Multi-time-slot wireless resource allocation method based on multi-cell distributed cooperation

Abstract

The invention discloses a multi-time-slot wireless resource allocation method based on multi-cell distributed cooperation, relates to the technical field of communication, and aims to solve the problem of cache dynamic reduction caused by the actual data service scale of a multi-cell cellular system in multi-time-slot transmission and continuous transmission in each scheduling period. The resource management problem of batch task multi-time-slot downloading under dense cellular network deployment is researched, the defect that an existing scheme only pays attention to a single time slot and an infinite cache hypothesis is overcome, and the characteristic that the amount of data to be transmitted is limited in actual application is considered; an original problem is converted into a series of online sub-problems with the maximum rate to decouple the multi-dimensional time coupling amount, then a multi-base-station parallel distributed algorithm is proposed to decompose a global calculation task to respective base stations, and the designed low-complexity algorithm can effectively reduce the transmission delay of the system. Therefore, the efficiency and the reliability of batch downloading of multi-user services are improved.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a multi-slot wireless resource allocation method based on multi-cell distributed collaboration. Background Art

[0002] With the rapid development of B5G wireless communication systems, the number of mobile devices and communication nodes has increased exponentially, resulting in a sharp increase in the demand for spectrum resources and a higher requirement for the immediacy of information transmission. One feasible solution is to increase data throughput by deploying more base stations. This network expansion method also highlights a core issue: how to efficiently manage and utilize existing wireless resources to stably support the growing demand for instant services, which has become a key technical problem that needs to be solved urgently.

[0003] Due to limited wireless resources, each single-cell cellular system cannot monopolize a specific spectrum, so it is necessary to implement spatial frequency reuse in multi-cell cellular networks. By transmitting synchronously on the same frequency band, the transmitter's transmit power will affect the signal strength of the target receiver and cause co-channel interference to non-target receivers. Specifically, the user's transmission rate depends not only on the wireless resources configured by the current base station, but also on the inter-cell interference from other base stations, which in turn is affected by their resource allocation. Therefore, in order to achieve effective transmission in full spectrum reuse, multi-cell cellular networks must consider efficient resource management and interference coordination strategies, so further analysis is necessary.

[0004] However, existing research on resource allocation in multi-cellular systems is mainly limited to the consideration of a single time slot and is based on the assumption of infinite cache. Its objective function is generally dominated by spectrum efficiency, energy consumption and energy efficiency, and constraints based on the concept of static single time slot are used to ensure service quality. In other words, the data scale of actual services and the dynamic reduction of cache caused by continuous transmission under each scheduling period (SP) are not considered here. This raises a key issue about multi-cell cellular systems, that is, how to minimize transmission delay while fully considering the data scale of business needs. Generally speaking, delay-related research is more common in the field of edge computing, which focuses on uploading and unloading tasks. All of these works are based on the same assumption that the transmission of data services can be completed within a channel coherence time. Due to the uncertain scale of data volume and the competition for resources among access users, business needs often need to be transmitted through multiple SPs, which requires considering changes in the communication environment and dynamically allocating resources. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a multi-slot wireless resource allocation method based on multi-cell distributed collaboration, comprising the following steps:

[0006] S1. Build a system model, including building a multi-cell OFDMA downlink network model, defining wireless resource allocation variables, building a calculation model for transmission delay, and modeling constraints of a multi-cell communication system;

[0007] S2, online conversion of multi-slot discrete optimization problems, including constructing a series of continuous problems based on the current cache constraints and rate maximization and analyzing the continuous conversion of the original problem;

[0008] S3. Perform distributed parallel resource allocation for multi-base station collaboration, including a low-complexity solution for wireless resource allocation in each cell and design of a multi-cell distributed collaboration framework.

[0009] The technical solution further defined in the present invention is:

[0010] Furthermore, in step S1, a multi-cell OFDMA downlink network model is established, which includes a cloud server, M base stations and K cellular users; each user downloads data from the server through the accessed base station, and the total amount of data of the kth user is D k bits, where k = 1, ..., K; the transmission duration of each scheduling period is defined as T 0 , where the wireless channel is a quasi-static flat fading channel, and its coefficient remains stable in each scheduling time slot; the multiple access mode of downlink transmission is OFDMA, and the total frequency bandwidth is NB 0 , where N is the number of carriers, B 0 is the bandwidth of each carrier.

[0011] As described above, in a multi-slot wireless resource allocation method based on multi-cell distributed collaboration, in step S1, a wireless resource allocation variable is defined, and the signal transmission power allocated by base station m to user k on the nth subcarrier in the tth scheduling period is defined as For all user-base-subcarrier association pairs (k, m, n), define the indicator variable If user k uses the nth subcarrier of base station m in the tth scheduling period, then otherwise

[0012] As described above, in a multi-slot wireless resource allocation method based on multi-cell distributed collaboration, in step S1, a calculation model for transmission delay is established, assuming that the transmitted signals of all users in the t-th scheduling period obey independent and identical distribution, and the signal-to-interference-to-noise ratio of the signal received by user k from base station m on the n-th subcarrier is Calculated as:

[0013]

[0014] in, represents the channel coefficient from base station m to user k on the nth subcarrier in the tth scheduling period; is the power spectral density of Gaussian white noise. According to Shannon’s theorem, the total achievable rate of user k in the tth scheduling period is

[0015]

[0016] In order to quantify the transmission delay T of user k k , the following conditions must be met at the same time: from t = 1 to t = T k , the size of the transmitted data in the queue is greater than or equal to the requirement of the kth service, and from t = 1 to t = T k -1, the amount of data transferred is less than D k , then there are the following relational constraints:

[0017]

[0018] So the objective function is expressed as:

[0019]

[0020] Among them, T k represents the transmission delay of user k.

[0021] As described above, in a multi-slot wireless resource allocation method based on multi-cell distributed collaboration, in step S1, the constraints of the multi-cell communication system are modeled, and the maximum transmission power of each base station is recorded as P max , the downlink power meets the following conditions:

[0022]

[0023] Each user can only access one base station at most. Due to the OFDMA mechanism, each subcarrier of each base station can be allocated to one user at most, and each user can be allocated multiple carriers in the same base station. Also need to meet:

[0024]

[0025] if but if but Therefore, the following equality conditions need to be satisfied:

[0026]

[0027] Because each user can only access one base station at most, The following constraints should be followed:

[0028]

[0029] As described above, in a multi-slot wireless resource allocation method based on multi-cell distributed collaboration, in step S2, online conversion of the multi-slot discrete optimization problem is realized, and the amount of data to be transmitted by user k in the tth scheduling period is defined as l k (t), and l k (0) = D k ; Since the transmission delay is discretized into the number of occupied scheduling cycles, the original problem is reconstructed as an online problem of maximizing the sum rate under the current cache constraint, that is, the problem of equation (4) is equivalently transformed into the problem about t=1,...,T max The online problem is as follows:

[0030]

[0031] As described above, in a multi-slot wireless resource allocation method based on multi-cell distributed collaboration, in step S2, the update process of the amount of data to be downloaded along with the scheduling cycle is expressed as:

[0032]

[0033]

[0034] As described above, in a multi-slot wireless resource allocation method based on multi-cell distributed collaboration, in step S2, the rationality of the online conversion of the original problem is analyzed. Assume It is not maximized under the constraint of limited cache, which means that there is at least one user k such that r k (t) does not reach the maximum, resulting in the amount of data that should be downloaded being left until the next scheduling cycle. In this case, by adjusting the carrier allocation variable x(t) and the power allocation variable p(t), r k (t) obtains the maximum value; then, l k (t+1) becomes smaller, making T k Continue to reduce, which is the same as minimizing It contradicts the original intention.

[0035] As described above, in a multi-time slot wireless resource allocation method based on multi-cell distributed collaboration, in step S3, a multi-base station parallel resource allocation method is firstly given, and a resource allocation framework is deployed in each cell; under a given base station selection, each base station completes the solution of its own resource allocation problem in a distributed manner; in the process of designing the method, an auxiliary variable is introduced And rewrite the nonlinear constraint of equation (7) as:

[0036]

[0037] Will Relax to continuous variables and introduce the following equivalent constraints to replace the integer constraints:

[0038]

[0039] By adopting the Lagrange dual method, the above constraints are written into the objective function as penalty terms, and we have:

[0040]

[0041] Among them, η is used as the penalty factor of the objective function to prevent It does not take a binary value; the transmission power of all subcarriers except base station m is defined as Then the interference suffered by user k under base station m on the nth subcarrier is Expressed as By splitting and integrating formula (14), it can be rewritten in the form of differential programming, namely:

[0042]

[0043] in, and They are defined as

[0044]

[0045]

[0046] because About Concave function, using continuous convex approximation to perform first-order Taylor expansion, expressed as:

[0047]

[0048] in, is the solution of the problem at the i-th iteration, About The gradient constant; the limited cache constraint constructs a non-convex mirror image, the left side of the inequality is a strict concave function, and the continuous convex approximation method is used to obtain the convex feasible domain function:

[0049]

[0050] Therefore, by substituting equations (17) and (18) into equation (9), we can obtain the convex approximation form of the original problem at the i+1th iteration. The problem is solved by the interior point method in convex optimization theory, so that the resource allocation scheme is continuously iteratively updated until it converges to a local optimal solution.

[0051] As described above, in a multi-slot wireless resource allocation method based on multi-cell distributed collaboration, in step S3, a collaborative framework for multi-cell distributed transmission is designed. At the beginning of each scheduling cycle, all base stations measure and broadcast corresponding channel state information and initial power allocation plans; then, each base station temporarily fixes the power variable of the interference link and calculates its current resource allocation plan in parallel; the base stations broadcast or send their updated plans through the backhaul link respectively until convergence or the number of iterations is reached; finally, each base station implements data transmission according to the final calculation result.

[0052] The beneficial effects of the present invention are:

[0053] In the present invention, a multiple base station cooperation distributed parallel (MBCDP) resource allocation method based on full spectrum multiplexing is mainly designed. Aiming at the actual data service scale of multi-time slot transmission in multi-cell cellular systems and the dynamic reduction problem of cache caused by continuous transmission in each scheduling cycle, the resource management problem of multi-time slot downloading of batch tasks under dense cellular network deployment is studied, which makes up for the defects of existing solutions that only focus on single time slot and infinite cache assumptions, and takes into account the feature that the amount of data to be transmitted in practical applications is limited; the multi-dimensional time coupling quantity is decoupled by converting the original problem into a series of online sub-problems with maximum sum and rate, and then a multi-base station parallel distributed algorithm is proposed to decompose the global computing tasks to their respective base stations. The designed low-complexity algorithm can effectively reduce the transmission delay of the system, thereby improving the efficiency and reliability of batch downloading of multi-user services. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A schematic diagram of a system model in an embodiment of the present invention;

[0055] Figure 2 It is a schematic diagram of the overall process of the present invention;

[0056] Figure 3 Schematic diagram of the relationship between different numbers of users and total transmission delay in an embodiment of the present invention;

[0057] Figure 4 Schematic diagram of the relationship between different transmission powers and total transmission delay in an embodiment of the present invention;

[0058] Figure 5Schematic diagram of the relationship between the total amount of data to be transmitted and the scheduling time slot in an embodiment of the present invention. DETAILED DESCRIPTION

[0059] This embodiment provides a multi-slot wireless resource allocation method based on multi-cell distributed collaboration, and uses a simulation experiment platform to verify the performance of the proposed algorithm. Figure 1 As shown, cellular users are evenly distributed within the coverage of multiple base stations, and the transmission delay performance of the present invention is evaluated by comparing the performance with similar algorithms in actual application scenarios.

[0060] like Figure 2 As shown, the multi-slot wireless resource allocation method based on multi-cell distributed collaboration provided in this embodiment includes the following steps:

[0061] S1. Construction of system model, including: S1.1. Establishment of network model; S1.2. Definition of wireless resource allocation variables; S1.3. Establishment of transmission delay model; S1.4. Establishment of system constraints.

[0062] S2. Online conversion of multi-slot discrete optimization problems, including: S2.1. Construction of a series of continuous problems of sum rate maximization under current cache constraints; S2.2. Analysis of the continuous conversion of the original problem.

[0063] S3. A multi-base station collaborative distributed parallel resource allocation method, comprising: S3.1. A low-complexity solution for wireless resource allocation in each cell; S3.2. A multi-cell distributed collaborative framework.

[0064] In step S1.1, a multi-cell orthogonal frequency division multiple access (OFDMA) downlink network model is established, which includes a cloud server, M base stations and K cellular users; each user downloads data from the server through the accessed base station, and the total amount of data of the kth user is D k bits, where k = 1, ..., K; the transmission duration of each scheduling period (SP) is defined as T 0 , where the wireless channel is a quasi-static flat fading channel, and its coefficient remains stable in each scheduling time slot; the multiple access mode of downlink transmission is OFDMA, and the total frequency bandwidth is NB 0 , where N is the number of carriers, B 0 The bandwidth of each carrier; each user can select multiple subcarriers for downlink transmission, and users accessing the same base station occupy different subcarriers to avoid interference within the cell.

[0065] In step S1.2, the wireless resource allocation variable is defined, and the signal transmission power allocated by base station m to user k on the nth subcarrier in the tth scheduling period is defined as For all user-base-subcarrier association pairs (k, m, n), define the indicator variable If user k uses the nth subcarrier of base station m in the tth scheduling period, then otherwise

[0066] In step S1.3, a calculation model for transmission delay is established. It is assumed that the transmitted signals of all users in the t-th scheduling period are independent and identically distributed. The signal-to-interference-noise ratio of the signal received by user k from base station m on the n-th subcarrier is Calculated as:

[0067]

[0068] in, represents the channel coefficient from base station m to user k on the nth subcarrier in the tth scheduling period; is the power spectral density of Gaussian white noise. According to Shannon’s theorem, the total achievable rate (bits / s) of user k in the tth scheduling period is

[0069]

[0070] In order to quantify the transmission delay T of user k k , the conditions that need to be met can be summarized as two points: First, from t = 1 to t = T k , the size of the transmitted data in the queue is greater than or equal to the requirement of the kth service; secondly, from t = 1 to t = T k -1, the amount of data transferred is less than D k ; Then there are the following relationship constraints:

[0071]

[0072] So the objective function is expressed as:

[0073]

[0074] Among them, T k represents the transmission delay of user k.

[0075] Step S1.4 models the constraints of the multi-cell communication system, and the maximum transmit power of each base station is denoted as P max , the downlink power meets the following conditions:

[0076]

[0077] Each user can only access one base station at most. Due to the OFDMA mechanism, each subcarrier of each base station can be allocated to one user at most, and each user can be allocated multiple carriers in the same base station. Also need to meet:

[0078]

[0079] if but if but Therefore, the following equality conditions need to be satisfied:

[0080]

[0081] Because each user can only access one base station at most, The following constraints should be followed:

[0082]

[0083] In step S2.1, the online conversion of the multi-slot discrete optimization problem is realized, and the amount of data to be transmitted by user k in the tth scheduling period is defined as l k (t), and l k (0) = D k The amount of data to be downloaded can be explicitly expressed as the update process of the scheduling cycle

[0084]

[0085]

[0086] Since the transmission delay is discretized into the number of occupied scheduling cycles, in order to facilitate the solution, the original problem is reconstructed into an online problem of maximizing the sum rate under the current cache constraint, that is, the problem of equation (23) is equivalently transformed into an online problem about t=1,...,T max The online problem is as follows:

[0087]

[0088] Step S2.2 analyzes the rationality of the online transformation of the original problem, assuming It is not maximized under the constraint of limited cache, which means that there is at least one user k such that r k (t) does not reach the maximum, resulting in the amount of data that should be downloaded being left until the next scheduling cycle. In this case, by adjusting the carrier allocation variable x(t) and the power allocation variable p(t), r k (t) obtains the maximum value; then, l k (t+1) becomes smaller, making T kContinue to reduce, which is the same as minimizing It contradicts the original intention.

[0089] In step S3.1, a multi-base station parallel resource allocation method is first given, and a resource allocation framework is deployed in each cell; under the heuristically given base station selection, each base station completes the solution of its own resource allocation problem in a distributed manner; in the process of designing the method, an auxiliary variable is introduced And rewrite the nonlinear constraint of equation (26) as:

[0090]

[0091] Will Relax to continuous variables and introduce the following equivalent constraints to replace the integer constraints:

[0092]

[0093] By adopting the Lagrange dual method, the above constraints are written into the objective function as penalty terms, and we have:

[0094]

[0095] Among them, η is used as the penalty factor of the objective function to prevent It does not take a binary value; the transmission power of all subcarriers except base station m is defined as Then the interference suffered by user k under base station m on the nth subcarrier is Expressed as By splitting and integrating formula (33), it can be rewritten in the form of differential programming, namely:

[0096]

[0097] in, and They are defined as

[0098]

[0099]

[0100] because About Concave function, using continuous convex approximation to perform first-order Taylor expansion, expressed as:

[0101]

[0102] in, is the solution of the problem at the i-th iteration, About The gradient constant; the limited cache constraint constructs a non-convex mirror image, the left side of the inequality is a strict concave function, and the continuous convex approximation method is used to obtain the convex feasible domain function:

[0103]

[0104] Therefore, by substituting equations (36) and (37) into equation (30), we can obtain the convex approximation form of the original problem at the i+1th iteration; the problem is solved by the interior point method in convex optimization theory, so that the resource allocation plan is continuously iteratively updated until it converges to a local optimal solution.

[0105] Step S3.2 designs a collaborative framework for multi-cell distributed transmission. At the beginning of each scheduling cycle, all base stations measure and broadcast the corresponding channel state information and the initial power allocation plan; then, each base station temporarily fixes the power variable of the interference link and calculates its current resource allocation plan in parallel; the base stations broadcast or send their updated plans through the backhaul link until convergence or the number of iterations is reached; finally, each base station implements data transmission according to the final calculation result.

[0106] In this embodiment, the algorithm performance is verified by simulation experiments. Figure 1 As shown in the figure, assuming that there are K users served by M base stations, the average number of users in each cell is expressed as K / M; the path loss that varies with distance is 120.9+37.6lgd, where d is the distance from the user to the base station in kilometers; shadow fading is represented by a log-normal random variable with a mean of zero and a variance of 8dB; the small-scale fading coefficient obeys a Rayleigh distribution with unit variance; each base station is located at the center of a cell with a radius of 300 meters, and users are evenly distributed in the cell.

[0107] Experiment (A): Compare the performance of similar algorithms under different parameters, increase the average number of users per cell from 2 to 9, and increase the maximum transmit power of each base station from 30dBm to 36dBm.

[0108] Experiment (B): The change in the total amount of data to be transmitted in the system as the number of scheduled time slots increases.

[0109] This example will consider two performance indicators, namely: 1. Comparison of system transmission delay with similar algorithms under different parameters; 2. Comparison of different algorithms in terms of the total amount of data to be transmitted as the number of transmission time slots increases; the experimental parameters are shown in Table 1 below.

[0110] Table 1

[0111]

[0112] The simulation results of this example are as follows:

[0113] 1. Comparison of system transmission delay with similar algorithms under different parameters:

[0114] like Figure 3 As shown, it can be seen that the total transmission delay in each scheme increases with the increase in the number of users; generally, the increase in the number of users in a cell directly leads to more intense resource competition, and some users therefore face higher transmission delays.

[0115] Compared with other benchmark methods, when K / M=6, the Multiple Base-station Cooperation Distributed Parallel (MBCDP) algorithm can achieve a maximum gain of 55.40% in the performance indicator of latency; in an environment with a small number of users, except for the Downlink-side Power Allocation for Rate Maximization (DPARM) algorithm, the other schemes show similar performance levels due to relatively abundant wireless resources and low interference levels; however, with the continuous increase in the number of users, the necessity of resource management algorithms in effectively regulating network interference and ensuring network performance has become increasingly prominent, and this trend has been further verified.

[0116] like Figure 4 As shown in the figure, as the power consumption budget of each base station increases, the total transmission delay of each algorithm decreases accordingly. The reason is that the increase in power gives the base station the opportunity to further improve the transmission rate to the user, thereby effectively shortening the overall service download time. Compared with other schemes, the performance of the algorithm in this embodiment is better in a multi-cell communication environment. This is mainly due to the efficient collaboration between base stations and the precise interference control strategy. These two factors work together to bring significant performance gains to the system.

[0117] Compared with the Joint User Association and PowerAllocation (JUAPA) algorithm and the improved iterated greedy (IIG) algorithm, the total transmission delay of the proposed distributed scheme MBCDP is reduced by 18.89% and 31.12% respectively; in addition, although the comparison method based on infinite cache tends to maximize the transmission rate, users with better channel quality do not necessarily match their remaining data volume, resulting in reduced transmission efficiency, which further demonstrates the effectiveness of the proposed algorithm.

[0118] 2. Comparison of different algorithms on the total amount of data to be transmitted as the number of transmission time slots increases:

[0119] like Figure 5 As shown in the figure, compared with the three benchmark schemes, the MBCDP algorithm significantly reduces the size of the remaining data, and its downward trend is steeper, which proves the excellent performance of the algorithm in maintaining network performance. This is partly due to the consideration of data backlog factors in the design of the proposed scheme, as well as the intelligent resource allocation strategy for users with non-empty caches, thereby achieving improved resource utilization efficiency.

[0120] The allocation method of this embodiment targets the actual data service scale of a multi-cell cellular system in multi-slot transmission and the problem of dynamic cache reduction caused by continuous transmission in each scheduling cycle, and studies the resource management problem of multi-slot download of batch tasks under dense cellular network deployment, which makes up for the defects of existing solutions that only focus on single time slot and infinite cache assumptions, and takes into account the characteristic that the amount of data to be transmitted in practical applications is limited; the multi-dimensional time coupling quantity is decoupled by converting the original problem into a series of online sub-problems with maximum sum and rate, and then a multi-base station parallel distributed algorithm is proposed to decompose the global computing tasks to their respective base stations. The designed low-complexity algorithm can effectively reduce the transmission delay of the system, thereby improving the efficiency and reliability of batch download of multi-user services.

[0121] In addition to the above embodiments, the present invention may also have other implementation modes. Any technical solution formed by equivalent replacement or equivalent transformation falls within the protection scope required by the present invention.

Claims

1. A multi-slot wireless resource allocation method based on multi-cell distributed collaboration, characterized in that: The following steps are involved: S1. Build a system model, including building a multi-cell OFDMA downlink network model, defining wireless resource allocation variables, building a calculation model for transmission delay, and modeling constraints of a multi-cell communication system; S2, online conversion of multi-slot discrete optimization problems, including constructing a series of continuous problems based on the current cache constraints and rate maximization and analyzing the continuous conversion of the original problem; S3. Perform distributed parallel resource allocation for multi-base station collaboration, including a low-complexity solution for wireless resource allocation in each cell and design of a multi-cell distributed collaboration framework.

2. The multi-slot wireless resource allocation method based on multi-cell distributed collaboration according to claim 1, characterized in that: In step S1, a multi-cell OFDMA downlink network model is established, which includes a cloud server, M base stations and K cellular users; each user downloads data from the server through the accessed base station, and the total amount of data of the kth user is D k bits, where k = 1, ..., K; define the transmission duration of each scheduling period as T0, where the wireless channel is a quasi-static flat fading channel, and its coefficient remains stable and unchanged in each scheduling time slot; stipulate that the multiple access mode of downlink transmission is OFDMA, and the total frequency bandwidth is NB0, where N is the number of carriers and B0 is the bandwidth of each carrier.

3. The multi-slot wireless resource allocation method based on multi-cell distributed collaboration according to claim 1, characterized in that: In step S1, a wireless resource allocation variable is defined, and the signal transmission power allocated to user k by base station m on the nth subcarrier in the tth scheduling period is defined as For all user-base-subcarrier association pairs (k, m, n), define the indicator variable If user k uses the nth subcarrier of base station m in the tth scheduling period, then otherwise 4. The multi-slot wireless resource allocation method based on multi-cell distributed collaboration according to claim 1, characterized in that: In step S1, a calculation model for transmission delay is established, assuming that the transmitted signals of all users in the t-th scheduling period are independent and identically distributed, and the signal to interference noise ratio of the signal received by user k from base station m on the n-th subcarrier is Calculated as: in, represents the channel coefficient from base station m to user k on the nth subcarrier in the tth scheduling period; is the power spectral density of Gaussian white noise. According to Shannon’s theorem, the total achievable rate of user k in the tth scheduling period is In order to quantify the transmission delay T of user k k , the following conditions must be met at the same time: from t = 1 to t = T k , the size of the transmitted data in the queue is greater than or equal to the requirement of the kth service, and from t = 1 to t = T k -1, the amount of data transferred is less than D k , then there are the following relational constraints: So the objective function is expressed as: Among them, T k represents the transmission delay of user k.

5. The multi-slot wireless resource allocation method based on multi-cell distributed collaboration according to claim 1, characterized in that: In step S1, the constraints of the multi-cell communication system are modeled, and the maximum transmission power of each base station is recorded as P max , the downlink power meets the following conditions: Each user can only access one base station at most. Due to the OFDMA mechanism, each subcarrier of each base station can be allocated to one user at most, and each user can be allocated multiple carriers in the same base station. Also need to meet: if but if but Therefore, the following equality conditions need to be satisfied: Because each user can only access one base station at most, The following constraints should be followed:

6. The multi-slot wireless resource allocation method based on multi-cell distributed collaboration according to claim 1, characterized in that: In step S2, the online conversion of the multi-slot discrete optimization problem is realized, and the amount of data to be transmitted by user k in the tth scheduling period is defined as l k (t), and l k (0) = D k ; Since the transmission delay is discretized into the number of occupied scheduling cycles, the original problem is reconstructed as an online problem of maximizing the sum rate under the current cache constraint, that is, the problem of equation (4) is equivalently transformed into the problem about t=1,...,T max The online problem is as follows:

7. The multi-slot wireless resource allocation method based on multi-cell distributed collaboration according to claim 6, characterized in that: In step S2, the update process of the amount of data to be downloaded along with the scheduling period is expressed as 8. The multi-slot wireless resource allocation method based on multi-cell distributed collaboration according to claim 1, characterized in that: In step S2, the rationality of online conversion of the original problem is analyzed, assuming It is not maximized under the constraint of limited cache, which means that there is at least one user k such that r k (t) does not reach the maximum, resulting in the amount of data that should be downloaded being left until the next scheduling cycle. In this case, by adjusting the carrier allocation variable x(t) and the power allocation variable p(t), r k (t) obtains the maximum value; then, l k (t+1) becomes smaller, making T k Continue to reduce, which is the same as minimizing It contradicts the original intention.

9. The multi-slot wireless resource allocation method based on multi-cell distributed collaboration according to claim 1, characterized in that: In step S3, a multi-base station parallel resource allocation method is firstly provided, and a resource allocation framework is deployed in each cell; under a given base station selection, each base station completes the solution of its own resource allocation problem in a distributed manner; in the process of designing the method, an auxiliary variable is introduced And rewrite the nonlinear constraint of equation (7) as: Will Relax to continuous variables and introduce the following equivalent constraints to replace the integer constraints: By adopting the Lagrange dual method, the above constraints are written into the objective function as penalty terms, and we have: Among them, η is used as the penalty factor of the objective function to prevent It does not take a binary value; the transmission power of all subcarriers except base station m is defined as Then the interference suffered by user k under base station m on the nth subcarrier is Expressed as By splitting and integrating formula (14), it can be rewritten in the form of differential programming, namely: in, and They are defined as because For about Concave function, using continuous convex approximation to perform first-order Taylor expansion, expressed as: in, is the solution of the problem at the i-th iteration, About The gradient constant; the limited cache constraint constructs a non-convex mirror image, the left side of the inequality is a strict concave function, and the continuous convex approximation method is used to obtain the convex feasible domain function: Therefore, by substituting equations (17) and (18) into equation (9), we can obtain the convex approximation form of the original problem at the i+1th iteration. The problem is solved by the interior point method in convex optimization theory, so that the resource allocation scheme is continuously iteratively updated until it converges to a local optimal solution.

10. The multi-slot wireless resource allocation method based on multi-cell distributed collaboration according to claim 1, characterized in that: In step S3, a collaborative framework for multi-cell distributed transmission is designed. At the beginning of each scheduling cycle, all base stations measure and broadcast corresponding channel state information and initial power allocation schemes. Then, each base station temporarily fixes the power variable of the interfering link and calculates its current resource allocation scheme in parallel. The base stations broadcast or send their updated solutions through the backhaul link until convergence or the number of iterations is reached; finally, each base station implements data transmission based on the final calculation results.

Images