Access network side network optimization method, system and device

Abstract

The present invention belongs to the field of mobile communication technology, and discloses a method, system and device for optimizing the access network side. The method of the present invention comprises: constructing a predefined exclusive codebook for each antenna: selecting a fixed-size subset from an initial optional radio frequency parameter configuration set according to the propagation environment around each antenna and in accordance with a pre-set criterion as the exclusive codebook for the antenna; and determining the radio frequency parameter configuration of each antenna based on the predefined exclusive codebook for the antenna. The present invention can reduce the execution time of the network optimization algorithm, thereby realizing large-scale network optimization.

Description

Technical Field

[0001] The present invention belongs to the technical field of mobile communications, and in particular relates to a method, system and device for optimizing a network at an access network side. Background Art

[0002] Network optimization is a key means to achieve efficient wireless resource management in mobile communication systems. It improves various key performance indicators of the network (such as call drop rate, power coverage, system capacity and other parameters that reflect the network status) by adjusting network parameters. The actual communication network includes access network, transmission network and core network. The number of network elements in the transmission network and the core network is small and the operating environment is stable, so the overall network communication quality is relatively good. The number of network elements in the wireless network is relatively large, the operating environment is relatively complex, and the customer distribution is not concentrated and dynamic, so the network communication quality is unstable. In summary, the optimization of mobile communication networks is actually mainly to optimize the access network, that is, to adjust the radio frequency parameters of the base stations in the network (such as antenna power, switching parameters, antenna direction angle, antenna downtilt angle, antenna radiation pattern and other parameters that can be adjusted by base station equipment). Considering that the tuning of one key performance indicator may lead to the decline of another key performance indicator, network optimization needs to balance between various performance indicators. For example, power coverage tends to maximize the area where a sufficiently strong signal is received, while capacity coverage depends on the signal-to-interference-plus-noise ratio (SINR), which is related not only to the strength of the target signal, but also to the strength of the interference signal. Increasing the transmission power can improve the power coverage of the cell, but such a power increase will also bring more interference to adjacent cells, resulting in a reduction in capacity coverage. Therefore, it is necessary to balance power coverage and capacity coverage through network optimization.

[0003] Generally speaking, network optimization is a combinatorial optimization problem. In some simplified scenarios, it is proven to be NP-hard, that is, any NP problem (non-deterministic polynomial time problem, such problem cannot be determined whether the answer can be found in polynomial time, but the correctness of the answer can be verified in polynomial time) can be reduced to a problem in polynomial time, which means that the problem cannot currently obtain the optimal solution in polynomial time. Therefore, even the solution of this problem at a medium scale is extremely difficult. For example, if the azimuth angle is discretized with a fine granularity of 5 degrees and the downtilt angle is discretized with a fine granularity of 1 degree, there are 13032 optional configuration combinations for a single antenna. Based on this, the combination scale of network configurations for a 10-station scale is approximately 10 120 , which is much higher than the total number of atoms in the currently observable universe, and it is difficult to find the optimal solution through exhaustive enumeration.

[0004] The main algorithms currently available can be divided into three types: random search algorithms, approximate algorithms, and reinforcement learning algorithms. Among them, random search algorithms include Tabu Search, Simulated Annealing, and Genetic Algorithms. These algorithms are simple to implement, but the computation time and solution quality cannot be guaranteed; approximate algorithms can obtain polynomial complexity algorithms with provable performance bounds, but the constraints that can be processed are extremely limited and cannot meet the requirements of actual network planning in most cases; reinforcement learning algorithms optimize RF parameter configuration based on a general Markov decision process, but they have a high demand for computing and storage resources.

[0005] What is particularly serious is that these three types of algorithms are unable to handle large-scale network optimization problems. With the rapid development of emerging applications and smart terminals, mobile communication networks are characterized by dense network deployment and complex device functions. These characteristics have led to a large number of network parameters and a sharp increase in the scale of network optimization problems, far exceeding the ability of existing algorithms to solve them under current computing power.

[0006] At present, the main method to deal with the rapid increase in the scale of network optimization problems is to divide the target area into multiple sub-areas and solve these sub-areas separately. This method can effectively control the scale of the problem, but it has a major flaw: the optimization of sub-areas is decoupled from each other, which makes it impossible to consider the impact of base station facilities in other sub-areas on the sub-area when optimizing a sub-area. Therefore, the final base station RF parameter configuration is often suboptimal.

[0007] The method proposed in the present invention is based on the following scientific fact: the solution space of the original problem contains a large number of redundant antenna RF parameter configurations. On the one hand, similar antenna RF parameter configurations correspond to similar network key performance indicators, which means that a small part of the RF parameter configurations is sufficient to cover all network scenarios, and there is no need to use all similar antenna RF parameter configurations as candidate antenna RF parameter configurations. On the other hand, affected by the propagation environment around the base station, part of the RF parameter configuration of a given antenna will not bring gains to the key performance indicators of the network under any circumstances, and traversing these "useless" RF parameter configurations is a waste of computing resources. In summary, efficient access network side network optimization can be achieved by eliminating these redundant RF parameter configurations.

[0008] Compression coding technology is a common method to eliminate redundancy. It uses fewer bits than the original number of bits for encoding. Traditional compression coding technology usually considers the information source level, and the goal is to make the information loss after compression and decompression steps small enough or even lossless. This evaluation standard is not applicable to the compression of RF parameter configuration space in the field of network optimization. Summary of the invention

[0009] The purpose of the present invention is to provide a network optimization method, system and device for access network side in response to the huge scale challenge of current network optimization problems, which can reduce the execution time of network optimization algorithms and thus efficiently improve the key performance indicators of large-scale networks.

[0010] Specifically, the present invention is implemented by adopting the following technical solutions.

[0011] On the one hand, the present invention provides an access network side network optimization method, comprising:

[0012] Constructing a predefined exclusive codebook for each antenna: According to the propagation environment around each antenna and in accordance with pre-set criteria, a fixed-size subset is selected from the initial optional RF parameter configuration set as the exclusive codebook for the antenna;

[0013] The interference between antennas is coordinated according to the optimization theory, and the radio frequency parameter configuration of each antenna is determined based on the dedicated codebook predefined for the antenna.

[0014] Further, the pre-set criterion is that the RF parameter configuration in the dedicated codebook of the antenna satisfies the maximum sum of the mutual differences of the RF parameter configurations, and the construction of the dedicated codebook predefined for each antenna specifically includes:

[0015] Step 1-1: Initialize the codebook Ψ to an empty set;

[0016] Step 1-2: Put the RF parameter configuration that maximizes the compliance rate into the codebook Ψ, that is:

[0017]

[0018] in,

[0019] i represents the number of the RF parameter configuration put into the codebook Ψ in the initial optional RF parameter configuration set, and the value range is {1, 2, 3, ..., D};

[0020] Indicates the i-th group of RF parameter configuration when the target area is served only by antenna a;

[0021] Indicates that the target area is only served by antenna a when the RF parameters are configured as Key performance indicators of the lower grid u;

[0022] Represents key performance indicators Whether the threshold value T is exceeded or not, the output is a binary variable, 1 means that the standard is met, that is, the threshold value T is exceeded, and 0 means that the standard is not met, that is, the threshold value T is not exceeded;

[0023] Indicates that the target area is only served by antenna a when the RF parameters are configured as The proportion of the lower standard grid;

[0024] Indicates the configuration with the largest number of grids meeting the standard The corresponding number is i;

[0025] ∪ represents the union operation;

[0026] Step 1-3: Add the RF parameter configuration with the largest sum of mutual differences with the RF parameter configuration in the current codebook to fill the codebook, that is:

[0027] Ψ=Ψ∪argmax i∈{1，2，3，...，D} ∑ ψ∈Ψ c i，ψ

[0028] in,

[0029] ψ∈Ψ, indicating that ψ is the RF parameter configuration in the current codebook Ψ;

[0030] i represents the number of the RF parameter configuration put into the current codebook Ψ in the initial optional RF parameter configuration set, and the value range is {1, 2, 3, ..., D};

[0031] c i，ψ , represents the mutual difference between the RF parameter configuration of the i-th group and the RF parameter configuration ψ in the current codebook Ψ;

[0032] argmax i∈{1，2，3，...，D} ∑ ψ∈Ψ c i，ψ , represents the number i corresponding to the RF parameter configuration with the largest sum of differences between the RF parameter configuration of the i-th group and the RF parameter configuration in the current codebook Ψ;

[0033] Step 1-4: Repeat steps 1-3 until the number of RF parameter configurations in the codebook reaches K, and the codebook is used as a predefined exclusive codebook for the antenna.

[0034] Further, the pre-set criterion is that the radio frequency parameter configuration in the dedicated codebook of the antenna satisfies that the sum of the coverage ranges corresponding to each radio frequency parameter configuration is maximum, and the construction of the dedicated codebook predefined for each antenna specifically includes:

[0035] Step 1-1: Initialize the codebook Ψ to an empty set;

[0036] Step 1-2: Put the RF parameter configuration that maximizes the compliance rate into the codebook Ψ, that is:

[0037]

[0038] in,

[0039] i represents the number of the RF parameter configuration put into the current codebook Ψ in the initial optional RF parameter configuration set, and the value range is {1, 2, 3, ..., D};

[0040] Indicates the i-th group of RF parameter configuration when the target area is served only by antenna a;

[0041] Indicates that the target area is served only by antenna a. When the RF parameters are configured as Key performance indicators of time grid u;

[0042] Represents key performance indicators Whether the threshold value T is exceeded or not, the output is a binary variable, 1 means that the standard is met, that is, the threshold value T is exceeded, and 0 means that the standard is not met, that is, the threshold value T is not exceeded;

[0043] Indicates that when the target area is served only by antenna a, when the RF parameters are configured as The proportion of grids that meet the standard;

[0044] Indicates that the RF parameter configuration with the largest number of grids meeting the standard is selected. The corresponding serial number i;

[0045] ∪ represents the union operation;

[0046] Step 1-3: Add the parameter configuration with the largest sum of coverage corresponding to the RF parameter configuration in the current codebook to the codebook Ψ to fill the codebook, that is:

[0047]

[0048] in,

[0049] Indicates the i-th group of RF parameter configuration when the target area is served only by antenna a;

[0050] Indicates that when the target area is served only by antenna a, when the RF parameters are configured as Key performance indicators of time grid u;

[0051] Represents key performance indicators Whether the threshold value T is exceeded or not, the output is a binary variable, 1 means that the standard is met, that is, the threshold value T is exceeded, and 0 means that the standard is not met, that is, the threshold value T is not exceeded;

[0052] 1 a，u，Ψ，T, which means that when the target area is only served by antenna a, when the RF parameter configuration is selected in Ψ, the key performance indicator of grid u exceeds the threshold T. The output is a binary variable, 1 means that the standard is met, that is, the threshold T is exceeded, and 0 means that the standard is not met, that is, the threshold T is not exceeded;

[0053] || represents OR operation;

[0054] Indicates that the target area is served only by antenna a, and the RF parameters are configured as Or when it belongs to Ψ, the proportion of grids that can meet the standard;

[0055] Indicates the RF parameter configuration when the proportion of the standard grid is the largest The corresponding number i;

[0056] ∪ represents the union operation;

[0057] Step 1-4: Repeat steps 1-3 until the number of RF parameter configurations in the codebook reaches K, and the codebook is used as a predefined exclusive codebook for the antenna.

[0058] Further, the coordinating interference between antennas according to the optimization theory and determining the radio frequency parameter configuration of each antenna based on the dedicated codebook predefined for the antenna includes:

[0059] Step 2-1: Define the vector composed of the RF parameters of each antenna in the network as the representation of the problem to be solved, define the fitness function as the achievement rate η(θ) of the key performance indicators of the target area, set the evolutionary generation counter q and the maximum evolutionary generation Q, q∈{0, 1, 2, 3, ..., Q}, and randomly generate M network RF parameter configurations based on the predefined exclusive codebook of each antenna. The set of these network RF parameter configurations is used as the population P(0) of the initial network RF parameter configuration;

[0060] Step 2-2: For the qth population P(q), calculate the fitness of each network RF parameter configuration therein;

[0061] Step 2-3: According to the fitness obtained in step 2-2 and the preset selection operator, select the best individuals from the population P(q) and eliminate the inferior individuals;

[0062] Step 2-4: Apply the crossover operator and mutation operator to the population to generate a population P(q+1) of the offspring network RF parameter configuration, and set q=q+1;

[0063] Step 2-5: Repeat steps 2-2 to 2-4 until q=Q, and use the network radio frequency parameter configuration with the maximum fitness obtained in this process as the optimal solution output and terminate the algorithm.

[0064] Furthermore, the basic steps of applying the crossover operator to the group include:

[0065] Randomly select two network radio frequency parameter configurations under the current group; exchange some radio frequency parameter configurations of the two network radio frequency parameter configurations to form a new radio frequency parameter configuration;

[0066] The basic steps of applying the mutation operator to the population include:

[0067] All individuals in the group are judged whether to mutate based on a pre-set mutation probability; for individuals that are to mutate, mutation sites are randomly selected for mutation.

[0068] Further, the coordinating interference between antennas according to the optimization theory and determining the radio frequency parameter configuration of each antenna based on the dedicated codebook predefined for the antenna includes:

[0069] Step 2-1: Randomly generate the initial network RF parameter configuration x based on the antenna predefined exclusive codebook, and set the taboo table to empty;

[0070] Step 2-2: Generate several neighborhood solutions according to the neighborhood function of the current network radio parameter configuration x, and the set X of these neighborhood solutions is used as the candidate network radio parameter configuration set;

[0071] Step 2-3: Determine whether the candidate network RF parameter configuration meets the contempt criteria. If yes, replace x with the optimal network RF parameter configuration y in the candidate network parameter configuration set X, and add the taboo object corresponding to y to the taboo table, and go to step 2-5; otherwise, go to step 2-4;

[0072] Step 2-4: Select the optimal non-taboo network parameter configuration in set X as the new current network RF parameter configuration, and add its taboo object to the taboo table;

[0073] Step 2-5: Determine whether the algorithm termination condition is met. If so, end the algorithm and output the optimal result. Otherwise, go to step 2-2.

[0074] Further, the neighborhood refers to a set of network RF parameter configurations close to the initial network RF parameter configuration x, that is, if the network RF parameter configuration in X belongs to the neighborhood of x, then the network RF parameter configuration in X is inconsistent with the configuration of x on only one antenna. The RF parameter configuration of the network RF parameter configuration in X in the antenna is selected from the exclusive codebook predefined for the antenna.

[0075] On the other hand, the present invention also provides an access network side network optimization system, including a computing unit, a collection unit and a transmission unit; to implement the access network side network optimization method;

[0076] The dedicated codebook predefined for each antenna is uniformly generated by the operation unit;

[0077] The collecting unit collects information related to all antennas and propagation environments;

[0078] The operation unit coordinates the interference between antennas according to the optimization theory, determines the antenna RF parameter configuration based on the dedicated code book predefined by the antenna, and sends the antenna RF parameter configuration to each antenna through the transmission unit for deployment configuration.

[0079] The present invention also provides an access network side network optimization system, comprising a collection unit, a calculation unit and a transmission unit; to implement the access network side network optimization method;

[0080] The computing unit also includes computing modules distributed on each antenna side;

[0081] The dedicated codebook predefined for each antenna is generated by the computing module on the antenna side;

[0082] The collecting unit collects antenna predefined exclusive codebooks generated by each antenna side operation module, and information related to all antennas and propagation environments;

[0083] The operation unit coordinates the interference between antennas according to the optimization theory, determines the antenna RF parameter configuration based on the dedicated code book predefined by the antenna, and sends the antenna RF parameter configuration to each antenna through the transmission unit for deployment configuration.

[0084] On the other hand, the present invention also provides an access network side network optimization device, the device includes a memory and a processor; the memory stores a computer program for implementing the access network side network optimization method, and the processor executes the computer program to implement the steps of the above method.

[0085] The access network side network optimization method, system and device of the present invention have the following beneficial effects:

[0086] The access network side network optimization method, system and device of the present invention are based on the idea of ​​compression coding, aiming at the huge scale challenge of current network optimization problems, and designing a set of codebook construction schemes suitable for access network side network optimization, and realizing efficient network optimization methods based on codebooks. Specifically, by limiting the optional parameter configuration space of each base station in the mobile communication network, the scale of the network optimization problem is compressed, thereby reducing the execution time of the network optimization algorithm, thereby realizing large-scale network optimization.

[0087] The maximum differentiation (i.e., the difference in coverage corresponding to each candidate antenna configuration in the codebook is the largest) or the maximum total coverage (i.e., the sum of the coverage corresponding to each candidate antenna configuration in the codebook is the largest) is used as the evaluation standard for the codebook quality that is exclusive to the network parameter optimization problem, and a heuristic algorithm is designed based on this standard to efficiently generate its own codebook for each antenna. In the codebook generation stage, each antenna generates its own exclusive codebook based on the propagation environment around it, and the antenna RF parameter configuration contained in the codebook is considered to be expected to improve the key performance indicators of the entire network.

[0088] Through operational optimization methods (such as genetic algorithms, taboo search algorithms, reinforcement learning algorithms, etc.), the interference between antennas is coordinated during the antenna configuration stage, and the configuration of the RF parameters of each antenna is determined based on the pre-defined exclusive codebook of each antenna to improve the key performance indicators of the network (such as improving system throughput, reducing interference, reducing latency, etc.), thereby achieving efficient optimization of the access network side.

[0089] The experimental results show that within a short period of time (300 seconds), the access network side network optimization method, system and device of the present invention, using a codebook-based network optimization algorithm (curve), can achieve performance far exceeding that of the traditional method (fork-shaped curve). BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 It is a schematic diagram of a mobile communication network of the present invention.

[0091] Figure 2 This is a flowchart of the access network side network optimization of this embodiment.

[0092] Figure 3 It is a relationship curve between the key performance indicator (effective coverage) of this embodiment and the algorithm execution time. DETAILED DESCRIPTION

[0093] The present invention is further described in detail below in conjunction with embodiments and with reference to the accompanying drawings.

[0094] The system model of the present invention is as follows:

[0095] The mobile communication network consists of several macro base stations, each of which contains three directional antennas, such as Figure 1 As shown in the figure, the shaded area around each antenna represents the coverage of the antenna. The overlap of the shaded areas indicates that the area is covered by multiple antennas, which is generally considered to cause interference. The adjustable range of the azimuth angle of all antennas is [0°, 360°), and the adjustable range of the downtilt angle is [-90°, 90°]. In order to accurately measure the key performance indicators of the mobile communication network, the target area is divided into square grids, such as 5m×5m square grids.

[0096] Preferably, the signal received at the center of the grid is used to replace the average signal in the corresponding grid. The signals at different positions in the grid may be inconsistent. Using the signal at the center of the grid to approximately replace the signals at each position of the grid is based on the prior knowledge that the signal strength in the grid is similar (within 3dB of difference). This solution can effectively reduce the scale of the problem, that is, when considering network performance, only the center of the grid is concerned and other positions are not concerned.

[0097] Let the set of antennas be A, and the size of the antenna set, i.e. the number of antennas, be |A|. The element a in A represents the antenna, which contains the longitude, latitude and altitude information of the antenna. Let the set of grids be U, and the size of the grid set, i.e. the number of grids, be |U|. The element u in U represents the grid, which contains the longitude, latitude, altitude and topography information of the grid. Define that for any grid u∈U, the key performance indicators (such as received signal strength, reference signal received power, reference signal received quality, etc.) corresponding to the signal received from antenna a∈A are f a，u (θ a ), where θ a Indicates the current network RF parameter configuration of antenna a.

[0098] It is stipulated that each grid can only be served by one antenna within a transmission time interval (in the LTE system, the transmission time interval is usually set to 1 millisecond). Then, within this transmission time interval, the key performance indicator f corresponding to the signal received by grid u from the base station antenna is u (θ) can be written as It indicates that the service received by the grid u comes from the antenna that can provide the maximum key performance indicator. This is a network RF parameter configuration θ = {θ1, θ2, ..., θ |U|}. The network RF parameter configuration refers to the combination of RF parameter configurations of all antennas in the network. The compliance rate η(θ) of the key performance indicators of the target area in the current state can be expressed as

[0099] η(θ)=∑ u∈U 1(f u (θ)>T) / |U| (1)

[0100] Among them, U is the set of grids, |U| is the size of the grid set, that is, the number of grids, u is the grid in U, and f u (θ) is the key performance indicator corresponding to the signal received by grid u from the base station antenna, and T is the set threshold, which is set in advance according to the average service quality of the receiving end user.

[0101] The above formula shows that the key performance indicator f u(θ) The ratio of the number of grids u exceeding the threshold T to the total number of grids. Where 1(x) is an indicator symbol, which takes 1 if x is true and 0 otherwise; the threshold T is usually given in advance based on the service quality feedback from the receiving end user (such as a mobile phone).

[0102] Embodiment 1:

[0103] One embodiment of the present invention is a method, system and device for access network side network optimization.

[0104] The access network side network optimization system of the present invention comprises a collection unit, a calculation unit and a transmission unit.

[0105] The exclusive codebook predefined for each antenna can be uniformly generated by the operation unit. At this time, the operation unit coordinates the interference between the antennas according to the optimization theory, and determines the RF parameter configuration of all antennas based on the exclusive codebook predefined for the antenna to coordinate the interference between the antennas. After determining the network RF parameter configuration scheme, the RF parameter configuration of the antenna is sent to each antenna through the transmission unit for deployment and configuration. The acquisition unit collects information related to the propagation environment of all antennas (such as the terrain and building distribution near the antenna), and this information is used to construct the reciprocity matrix C between the RF parameter configurations in the codebook generation stage.

[0106] The predefined exclusive codebook for each antenna can also be generated by the antenna side. At this time, the operation unit also includes an operation module distributed on each antenna side, and the predefined exclusive codebook for each antenna is generated by the operation module on the antenna side. The acquisition unit acquires the predefined exclusive codebook for each antenna generated by the operation module on each antenna side, as well as information related to the propagation environment of all antennas (such as the terrain and building distribution near the antenna). The operation unit coordinates the interference between antennas according to the optimization theory, and determines the antenna RF parameter configuration based on the predefined exclusive codebook for the antenna, that is, the RF parameter configuration of all antennas, to coordinate the interference between antennas. After determining the network RF parameter configuration scheme, the RF parameter configuration of the antenna is sent to each antenna through the transmission unit for deployment and configuration.

[0107] The access network side network optimization method of the present invention comprises the following steps: Figure 2 shown.

[0108] Step 1: Construct a predefined exclusive codebook for each antenna.

[0109] According to the propagation environment around each antenna, based on the criterion of maximum difference between candidate RF parameter configurations in the codebook, a fixed-size subset is selected from the initial optional RF parameter configuration set, and this subset is used as the exclusive codebook for the antenna. The initial optional RF parameter configuration set contains all feasible parameter configurations for an antenna. The maximum difference between candidate RF parameter configurations in the codebook means that the parameter configuration in the antenna exclusive codebook satisfies the maximum sum of the mutual differences of each parameter, and the antenna RF parameters contained in the codebook are considered to be expected to improve the key performance indicators of the entire network.

[0110] For any antenna, the initial optional RF parameter configuration set when redundant RF parameter configuration is not eliminated is: in represents the i-th group of optional RF parameter configurations for antenna a, and D represents the size of the initial optional RF parameter configuration set.

[0111] For example, if it is assumed that the azimuth angle of the antenna in the base station is discretized in fine granularity of 5° and the downtilt angle is discretized in fine granularity of 1°, then the set of initial azimuth angle and downtilt angle configuration is {(0°, -90°), (0°, -89°), ... (0°, 90°), (5°, -90°), (5°, -89°), ..., (5°, 90°), ..., (355°, -90°), (355°, -89°), ..., (355°, 90°)}. The set contains a total of D = 72 × 181 = 13032 elements. The size of the initial optional RF parameter configuration set is related to the computing and storage capabilities of the computing platform.

[0112] Defines the optional i-th group of RF parameter configurations for antenna a and antenna a optional j-th group of RF parameter configuration The mutual difference between them is c i，j :

[0113]

[0114] Where u is the set of grids in the target area, u is the grid in u, is the i-th group of RF parameter configuration corresponding to the signal received by grid u from antenna a, T is the set threshold, is the XOR symbol, is the jth set of RF parameter configuration corresponding to the signal from antenna a received by grid u, and |U| is the size of the grid set.

[0115] The above formula indicates that when there is only one antenna a in the network, the RF parameter configuration is and The grids with different key performance indicators (such as the key performance indicators of grid u) Under the standard, If the standard is not met, the grid is considered and The proportion of the total grid U (under different compliance conditions) is

[0116] Generally speaking, the mathematical expression of this problem can be written as follows:

[0117]

[0118] ste T x=K (4)

[0119] x={0,1} D (5)

[0120] The first line represents the optimization objective and optimization variable of the problem, where x is the optimization variable, indicating whether the parameters in the initial optional RF parameter configuration set are put into the codebook; C represents the optimization variable of c i，j The matrix is ​​composed of: x is a D-dimensional 0-1 vector, which is used to indicate whether the parameters in the initial optional RF parameter configuration set are placed in the codebook. i Represents from the set The i-th parameter selected in j Represents from the set The jth parameter selected in .

[0121] The second and third rows represent the constraints of the problem, where the second row indicates that the number of parameters in the codebook is a pre-given constant K, which is usually given in advance based on the storage capacity of the base station in the mobile communication system; the third row indicates that x is a D-dimensional 0-1 vector, each element of which represents whether the parameter configuration at the corresponding position in the original optional RF parameter configuration set is placed in the exclusive codebook (1 means placed, 0 means not placed).

[0122] This is a 0-1 quadratic programming problem with linear constraints, and it is currently impossible to find the optimal solution in polynomial time.

[0123] The access network side network optimization method provided by the present invention can find a high-quality feasible solution in a short time, that is, select parameters to be put into the codebook from the initial optional radio frequency parameter configuration set, which is as follows:

[0124] Step 1-1: Initialize the codebook Ψ to (empty set).

[0125] Step 1-2: Put the RF parameter configuration that maximizes the compliance rate into the codebook Ψ, that is:

[0126]

[0127] in,

[0128] i represents the number of the RF parameter configuration put into the codebook Ψ in the initial optional RF parameter configuration set, and the value range is {1, 2, 3, ..., D};

[0129] Indicates the i-th group of RF parameter configuration when there is only antenna a (i.e., only served by antenna a);

[0130] Indicates that the mobile communication network is only served by antenna a when the RF parameters are configured as Key performance indicators of the lower grid u;

[0131] Represents key performance indicators Whether the threshold value T is exceeded or not, the output is a binary variable, 1 means that the standard is met, that is, the threshold value T is exceeded, and 0 means that the standard is not met, that is, the threshold value T is not exceeded;

[0132] Indicates that only antenna a is used for service and the RF parameters are configured as The proportion of the lower standard grid;

[0133] Indicates that the configuration with the largest number of qualified grids is selected The corresponding serial number i;

[0134] ∪ represents the union operation.

[0135] Step 1-3: Add the RF parameter configuration with the largest sum of mutual differences with the parameters in the current codebook to fill the codebook, that is:

[0136] Ψ=Ψ∪argmax i∈{1，2，3，...，D} ∑ ψ∈Ψ c i，ψ (7)

[0137] in,

[0138] ψ∈Ψ, indicating that ψ is the RF parameter configuration in the current codebook Ψ;

[0139] i represents the number corresponding to the parameter (i.e., RF parameter configuration) currently to be put into the current codebook Ψ, and the value range is {1, 2, 3, ..., D};

[0140] c i，ψ , represents the mutual difference between the RF parameter configuration of the i-th group and the RF parameter configuration ψ in the current codebook Ψ;

[0141] argmax i∈{1，2，3，...，D} ∑ ψ∈Ψ c i，ψ, represents the RF parameter configuration with the largest sum of mutual differences between the i-th group of RF parameter configurations and the parameters in the current codebook Ψ.

[0142] Step 1-4: Repeat steps 1-3 until the codebook size reaches a given number of parameters K in the codebook, and the initialized codebook is used as a predefined exclusive codebook for the antenna.

[0143] Step 2: Determine the RF parameter configuration based on the predefined dedicated codebook for each antenna

[0144] The access network side network optimization system of the present invention is based on the exclusive codebook predefined for each antenna, coordinates the interference between antennas according to the optimization theory, and determines the antenna RF parameter configuration based on the exclusive codebook predefined for each antenna to improve various key performance indicators of the network (such as improving system throughput, reducing interference, reducing latency, etc.) and achieve efficient network optimization.

[0145] Here, a genetic algorithm is used as an example to introduce a specific method for determining the RF parameter configuration of an antenna based on a predefined exclusive codebook for the antenna.

[0146] Genetic algorithm is a meta-heuristic method that simulates the evolution process of organisms in nature. It uses mathematical methods and computer simulation operations to convert the problem-solving process into a process similar to the crossover and mutation of chromosome genes in biological evolution.

[0147] Since genetic algorithms cannot directly process the parameters of the problem space, the problem to be solved must be converted into chromosomes or individuals in the genetic algorithm (GA) space through representation. The following three specifications are often used to evaluate the representation strategy:

[0148] Completeness: All points (candidate solutions) in the problem space can be represented as points (chromosomes) in the GA space.

[0149] Soundness: The chromosomes in the GA space can correspond to all candidate solutions in the problem space.

[0150] Non-redundancy: There is a one-to-one correspondence between chromosomes and candidate solutions.

[0151] Fitness in the theory of evolution refers to an individual's ability to adapt to the environment and its ability to reproduce. The fitness function of a genetic algorithm is also called an evaluation function, which is an indicator used to judge the quality of individuals in a group. It is evaluated based on the objective function of the problem being solved. Genetic algorithms generally do not require other external information during the search and evolution process. They only use evaluation functions to evaluate the quality of individuals or solutions and serve as the basis for subsequent genetic operations.

[0152] The specific method for determining the antenna RF parameter configuration based on the dedicated codebook predefined for each antenna is as follows:

[0153] Step 2-1: Define the vector of RF parameters of each antenna in the network as the representation of the problem to be solved, and define the fitness function as the achievement rate η(θ) of the key performance indicators of the target area. Set the evolutionary algebra counter q=0 and the maximum evolutionary algebra Q, and randomly generate M network RF parameter configurations based on the predefined exclusive codebook of each antenna. The set of these network RF parameter configurations is used as the population P(0) of the initial parameter configuration.

[0154] Wherein, I is selected according to the computing and storage capabilities of the computer, and is 100 in this embodiment; M is selected according to the computing and storage capabilities of the computer, and is 200 in this embodiment; the group P(0) includes M network radio frequency parameter configuration sets, which are represented as {p1, p2, p3, ..., p M}.

[0155] Step 2-2: Calculate the fitness of each individual in the group, that is, for the qth group P(q), calculate the fitness of each parameter configuration therein.

[0156] A group refers to a collection of currently generated network RF parameter configurations, and an individual refers to a network RF parameter configuration in the collection.

[0157] In this embodiment, the fitness can be represented by key performance indicators (obtained through simulation software) of the network under the current radio frequency parameter configuration.

[0158] Step 2-3: Eliminate some individuals according to the fitness obtained in step 2-2 and the preset selection operator. That is, according to the fitness function, use the preset selection operator to select the best individuals from the population P(q) and eliminate the inferior individuals.

[0159] There are multiple selection schemes for the selection operator. Preferably, this embodiment adopts a proportional selection operator, the basic idea of ​​which is that the probability of each individual in the group being selected is proportional to its fitness.

[0160] Step 2-4: Apply the crossover operator and mutation operator to the population to generate a population P(q+1) of offspring parameter configurations, and set q=q+1.

[0161] The basic steps of applying the crossover operator to the group include: 1) randomly selecting two network radio frequency parameter configurations under the current group; 2) exchanging some parameters in the two network radio frequency parameter configurations to form a new parameter configuration.

[0162] The basic steps of applying the mutation operator to a group include: 1) judging whether to mutate all individuals in the group with a pre-set mutation probability; 2) randomly selecting mutation sites for the individuals to be mutated.

[0163] There are multiple selection schemes for the crossover operator. Preferably, this embodiment adopts single-point crossover, that is, two individuals are selected, split at randomly selected positions and the second half of the two individuals are exchanged, thereby obtaining two new individuals.

[0164] The mutation operator can also have multiple selection schemes. Preferably, this embodiment adopts basic bit mutation, that is, for the RF parameter configuration of each antenna of an individual, another parameter configuration is selected from its corresponding codebook to replace the current configuration according to the mutation probability. M is a hyperparameter that reflects the stability of the environment and is usually fixed at P M =10 -6 .

[0165] Step 2-5: Repeat steps 2-2 to 2-4 until q = Q, and output the network parameter configuration with the maximum fitness obtained in this process as the optimal solution and terminate the algorithm

[0166] Embodiment 2:

[0167] Another embodiment of the present invention is a method, system and device for access network side network optimization. The difference from the first embodiment is that the step 2 of the present embodiment adopts a traditional taboo search method.

[0168] Step 1: Construct a predefined exclusive codebook for each antenna.

[0169] Step 1-1: Initialize the codebook Ψ to (empty set).

[0170] Step 1-2: Put the RF parameter configuration that maximizes the compliance rate into the codebook Ψ, that is:

[0171]

[0172] in,

[0173] i represents the number of the RF parameter configuration put into the codebook Ψ in the initial optional RF parameter configuration set, and the value range is {1, 2, 3, ..., D};

[0174] Indicates the i-th group of RF parameter configuration when the target area is only served by antenna a.

[0175] Indicates that the target area is only served by antenna a when the RF parameters are configured as The key bear indicators of the lower grid u;

[0176] Represents key performance indicators Whether the threshold value T is exceeded or not, the output is a binary variable, 1 means that the standard is met, that is, the threshold value T is exceeded, and 0 means that the standard is not met, that is, the threshold value T is not exceeded;

[0177] Indicates that the target area is only served by antenna a when the RF parameters are configured as The proportion of the lower standard grid;

[0178] Indicates that the configuration with the largest number of qualified grids is selected The corresponding serial number i;

[0179] ∪ represents the union operation.

[0180] Step 1-3: Add the parameter configuration with the largest sum of mutual differences with the parameters in the current codebook to fill the codebook, that is:

[0181] Ψ=Ψ∪argmax i∈{1，2，3，...，D} ∑ ψ∈Ψ c i，ψ (7)

[0182] in,

[0183] ψ∈Ψ, indicating that ψ is the RF parameter configuration in the current codebook Ψ;

[0184] i represents the number of the RF parameter configuration in the initial optional RF parameter configuration set that is placed in the current codebook Ψ, and the value range is {1, 2, 3, ..., D};

[0185] c i，ψ , represents the mutual difference between the RF parameter configuration of the i-th group and the RF parameter configuration ψ in the current codebook Ψ;

[0186] argmax i∈{1，2，3，...，D} ∑ ψ∈Ψ c i，ψ , represents the parameter configuration with the largest sum of mutual differences between the i-th group of RF parameter configurations and the parameters in the current codebook Ψ.

[0187] Step 1-4: Repeat steps 1-3 until the number of RF parameter configurations in the codebook reaches K, and the codebook is used as a predefined exclusive codebook for the antenna.

[0188] Step 2: Determine the RF parameter configuration based on the predefined dedicated codebook for each antenna

[0189] The access network side network optimization system of the present invention is based on the exclusive codebook predefined for each antenna, coordinates the interference between antennas according to the optimization theory, and determines the antenna RF parameter configuration based on the exclusive codebook predefined for each antenna to improve various key performance indicators of the network (such as improving system throughput, reducing interference, reducing latency, etc.) and achieve efficient network optimization.

[0190] Here, taking the traditional taboo search method as an example, a specific method for determining the network RF parameter configuration based on the antenna predefined exclusive codebook is introduced.

[0191] The specific method for determining the antenna RF parameter configuration based on the dedicated codebook predefined for each antenna is as follows:

[0192] Step 2-1: Randomly generate an initial network RF parameter configuration x based on the antenna predefined exclusive codebook, and set the taboo table to empty.

[0193] Step 2-2: Generate several neighborhood solutions according to the neighborhood function of the current network RF parameter configuration x, and the set X of these neighborhood solutions is used as the candidate network RF parameter configuration set. Neighborhood refers to a set of parameter configurations close to the initial RF parameter configuration x. In this embodiment, a first-order neighborhood is used, that is, if the network RF parameter configuration in X belongs to the neighborhood of x, the network RF parameter configuration in X is inconsistent with the configuration of x only on one antenna, and the RF parameter configuration of the network RF parameter configuration in X at the antenna is selected from the exclusive codebook predefined for the antenna.

[0194] Step 2-3: Determine whether the RF parameter configuration in set X meets the contempt criterion. If so, replace x with the optimal network RF parameter configuration y in X, add the taboo object corresponding to y to the taboo table, and go to step 2-5; otherwise, go to step 2-4.

[0195] Step 2-4: Select the optimal non-taboo network RF parameter configuration in X as the new current network RF parameter configuration, and add its taboo object to the taboo table.

[0196] Step 2-5: Determine whether the algorithm termination condition is met. If so, end the algorithm and output the optimal result. Otherwise, go to step 2-2.

[0197] It can be understood that for step 2 of the present invention, traditional random search algorithms (tabu search, simulated annealing, etc.) and reinforcement learning algorithms (Q learning, Monte Carlo tree search, etc.) can be combined with the idea of ​​codebooks to design base station RF parameter configuration schemes. It is only necessary to change the solution space from the space composed of all feasible antenna configurations to the space composed of codebooks.

[0198] Embodiment 3:

[0199] Another embodiment of the present invention is a method, system and device for optimizing access network side network. The difference from embodiments 1 and 2 is that the criterion for constructing the antenna's exclusive codebook in step 1 of this embodiment adopts maximizing total coverage (i.e., the RF parameter configuration in the antenna's exclusive codebook satisfies the maximum sum of the coverage ranges corresponding to each RF parameter configuration).

[0200] Step 1: Construct a predefined exclusive codebook for each antenna.

[0201] Step 1-1: Initialize the codebook Ψ to (empty set).

[0202] Step 1-2: Put the RF parameter configuration that maximizes the compliance rate into the codebook Ψ, that is:

[0203]

[0204] in,

[0205] i represents the number of the RF parameter configuration put into the codebook Ψ in the initial optional RF parameter configuration set, and the value range is {1, 2, 3, ..., D};

[0206] Indicates the i-th group of RF parameter configuration when the target area is only served by antenna a.

[0207] Indicates that the target area is served only by antenna a. When the RF parameters are configured as Key performance indicators of time grid u;

[0208] Represents key performance indicators Whether the threshold value T is exceeded or not, the output is a binary variable, 1 means that the standard is met, that is, the threshold value T is exceeded, and 0 means that the standard is not met, that is, the threshold value T is not exceeded;

[0209] Indicates that when the target area is served only by antenna a, when the RF parameters are configured as The proportion of grids that meet the standard;

[0210] Indicates that the configuration with the largest number of qualified grids is selected The corresponding serial number i;

[0211] ∪ represents the union operation.

[0212] Step 1-3: Add the RF parameter configuration with the largest sum of coverage corresponding to the elements in the current codebook to the codebook Ψ to fill the codebook, that is:

[0213]

[0214] in,

[0215] i represents the number of the RF parameter configuration put into the codebook Ψ in the initial optional RF parameter configuration set, and the value range is {1, 2, 3, ..., D};

[0216] Indicates the i-th group of RF parameter configuration when the target area is served only by antenna a;

[0217] Indicates that the target area is served by antenna a only, when the RF parameters are configured as Key performance indicators of time grid u;

[0218] Represents key performance indicators Whether the threshold value T is exceeded or not, the output is a binary variable, 1 means that the standard is met, that is, the threshold value T is exceeded, and 0 means that the standard is not met, that is, the threshold value T is not exceeded;

[0219] 1 a，u，Ψ，T , which means that when the target area is only served by antenna a, when the RF parameter configuration is selected in Ψ, the key performance indicator of grid u exceeds the threshold T. The output is a binary variable, 1 means that the standard is met, that is, the threshold T is exceeded, and 0 means that the standard is not met, that is, the threshold T is not exceeded;

[0220] || represents OR operation;

[0221] Indicates that when only antenna a is used in the mobile communication network, the RF parameters are configured as Or when it belongs to Ψ, the proportion of grids that can meet the standard;

[0222] Indicates that the proportion of the standard grid is the maximum when the configuration The corresponding number i;

[0223] ∪ represents the union operation;

[0224] Step 1-4: Repeat steps 1-3 until the number of RF parameter configurations in the codebook reaches K, and the codebook is used as a predefined exclusive codebook for the antenna.

[0225] Step 2: Determine the RF parameter configuration based on the predefined dedicated codebook for each antenna

[0226] Same as step 2 of embodiment 1 or step 2 of embodiment 2.

[0227] The access network side network optimization method, system and device of the present invention are based on the idea of ​​compression coding, aiming at the huge scale challenge of current network optimization problems, and designing a set of codebook construction schemes suitable for access network side network optimization, and realizing efficient network optimization methods based on codebooks. Specifically, by limiting the space of optional radio frequency parameter configurations of each base station in the mobile communication network, the scale of network optimization problems is compressed, thereby reducing the execution time of network optimization algorithms, thereby realizing large-scale network optimization.

[0228] The maximum differentiation (i.e., the difference in coverage corresponding to each candidate antenna RF parameter configuration in the codebook is the largest) or the maximum total coverage (i.e., the sum of the coverage corresponding to each candidate antenna RF parameter configuration in the codebook is the largest) is used as the evaluation standard for the codebook quality that is exclusive to the network parameter optimization problem, and a heuristic algorithm is designed based on this standard to efficiently generate its own codebook for each antenna. In the codebook generation stage, each antenna generates its own exclusive codebook based on the propagation environment around it, and the antenna RF parameter configuration contained in the codebook is considered to be expected to improve the key performance indicators of the entire network.

[0229] Through optimization methods (such as genetic algorithms, taboo search algorithms, reinforcement learning algorithms, etc.), the interference between antennas is coordinated during the antenna configuration stage, and the configuration of the RF parameters of each antenna is determined based on the pre-defined exclusive codebook of each antenna to improve the key performance indicators of the network (such as improving system throughput, reducing interference, reducing latency, etc.), thereby achieving efficient optimization of the access network side.

[0230] The experimental results show that within a short period of time (300 seconds), the access network side network optimization method, system and device of the present invention, using the codebook-based network optimization algorithm (curve), can achieve performance far exceeding that of the traditional method (fork-shaped curve), such as Figure 3 shown.

[0231] In some embodiments, certain aspects of the above-mentioned techniques may be implemented by one or more processors of a processing system executing software. The software includes one or more executable instruction sets stored or otherwise tangibly implemented on a non-transitory computer-readable storage medium. The software may include instructions and certain data that manipulate one or more processors to perform one or more aspects of the above-mentioned techniques when executed by one or more processors. Non-transitory computer-readable storage media may include, for example, magnetic or optical disk storage devices, solid-state storage devices such as flash memory, cache, random access memory (RAM), or other non-volatile memory devices. The executable instructions stored on the non-transitory computer-readable storage medium may be source code, assembly language code, object code, or other instruction formats interpreted or otherwise executed by one or more processors.

[0232] Computer-readable storage media may include any storage medium or combination of storage media that can be accessed by a computer system during use to provide instructions and / or data to the computer system. Such storage media may include, but are not limited to, optical media (e.g., compact disks (CDs), digital versatile disks (DVDs), Blu-ray disks), magnetic media (e.g., floppy disks, tapes, or magnetic hard drives), volatile memory (e.g., random access memory (RAM) or cache), non-volatile memory (e.g., read-only memory (ROM) or flash memory), or micro-electromechanical system (MEMS)-based storage media. Computer-readable storage media may be embedded in a computing system (e.g., system RAM or ROM), fixedly attached to a computing system (e.g., a magnetic hard drive), removably attached to a computing system (e.g., an optical disk or a universal serial bus (USB)-based flash memory), or coupled to a computer system via a wired or wireless network (e.g., network accessible storage (NAS)).

[0233] Please note that not all activities or elements in the general description above are required, that a portion of a particular activity or device may not be required, and that one or more further activities or included elements may be performed in addition to those described. Further, the order in which the activities are listed is not necessarily the order in which they are performed. Moreover, these concepts have been described with reference to specific embodiments. However, those of ordinary skill in the art recognize that various modifications and changes may be made without departing from the scope of the present disclosure as set forth in the following claims. Therefore, the specification and drawings are to be regarded as illustrative rather than restrictive, and all such modifications are included within the scope of the present disclosure.

[0234] Benefits, other advantages, and solutions to problems have been described above with respect to specific embodiments. However, benefits, advantages, solutions to problems, and any features that may cause any benefit, advantage, or solution to occur or become more apparent should not be construed as key, essential, or essential features of any or all claims in any or other respects. In addition, the specific embodiments disclosed above are merely illustrative, as the disclosed subject matter may be modified and implemented in different but equivalent ways that are apparent to those skilled in the art having the benefit of the teachings herein. There is no intention to limit the details of construction or design shown herein, except as described in the claims. It is therefore apparent that the specific embodiments disclosed above may be altered or modified, and all such variations are considered to be within the scope of the disclosed subject matter.

Claims

1. A network optimization method at the access network side, characterized in that: include: Constructing a predefined exclusive codebook for each antenna: According to the propagation environment around each antenna and in accordance with pre-set criteria, a fixed-size subset is selected from the initial optional RF parameter configuration set as the exclusive codebook for the antenna; Coordinate the interference between antennas according to optimization theory, and determine the radio frequency parameter configuration of each antenna based on a predefined exclusive codebook for the antenna; The pre-set criterion is that the RF parameter configuration in the dedicated codebook of the antenna satisfies the maximum sum of the mutual differences of the RF parameter configurations, and the construction of the dedicated codebook predefined for each antenna specifically includes: Step 1-1: Codebook Initialize to an empty set; Step 1-2: In the codebook Put the RF parameter configuration that maximizes the compliance rate into it, that is: ,in, Indicates that the codebook is placed in the initial optional radio frequency parameter configuration set The number of the radio parameter configuration, the value range is ; Indicates that the target area is only covered by the antenna Service Group RF parameter configuration; , indicating that the target area is only covered by the antenna When the service is configured in the radio frequency parameters Lower grid Key performance indicators; , which represents the key performance indicator Whether it exceeds the threshold The output is a binary variable, 1 means reaching the target, that is, exceeding the threshold , 0 means not meeting the standard, that is, not exceeding the threshold ; , indicating that the target area is only covered by the antenna When the service is configured in the radio frequency parameters The proportion of the lower standard grid; , indicating the configuration with the largest number of grids meeting the standard The corresponding number is ; Represents a union operation; Step 1-3: In the codebook Add the RF parameter configuration with the largest sum of differences from the RF parameter configuration in the current codebook to fill the codebook, that is: ， in, ,express is the current codebook RF parameter configuration in; Indicates that the current codebook is placed in the initial optional radio frequency parameter configuration set The number of the radio parameter configuration, the value range is ; , indicating the Group RF parameter configuration and current codebook RF parameter configuration in The mutual differences between , indicating the Group RF parameter configuration and current codebook The number of the RF parameter configuration with the largest sum of RF parameter configuration differences ; Step 1-4: Repeat steps 1-3 until the number of RF parameter configurations in the codebook reaches , then the codebook is used as the exclusive codebook predefined for the antenna; The step of coordinating interference between antennas according to the optimization theory and determining the radio frequency parameter configuration of each antenna based on a dedicated codebook predefined for the antenna includes: Step 2-1: Define the vector composed of the RF parameters of each antenna in the network as the representation of the problem to be solved, and define the fitness function as the achievement rate of the key performance indicators in the target area , set the evolution generation counter and the maximum evolutionary number , , and randomly generate based on the exclusive codebook predefined for each antenna The set of network RF parameter configurations is used as the initial network RF parameter configuration group. ; Step 2-2: For Groups , calculate the fitness of each network RF parameter configuration; Step 2-3: According to the fitness obtained in step 2-2 and the preset selection operator, select Select the best individuals and eliminate the inferior ones; Step 2-4: Apply the crossover operator and mutation operator to the population to generate a population of offspring network RF parameter configurations , and order ; Step 2-5: Repeat steps 2-2 to 2-4 The network RF parameter configuration with the maximum fitness obtained in this process is output as the optimal solution and the algorithm is terminated.

2. A network optimization method at the access network side, characterized in that: include: Constructing a predefined exclusive codebook for each antenna: According to the propagation environment around each antenna and in accordance with pre-set criteria, a fixed-size subset is selected from the initial optional RF parameter configuration set as the exclusive codebook for the antenna; Coordinate the interference between antennas according to optimization theory, and determine the radio frequency parameter configuration of each antenna based on a predefined exclusive codebook for the antenna; The pre-set criterion is that the RF parameter configuration in the dedicated codebook of the antenna satisfies the maximum sum of the mutual differences of the RF parameter configurations, and the construction of the dedicated codebook predefined for each antenna specifically includes: Step 1-1: Codebook Initialize to an empty set; Step 1-2: In the codebook Put the RF parameter configuration that maximizes the compliance rate into it, that is: ,in, Indicates that the codebook is placed in the initial optional radio frequency parameter configuration set The number of the radio parameter configuration, the value range is ; Indicates that the target area is only covered by the antenna Service Group RF parameter configuration; , indicating that the target area is only covered by the antenna When the service is configured in the radio frequency parameters Lower grid Key performance indicators; , which represents the key performance indicator Whether it exceeds the threshold The output is a binary variable, 1 means reaching the target, that is, exceeding the threshold , 0 means not meeting the standard, that is, not exceeding the threshold ; , indicating that the target area is only covered by the antenna When the service is configured in the radio frequency parameters The proportion of the lower standard grid; , indicating the configuration with the largest number of grids meeting the standard The corresponding number is ; Represents a union operation; Step 1-3: In the codebook Add the RF parameter configuration with the largest sum of differences from the RF parameter configuration in the current codebook to fill the codebook, that is: ， in, ,express is the current codebook RF parameter configuration in; Indicates that the current codebook is placed in the initial optional radio frequency parameter configuration set The number of the radio parameter configuration, the value range is ; , indicating the Group RF parameter configuration and current codebook RF parameter configuration in The mutual differences between , indicating the Group RF parameter configuration and current codebook The number of the RF parameter configuration with the largest sum of RF parameter configuration differences ; Step 1-4: Repeat steps 1-3 until the number of RF parameter configurations in the codebook reaches , then the codebook is used as the exclusive codebook predefined for the antenna; The step of coordinating interference between antennas according to the optimization theory and determining the radio frequency parameter configuration of each antenna based on a dedicated codebook predefined for the antenna includes: Step 2-1: Randomly generate the initial network RF parameter configuration x based on the antenna predefined exclusive codebook, and set the taboo table to empty; Step 2-2: Generate several neighborhood solutions according to the neighborhood function of the current network radio parameter configuration x, and the set X of these neighborhood solutions is used as the candidate network radio parameter configuration set; Step 2-3: Determine whether the network RF parameter configuration in set X meets the contempt criterion. If so, replace x with the optimal network RF parameter configuration y in set X, add the taboo object corresponding to y to the taboo table, and go to step 2-5; otherwise, go to step 2-4; Step 2-4: Select the optimal non-taboo network parameter configuration in set X as the new current network RF parameter configuration, and add its taboo object to the taboo table; Step 2-5: Determine whether the algorithm termination condition is met. If so, end the algorithm and output the optimal result. Otherwise, go to step 2-2.

3. A network optimization method at the access network side, characterized in that: include: Constructing a predefined exclusive codebook for each antenna: According to the propagation environment around each antenna and in accordance with pre-set criteria, a fixed-size subset is selected from the initial optional RF parameter configuration set as the exclusive codebook for the antenna; Coordinate the interference between antennas according to optimization theory, and determine the radio frequency parameter configuration of each antenna based on a predefined exclusive codebook for the antenna; The pre-set criterion is that the radio frequency parameter configuration in the dedicated codebook of the antenna satisfies the maximum sum of the coverage ranges corresponding to each radio frequency parameter configuration, and the construction of the dedicated codebook predefined for each antenna specifically includes: Step 1-1: Codebook Initialize to an empty set; Step 1-2: In the codebook Put the RF parameter configuration that maximizes the compliance rate into it, that is: ,in, Indicates that the current codebook is placed in the initial optional radio frequency parameter configuration set The number of the radio parameter configuration, the value range is ; Indicates that the target area is only covered by the antenna Service Group RF parameter configuration; , indicating that the target area is only covered by the antenna When the radio parameters are configured as Time grid Key performance indicators; , which represents the key performance indicator Whether it exceeds the threshold The output is a binary variable, 1 means reaching the target, that is, exceeding the threshold , 0 means not meeting the standard, that is, not exceeding the threshold ; , indicating that only the antenna When the radio parameters are configured as The proportion of grids that meet the standard; , indicating that the RF parameter configuration with the largest number of grids meeting the standard is selected Corresponding serial number ; Represents a union operation; Step 1-3: In the codebook Add the parameter configuration with the largest sum of coverage corresponding to the RF parameter configuration in the current codebook to fill the codebook, that is: ,in, Indicates that the target area is only covered by the antenna Service Group RF parameter configuration; , indicating that only the antenna When the radio parameters are configured as Time grid Key performance indicators; , which represents the key performance indicator Whether it exceeds the threshold The output is a binary variable, 1 means reaching the target, that is, exceeding the threshold , 0 means not meeting the standard, that is, not exceeding the threshold ; , indicating that the target area is only covered by the antenna When serving, Select the RF parameter configuration time grid Whether the key performance indicators exceed the threshold The output is a binary variable, 1 means reaching the target, that is, exceeding the threshold , 0 means not meeting the standard, that is, not exceeding the threshold ; Represents or operation; , indicating that the target area is only covered by the antenna When serving, the radio frequency parameters are configured as or belongs to The proportion of grids that can meet the standard when , indicating the RF parameter configuration when the proportion of the standard grid is the largest Corresponding number ; Represents a union operation; Step 1-4: Repeat steps 1-3 until the number of RF parameter configurations in the codebook is , then the codebook is used as the exclusive codebook predefined for the antenna; The step of coordinating interference between antennas according to the optimization theory and determining the radio frequency parameter configuration of each antenna based on a dedicated codebook predefined for the antenna includes: Step 2-1: Define the vector composed of the RF parameters of each antenna in the network as the representation of the problem to be solved, and define the fitness function as the achievement rate of the key performance indicators in the target area , set the evolution generation counter and the maximum evolutionary number , , and randomly generate based on the exclusive codebook predefined for each antenna The set of network RF parameter configurations is used as the initial network RF parameter configuration group. ; Step 2-2: For Groups , calculate the fitness of each network RF parameter configuration; Step 2-3: According to the fitness obtained in step 2-2 and the preset selection operator, select Select the best individuals and eliminate the inferior ones; Step 2-4: Apply the crossover operator and mutation operator to the population to generate a population of offspring network RF parameter configurations , and order ; Step 2-5: Repeat steps 2-2 to 2-4 The network RF parameter configuration with the maximum fitness obtained in this process is output as the optimal solution and the algorithm is terminated.

4. A network optimization method at the access network side, characterized in that: include: Constructing a predefined exclusive codebook for each antenna: According to the propagation environment around each antenna and in accordance with pre-set criteria, a fixed-size subset is selected from the initial optional RF parameter configuration set as the exclusive codebook for the antenna; Coordinate the interference between antennas according to optimization theory, and determine the radio frequency parameter configuration of each antenna based on a predefined exclusive codebook for the antenna; The pre-set criterion is that the radio frequency parameter configuration in the dedicated codebook of the antenna satisfies the maximum sum of the coverage ranges corresponding to each radio frequency parameter configuration, and the construction of the dedicated codebook predefined for each antenna specifically includes: Step 1-1: Codebook Initialize to an empty set; Step 1-2: In the codebook Put the RF parameter configuration that maximizes the compliance rate into it, that is: ,in, Indicates that the current codebook is placed in the initial optional radio frequency parameter configuration set The number of the radio parameter configuration, the value range is ; Indicates that the target area is only covered by the antenna Service Group RF parameter configuration; , indicating that the target area is only covered by the antenna When the radio parameters are configured as Time grid Key performance indicators; , which represents the key performance indicator Whether it exceeds the threshold The output is a binary variable, 1 means reaching the target, that is, exceeding the threshold , 0 means not meeting the standard, that is, not exceeding the threshold ; , indicating that only the antenna When the radio parameters are configured as The proportion of grids that meet the standard; , indicating that the RF parameter configuration with the largest number of grids meeting the standard is selected Corresponding serial number ; Represents a union operation; Step 1-3: In the codebook Add the parameter configuration with the largest sum of coverage corresponding to the RF parameter configuration in the current codebook to fill the codebook, that is: in, Indicates that the target area is only covered by the antenna Service Group RF parameter configuration; , indicating that only the antenna When the radio parameters are configured as Time grid Key performance indicators; , which represents the key performance indicator Whether it exceeds the threshold The output is a binary variable, 1 means reaching the target, that is, exceeding the threshold , 0 means not meeting the standard, that is, not exceeding the threshold ; , indicating that the target area is only covered by the antenna When serving, Select the RF parameter configuration time grid Whether the key performance indicators exceed the threshold The output is a binary variable, 1 means reaching the target, that is, exceeding the threshold , 0 means not meeting the standard, that is, not exceeding the threshold ; Represents or operation; , indicating that the target area is only covered by the antenna When serving, the radio frequency parameters are configured as or belongs to The proportion of grids that can meet the standard when , indicating the RF parameter configuration when the proportion of the standard grid is the largest Corresponding number ; Represents a union operation; Step 1-4: Repeat steps 1-3 until the number of RF parameter configurations in the codebook is , then the codebook is used as the exclusive codebook predefined for the antenna; The step of coordinating interference between antennas according to the optimization theory and determining the radio frequency parameter configuration of each antenna based on a dedicated codebook predefined for the antenna includes: Step 2-1: Randomly generate the initial network RF parameter configuration x based on the antenna predefined exclusive codebook, and set the taboo table to empty; Step 2-2: Generate several neighborhood solutions according to the neighborhood function of the current network radio parameter configuration x, and the set X of these neighborhood solutions is used as the candidate network radio parameter configuration set; Step 2-3: Determine whether the network RF parameter configuration in set X meets the contempt criterion. If so, replace x with the optimal network RF parameter configuration y in set X, add the taboo object corresponding to y to the taboo table, and go to step 2-5; otherwise, go to step 2-4; Step 2-4: Select the optimal non-taboo network parameter configuration in set X as the new current network RF parameter configuration, and add its taboo object to the taboo table; Step 2-5: Determine whether the algorithm termination condition is met. If so, end the algorithm and output the optimal result. Otherwise, go to step 2-2.

5. The access network side network optimization method according to claim 1 or 3, characterized in that: The basic steps of applying the crossover operator to a population include: Randomly select two network radio frequency parameter configurations in the current group; exchange some radio frequency parameter configurations in the two network radio frequency parameter configurations to form a new radio frequency parameter configuration; The basic steps of applying the mutation operator to the population include: All individuals in the group are judged whether to mutate based on a pre-set mutation probability; for individuals that are to mutate, mutation sites are randomly selected for mutation.

6. The access network side network optimization method according to claim 2 or 4, characterized in that: The neighborhood refers to a set of network RF parameter configurations that are close to the initial network RF parameter configuration x, that is, if the network RF parameter configurations in X belong to the neighborhood of x, then these network RF parameter configurations are inconsistent with the configuration of x on only one antenna; the network RF parameter configuration in X is selected from the RF parameter configuration of the antenna predefined in the exclusive codebook of the antenna.

7. A network optimization system at the access network side, characterized in that: It comprises a computing unit, a collection unit and a transmission unit; and implements the access network side network optimization method as described in any one of claims 1 to 6; The dedicated codebook predefined for each antenna is uniformly generated by the operation unit; The collecting unit collects information related to all antennas and propagation environments; The operation unit coordinates the interference between antennas according to the optimization theory, determines the antenna RF parameter configuration based on the dedicated code book predefined by the antenna, and sends the antenna RF parameter configuration to each antenna through the transmission unit for deployment configuration.

8. A network optimization system at the access network side, characterized in that: It includes a collection unit, a computing unit and a transmission unit; it implements the access network side network optimization method as described in any one of claims 1 to 6; The computing unit also includes computing modules distributed on each antenna side; The dedicated codebook predefined for each antenna is generated by the computing module on the antenna side; The collecting unit collects antenna predefined exclusive codebooks generated by each antenna side operation module, and information related to all antennas and propagation environments; The operation unit coordinates the interference between antennas according to the optimization theory, determines the antenna RF parameter configuration based on the dedicated code book predefined by the antenna, and sends the antenna RF parameter configuration to each antenna through the transmission unit for deployment configuration.

9. A network optimization device at the access network side, characterized in that: The device includes a memory and a processor; the memory stores a computer program for implementing the access network side network optimization method, and the processor executes the computer program to implement the steps of any method according to claims 1-6.

Images