Robust network optimization methods, related apparatuses, and media

Abstract

Embodiments of the present disclosure provide a robust network optimization method, related apparatus and medium. The method comprises: obtaining network sample data of a target network and antenna parameters of each cell in the target network; based on the network sample data and the antenna parameters, performing gradient estimation on a second optimization objective function by a zero-order estimation algorithm to obtain an estimated gradient value; updating the antenna parameters by gradient descent according to the estimated gradient value; calculating a first value of a first optimization objective function and a second value of the second optimization objective function respectively according to the updated antenna parameters; jumping to performing gradient estimation on the second optimization objective function by the zero-order estimation algorithm based on the network sample data and the antenna parameters to obtain an estimated gradient value until a preset condition is met, so as to obtain a plurality of first values; and taking the antenna parameters corresponding to the smallest first value in the plurality of first values as target antenna parameters. The embodiments of the present disclosure can improve the efficiency and robustness of network optimization.

Description

Technical Field

[0001] This disclosure relates to the field of communications, and in particular to a robust network optimization method, related apparatus and medium. Background Technology

[0002] With the rapid development of fifth-generation (5G) wireless cellular communication, network optimization plays a crucial role in improving 5G network performance. 5G networks contain numerous adjustable network parameters, such as antenna parameters, beam parameters, and transmit power. These parameters need to be appropriately adjusted to align with the actual environment, thereby maximizing network performance.

[0003] In related technologies, models of real-world wireless networks are constructed, and then network parameters are simulated and optimized based on these models, such as digital twin networks. However, current network optimization algorithms do not fully utilize the randomness of antenna networks, resulting in poor robustness. How to improve the efficiency and robustness of network optimization is an urgent issue to be discussed. Summary of the Invention

[0004] This disclosure provides a robust network optimization method, related apparatus, and medium, which aim to improve the efficiency and robustness of network optimization.

[0005] In a first aspect, embodiments of this disclosure provide a robust network optimization method, the method comprising:

[0006] Obtain network sample data of the target network and antenna parameters of each cell in the target network, wherein the network sample data includes the spectral efficiency performance of a first number of grids of the target network at a second number of time points;

[0007] Based on the network sample data and the antenna parameters, the gradient of the second optimization objective function is estimated using a zero-order estimation algorithm to obtain the estimated gradient value; wherein, the first optimization objective function is constructed using a first quantile and the network efficiency optimization expectation, the first quantile is used to select the grid to be optimized, and the network efficiency optimization expectation is used to represent the optimization mapping relationship between the spectral efficiency performance of the target network and the antenna parameters;

[0008] The antenna parameters are updated using gradient descent based on the estimated gradient value.

[0009] Based on the updated antenna parameters, the first value of the first optimization objective function and the second value of the second optimization objective function are calculated respectively. The second optimization objective function is constructed by the second quantile and the network efficiency optimization expectation. The second quantile is used to select the grid that needs to be optimized.

[0010] Jump to the step of performing gradient estimation on the second optimization objective function based on the network sample data and the antenna parameters using a zero-order estimation algorithm to obtain estimated gradient values ​​until a preset condition is met, so as to obtain multiple first values;

[0011] The antenna parameter corresponding to the smallest of the multiple first values ​​is taken as the target antenna parameter.

[0012] Secondly, embodiments of this disclosure provide a robust network optimization apparatus, comprising:

[0013] The parameter acquisition unit is used to acquire network sample data of the target network and antenna parameters of each cell in the target network, wherein the network sample data includes the spectral efficiency performance of a first number of grids of the target network at a second number of time points;

[0014] The gradient estimation unit is used to estimate the gradient of the second optimization objective function based on the network sample data and the antenna parameters using a zero-order estimation algorithm to obtain the estimated gradient value; wherein, the first optimization objective function is constructed using a first quantile and a network efficiency optimization expectation, the first quantile is used to select the grid to be optimized, and the network efficiency optimization expectation is used to represent the optimization mapping relationship between the spectral efficiency performance of the target network and the antenna parameters;

[0015] A gradient update unit is used to update the antenna parameters by gradient descent based on the estimated gradient value.

[0016] The optimization calculation unit is used to calculate the first value of the first optimization objective function and the second value of the second optimization objective function according to the updated antenna parameters, wherein the second optimization objective function is constructed by the second quantile and the network efficiency optimization expectation, and the second quantile is used to select the grid that needs to be optimized;

[0017] An iterative jump unit is used to jump to the step of performing gradient estimation on the second optimization objective function based on the network sample data and the antenna parameters, obtaining the estimated gradient value, until a preset condition is met, so as to obtain multiple first values;

[0018] The parameter determination unit is used to take the antenna parameter corresponding to the smallest first value among a plurality of first values ​​as the target antenna parameter.

[0019] Optionally, the robust network optimization device further includes a first quantile determination unit, which is used to:

[0020] Obtain a preset first loss function, wherein the first loss function is a non-smooth function;

[0021] Based on the first loss function, a first quantile estimation function is constructed, wherein the first quantile estimation function is used to output the first optimization variable that minimizes the mathematical expectation corresponding to the first loss function;

[0022] The network sample data is input into the first quantile estimation function to obtain the first optimization variable;

[0023] The first optimization variable is used as the first quantile.

[0024] Optionally, the robust network optimization device further includes a second quantile determination unit, which is used for:

[0025] Obtain a preset second loss function, wherein the second loss function is the upper bound proxy function corresponding to the first loss function;

[0026] Based on the second loss function, a second quantile estimation function is constructed, wherein the second quantile estimation function is used to output a second optimization variable that minimizes the mathematical expectation corresponding to the second loss function;

[0027] The network sample data is input into the first quantile estimation function to obtain the second optimization variable;

[0028] The second optimization variable is used as the second quantile.

[0029] Optionally, the robust network optimization device further includes a network performance calculation unit, wherein the network performance calculation is used for:

[0030] The spectral efficiency performance of the target network is determined based on the average spectral efficiency performance of a first number of grids at a second number of time points.

[0031] Optionally, the first optimization objective function is as follows:

[0032]

[0033] in, Let the first optimization objective function be... Let be the first quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell The set of cells in the target network. Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L For the first number, S For the second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively.

[0034] Optionally, the second optimization objective function is as follows:

[0035]

[0036] in, Let the second optimization objective function be... This is the second quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell The set of cells in the target network. Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L For the first number, S For the second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively.

[0037] Thirdly, embodiments of this disclosure provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the robust network optimization method as described in the first aspect.

[0038] Fourthly, embodiments of this disclosure provide a computer-readable storage medium storing computer-executable instructions, wherein when a computer executes the computer program, it implements the robust network optimization method as described in the first aspect.

[0039] Fifthly, embodiments of this disclosure provide a computer program product, including a computer program or computer instructions, the computer program or computer instructions being stored in a computer-readable storage medium, a processor of a computer device reading the computer program or computer instructions from the computer-readable storage medium, and the processor executing the computer program or computer instructions to cause the computer device to perform the robust network optimization method as described in the first aspect.

[0040] This disclosure collects network sample data of a target network with two dimensions: spatial (grid) and temporal (moment). Based on this dual-dimensional network sample data, an optimization objective function is constructed using quantiles and network efficiency optimization expectations. The network efficiency optimization expectation addresses the statistical characteristics of spectral efficiency performance over time, thereby mitigating temporal fluctuations in network performance. The quantiles address the statistical characteristics of spectral efficiency performance in space, optimizing graticules with lower spectral efficiency performance from a first set of graticules, ensuring that most graticules in space achieve relatively good spectral efficiency performance. Iterative optimization is performed on a second optimization objective function with a second quantile, and then the target antenna parameters are determined based on the results of the second optimization objective function using a first optimization objective function with a first quantile. This ensures that the final optimized target antenna parameters better match the actual situation of the target network, improving optimization efficiency. This robust network optimization method not only mitigates temporal fluctuations but also ensures satisfactory spectral efficiency performance in most geographical graticules, improving the efficiency and robustness of network optimization.

[0041] Other features and advantages of this disclosure will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the disclosure. The objectives and other advantages of this disclosure may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0042] Figure 1 A system architecture diagram for the robust network optimization method applied according to embodiments of this disclosure;

[0043] Figure 2 A schematic diagram illustrating a scenario of 5G network optimization in a certain area, as provided as an example of this disclosure;

[0044] Figure 3 This is a flowchart of a robust network optimization method according to an embodiment of the present disclosure;

[0045] Figure 4 A schematic diagram of network optimization simulation results provided as an example of this disclosure;

[0046] Figure 5 A schematic diagram of network optimization simulation results provided as another example of this disclosure;

[0047] Figure 6 This is a block diagram of a robust network optimization apparatus according to embodiments of the present disclosure;

[0048] Figure 7 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0049] To make the objectives, technical methods, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this disclosure.

[0050] It should be noted that although a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0051] In the description of the embodiments of this disclosure, unless otherwise expressly limited, terms such as "setting," "installing," and "connecting" should be interpreted broadly. Those skilled in the art can reasonably determine the specific meaning of the above terms in the embodiments of this disclosure in conjunction with the specific content of the technical solution. In the embodiments of this disclosure, terms such as "furthermore," "exemplarily," or "optionally" are used to indicate examples, illustrations, or explanations, and should not be construed as being more preferred or more advantageous than other embodiments or design solutions. The use of terms such as "furthermore," "exemplarily," or "optionally" is intended to present the relevant concepts in a specific manner.

[0052] With the rapid development of fifth-generation (5G) wireless cellular communication, network optimization plays a crucial role in improving 5G network performance. 5G networks contain numerous adjustable network parameters, such as antenna parameters, beam parameters, and transmit power. These parameters need to be properly adjusted to align with the real-world environment, thereby maximizing network performance. Since inappropriate network parameters can significantly degrade network performance and lead to severe network outages, network optimization often requires an offline approach. Traditionally, offline optimization of these network parameters relies on expertise and engineering experience, primarily achieved through trial-and-error drive tests. Network experts typically experiment with various network parameter configurations and conduct multiple rounds of drive test measurements during the optimization process; this method is time-consuming, labor-intensive, and fails to achieve optimal network performance. Therefore, network optimization requires automated and intelligent strategies that eliminate the need for human intervention. A promising approach is to construct models of real-world wireless networks and then use these models for simulation and parameter optimization, such as digital twin networks. Inspired by this approach, current offline wireless network optimization methods are based on mathematical models and advanced optimization algorithms.

[0053] Among related technologies, some advanced algorithms already applied in network optimization include the following: Method 1: Adjusting the antenna's downtilt and horizontal angles to optimize coverage by first making the objective function differentiable and then using gradient-based methods for optimization; Method 2: Adjusting the antenna's horizontal angle to optimize coverage by defining a pseudo-metric horizontal angle solution space and then using a metaheuristic evolutionary algorithm to simulate natural evolution to find a suitable solution; Method 3: Adjusting the antenna's downtilt angle to optimize coverage and capacity by using the deep deterministic policy gradient (DPR) reinforcement learning algorithm to optimize the network.

[0054] However, the above methods do not fully utilize the randomness of wireless networks and do not consider robustness optimization, resulting in poor robustness. How to improve the efficiency and robustness of network optimization is an urgent issue to be discussed.

[0055] Based on this, embodiments of this disclosure provide a robust network optimization method, related apparatus, and medium. Embodiments of this disclosure collect network sample data of a target network with two dimensions: spatial (grid) and temporal (moment). Based on the two-dimensional network sample data, an optimization objective function is constructed by setting quantiles and network efficiency optimization expectations. The network efficiency optimization expectations target the statistical characteristics of spectral efficiency performance over time, thereby mitigating temporal fluctuations in network performance. The quantiles target the statistical characteristics of spectral efficiency performance in space, optimizing grids with lower spectral efficiency performance from a first set of grids, thus ensuring that most grids in space achieve relatively good spectral efficiency performance. By iteratively optimizing a second optimization objective function with a second quantile, and then using a first optimization objective function with a first quantile based on the result of the second optimization objective function, the target antenna parameters are determined, making the finally optimized target antenna parameters more consistent with the actual situation of the target network, thus improving optimization efficiency. The robust network optimization method of this disclosure not only mitigates temporal fluctuations but also ensures satisfactory spectral efficiency performance in most geographical grids, improving the efficiency and robustness of network optimization.

[0056] System architecture and scenario description of the embodiments disclosed herein

[0057] Figure 1 This is a system architecture diagram of the robust network optimization method applied according to embodiments of the present disclosure. It includes a server 110, an Internet 120, and an object terminal 130, etc.

[0058] Server 110 refers to a computer system capable of providing network optimization processing services. For example, it can obtain optimized target antenna parameters based on network sample data from the target network and the initial antenna parameters of each cell in the target network, using the robust network optimization method provided in the embodiments of this disclosure, and output these optimized target antenna parameters so that each cell in the target network can adjust its own antenna according to the target antenna parameters. Compared to the target terminal 130, server 110 has higher requirements in terms of stability, security, and performance. Server 110 can be a high-performance computer in a network platform, a cluster of multiple high-performance computers, a portion of a high-performance computer (e.g., a virtual machine), or a combination of portions of multiple high-performance computers (e.g., virtual machines). Server 110 can also communicate with the Internet 120 via wired or wireless means to exchange data.

[0059] The target terminal 130 is a device that displays network optimization results and a sample selection / input interface for network optimization to the target, allowing the target to select a target network and set sampling conditions. It can take various forms, including desktop computers, laptops, PDAs (Personal Digital Assistants), mobile phones, vehicle terminals, home theater terminals, and dedicated terminals. Furthermore, it can be a single device or a collection of multiple devices. For example, multiple devices can be connected via a local area network, sharing a single display device to work collaboratively, forming a single terminal. The terminal can also communicate with the Internet 120 via wired or wireless means to exchange data.

[0060] The embodiments disclosed herein can be applied in various scenarios, such as in... Figure 2 The diagram shows a scenario of 5G network optimization in a certain area.

[0061] For example, such as Figure 2 As shown, the 5G network in this area is taken as the target network. It is assumed that the target network consists of Q cells 210, and the coverage area of ​​the target network is divided into L grids.

[0062] The optimization process for the target network is as follows:

[0063] Step 1: Obtain the spectral efficiency performance of L grids in the target network at S time points as network sample data, i.e., the number of network sample data is U=L at this time. S cells. And obtain the initial antenna parameters of Q cells in the target network.

[0064] Step 2: Based on U network sample data and Q antenna parameters, the gradient of the second optimization objective function is estimated using the zero-order estimation algorithm to obtain the estimated gradient value. The second optimization objective function is constructed using the second quantile and the network efficiency optimization expectation. The second quantile is used to select the grids to be optimized, and the network efficiency optimization expectation is used to represent the optimization mapping relationship between the spectral efficiency performance of the target network and the antenna parameters.

[0065] Step 3: Update the antenna parameters using gradient descent based on the estimated gradient values;

[0066] Step 4: Based on the updated antenna parameters, calculate the first value of the first optimization objective function and the second value of the second optimization objective function, respectively. The first optimization objective function is constructed by the first quantile and the network efficiency optimization expectation. The first quantile is used to select the grids that need to be optimized.

[0067] Step 5: Store the obtained first and second values; the second value will be used in the next gradient descent iteration.

[0068] Step 6: Repeat steps 2 to 5 for a preset number of times T to obtain multiple first values; take the antenna parameter corresponding to the smallest first value among the multiple first values ​​as the target antenna parameter;

[0069] Step 7: Based on the target antenna parameters, adjust the antennas of Q cells in the target network to maximize the 5G network performance in the area.

[0070] It should be noted that steps 1-6 can be performed by server 220. Step 7 can be performed by server 220 sending the target antenna parameters to the corresponding cell so that the cell's base station adjusts its own parameters; or it can be performed by maintenance personnel adjusting the antennas of the base stations of Q cells according to the target antenna parameters.

[0071] In the above example, network sample data of the target network with two dimensions—spatial (grid) and temporal (moment)—is collected. Based on this dual-dimensional network sample data, an optimization objective function is constructed using quantiles and network efficiency optimization expectations. The network efficiency optimization expectation addresses the statistical characteristics of spectral efficiency performance over time, thereby mitigating temporal fluctuations in network performance. The quantiles address the statistical characteristics of spectral efficiency performance in space, optimizing graticules with lower spectral efficiency performance from a first set of graticules, ensuring that most graticules in space achieve relatively good spectral efficiency performance. Iterative optimization is performed on a second optimization objective function with a second quantile, and then the target antenna parameters are determined based on the results of the second optimization objective function using a first optimization objective function with a first quantile. This ensures that the final optimized target antenna parameters better match the actual situation of the target network, improving optimization efficiency. The robust network optimization method disclosed in this paper not only mitigates temporal fluctuations but also ensures satisfactory spectral efficiency performance in most geographic graticules, improving the efficiency and robustness of network optimization.

[0072] It should be understood that the above description only illustrates some application scenarios of this disclosure. The business scenarios to which this disclosure can be applied may include, but are not limited to, the specific embodiments described above.

[0073] General Description of Embodiments in this Disclosure

[0074] It should be emphasized that the embodiments of this disclosure are applicable to various application scenarios, such as 5G network optimization and regional network optimization. In related technologies, current network optimization algorithms do not fully utilize the randomness of antenna networks, resulting in poor robustness. Some embodiments of this disclosure provide a robust network optimization method, related apparatus, and medium, aiming to improve the efficiency and robustness of network optimization.

[0075] The robust network optimization method of this disclosure can be executed on a server, on a terminal, or partially on a server and partially on a terminal.

[0076] like Figure 3 As shown, according to one embodiment of this disclosure, the robust network optimization method includes, but is not limited to:

[0077] Step 310: Obtain network sample data of the target network and antenna parameters of each cell in the target network, wherein the network sample data includes the spectral efficiency performance of the first number of grids of the target network at the second number of time points;

[0078] Step 320: Based on network sample data and antenna parameters, perform gradient estimation on the second optimization objective function using a zero-order estimation algorithm to obtain the estimated gradient value; wherein, the second optimization objective function is constructed using the second quantile and the network efficiency optimization expectation, the second quantile is used to select the grid to be optimized, and the network efficiency optimization expectation is used to represent the optimization mapping relationship between the spectral efficiency performance of the target network and the antenna parameters;

[0079] Step 330: Update the antenna parameters using gradient descent based on the estimated gradient values;

[0080] Step 340: Based on the updated antenna parameters, calculate the first value of the first optimization objective function and the second value of the second optimization objective function, respectively. The first optimization objective function is constructed by the first quantile and the network efficiency optimization expectation. The first quantile is used to select the grids that need to be optimized.

[0081] Step 350: Jump to the second optimization objective function based on network sample data and antenna parameters, and perform gradient estimation on the second optimization objective function using the zero-order estimation algorithm to obtain the estimated gradient value (step 320) until the preset conditions are met to obtain multiple first values;

[0082] Step 360: Take the antenna parameter corresponding to the smallest first value among multiple first values ​​as the target antenna parameter.

[0083] Steps 310 to 360 are briefly described below.

[0084] In step 310, the target network refers to the network covered by the area requiring network optimization. A grid refers to multiple geographical grids obtained by geographically dividing the area requiring network optimization. A cell refers to each base station in the target network; in this disclosure, a cell can be understood as a base station within that cell. Antenna parameters include the antenna's downtilt angle and horizontal angle. Spectral efficiency (SE) is one of the important indicators for measuring the performance of a wireless communication system. It is defined as the transmission rate per unit bandwidth, that is, how many bits of data can be transmitted per hertz (Hz) of bandwidth.

[0085] For example, suppose we use To represent spectral efficiency performance, use This refers to the antenna parameters. In this disclosure, the task of network optimization is to adjust the antenna parameters. To optimize the spectral efficiency performance of the target network. Antenna parameters of the target network. It can be represented as:

[0086] (1)

[0087] in, and The first Antenna downtilt angle and horizontal angle for each cell For the set of cells in the target network, ∈ .

[0088] In step 320, the zero-order estimation algorithm is an optimization method that does not require gradient information and is suitable for black-box optimization scenarios, i.e., situations where the objective function is a black box and its gradient information cannot be obtained. In this embodiment, the zero-order estimation algorithm is used to obtain estimated gradient values ​​through function evaluation, thereby updating the antenna parameters.

[0089] The second quantile refers to a parameter relating to the spatial dimension of the target network's grids, i.e., the network sample data. By setting the second quantile, optimization is performed on the grids with lower spectral efficiency performance from the first few grids of the target network, rather than directly optimizing all grids together. This results in better and more uniform spectral efficiency performance for the overall target network, improving overall optimization efficiency. It should be noted that the specific definition of the second optimization objective function is illustrated in the following embodiments and will not be elaborated upon here.

[0090] For example, the desired network efficiency optimization can be expressed by the following formula:

[0091] (2)

[0092] in, Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L As the first number, S The second number; to avoid antenna collisions and over-coverage, set and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively.

[0093] Formula (2) is mainly optimized by maximizing the mean of the samples, where, Defined as finding the mean, specifically:

[0094] (3)

[0095] in, U This represents the total number of network sample data. U=L S Formula (3) is for U The average of the spectral efficiency performance is calculated.

[0096] However, even with high average spectral efficiency, some geographic grids at cell edges may exhibit poor spectral efficiency. To account for the poor spectral efficiency of these grids, quantiles are used to incorporate spatial information (the impact of grid geographic location on spectral efficiency) into the optimization. Quantiles provide more information about spectral efficiency performance than the network efficiency optimization expectation. Information about the distribution, especially characterizing random variables (spectral efficiency performance) The tail behavior of the distribution. Quantiles can be expressed as random variables (spectral efficiency performance) It is the inverse function of the cumulative probability density function.

[0097] For spectral efficiency performance Its cumulative probability density function can be expressed as:

[0098] (4)

[0099] in, For spectral efficiency performance The cumulative probability density function, This indicates an operation that takes a probability. For spectral efficiency performance The sampled value at a specific moment.

[0100] Without loss of generality, combining the cumulative probability density function of formula (4), for any 0 < <1, -Quantiles can be represented as:

[0101] (5)

[0102] For example, when When =0.1, - Quantiles represent a value where 10% of all network sample data are less than a certain threshold. - The quantile value is greater than 90% of the samples. - The value of the quantile.

[0103] In step 330, the estimated gradient value refers to the gradient value obtained by performing gradient estimation on the second optimization objective function using a zero-order estimation algorithm.

[0104] In step 340, the first value refers to the objective function value calculated using the first optimization objective function based on the updated antenna parameters. The second value refers to the objective function value calculated using the second optimization objective function based on the updated antenna parameters.

[0105] The difference between the second and first optimization objective functions lies in their quantiles, which are the first quantile and the second quantile, respectively. The definitions and explanations of quantiles in the examples above can be referenced for both the first and second quantiles. Their specific differences will be described in detail in subsequent embodiments and will not be repeated here.

[0106] In step 350, the preset condition refers to the condition that causes the antenna parameter optimization iteration to converge. For example, the preset condition could be that the loop terminates after being executed T times, where T is a positive integer set according to requirements. The preset condition could also be that the loop terminates when the first value of the first optimization objective function is less than a preset threshold.

[0107] In step 360, based on the settings of the first / second optimization objective function of this disclosure, the larger the value of the optimization objective function, the greater the spectral efficiency performance of the target network. Therefore, the smallest first value is the value with the largest spectral efficiency performance in multiple iterations. Thus, the antenna parameters corresponding to the smallest first value are taken as the optimal solution for the current optimization and used as the target antenna parameters. After obtaining the target antenna parameters, the antennas of each cell in the target network can be set according to the target antenna parameters to achieve the optimal spectral efficiency performance of the target network.

[0108] In the embodiments of steps 310 to 360 above, network sample data of the target network with two dimensions, spatial (grid) and temporal (moment), are collected. Based on the two-dimensional network sample data, an optimization objective function is constructed using quantiles and network efficiency optimization expectations. The network efficiency optimization expectations target the statistical characteristics of spectral efficiency performance over time, thereby mitigating temporal fluctuations in network performance. The quantiles target the statistical characteristics of spectral efficiency performance in space, optimizing grids with lower spectral efficiency performance from a first set of grids, thus ensuring that most grids in space achieve relatively good spectral efficiency performance. By iteratively optimizing the second optimization objective function with a second quantile, and then using the first optimization objective function with a first quantile based on the result of the second optimization objective function, the target antenna parameters are determined, making the finally optimized target antenna parameters more consistent with the actual situation of the target network and improving optimization efficiency. The network optimization method disclosed herein not only mitigates temporal fluctuations but also ensures satisfactory spectral efficiency performance in most geographical grids, improving the efficiency and robustness of network optimization.

[0109] The above is a general description of steps 310 to 360. The specific implementation process of steps 310 to 360 will be described in detail below through various embodiments.

[0110] In one embodiment, the robust network optimization method further includes:

[0111] The spectral efficiency performance of the target network is determined based on the average spectral efficiency performance of the first number of grids of the target network at the second number of time points. For example, the spectral efficiency performance of the target network can be obtained using formula (3) provided in the example above.

[0112] In one embodiment, the first quantile is obtained through the following steps:

[0113] Obtain a preset first loss function, wherein the first loss function is a non-smooth function;

[0114] Based on the first loss function, a first quantile estimation function is constructed, wherein the first quantile estimation function is used to output the first optimization variable that minimizes the mathematical expectation corresponding to the first loss function;

[0115] Input the network sample data into the first quantile estimation function to obtain the first optimization variable;

[0116] The first optimization variable is used as the first quantile.

[0117] In this embodiment, due to the highly complex communication environment of a real 5G network, the cumulative density function for spectral efficiency performance in formula (4) cannot be explicitly obtained. Therefore, the value of the first quantile cannot be directly calculated using formula (5). To calculate a value of the first quantile that better reflects the actual situation, a first loss function is first defined. Specifically:

[0118] (6)

[0119] Among them, the first loss function It is a non-smooth function. c is the independent variable, i.e., the input of formula (6).

[0120] Furthermore, according to the first loss function Constructing the first quantile estimation function To calculate the value of the first quantile, the specific steps are as follows:

[0121] (7)

[0122] in, To optimize variables.

[0123] It can be proven that, by modifying formula (7) Taking the derivative and setting the partial derivative to zero, we can obtain... If the quantile is one of the solutions, then we can deduce that:

[0124] (8)

[0125] Finally, formula (8) is equivalent to:

[0126] (9)

[0127] Based on formulas (8) and (9), it can be determined that formula (7) can calculate the quantiles defined by formula (5), that is, it can calculate the optimization variables obtained by formula (7). As the first quantile.

[0128] After obtaining the first quantile estimation function, i.e., formula (7), the network sample data is... Substituting into formula (7), where, u∈U , U Given the number of network sample data, we can calculate the estimated value of the first quantile corresponding to the target network to be optimized, as shown below:

[0129] (10)

[0130] In one embodiment, based on the operation of calculating the mean of network sample data in formula (3) and the operation of calculating the first quantile in formula (10) provided in the above embodiment, a two-layer first optimization objective function based on expectation and quantile is designed, as follows:

[0131] (11)

[0132] in, The first optimization objective function is... Let be the first quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell For the set of cells in the target network, Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L As the first number, S The second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively.

[0133] In Equation (11), the inner layer calculates the network efficiency optimization expectation, and the outer layer calculates the first quantile. The inner layer expectation mainly focuses on the statistical characteristics of spectral efficiency over time, while the outer layer quantile mainly focuses on the statistical characteristics of spectral efficiency over space. This robust design not only mitigates temporal fluctuations but also ensures satisfactory spectral efficiency performance in most grids.

[0134] In one embodiment, the second quantile is obtained through the following steps:

[0135] Obtain a preset second loss function, wherein the second loss function is the upper bound proxy function corresponding to the first loss function;

[0136] Based on the second loss function, a second quantile estimation function is constructed, whereby the second quantile estimation function is used to output the second optimization variable that minimizes the mathematical expectation corresponding to the second loss function;

[0137] The network sample data is input into the first quantile estimation function to obtain the second optimization variable;

[0138] The second optimization variable is used as the second quantile.

[0139] In this embodiment, since the quantiles in formula (10) are not smooth, this will affect the network optimization performance. Therefore, in this embodiment, a smooth second loss function is set to obtain a smooth second quantile estimation function for calculating the quantiles, thereby enabling the first optimization objective function to be rewritten as a smooth second optimization objective function.

[0140] First, based on the first loss function (Equation (6)), an upper bound surrogate function, namely the second loss function, is set for Equation (6) to make Equation (6) smooth. Specifically, the second loss function... for:

[0141] (12)

[0142] in, It is a non-negative smoothness factor. The larger the value, the greater the smoothness introduced. The upper bound surrogate function defined by formula (12) is greater than or equal to the first loss function defined by formula (6). When As it approaches zero, the upper bound surrogate function defined by formula (12) approaches the first loss function defined by formula (6). Furthermore, it can be defined as follows: - A smoothed surrogate function for quantiles, i.e., a smoothed second quantile estimation function. :

[0143] (13)

[0144] in, For network sample data, To optimize variables. From formulas (10) and (13), it can be seen that formula (13) can improve the smoothness of network optimization, thereby improving optimization performance.

[0145] In one embodiment, the operation of calculating the mean of network sample data in formula (3) and the operation of calculating the second quantile after smoothing in formula (13) provided in the above embodiments can be used to change the first optimization objective function to the second optimization objective function as shown below:

[0146] (14)

[0147] in, For the second optimization objective function, This is the second quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell For the set of cells in the target network, Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L As the first number, S The second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively.

[0148] In one embodiment, in order to solve formula (14), since the relationship between antenna parameters and spectral efficiency performance is a black box mapping, a zero-order estimation algorithm is used to estimate the gradient, as shown in the following formula (15):

[0149] (15)

[0150] in, It is an estimate of the radius, and .

[0151] Based on formula (15), the antenna parameters are updated using gradient descent, as shown in formula (16) below:

[0152] (16)

[0153] in, It is the set of feasible regions that satisfy the constraints, namely that the antenna's downtilt angle and horizontal angle need to be within the corresponding adjustment threshold range.

[0154] For example, such as Figure 4 and Figure 5As shown, four cases were compared: before optimization, expectation optimization, joint expectation and quantile optimization (without smoothing), and joint expectation and quantile optimization (after smoothing). First, it can be seen that compared to before optimization, the three optimized cases—expectation optimization, joint expectation and quantile optimization (without smoothing), and joint expectation and quantile optimization (after smoothing)—significantly reduce the difference in spectral efficiency performance of the geographic raster. The robust network optimization method based on a two-layer expectation and quantile design in this disclosure has a curve below all curves after smoothing, indicating that the optimization performance of the robust network optimization method based on a two-layer expectation and quantile design in this disclosure is the best after smoothing. Furthermore, =0.1 and Both parameter settings of 0.3 and 0.3 can meet the robust optimization requirements of 5G networks in practice.

[0155] Description of apparatus and devices according to embodiments of this disclosure

[0156] It is understood that although the steps in the above flowcharts are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated in this embodiment, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the above flowcharts may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps.

[0157] Figure 6 This disclosure provides a schematic diagram of the robust network optimization device 600. The robust network optimization device 600 includes:

[0158] The parameter acquisition unit 610 is used to acquire network sample data of the target network and antenna parameters of each cell in the target network. The network sample data includes the spectral efficiency performance of a first number of grids of the target network at a second number of time points.

[0159] The gradient estimation unit 620 is used to estimate the gradient of the second optimization objective function based on network sample data and antenna parameters using a zero-order estimation algorithm to obtain the estimated gradient value. The first optimization objective function is constructed using a first quantile and the network efficiency optimization expectation. The first quantile is used to select the grid to be optimized, and the network efficiency optimization expectation is used to represent the optimization mapping relationship between the spectral efficiency performance of the target network and the antenna parameters.

[0160] The gradient update unit 630 is used to update the antenna parameters by gradient descent based on the estimated gradient value.

[0161] The optimization calculation unit 640 is used to calculate the first value of the first optimization objective function and the second value of the second optimization objective function according to the updated antenna parameters. The second optimization objective function is constructed by the second quantile and the network efficiency optimization expectation. The second quantile is used to select the grid that needs to be optimized.

[0162] The iterative jump unit 650 is used to jump to the second optimization objective function based on network sample data and antenna parameters, and to estimate the gradient value by using a zero-order estimation algorithm until a preset condition is met, so as to obtain multiple first values.

[0163] The parameter determination unit 660 is used to take the antenna parameter corresponding to the smallest first value among multiple first values ​​as the target antenna parameter.

[0164] Optionally, the robust network optimization device 600 further includes a first quantile determination unit (not shown in the figure), which is used for:

[0165] Obtain a preset first loss function, wherein the first loss function is a non-smooth function;

[0166] Based on the first loss function, a first quantile estimation function is constructed, wherein the first quantile estimation function is used to output the first optimization variable that minimizes the mathematical expectation corresponding to the first loss function;

[0167] Input the network sample data into the first quantile estimation function to obtain the first optimization variable;

[0168] The first optimization variable is used as the first quantile.

[0169] Optionally, the robust network optimization device 600 further includes a second quantile determination unit (not shown in the figure), which is used for:

[0170] Obtain a preset second loss function, wherein the second loss function is the upper bound proxy function corresponding to the first loss function;

[0171] Based on the second loss function, a second quantile estimation function is constructed, whereby the second quantile estimation function is used to output the second optimization variable that minimizes the mathematical expectation corresponding to the second loss function;

[0172] The network sample data is input into the first quantile estimation function to obtain the second optimization variable;

[0173] The second optimization variable is used as the second quantile.

[0174] Optionally, the robust network optimization device 600 also includes a network performance calculation unit (not shown in the figure), which calculates network performance for:

[0175] The spectral efficiency performance of the target network is determined by the average of the spectral efficiency performance of the first number of grids of the target network at the second number of time points.

[0176] Optionally, the first optimization objective function is as follows:

[0177]

[0178] in, The first optimization objective function is... Let be the first quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell For the set of cells in the target network, Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L As the first number, S The second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively.

[0179] Optionally, the second optimization objective function is as follows:

[0180]

[0181] in, For the second optimization objective function, This is the second quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell For the set of cells in the target network, Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L As the first number, S The second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively.

[0182] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. For example... Figure 6As shown, the electronic device 2000 includes a memory 2101 and a processor 2201. The number of memory 2101 and processor 2201 can be one or more. Figure 6 Taking a memory 2101 and a processor 2201 as an example; the memory 2101 and processor 2201 in the network device can be connected via a bus or other means. Figure 6 Taking the example of a connection between China and Israel via a bus.

[0183] The memory 2101, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as program instructions / modules corresponding to the methods provided in any embodiment of this disclosure. The processor 2201 implements the robust network optimization method provided in any of the above embodiments by running the software programs, instructions, and modules stored in the memory 2101.

[0184] Memory 2101 may primarily include a program storage area and a data storage area, wherein the program storage area may store the operating system and application programs required for at least one function. Furthermore, memory 2101 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, memory 2101 further includes memory remotely located relative to processor 2201, and this remote memory can be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0185] An embodiment of this disclosure also provides a computer-readable storage medium storing computer-executable instructions for performing robust network optimization methods as provided in any embodiment of this disclosure.

[0186] An embodiment of this disclosure also provides a computer program product, including a computer program or computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer program or computer instructions from the computer-readable storage medium and executes the computer program or computer instructions, causing the computer device to perform the robust network optimization method provided in any embodiment of this disclosure.

[0187] The system architecture and application scenarios described in this disclosure are intended to more clearly illustrate the technical solutions of this disclosure and do not constitute a limitation on the technical solutions provided in this disclosure. As those skilled in the art will know, with the evolution of system architecture and the emergence of new application scenarios, the technical solutions provided in this disclosure are also applicable to similar technical problems.

[0188] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0189] In hardware implementations, the division between functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0190] As used in this specification, the terms "component," "module," "system," etc., are used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program, or a computer. As illustrated, applications running on computing devices and computing devices can both be components. One or more components may reside in a process or execution thread, and components may be located on a single computer or distributed among two or more computers. Furthermore, these components can be executed from various computer-readable media on which various data structures are stored. Components can communicate, for example, via local or remote processes based on signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system, or a network, such as the Internet interacting with other systems via signals).

[0191] The above description, with reference to the accompanying drawings, illustrates some embodiments of this disclosure, but does not limit the scope of this disclosure. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of this disclosure shall be within the scope of this disclosure.

Claims

1. A robust network optimization method, characterized in that, The method includes: Obtain network sample data of the target network and antenna parameters of each cell in the target network, wherein the network sample data includes the spectral efficiency performance of a first number of grids of the target network at a second number of time points; Based on the network sample data and the antenna parameters, the gradient of the second optimization objective function is estimated using a zero-order estimation algorithm to obtain the estimated gradient value. The second optimization objective function is constructed using a second quantile and the network efficiency optimization expectation. The second quantile is used to select the grids to be optimized, and the network efficiency optimization expectation is used to represent the optimization mapping relationship between the spectral efficiency performance of the target network and the antenna parameters. The antenna parameters are updated using gradient descent based on the estimated gradient value. Based on the updated antenna parameters, the first value of the first optimization objective function and the second value of the second optimization objective function are calculated respectively. The first optimization objective function is constructed by the first quantile and the network efficiency optimization expectation. The first quantile is used to select the grid that needs to be optimized. Jump to the step of performing gradient estimation on the second optimization objective function based on the network sample data and the antenna parameters using a zero-order estimation algorithm to obtain estimated gradient values ​​until a preset condition is met, so as to obtain multiple first values; The antenna parameter corresponding to the smallest first value among multiple first values ​​is taken as the target antenna parameter; The first optimization objective function is as follows: in, Let the first optimization objective function be... Let be the first quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell The set of cells in the target network. Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L As the first number, S The second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively; The second optimization objective function is as follows: in, Let the second optimization objective function be... This is the second quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell The set of cells in the target network. Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L As the first number, S The second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively.

2. The method according to claim 1, characterized in that, The first quantile is obtained through the following steps: Obtain a preset first loss function, wherein the first loss function is a non-smooth function; Based on the first loss function, a first quantile estimation function is constructed, wherein the first quantile estimation function is used to output the first optimization variable that minimizes the mathematical expectation corresponding to the first loss function; The network sample data is input into the first quantile estimation function to obtain the first optimization variable; The first optimization variable is used as the first quantile.

3. The method according to claim 2, characterized in that, The second quantile is obtained through the following steps: Obtain a preset second loss function, wherein the second loss function is the upper bound proxy function corresponding to the first loss function; Based on the second loss function, a second quantile estimation function is constructed, wherein the second quantile estimation function is used to output a second optimization variable that minimizes the mathematical expectation corresponding to the second loss function; The network sample data is input into the first quantile estimation function to obtain the second optimization variable; The second optimization variable is used as the second quantile.

4. The method according to claim 1, characterized in that, The method further includes: The spectral efficiency performance of the target network is determined based on the average spectral efficiency performance of a first number of grids at a second number of time points.

5. A robust network optimization device, characterized in that, include: The parameter acquisition unit is used to acquire network sample data of the target network and antenna parameters of each cell in the target network, wherein the network sample data includes the spectral efficiency performance of a first number of grids of the target network at a second number of time points; The gradient estimation unit is used to estimate the gradient of the second optimization objective function based on the network sample data and the antenna parameters using a zero-order estimation algorithm, and obtain the estimated gradient value. The second optimization objective function is constructed using a second quantile and the network efficiency optimization expectation. The second quantile is used to select the grid to be optimized, and the network efficiency optimization expectation is used to represent the optimization mapping relationship between the spectral efficiency performance of the target network and the antenna parameters. A gradient update unit is used to update the antenna parameters by gradient descent based on the estimated gradient value. The optimization calculation unit is used to calculate the first value of the first optimization objective function and the second value of the second optimization objective function according to the updated antenna parameters, wherein the first optimization objective function is constructed by the first quantile and the network efficiency optimization expectation, and the first quantile is used to select the grid to be optimized; An iterative jump unit is used to jump to the step of performing gradient estimation on the second optimization objective function based on the network sample data and the antenna parameters, obtaining the estimated gradient value, until a preset condition is met, so as to obtain multiple first values; A parameter determination unit is used to take the antenna parameter corresponding to the smallest first value among a plurality of first values ​​as the target antenna parameter; The first optimization objective function is as follows: in, Let the first optimization objective function be... Let be the first quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell The set of cells in the target network. Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L As the first number, S The second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively; The second optimization objective function is as follows: in, Let the second optimization objective function be... This is the second quantile estimation function. and The first Antenna downtilt angle and horizontal angle for each cell The set of cells in the target network. Indicates the first The grid in the first The spectral efficiency performance measured at each time point. L As the first number, S The second number, Used to calculate the mean. and These are the preset adjustment thresholds for the tilt angle and horizontal angle, respectively.

6. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor, when executing the computer program, implements the robust network optimization method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a computer, implement the robust network optimization method as described in any one of claims 1 to 4.

8. A computer program product, comprising a computer program or computer instructions, characterized in that, The computer program or the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer program or the computer instructions from the computer-readable storage medium and executes the computer program or the computer instructions, causing the computer device to perform the robust network optimization method as described in any one of claims 1 to 4.

Images