimization sub-model.

[0011] The specific process of convexification and iterative solution of the user scheduling sub-model in step 3 includes: Step 3.1: For the user scheduling sub-model in Step 2.4, multiple sets of auxiliary variables are introduced. , , , The non-convex objective function and constraints are equivalently transformed by replacing the non-convex expressions in the signal-to-interference-plus-noise ratio term, interference plus noise term, and noise term with the corresponding auxiliary variable forms, and transforming the bilinear coupling term in the low-speed user throughput constraint into a second-order cone constraint form to obtain the intermediate non-convex model. Step 3.2: For each non-convex term in the intermediate non-convex model obtained in Step 3.1, construct... , real part virtual part Sum of squares The first-order Taylor lower bound is approximately: , , and ,in, This indicates the operation of taking the real part. This represents the operation of taking the imaginary part and continuously relaxing the Boolean user scheduling variables, transforming the intermediate non-convex model obtained in step 3.1 into a solvable convex model, including: the objective function, expressed as: (7) The constraints include: convex constraints that simplify the objective: (8) (9) (10) Simplify the convex constraints of throughput requirements: (11) (12) Boolean variable constraints between user groups, and relaxation of them. ;in, , , and Auxiliary variables , , and A set; Step 3.3: Initialize the iteration parameters, including the iteration number index, convergence threshold, and target judgment value; Step 3.4: Based on the system parameters obtained in Step 1.1, randomly generate user matching schemes, and use the DD domain channel matrix obtained in Step 2.1 to calculate the feasible points required for the Taylor expansion in Step 3.2; Step 3.5: Using the interior point method, call the convex optimization solver to solve the convex model obtained in step 3.2, and obtain the optimal target value and user matching scheme; Step 3.6: Based on the target judgment value, if the optimal target value obtained in Step 3.5 does not meet the convergence threshold condition, then update the target judgment value and the feasible points of Taylor expansion and return to Step 3.5; if the convergence threshold condition is met, then terminate the iteration and output the optimal user matching scheme for hybrid mobile users and channels.

[0012] The specific process of convexification and iterative solution of the joint optimization sub-model of beamwidth and transmit power in step 4 includes: Step 4.1: For the joint optimization sub-model of beamwidth and transmit power in Step 2.4, variable substitution and the introduction of multiple sets of auxiliary variables are used. , , , , , By performing an equivalent transformation on the non-convex objective function and constraints, the power variable is rewritten in its reciprocal form. and The nonlinear expressions in the signal-to-interference-plus-noise ratio (SINR) and interference-plus-noise (INR) terms are replaced with the corresponding auxiliary variable forms to obtain the intermediate non-convex model. Step 4.2: For each non-convex term in the intermediate non-convex model obtained in Step 4.1, linear boundary approximations are performed on the power reciprocal term, beamwidth related term, and logarithmic term of auxiliary variables through feasible points of given variables. Specifically, this is achieved by constructing... , First-order Taylor upper bound approximation , and construct respectively , , , , , First-order Taylor lower bound approximation , , , , , The intermediate non-convex model obtained in step 4.1 is transformed into a solvable convex model, including: the objective function, expressed as: (13) The constraints include: convex constraints to simplify the positive part of the objective function: (14) (15) Simplify the convex constraints of the negative part of the objective function: (16) (17) (18) Simplify the convex constraints of throughput requirements: (19) (20) Total power convex constraint after variable substitution: and NOMA power convex constraint: And total time constraints and beamwidth constraints; where the constant expression , , ;in , , , , and Auxiliary variables , , , , and A set; Step 4.3: Initialize the iteration parameters, including the iteration number index, convergence threshold, and target judgment value; Step 4.4: Based on the system parameters obtained in Step 1.1 and the optimal user matching scheme obtained in Step 3, use the DD domain channel matrix obtained in Step 2.1 to calculate the feasible points required for the Taylor expansion in Step 4.2; Step 4.5: Using the interior point method, call the convex optimization solver to solve the convex model obtained in step 4.2 to obtain the optimal target value and beamwidth and transmit power allocation scheme; Step 4.6: Based on the target judgment value, if the optimal target value obtained in Step 4.5 does not meet the convergence threshold condition, update the target judgment value and the feasible points of Taylor expansion and return to Step 4.5; if the convergence threshold condition is met, terminate the iteration and output the optimal beamwidth and transmit power allocation scheme.

[0013] The specific process of transmission efficiency evaluation and feedback in step 5 is as follows: Based on the optimal user matching scheme obtained in step 3 and the optimal beamwidth and transmit power allocation scheme obtained in step 4, the global objective value of the joint optimization model under the current iteration is calculated according to the objective function of the sum of average spectral efficiencies defined in step 2.3. Based on the set objective evaluation value, it is determined whether the obtained global objective value satisfies the convergence condition: if it does, the final resource management scheme for user scheduling, beamwidth, and transmit power is output; if it does not, the objective evaluation value, the initial user matching scheme, and the beamwidth and power allocation scheme are updated, and the process returns to step 3 to continue iterative optimization.

[0014] The present invention also provides a radio resource management system for hybrid mobile communication networks based on convex optimization, comprising: Transmission scenario construction module: used to initialize parameters of the mobile communication network system and establish its downlink transmission scenario model; Resource Management Module: This module is used to calculate the equivalent delay-Doppler DD domain channel matrix and the spectral efficiency of high-speed mobile users and low-speed mobile users based on the downlink transmission scenario model. It also constructs a joint optimization model for user scheduling, beamwidth, and transmit power, and decomposes it into a user scheduling sub-model and a beamwidth and transmit power joint optimization sub-model. Model solving module: used to perform convexification and iterative solution on the user scheduling sub-model to obtain the optimal user matching scheme for hybrid mobile users and channels; and based on the optimal user matching scheme, to perform convexification and iterative solution on the beamwidth and transmit power joint optimization sub-model to obtain the optimal beamwidth and transmit power allocation scheme. Performance evaluation module: used to evaluate and provide feedback on transmission efficiency based on the optimal user matching scheme and the optimal beamwidth and transmit power allocation scheme.

[0015] The present invention also provides a radio resource management device for hybrid mobile communication networks based on convex optimization, comprising: Memory: A computer program that stores the above-described method for managing radio resources in a hybrid mobile communication network based on convex optimization, and is a computer-readable device; Processor: Used to implement the aforementioned method for managing radio resources in a hybrid mobile communication network based on convex optimization when executing the computer program.

[0016] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, enables the implementation of the aforementioned method for managing wireless resources in a hybrid mobile communication network based on convex optimization.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention uses a time-varying channel model based on Nakagami fading distribution, considers the delay and Doppler of multipath channels in hybrid mobile communication networks, establishes an equivalent channel matrix in the DD domain based on the OTFS modulation module, and utilizes the full diversity gain of multipath channels to achieve more accurate and reliable wireless transmission scenario modeling.

[0018] 2. This invention considers the user scheduling sub-model and proposes an optimal user matching scheme for hybrid mobile communication networks based on convex optimization. By coordinating the matching relationship between high-speed mobile user groups, low-speed mobile user groups and available frequency bands, it ensures the minimum throughput of low-speed users and improves the spectrum efficiency of high-speed users.

[0019] 3. This invention considers a joint optimization sub-model of beamwidth and transmit power, and proposes an optimal beamwidth and transmit power allocation scheme for hybrid mobile communication networks based on convex optimization. By jointly optimizing beamwidth and transmit power, it ensures the minimum throughput for low-speed users and improves the spectrum efficiency for high-speed users.

[0020] 4. Based on solving the user scheduling sub-model and the beamwidth and transmit power joint optimization sub-model separately, this invention establishes a joint optimization model for user matching, beamwidth and transmit power. Through wireless resource management in the time-spectrum-power domain, it meets the minimum throughput requirements of low-speed users and ensures high-reliability and high-speed transmission for high-speed users.

[0021] In summary, this invention achieves more accurate and reliable modeling of high-speed mobile scenarios by introducing the Nakagami time-varying channel model and OTFS modulation; it proposes a convex optimization-based method for radio resource management in hybrid mobile communication networks, which improves the spectrum efficiency of high-speed users while ensuring the minimum throughput for low-speed users. Attached Figure Description

[0022] Figure 1 This is a flowchart of the method of the present invention.

[0023] Figure 2 This is a graph showing the number of iterations for convex optimization under different cases.

[0024] Figure 3 This is a comparison chart of convex optimization and average distribution spectral efficiency under different power levels.

[0025] Figure 4This is a comparison chart of the bit error rates of different modulation schemes under different power levels.

[0026] Figure 5 This is a comparison chart of throughput of different multiple access schemes with convex optimization at different power levels.

[0027] Figure 6 This is a comparison chart of the spectral efficiency of convex optimization, greedy algorithm, and random scheduling under different numbers of NOMA groups. Detailed Implementation

[0028] The present invention will now be described in further detail with reference to the accompanying drawings.

[0029] like Figure 1 As shown, the radio resource management method for hybrid mobile communication networks based on convex optimization specifically includes the following steps: Step 1: Initialize the parameters of the mobile communication network system and establish its downlink transmission scenario model; the specific process includes: Step 1.1: Obtain the system parameters of the mobile communication network, including network configuration parameters: the number of time intervals N and the number of subcarriers M in orthogonal time-frequency space-time (OTFS) modulation, and the symbol duration. Subcarrier spacing Preset sector angle, base station location, pilot transmission time Total working hours User parameters: number of users, initial user location, user speed, multipath delay. With Doppler frequency shift Channel parameters: Number of available frequency bands Q The path loss factor and Nakagami distribution parameters are used to initialize system parameters to support subsequent transmission scenario modeling; among them, the subscripts... i Indicates the first i A high-speed mobile user non-orthogonal multiple access (NOMA) group, subscript k Indicates the number of NOMA groups contained in a set. k A high-speed or low-speed mobile user, subscript p Indicates the first p One transmission path; Step 1.2: Based on the system parameters obtained in Step 1.1, implement NOMA access user grouping and resource block allocation, and establish a downlink transmission scenario model for the hybrid mobile communication network. The specific process is as follows: By employing a phased scanning mechanism to reduce millimeter-wave beam alignment time overhead, NOMA access user groups are first grouped using sector-level scanning, followed by beam-level scanning to establish an efficient communication link between the base station and the mobile terminal. Sector partitioning for the hybrid mobile communication network downlink transmission scenario model is completed based on the number of users, initial user locations, base station locations, and preset sector angles. The number of NOMA access groups within each sector is determined based on the available frequency bands, and a corresponding number of high-speed and low-speed mobile users are configured for each NOMA user group according to user speed and distribution. Simultaneously, the base station transmits narrow beams to users within the group for precise alignment. Resource blocks in the DD domain are allocated to the high-speed mobile user group and resource blocks in the TF domain are allocated to the low-speed mobile user group according to network configuration parameters. User information is superimposed using inverse sine Fourier transform and power domain NOMA access technology, completing the NOMA access user grouping and resource block partitioning, and establishing the hybrid mobile communication network downlink transmission scenario model.

[0030] Step 2: Based on the downlink transmission scenario model, calculate the equivalent delay-Doppler DD domain channel matrix and the spectral efficiency for high-speed and low-speed mobile users. Construct a joint optimization model for user scheduling, beamwidth, and transmit power, and decompose it into a user scheduling sub-model and a joint optimization sub-model for beamwidth and transmit power. The specific process includes: Step 2.1: Based on the equivalent DD domain channel matrix of each user in the downlink transmission scenario model obtained in Step 1, the specific process is as follows: Calculate the large-scale fading of the channel using the path loss factor obtained in step 1.1, and calculate the small-scale fading of the channel using the Nakagami distribution parameters. Multiply the large-scale fading and the small-scale fading by the total fading coefficient of the channel to obtain the total fading coefficient. subscript q Indicates the first q Available frequency bands; using the multipath delay obtained in step 1.1 With Doppler frequency shift It is then converted into a DD-domain tap index through discretization sampling, including Doppler taps and time-delay taps, wherein the Doppler taps include an integer part. and fractional part , represented as The delay tap is ; Based on the NOMA access user packets and resource block partitioning obtained in step 1.2, the equivalent DD domain channel matrix for each path is constructed using the DD domain tap index through block cyclic matrix operations, and then all... The channel matrix of the i-th path is calculated. i In the NOMA group, the first k The high-speed user in the first q Equivalent DD domain channel matrix on each frequency band , represented as: ,(twenty one) in, For the fractional Doppler case, the approximate constant is... d This indicates the selection tap acting on the Doppler approximation. Represent a A cyclic matrix is ​​created by cyclically shifting each row of the identity matrix. get, Represents modulo N operation. Indicates the integer part of the Doppler tap. This represents the fractional part of the Doppler tap. Indicates a delayed tap. Representing the Kronecker product; the resulting equivalent DD-domain channel matrix helps in calculating the spectral efficiency of high-speed mobile users; Step 2.2: Based on the downlink transmission scenario model and the equivalent DD domain channel matrix, construct the user signal reception model and calculate the spectral efficiency expression. The specific process is as follows: Based on the downlink transmission scenario model parameters obtained in step 1, the beam alignment time is calculated as follows: ,(twenty two) in, I Indicates the number of NOMA groups for high-speed mobile users. J This indicates the number of NOMA groups for low-speed mobile users and defines... I=J=Q To ensure a one-to-one matching relationship between high-speed mobile users, low-speed mobile users, and available frequency bands, K This indicates the number of high-speed or low-speed mobile users included in a NOMA group. For pilot transmission time, To fix the sector angle value, and These are the beamwidth variables related to high-speed mobile users and low-speed mobile users, respectively, where the subscripts... j Indicates the first j For each low-speed mobile user NOMA group, the effective transmission time is then calculated. The effective antenna gain in the main lobe direction for high-speed and low-speed mobile users was calculated respectively. and ,in Indicates sidelobe gain; For high-speed mobile users, zero-forcing equalization detection in the DD domain is employed. The equivalent DD domain channel matrix is ​​diagonalized using the discrete Fourier transform matrix. ,use 3D discrete Fourier transform matrix The diagonal matrix of the equivalent DD domain channel matrix is ​​obtained. N and M are the dimensions of the OTFS frame, combined with the power of high-speed mobile users. Low-speed mobile user power and Boolean scheduling variables and Receiver antenna gain Noise power and diagonal matrix The One main diagonal element A spectrum efficiency expression for high-speed mobile users is constructed using Shannon's formula. , means as follows: ;(twenty three) For low-speed mobile users, frequency-domain zero-forcing equalization detection combined with continuous interference cancellation technology is adopted, based on the power of the low-speed mobile user. Receiver antenna gain Noise power and the TF domain channel matrix for low-speed mobile users The expression for the spectral efficiency of low-speed mobile users is constructed as follows; ,(twenty four) in, This indicates the number of subcarriers allocated to each user within the NOMA group; Step 2.3: Based on the spectral efficiency expressions for high-speed mobile users and low-speed mobile users, construct a set of user scheduling variables. Beamwidth variable set With the set of transmit power variables The joint optimization model includes: setting the objective function as the sum of the average spectral efficiencies of all high-speed mobile users, expressed as: (25) The constraints include: low-speed user throughput requirements, expressed as: (26) in, Indicates the subcarrier width. This indicates the minimum throughput rate limit; the system power limit is expressed as: , Indicates the base station's maximum transmit power; NOMA decoding constraints: Total time constraint: Beamwidth constraints: , Indicates the minimum beamwidth. Indicates the maximum beamwidth; Boolean variable constraints between user groups: and ,in Since the original joint optimization model is non-convex, the computational complexity of solving it directly is extremely high. Step 2.4: Decompose the joint optimization model constructed in Step 2.3 into two sub-models: a user scheduling sub-model obtained by fixing beamwidth and transmit power, and a joint optimization sub-model of beamwidth and transmit power obtained by fixing user scheduling scheme. Specifically, when fixing beamwidth and transmit power, the beamwidth is initialized to the arithmetic mean of the adjustable maximum and minimum beamwidths of the base station, and the transmit power is initialized to the system's preset average transmit power. After solving the user scheduling sub-model, the optimal user matching scheme output by the user scheduling sub-model is used as the fixed user scheduling variable of the joint optimization sub-model.

[0031] Step 3: Perform convexification and iterative solution on the user scheduling sub-model obtained in Step 2 to obtain the optimal matching scheme between hybrid mobile users and channels; The specific process of convexification and iterative solution of the user scheduling sub-model includes: Step 3.1: For the user scheduling sub-model in Step 2.4, multiple sets of auxiliary variables are introduced. , , , The non-convex objective function and constraints are equivalently transformed by replacing the non-convex expressions in the signal-to-interference-plus-noise ratio term, interference plus noise term, and noise term with the corresponding auxiliary variable forms, and transforming the bilinear coupling term in the low-speed user throughput constraint into a second-order cone constraint form to obtain the intermediate non-convex model. Step 3.2: For each non-convex term in the intermediate non-convex model obtained in Step 3.1, construct... , real part virtual part Sum of squares The first-order Taylor lower bound is approximately: , , and ,in, This indicates the operation of taking the real part. This represents the operation of taking the imaginary part and continuously relaxing the Boolean user scheduling variables, transforming the intermediate non-convex model obtained in step 3.1 into a solvable convex model, including: the objective function, expressed as: (27) The constraints include: convex constraints that simplify the objective: (28) (29) (30) Simplify the convex constraints of throughput requirements: (31) (32) Boolean variable constraints between user groups, and relaxation of them. ;in, , , and Auxiliary variables , , and The set of user scheduling sub-models; the above convex model transformed into the user scheduling sub-model can be solved efficiently using the interior point method, and the optimal user matching scheme can be obtained through iterative solution using the continuous convex approximation method. Step 3.3: Initialize the iteration parameters, including the iteration number index, convergence threshold, and target judgment value; Step 3.4: Based on the system parameters obtained in Step 1.1, randomly generate user matching schemes, and use the DD domain channel matrix obtained in Step 2.1 to calculate the feasible points required for the Taylor expansion in Step 3.2; Step 3.5: Using the interior point method, call the convex optimization solver to solve the convex model obtained in step 3.2, and obtain the optimal target value and user matching scheme; Step 3.6: Based on the target judgment value, if the optimal target value obtained in Step 3.5 does not meet the convergence threshold condition, then update the target judgment value and the feasible points of Taylor expansion and return to Step 3.5; if the convergence threshold condition is met, then terminate the iteration and output the optimal user matching scheme for hybrid mobile users and channels.

[0032] Step 4: Based on the optimal user matching scheme obtained in Step 3, the beamwidth and transmit power joint optimization sub-model obtained in Step 2 is convexized and iteratively solved to obtain the optimal beamwidth and transmit power allocation scheme. The specific process of convexification and iterative solution of the joint optimization sub-model of beamwidth and transmit power includes: Step 4.1: For the joint optimization sub-model of beamwidth and transmit power in Step 2.4, variable substitution and the introduction of multiple sets of auxiliary variables are used. , , , , , By performing an equivalent transformation on the non-convex objective function and constraints, the power variable is rewritten in its reciprocal form. and The nonlinear expressions in the signal-to-interference-plus-noise ratio (SINR) and interference-plus-noise (INR) terms are replaced with the corresponding auxiliary variable forms to obtain the intermediate non-convex model. Step 4.2: For each non-convex term in the intermediate non-convex model obtained in Step 4.1, linear boundary approximations are performed on the power reciprocal term, beamwidth related term, and logarithmic term of auxiliary variables through feasible points of given variables. Specifically, this is achieved by constructing... , First-order Taylor upper bound approximation , and construct respectively , , , , , First-order Taylor lower bound approximation , , , , , The intermediate non-convex model obtained in step 4.1 is transformed into a solvable convex model, including: the objective function, expressed as: (33) The constraints include: convex constraints to simplify the positive part of the objective function: (34) (35) Simplify the convex constraints of the negative part of the objective function: (36) (37) (38) Simplify the convex constraints of throughput requirements: (39) (40) Total power convex constraint after variable substitution: and NOMA power convex constraint: And total time constraints and beamwidth constraints; where the constant expression , , ;in , , , , and Auxiliary variables , , , , and The set of beamwidth and transmit power joint optimization sub-models can be transformed into the above convex model, which can be solved efficiently using the interior point method, and the optimal beamwidth and transmit power allocation scheme can be obtained by iterative solution using the continuous convex approximation method. Step 4.3: Initialize the iteration parameters, including the iteration number index, convergence threshold, and target judgment value; Step 4.4: Based on the system parameters obtained in Step 1.1 and the optimal user matching scheme obtained in Step 3, use the DD domain channel matrix obtained in Step 2.1 to calculate the feasible points required for the Taylor expansion in Step 4.2; Step 4.5: Using the interior point method, call the convex optimization solver to solve the convex model obtained in step 4.2 to obtain the optimal target value and beamwidth and transmit power allocation scheme; Step 4.6: Based on the target judgment value, if the optimal target value obtained in Step 4.5 does not meet the convergence threshold condition, update the target judgment value and the feasible points of Taylor expansion and return to Step 4.5; if the convergence threshold condition is met, terminate the iteration and output the optimal beamwidth and transmit power allocation scheme.

[0033] Step 5: Based on the optimal user matching scheme and the optimal beamwidth and transmit power allocation scheme, perform transmission efficiency evaluation and feedback.

[0034] The specific process for transmission efficiency evaluation and feedback is as follows: Based on the optimal user matching scheme obtained in step 3 and the optimal beamwidth and transmit power allocation scheme obtained in step 4, the global objective value of the joint optimization model under the current iteration is calculated according to the objective function of the sum of average spectral efficiencies defined in step 2.3. Based on the set objective evaluation value, it is determined whether the obtained global objective value satisfies the convergence condition: if it does, the final resource management scheme for user scheduling, beamwidth, and transmit power is output; if it does not, the objective evaluation value, the initial user matching scheme, and the beamwidth and power allocation scheme are updated, and the process returns to step 3 to continue iterative optimization.

[0035] The present invention also provides a radio resource management system for hybrid mobile communication networks based on convex optimization, comprising: Transmission scenario construction module: used to initialize parameters of the mobile communication network system and establish its downlink transmission scenario model; Resource Management Module: This module is used to calculate the equivalent delay-Doppler DD domain channel matrix and the spectral efficiency of high-speed mobile users and low-speed mobile users based on the downlink transmission scenario model. It also constructs a joint optimization model for user scheduling, beamwidth, and transmit power, and decomposes it into a user scheduling sub-model and a beamwidth and transmit power joint optimization sub-model. Model solving module: used to perform convexification and iterative solution on the user scheduling sub-model to obtain the optimal user matching scheme for hybrid mobile users and channels; and based on the optimal user matching scheme, to perform convexification and iterative solution on the beamwidth and transmit power joint optimization sub-model to obtain the optimal beamwidth and transmit power allocation scheme. Performance evaluation module: used to evaluate and provide feedback on transmission efficiency based on the optimal user matching scheme and the optimal beamwidth and transmit power allocation scheme.

[0036] The present invention also provides a radio resource management device for hybrid mobile communication networks based on convex optimization, comprising: Memory: A computer program that stores the above-described method for managing radio resources in a hybrid mobile communication network based on convex optimization, and is a computer-readable device; Processor: Used to implement the aforementioned method for managing radio resources in a hybrid mobile communication network based on convex optimization when executing the computer program.

[0037] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, enables the implementation of the aforementioned method for managing wireless resources in a hybrid mobile communication network based on convex optimization.

[0038] The invention will be explained in detail below with simulation examples: 1. Simulation conditions: In this communication scenario, the carrier frequency is set to 30 GHz, the maximum user rate is 150 m / s, there are 4 propagation paths, the path loss baseline is -30 dB, the exponent is 2.5, the adjustable beamwidth has a lower limit of π / 36 and an upper limit of π / 6, the noise variance is 10^-7, the throughput requirement for low-speed users is 240 Kbps, 4-QAM modulation is used, the OTFS frame size is N=16, M=16, the number of available frequency bands is 4, and the number of high-speed / low-speed users in the NOMA group is 2.

[0039] 2. Simulation content: Figure 2The convergence performance of the user scheduling algorithm and the beamwidth-power joint optimization algorithm under different channel conditions is demonstrated. Simulation results show that the proposed algorithm can converge quickly within a few iterations, demonstrating good real-time performance; at the same time, the objective function value continuously increases with iteration in each case, verifying the stable convergence characteristics of the algorithm in hybrid mobile communication networks.

[0040] Figure 3 The spectral efficiency of the convex optimization scheme and the average allocation scheme under different power conditions was compared. Simulation results show that, under the minimum OTFS frame structure, the proposed scheme improves performance by approximately 25% compared to the average allocation scheme; as the OTFS frame size increases, the performance gain can reach up to 50%, significantly improving system transmission efficiency while ensuring high-reliability communication.

[0041] Figure 4 A comparison of the bit error rate (BER) performance of OTFS and OFDM modulation schemes under different power conditions is presented. The resource management scheme using OTFS modulation in this invention improves the BER performance by 2–3 orders of magnitude over traditional OFDM in the high signal-to-noise ratio range, indicating that it has better communication reliability in high-speed mobile environments.

[0042] Figure 5 Throughput performance was compared between NOMA and Orthogonal Multiple Access (OMA) access methods. Simulation results show that as the throughput constraint for low-speed users is relaxed, the average throughput for high-speed users increases. In the high signal-to-noise ratio range, the convex optimization scheme of NOMA in this invention improves throughput by approximately 30% compared to the OMA scheme, verifying the effectiveness of NOMA technology in improving the transmission performance of hybrid networks.

[0043] Figure 6 The spectral efficiency of three scheduling schemes under different numbers of NOMA groups was analyzed. As limited communication resources are allocated to more NOMA groups, the system spectral efficiency decreases. The convex optimization scheme proposed in this invention achieves a spectral efficiency improvement of 23% and 150% compared to the greedy algorithm and random scheduling scheme, respectively, when the scheduling space increases, indicating that the convex optimization scheme can maintain superior performance in high-dimensional scheduling.

[0044] In summary, the proposed convex optimization-based radio resource management method for hybrid mobile communication networks effectively improves the spectrum efficiency for high-speed users while ensuring the quality of service for low-speed users. Simulation results consistently demonstrate that the constructed joint optimization scheme significantly outperforms traditional allocation schemes in terms of transmission efficiency, reliability, and resource utilization, providing highly reliable and high-speed wireless transmission services for hybrid mobile scenarios.

Claims

1. A method for managing radio resources in a hybrid mobile communication network based on convex optimization, characterized in that, Includes the following steps: Step 1: Initialize the parameters of the mobile communication network system and establish its downlink transmission scenario model; Step 2: Based on the downlink transmission scenario model, calculate the equivalent delay-Doppler DD domain channel matrix and the spectral efficiency of high-speed mobile users and low-speed mobile users, construct a joint optimization model for user scheduling, beamwidth and transmit power, and decompose it into a user scheduling sub-model and a beamwidth and transmit power joint optimization sub-model. Step 3: Perform convexification and iterative solution on the user scheduling sub-model to obtain the optimal user matching scheme for hybrid mobile users and channels; Step 4: Based on the optimal user matching scheme, perform convexification and iterative solution on the joint optimization sub-model of beamwidth and transmit power to obtain the optimal beamwidth and transmit power allocation scheme; Step 5: Based on the optimal user matching scheme and the optimal beamwidth and transmit power allocation scheme, perform transmission efficiency evaluation and feedback.

2. The method for managing radio resources in a hybrid mobile communication network based on convex optimization according to claim 1, characterized in that, The specific process of step 1 includes: Step 1.1: Obtain the system parameters of the mobile communication network, including network configuration parameters: the number of time intervals N and the number of subcarriers M in orthogonal time-frequency space-time (OTFS) modulation, and the symbol duration. Subcarrier spacing Preset sector angle, base station location, pilot transmission time Total working hours User parameters: number of users, initial user location, user speed, multipath delay. With Doppler frequency shift Channel parameters: Number of available frequency bands Q The system parameters are initialized using path loss factor and Nakagami distribution parameters; where the subscript... i Indicates the first i A high-speed mobile user non-orthogonal multiple access (NOMA) group, subscript k Indicates the number of NOMA groups contained in a set. k A high-speed or low-speed mobile user, subscript p Indicates the first p One transmission path; Step 1.2: Based on the system parameters obtained in Step 1.1, implement NOMA access user grouping and resource block allocation, and establish a downlink transmission scenario model for the hybrid mobile communication network. The specific process is as follows: The downlink transmission scenario model is divided into sectors based on the number of users, initial user locations, base station locations, and preset sector angles. The number of NOMA access groups within each sector is determined based on the available frequency bands. Based on user speed, a corresponding number of high-speed mobile users and low-speed mobile users are configured for each NOMA user group according to user distribution. Simultaneously, the base station transmits narrow beams to users within the group for alignment. Resource blocks in the DD domain are allocated to the high-speed mobile user group according to network configuration parameters, and resource blocks in the TF domain are allocated to the low-speed mobile user group. User information is overlaid using inverse sine Fourier transform and power domain NOMA access technology, completing the NOMA access user grouping and resource block division.

3. The method for managing radio resources in a hybrid mobile communication network based on convex optimization according to claim 2, characterized in that, The specific process of step 2 includes: Step 2.1: Calculate the equivalent DD domain channel matrix for each user in the downlink transmission scenario model obtained in Step 1. The specific process is as follows: Calculate the large-scale fading of the channel using the path loss factor obtained in step 1.1, and calculate the small-scale fading of the channel using the Nakagami distribution parameters. Multiply the large-scale fading and the small-scale fading by the total fading coefficient of the channel to obtain the total fading coefficient. subscript q Indicates the first q Available frequency bands; using the multipath delay obtained in step 1.1 With Doppler frequency shift It is then converted into a DD-domain tap index through discretization sampling, including Doppler taps and time-delay taps, wherein the Doppler taps include an integer part. and fractional part , represented as The delay tap is ; Based on the NOMA access user packets and resource block partitioning obtained in step 1.2, the equivalent DD domain channel matrix for each path is constructed using the DD domain tap index through block cyclic matrix operations, and then all... The channel matrix of the i-th path is calculated. i In the NOMA group, the first k The high-speed user in the first q Equivalent DD domain channel matrix on each frequency band , is represented as: ，(1) in, For the fractional Doppler case, the approximate constant is... d This indicates the selection tap acting on the Doppler approximation. Represent a A cyclic matrix is ​​created by cyclically shifting each row of the identity matrix. get, Represents modulo N operation. Indicates the integer part of the Doppler tap. This represents the fractional part of the Doppler tap. Indicates a delayed tap. Indicates the Kronecker product; Step 2.2: Based on the downlink transmission scenario model and the equivalent DD domain channel matrix, construct the user signal reception model and calculate the spectral efficiency expression. The specific process is as follows: Based on the downlink transmission scenario model parameters obtained in step 1, the beam alignment time is calculated as follows: ，(2) in, I Indicates the number of NOMA groups for high-speed mobile users. J This indicates the number of NOMA groups for low-speed mobile users and defines... I= J=Q , K This indicates the number of high-speed or low-speed mobile users included in a NOMA group. For pilot transmission time, To fix the sector angle value, and These are the beamwidth variables related to high-speed mobile users and low-speed mobile users, respectively, where the subscripts... j Indicates the first j For each low-speed mobile user NOMA group, the effective transmission time is then calculated. The effective antenna gain in the main lobe direction for high-speed and low-speed mobile users was calculated respectively. and ,in Indicates sidelobe gain; For high-speed mobile users, zero-forcing equalization detection in the DD domain is employed. The equivalent DD domain channel matrix is ​​diagonalized using the discrete Fourier transform matrix. ,use dimensional discrete Fourier transform matrix The diagonal matrix of the equivalent DD domain channel matrix is ​​obtained. N and M are the dimensions of the OTFS frame, combined with the power of high-speed mobile users. Low-speed mobile user power and Boolean scheduling variables and Receiver antenna gain Noise power and diagonal matrix The One main diagonal element A spectrum efficiency expression for high-speed mobile users is constructed using Shannon's formula. , means as follows: ；(3) For low-speed mobile users, frequency-domain zero-forcing equalization detection combined with continuous interference cancellation technology is adopted, based on the power of the low-speed mobile user. Receiver antenna gain Noise power and the TF domain channel matrix for low-speed mobile users The expression for the spectral efficiency of low-speed mobile users is constructed as follows; ，(4) in, This indicates the number of subcarriers allocated to each user within the NOMA group; Step 2.3: Based on the spectral efficiency expressions for high-speed mobile users and low-speed mobile users, construct a set of user scheduling variables. Beamwidth variable set With the set of transmit power variables The joint optimization model includes: setting the objective function as the sum of the average spectral efficiencies of all high-speed mobile users, expressed as: ；(5) The constraints include: low-speed user throughput requirements, expressed as: ,(6) in, Indicates the subcarrier width. The minimum throughput rate limit is indicated by: The system power limit is indicated by: , Indicates the base station's maximum transmit power; NOMA decoding constraints: Total time constraint: Beamwidth constraints: , Indicates the minimum beamwidth. Indicates the maximum beamwidth; Boolean variable constraints between user groups: and ,in ; Step 2.4: Decompose the joint optimization model constructed in Step 2.3 into two sub-models: a user scheduling sub-model obtained by fixing the beamwidth and transmit power, and a joint optimization sub-model of beamwidth and transmit power obtained by fixing the user scheduling scheme.

4. The method for managing radio resources in a hybrid mobile communication network based on convex optimization according to claim 3, characterized in that, In step 2.4, when fixing the beamwidth and transmit power, the beamwidth is initialized to the arithmetic mean of the adjustable maximum beamwidth and minimum beamwidth of the base station, and the transmit power is initialized to the system's preset average transmit power. After solving the user scheduling sub-model, the optimal user matching scheme output by the user scheduling sub-model is used as the fixed user scheduling variable of the joint optimization sub-model.

5. The method for managing radio resources in a hybrid mobile communication network based on convex optimization according to claim 3, characterized in that, The specific process of convexifying and iteratively solving the user scheduling sub-model includes: Step 3.1: For the user scheduling sub-model in Step 2.4, multiple sets of auxiliary variables are introduced. , , , The non-convex objective function and constraints are equivalently transformed by replacing the non-convex expressions in the signal-to-interference-plus-noise ratio term, interference plus noise term, and noise term with the corresponding auxiliary variable forms, and transforming the bilinear coupling term in the low-speed user throughput constraint into a second-order cone constraint form to obtain the intermediate non-convex model. Step 3.2: For each non-convex term in the intermediate non-convex model obtained in Step 3.1, construct [the model] using the feasible points of the given variables. , real part virtual part Sum of squares The first-order Taylor lower bound is approximately: , , and ,in, This indicates the operation of taking the real part. This represents the operation of taking the imaginary part and continuously relaxing the Boolean user scheduling variables, transforming the intermediate non-convex model obtained in step 3.1 into a solvable convex model, including: the objective function, expressed as: ；(7) The constraints include: convex constraints that simplify the objective: ,(8) ,(9) ；(10) Simplify the convex constraints of throughput requirements: ，(11) ；(12) Boolean variable constraints between user groups, and relaxation of them. ;in, , , and Auxiliary variables , , and A set; Step 3.3: Initialize the iteration parameters, including the iteration number index, convergence threshold, and target judgment value; Step 3.4: Based on the system parameters obtained in Step 1.1, randomly generate user matching schemes, and use the DD domain channel matrix obtained in Step 2.1 to calculate the feasible points required for the Taylor expansion in Step 3.2; Step 3.5: Using the interior point method, call the convex optimization solver to solve the convex model obtained in step 3.2, and obtain the optimal target value and user matching scheme; Step 3.6: Based on the target judgment value, if the optimal target value obtained in Step 3.5 does not meet the convergence threshold condition, then update the target judgment value and the feasible points of Taylor expansion and return to Step 3.5; if the convergence threshold condition is met, then terminate the iteration and output the optimal user matching scheme for hybrid mobile users and channels.

6. The method for managing radio resources in a hybrid mobile communication network based on convex optimization according to claim 5, characterized in that, The specific process of convexifying and iteratively solving the joint optimization sub-model of beamwidth and transmit power includes: Step 4.1: For the joint optimization sub-model of beamwidth and transmit power in Step 2.4, variable substitution and the introduction of multiple sets of auxiliary variables are used. , , , , , By performing an equivalent transformation on the non-convex objective function and constraints, the power variable is rewritten in its reciprocal form. and The nonlinear expressions in the signal-to-interference-plus-noise ratio (SINR) and interference-plus-noise (INR) terms are replaced with the corresponding auxiliary variable forms to obtain the intermediate non-convex model. Step 4.2: For each non-convex term in the intermediate non-convex model obtained in Step 4.1, linear boundary approximations are performed on the power reciprocal term, beamwidth related term, and logarithmic term of auxiliary variables through feasible points of given variables. Specifically, this is achieved by constructing... , First-order Taylor upper bound approximation , and construct respectively , , , , , First-order Taylor lower bound approximation , , , , , The intermediate non-convex model obtained in step 4.1 is transformed into a solvable convex model, including: the objective function, expressed as: ，(13) The constraints include: convex constraints to simplify the positive part of the objective function: ， (14) ；(15) Simplify the convex constraints of the negative part of the objective function: ，(16) ， (17) ；(18) Simplify the convex constraints of throughput requirements: ，(19) ；(20) Total power convex constraint after variable substitution: and NOMA power convex constraint: And total time constraints and beamwidth constraints; where the constant expression , , ;in , , , , and Auxiliary variables , , , , and A set; Step 4.3: Initialize the iteration parameters, including the iteration number index, convergence threshold, and target judgment value; Step 4.4: Based on the system parameters obtained in Step 1.1 and the optimal user matching scheme obtained in Step 3, use the DD domain channel matrix obtained in Step 2.1 to calculate the feasible points required for the Taylor expansion in Step 4.2; Step 4.5: Using the interior point method, call the convex optimization solver to solve the convex model obtained in step 4.2 to obtain the optimal target value and beamwidth and transmit power allocation scheme; Step 4.6: Based on the target judgment value, if the optimal target value obtained in Step 4.5 does not meet the convergence threshold condition, update the target judgment value and the feasible points of Taylor expansion and return to Step 4.5; if the convergence threshold condition is met, terminate the iteration and output the optimal beamwidth and transmit power allocation scheme.

7. The method for managing radio resources in a hybrid mobile communication network based on convex optimization according to claim 6, characterized in that, The specific process of transmission efficiency evaluation and feedback in step 5 is as follows: Based on the optimal user matching scheme obtained in step 3 and the optimal beamwidth and transmit power allocation scheme obtained in step 4, the global objective value of the joint optimization model under the current iteration is calculated according to the objective function of the sum of average spectral efficiencies defined in step 2.

3. Based on the set objective evaluation value, it is determined whether the obtained global objective value satisfies the convergence condition: if it does, the final resource management scheme for user scheduling, beamwidth and transmit power is output. If the target evaluation value, initial user matching scheme, and beam and power allocation scheme are not met, the process is updated, and the process returns to step 3 to continue iterative optimization.

8. A radio resource management system for hybrid mobile communication networks based on convex optimization, used to implement the method of claim 1, characterized in that, include: Transmission scenario construction module: used to initialize parameters of hybrid mobile communication network systems and establish their downlink transmission scenario models; Resource Management Module: This module is used to calculate the equivalent delay-Doppler DD domain channel matrix and the spectral efficiency of high-speed mobile users and low-speed mobile users based on the downlink transmission scenario model. It also constructs a joint optimization model for user scheduling, beamwidth, and transmit power, and decomposes it into a user scheduling sub-model and a beamwidth and transmit power joint optimization sub-model. Model solving module: used to perform convexification and iterative solution on the user scheduling sub-model to obtain the optimal user matching scheme for hybrid mobile users and channels; and based on the optimal user matching scheme, to perform convexification and iterative solution on the beamwidth and transmit power joint optimization sub-model to obtain the optimal beamwidth and transmit power allocation scheme. Performance evaluation module: used to evaluate and provide feedback on transmission efficiency based on the optimal user matching scheme and the optimal beamwidth and transmit power allocation scheme.

9. A wireless resource management device for a hybrid mobile communication network based on convex optimization, characterized in that, include: Memory: A computer program for a method of wireless resource management for a hybrid mobile communication network based on convex optimization as described in any one of claims 1-7, and is a computer-readable device; Processor: Used to implement the wireless resource management method for hybrid mobile communication networks based on convex optimization as described in any one of claims 1-7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the wireless resource management method for a hybrid mobile communication network based on convex optimization as described in any one of claims 1-7.

Images