Network capacity planning method, device, equipment and medium based on Bayesian optimization

Abstract

The present disclosure provides a network capacity planning method, apparatus, device and medium based on Bayesian optimization, aiming to overcome the problems of high complexity, low versatility and poor generalization ability of the methods provided by related technologies. The method comprises: determining a search space for the resource capacity of each node and each link in the network based on the distribution of business requests; determining a network capacity solution for each round of iteration from the search space based on a historical evaluation point set; the predicted network capacity solution characterizes: allocating specific resource capacity to each node and each link in the network topology, the historical evaluation point set including: each network capacity solution determined by historical iterations and the performance indicators achieved; performing a performance evaluation on the network capacity solution to obtain a performance indicator; and, when an iteration end condition is met, determining the network capacity solution corresponding to the optimal performance indicator in the historical iteration as the network capacity solution corresponding to the business request distribution.

Description

Technical Field

[0001] The present disclosure relates to the technical field of software-defined networks, and in particular to a network capacity planning method, apparatus, device, and medium based on Bayesian optimization. Background Art

[0002] In software-defined networking (SDN), infrastructure providers (InPs) provide various physical network resources, such as computing, storage, and bandwidth. Internet service providers (ISPs) lease these physical network resources from InPs and offer a variety of application services to users. Given a physical network topology, ISPs need to plan the capacity of leased network resources at each node and link, known as a network capacity plan.

[0003] In software-defined networks, user business demands are highly dynamic. When the distribution of business requests changes significantly, ISPs need to adjust network capacity plans accordingly to allow the network to adapt to changing business demands and ensure service quality.

[0004] Related technologies have proposed methods based on various machine learning methods or deep reinforcement learning training, using performance metrics of interest to ISPs as reward functions and simulating the interaction between agents and environments to achieve optimal network capacity solutions. However, these methods require model training within a given network environment and service deployment scenario, resulting in poor generalization capabilities. Furthermore, these methods typically involve a large number of model parameters, resulting in a significant computational overhead. Summary of the Invention

[0005] To overcome the problems existing in related technologies, the present disclosure provides a network capacity planning method, apparatus, device, and medium based on Bayesian optimization. The technical solutions of the present disclosure are as follows:

[0006] According to a first aspect of an embodiment of the present disclosure, a network capacity planning method based on Bayesian optimization is provided, comprising:

[0007] Determining a search space for resource capacity of each node and each link in the network based on the service request distribution; the search space represents a selection range of the resource capacity;

[0008] Determining a network capacity solution for each iteration from the search space based on a set of historical evaluation points; the predicted network capacity solution represents: allocating specific resource capacity to each node and each link in the network topology; the set of historical evaluation points includes: each network capacity solution determined in historical iterations and the performance indicators achieved;

[0009] Performing performance evaluation on the network capacity solution to obtain performance indicators;

[0010] When the iteration end condition is met, the network capacity solution corresponding to the optimal performance indicator in the historical iteration is determined as the network capacity solution corresponding to the service request distribution.

[0011] Optionally, determining a search space for resource capacity of each node and each link in the network based on service request distribution includes:

[0012] Allocating a corresponding initial search space of resource capacity to each node and each link in the network topology;

[0013] Based on the service request distribution and the network topology, determining the shortest path between the starting and ending points of each group of requests according to a target routing algorithm;

[0014] Determine each of the shortest paths as a shortest path set;

[0015] Determining the frequency of each link in the shortest path set in the network topology and determining a corresponding frequency ratio; the frequency ratio represents: the frequency with which the corresponding link is selected in the route used for the service request distribution;

[0016] reallocating a search space of corresponding optimized resource capacity for each node and each link in the network topology according to the frequency ratio;

[0017] The search space of resource capacity of each node and each link in the network is determined according to the optimized search space of resource capacity.

[0018] Optionally, determining a network capacity solution for each iteration from the search space based on a set of historical evaluation points includes:

[0019] Determine a set of historical evaluation points; any evaluation point in the set of historical evaluation points represents a performance indicator under a certain network capacity solution;

[0020] Determining a proxy model corresponding to this iteration based on the historical evaluation point set; the proxy model represents an implicit functional relationship between the network capacity solution and the performance indicator;

[0021] Determining an acquisition function based on the proxy model; the acquisition function represents: the exploration value of each network capacity solution determined from the search space based on the proxy model;

[0022] We will explore the network capacity solution with the greatest value and determine it as the network capacity solution for this round of iteration.

[0023] Optionally, it also includes:

[0024] Determining a historical service request distribution corresponding to the service request distribution based on historical changes in the request distribution;

[0025] Determining historical iteration data of the historical service request distribution; the historical iteration data is historical iteration data optimized for a network capacity solution;

[0026] Through the transfer learning algorithm, the agent model is pre-trained before the iteration begins based on the historical iteration data.

[0027] Optionally, a set of historical evaluation points is determined, including:

[0028] Based on the service deployment history, an initial evaluation point set is obtained; the initial evaluation point set at least includes evaluation points consisting of an initial network capacity solution and corresponding performance indicators;

[0029] Combine the network capacity solutions and corresponding performance indicators obtained in the previous iterations to obtain various evaluation points.

[0030] The evaluation point is added to the initial evaluation point set to obtain the historical evaluation point set.

[0031] Optionally, a performance evaluation is performed on the network capacity solution to obtain performance indicators, including:

[0032] Perform a preset number of independent simulations based on the actual service deployment strategy to obtain corresponding performance indicators. A set of requests is generated for each simulation based on the business request distribution.

[0033] Determine, based on each of the performance indicators, a mean performance indicator and a variance performance indicator corresponding to the predetermined number of independent simulations;

[0034] A performance indicator of the network capacity solution is determined according to the mean performance indicator and the variance performance indicator.

[0035] Optionally, determining a network capacity solution to be explored in each iteration from the search space includes:

[0036] Determine constraints based on actual needs, where the constraints represent constraints of the network capacity solution;

[0037] Based on the constraints, a network capacity solution that needs to be explored in each iteration is determined from the search space.

[0038] According to a second aspect of an embodiment of the present disclosure, a network capacity planning device based on Bayesian optimization is provided, comprising:

[0039] A search space setting module is configured to determine a search space for resource capacity of each node and each link in the network according to the distribution of service requests; the search space represents a selection range of the resource capacity;

[0040] a solution determination module, configured to determine, from the search space, a network capacity solution for each iteration based on a set of historical evaluation points; wherein the predicted network capacity solution represents: allocating specific resource capacity to each node and each link in the network topology; and wherein the set of historical evaluation points includes: each network capacity solution determined in historical iterations and the performance indicators achieved therefrom;

[0041] A performance evaluation module, configured to perform a performance evaluation on the network capacity solution to obtain performance indicators;

[0042] The iteration end module is used to determine the network capacity solution corresponding to the optimal performance indicator in the historical iteration as the network capacity solution corresponding to the business request distribution when the iteration end condition is met.

[0043] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of the network capacity planning method based on Bayesian optimization as described in the first aspect are implemented.

[0044] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the network capacity planning method based on Bayesian optimization as described in the first aspect are implemented.

[0045] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the steps of the network capacity planning method based on Bayesian optimization as described in the first aspect.

[0046] The present disclosure dynamically determines the search space for network topology based on the distribution of service requests, and can adapt to changes in network capacity requirements at different times and for different user needs. The present disclosure does not rely on a specific network environment or service deployment solution for model training. Instead, it utilizes a Bayesian optimization algorithm to gradually approach the optimal solution through iterative search and performance evaluation, resulting in stronger generalization capabilities. The present disclosure utilizes a set of historical evaluation points to guide the search process, avoiding a comprehensive search across the entire search space and effectively reducing computational complexity. The present disclosure can use performance indicators of interest to ISPs as optimization targets based on actual service needs and network environments, flexibly selecting and optimizing performance indicators to provide higher-quality services. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following briefly introduces the drawings required for use in the description of the embodiments of the present disclosure. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0048] Figure 1 This is a schematic diagram of the steps of a network capacity planning method based on Bayesian optimization shown in an embodiment of the present disclosure;

[0049] Figure 2 is a schematic diagram of a test network topology shown in an embodiment of the present disclosure;

[0050] Figure 3 1 is a schematic diagram of a Bayesian optimization convergence performance comparison diagram shown in an embodiment of the present disclosure;

[0051] Figure 4 is a block diagram of a network capacity planning device based on Bayesian optimization shown in an embodiment of the present disclosure;

[0052] Figure 5 Schematic diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0053] The following will clearly and completely describe the technical solutions in the embodiments of the present disclosure in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present disclosure without making any creative efforts shall fall within the scope of protection of the present disclosure.

[0054] The terms "first", "second", etc. in the specification and claims of the present disclosure are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present disclosure can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally of the same type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects related to each other are in an "or" relationship.

[0055] The difference between network capacity planning and network resource allocation is that the former uses the resource capacity of each part of the network as an optimization variable, while the latter mainly optimizes the deployment plan of services, in which the resource capacity of each node and link in the network is fixed.

[0056] To address the issue of network capacity planning, various machine learning methods have been proposed. By analyzing service request histories and network resource allocation history, neural network models are trained to predict the distribution of future service requests and network resource demands. Alternatively, deep reinforcement learning (DRL) is used to train performance metrics of interest to ISPs as reward functions. By simulating the interaction between agents and the environment, the optimal network capacity solution is obtained.

[0057] However, the methods proposed in the related art have certain problems.

[0058] Problem 1: Methods proposed in related technologies typically require model training within a given network environment and service deployment scenario, resulting in poor generalization. Model performance can significantly decline when service deployment methods or performance metrics targeted by ISPs change. This results in low generalizability of existing solutions and makes them difficult to adapt to diverse network environments and service deployment scenarios. Furthermore, for more complex service deployment processes, the accuracy of network resource demand prediction using existing methods can decrease, impacting the performance of network capacity planning.

[0059] Problem 2: The methods proposed in related technologies often involve a large number of model parameters that must be optimized during the training process, resulting in a significant computational burden. In real-world applications, ISPs may need to rapidly respond to changing business needs. The high computational complexity of existing methods can hinder the model's ability to make decisions quickly. Furthermore, deep learning-based models may require high-performance GPUs (Graphics Processing Units) to accelerate training, which not only increases hardware costs but also limits the applicability of existing methods in scenarios with limited computing resources.

[0060] In order to solve the above technical problems, the present disclosure proposes a network capacity planning method based on Bayesian optimization. The method adopts a network capacity planning algorithm based on Bayesian optimization, thereby effectively solving the problems of low universality and poor generalization of the methods proposed in related technologies.

[0061] Figure 1 This is a schematic diagram of the steps of a network capacity planning method based on Bayesian optimization shown in an embodiment of the present disclosure. Figure 1 As shown, the method may specifically include the following steps:

[0062] Step S11: determining a search space for resource capacity of each node and each link in the network according to the service request distribution; the search space represents a selection range of the resource capacity.

[0063] Service request distribution refers to the frequency and pattern of various service requests received within a specific time period.

[0064] The search space representation for determining the resource capacity of each node and link in the network. The search space represents the range of resource capacity options, such as the minimum and maximum bandwidth on a link.

[0065] The network topology represents the connection relationship between each node and link in the network. The network topology can be expressed as , the total number of nodes and links are V and E respectively. The search space can be expressed as , which represents the upper and lower bounds of the resource capacity of each node and link in the network.

[0066] Collect business request distribution data, analyze traffic patterns and user needs in different time periods, and determine the search space for network topology based on user needs.

[0067] Step S12: Based on the historical evaluation point set, determine the network capacity solution for each iteration from the search space; the predicted network capacity solution represents: allocating specific resource capacity to each node and each link in the network topology, and the historical evaluation point set includes: each network capacity solution determined by historical iterations and the performance indicators achieved therein.

[0068] The historical evaluation point set, obtained during the initialization phase, can be used to store the network capacity solutions determined in each iteration and their corresponding performance indicators. Specifically, several evaluation points are generated during the historical iterations, each representing a network capacity solution that achieves a certain performance indicator. The network capacity solution and its corresponding performance indicator are considered as an evaluation point and added to the historical evaluation point set.

[0069] Performance indicators can be indicators such as latency, throughput, packet loss rate, etc. that can reflect the effectiveness of network services.

[0070] During each round of iteration, the network capacity solution obtained in the current round of iteration is determined from the search space based on the historical evaluation point set.

[0071] The performance indicators corresponding to the evaluation points in the historical evaluation point set are indicators that the ISP is currently focusing on and needs to optimize. For example, if the indicator that the ISP is focusing on and needs to optimize is the request rejection rate, the performance indicator may represent the number of rejected requests.

[0072] In the initialization phase of the network capacity planning algorithm, an initial network capacity plan can be set as the input of the first round of iteration, and the performance of the initial network capacity plan can be evaluated.

[0073] Step S13: Perform performance evaluation on the network capacity solution to obtain performance indicators.

[0074] After each round of iteration, a network capacity plan is generated. This involves conducting simulation tests on the network configured according to the plan, performing request simulation deployments, and evaluating its performance under a specific service request distribution. This results in performance indicators for the current network capacity plan. These indicators are currently the focus of ISPs and require optimization.

[0075] Performance indicator data can be used to analyze network performance under different network capacity solutions, including service quality, resource utilization, etc.

[0076] Step S14: When the iteration end condition is met, the network capacity solution corresponding to the optimal performance indicator in the historical iteration is determined as the network capacity solution corresponding to the service request distribution.

[0077] Iteration end conditions can be set. These conditions can be set to meet preset performance indicators, complete a certain number of iterations, or have no significant improvement in performance indicators over several consecutive iterations.

[0078] At the end of each iteration, check whether the set iteration end condition is met. If the iteration end condition is met, end the iteration and take the current iteration as the last iteration.

[0079] Across all historical iterations, the performance metrics of various network capacity solutions are compared to identify the optimal network capacity solution. This optimal network capacity solution is then determined as the final network capacity solution for the service request distribution and is ready for implementation.

[0080] After the network capacity plan is implemented, network performance can be continuously monitored to ensure that it meets business needs and dynamically adjusted according to actual conditions.

[0081] By employing the embodiments of the present disclosure, network capacity solutions are determined based on the distribution of service requests, ensuring an optimal match between the network capacity solution and the service request distribution, thereby improving network resource utilization. By determining the network capacity solution for each iteration based on a set of historical evaluation points, prior information can be fully utilized to iteratively search for a set of optimal network capacity solutions. By continuously optimizing the network capacity solution through iteration, changes in the service request distribution can be addressed, enabling flexible adjustments based on the actual needs of ISPs and unified optimization of heterogeneous network resources such as computing, storage, and bandwidth.

[0082] In some cases, due to upper limits on physical resources of nodes and links and to accelerate the convergence of Bayesian optimization, constraints can be set and network capacity solutions can be determined based on the constraints. The following is a specific example of applying the constraints.

[0083] In an optional embodiment, based on a set of historical evaluation points, a network capacity solution for each iteration is determined from the search space, including: determining constraints according to actual needs, wherein the constraints represent the constraints of the network capacity solution; and based on the constraints, determining the network capacity solution that needs to be explored in each iteration from the search space.

[0084] Constraints can be defined based on the actual needs of the ISP In practical implementation, resource rental cost can usually be used as a constraint. Assume that the price of renting a computing resource per unit resource on each node is , the price of renting a unit bandwidth resource on each link is , the total cost of ISP leasing network resources is capped at , the amount of resources leased by ISP on each node and link is expressed as , then the constraint It can be expressed as: .

[0085] It should be noted that in addition to cost constraints, performance-related constraints can also be considered, such as request response time, service quality, etc.

[0086] Integrate the constraints into the optimization model of network capacity planning to ensure that the network capacity solution obtained in each iteration satisfies these constraints during the process of solving the optimal solution.

[0087] By using the embodiments of the present disclosure, by selecting the network capacity solutions to be explored from the search space based on constraints in each iteration, the exploration of invalid solutions can be avoided, reducing unnecessary computation and simulation time. In actual applications, service requests and resource prices may change. By setting constraints, ISPs can flexibly adjust the restrictions of network capacity solutions based on actual conditions, ensuring rapid response to changes in service request distribution.

[0088] In each iteration, for a given service request distribution, evaluating the performance of the network capacity solution obtained in the current iteration is a challenge. To address this issue, the Monte Carlo method can be used to evaluate the performance of a network capacity solution through multiple simulations. The following are specific examples of simulations proposed in this disclosure.

[0089] In an optional embodiment, a performance evaluation is performed on the network capacity solution to obtain performance indicators, including: performing a preset number of independent simulations based on the service deployment strategy actually adopted to obtain corresponding performance indicators; generating a group of requests for each simulation based on the business request distribution; determining the mean performance indicator and variance performance indicator corresponding to the preset number of independent simulations based on each of the performance indicators; and determining the performance indicator of the network capacity solution based on the mean performance indicator and the variance performance indicator.

[0090] Determine the actual service deployment strategy to be adopted. The service deployment strategy can include network node configuration, link load balancing, and other aspects. Based on the service deployment strategy, build a simulation system that can simulate the actual network environment.

[0091] Using the Monte Carlo method, the simulation system is simulated for a preset number of times, for example, Q independent simulations are performed. Each simulation generates a set of requests based on the business request distribution, which can be expressed as ,The request can be implemented through a random number generator.,Each request contains necessary features such as the requested service type,,request start and end points, traffic size, etc., so as to be used in subsequent simulations.

[0092] Simulate and deploy these requests in a simulation environment. During each simulation, collect the performance indicators that the current ISP needs to optimize, such as response time and success rate. After each simulation, record the corresponding performance indicators and store them as structured data. For example, the performance indicator corresponding to the qth simulation in the nth iteration can be recorded as .

[0093] Perform statistical analysis on the performance indicators corresponding to all independent simulations and calculate the mean performance indicator and variance performance indicator. The mean performance indicator reflects the performance of the network capacity solution under average conditions, while the variance performance indicator reveals the fluctuation of the performance indicator. For the nth iteration, the mean performance indicator can be expressed as , the variance performance index can be expressed as .

[0094] Based on the mean performance index and variance performance index, a comprehensive performance evaluation of the network capacity solution is conducted. Specifically, the evaluation can be performed using the following formula:

[0095]

[0096] in, is an adjustable weight parameter. Adjustments are made to balance the influence of mean and variance in the final performance evaluation.

[0097] The meaning is: the network capacity solution obtained in this round of iteration Performance under .

[0098] In getting Performance under After that, it is taken as a tuple and added to the evaluation point set to construct a new evaluation point set to provide data support for the next round of iteration.

[0099] By adopting the embodiments of the present disclosure, through multiple independent simulations based on the service deployment strategy actually adopted, the performance of the network capacity solution in actual applications can be more realistically reflected, thereby improving the reliability and accuracy of the evaluation results. A set of requests is generated for each simulation based on the business request distribution, so that the evaluation process can take into account the randomness and uncertainty of the business requests, and can better simulate the request situation in the real environment, thereby improving the effectiveness of the performance evaluation. In addition, by simulating the service deployment process to evaluate network performance, it is not limited by the specific algorithm of service deployment and can be used in a variety of network environments and service deployment scenarios. By presetting multiple independent simulations, the performance of the network capacity solution under different circumstances can be captured, the randomness of a single simulation result can be reduced, and the robustness of the evaluation results can be enhanced.

[0100] Among them, in an optional embodiment, based on the historical evaluation point set, the network capacity plan for each round of iteration is determined from the search space, including: determining the historical evaluation point set; any evaluation point in the historical evaluation point set represents the performance index under a certain network capacity plan; determining the proxy model corresponding to this round of iteration based on the historical evaluation point set; the proxy model represents the implicit functional relationship between the network capacity plan and the performance index; based on the proxy model, determining the acquisition function; the acquisition function represents: based on the proxy model, the exploration value of each network capacity plan determined from the search space; the network capacity plan with the largest exploration value is determined as the network capacity plan for this round of iteration.

[0101] Collect the network capacity plans and corresponding performance indicators obtained from historical iterations, and regard each historical capacity plan and its corresponding performance indicator as an evaluation point.

[0102] Select an appropriate proxy model based on the historical evaluation point set and use the data from that set to train the proxy model. Specifically, by fitting the model parameters, the proxy model can accurately reflect the implicit functional relationship between the network capacity solution and the performance indicator. By fitting the existing evaluation point set, the performance of unevaluated solutions can be predicted.

[0103] Design an appropriate acquisition function based on the characteristics of the surrogate model and the search space. This acquisition function is used to evaluate the exploration value of each network capacity solution identified from the search space under the surrogate model. This acquisition function helps Bayesian optimization-based network capacity planning algorithms determine which solution to evaluate next.

[0104] In each iteration, the network capacity solution with the highest exploration value is found by maximizing the acquisition function. This solution represents the one most likely to improve performance under the current proxy model. This solution is considered the solution for this iteration.

[0105] In the embodiments of the present disclosure, a proxy model is used to approximate the functional relationship between complex network capacity solutions and performance indicators. This proxy model allows for rapid prediction of performance indicators for different network capacity solutions without direct actual evaluation, reducing computational resources and time consumption and improving iteration efficiency. The design of the acquisition function enables the algorithm to select solutions that are most likely to improve performance under the current proxy model, effectively exploring the search space.

[0106] In one optional embodiment, determining a historical evaluation point set includes: obtaining an initial evaluation point set based on a service deployment history; the initial evaluation point set includes at least an initial network capacity plan and evaluation points consisting of corresponding performance indicators; combining the network capacity plans and corresponding performance indicators obtained in each round of iteration before the current round of iteration to obtain each evaluation point; and adding the evaluation points to the initial evaluation point set to obtain the historical evaluation point set.

[0107] Historical data can be collected from the ISP's service deployment system, including detailed records of each service deployment. This includes information such as network capacity plans, the start and end points of service requests, deployment time, and performance indicators. Performance indicators can include service quality, request acceptance rate, and total deployment revenue.

[0108] Based on the actual needs of the ISP, you can determine the performance indicators that require attention. These could be performance indicators related to service quality, such as latency, bandwidth utilization, and packet loss rate; performance indicators related to request acceptance rate, such as the ratio of successfully processed requests to total requests; or performance indicators related to total deployment revenue, such as the difference between the ISP's cost of leasing resources and the revenue from successfully deployed services.

[0109] The collected historical data is organized into an evaluation point set. Each evaluation point in the evaluation point set can be represented as a tuple, which contains the network capacity plan and the corresponding performance indicator. The evaluation point can be represented as , The meaning is that in a certain network capacity solution Performance indicators under .

[0110] In the initial stage before Bayesian optimization, a set of network capacity solutions can be set based on historical data, business request distribution, and network requirements. These solutions can be used as initial solutions and as input for the first iteration of Bayesian optimization. The performance indicators of the initial solutions can then be obtained. The initial solutions and the corresponding performance indicators can be used as evaluation points and added to the initial evaluation point set.

[0111] In addition, in each round of iteration, the network capacity plan corresponding to each round of iteration and the performance indicators corresponding to the plan will be obtained. Based on the network capacity plan and corresponding performance indicators obtained in each round of iteration, new evaluation points are formed and added to the initial evaluation point set, thereby continuously enriching and improving the evaluation point set.

[0112] By using the embodiments of the present disclosure and utilizing historical service deployment data, the initialization of the evaluation point set can provide the model with a realistic, past-performance-based reference, helping to improve the accuracy of the proxy model in the initial stages and reducing the time and resource consumption of model training. The initialization of the evaluation point set can provide a better starting point for Bayesian optimization, thereby accelerating the convergence process and reducing the number of iterations required to reach the optimal solution. By adding the evaluation points generated during the iterative process to the initial evaluation point set, a more comprehensive and rich historical evaluation point set can be constructed, and because the evaluation point set is dynamically updated, the algorithm can continuously learn and improve its strategy.

[0113]

[0114] Table 1

[0115] Table 1 shows the steps of a Bayesian optimization-based network capacity planning algorithm according to an embodiment of the present disclosure. As shown in Table 1, steps 1-3 are the initialization phase, and steps 4-14 are the iterative optimization phase, which includes multiple rounds of iterations. Each round includes an evaluation phase (steps 8-11), which includes multiple independent simulations.

[0116] Step 1: Define the search space and constraints .

[0117] Step 2: Initialize the network capacity plan .

[0118] set up Set a set of network capacity plans for the amount of resource leased by the ISP on each node and link as an initial solution.

[0119] Step 3: Initialize the evaluation point set .

[0120] Initialize the evaluation point set based on service deployment history , where each evaluation point can be expressed as , that is, in a certain network capacity solution Performance indicators under ,This performance indicator is determined according to the actual needs of the ISP and can be set as ,quality of service, request acceptance rate, total deployment revenue, etc.

[0121] Step 4-14: After initialization, the iterative process of Bayesian optimization begins. Steps 4-14 are executed once in each iteration.

[0122] Step 5: According to Building a proxy model .

[0123] In each iteration, a proxy model is built based on the existing set of evaluation points , surrogate models are used to estimate network capacity solutions and performance indicators The implicit functional relationship between them.

[0124] Step 6: Construct a sampling function based on the surrogate model .

[0125] Compute the acquisition function based on the proxy model , the acquisition function can use models such as expected improvement or probability improvement to measure the search space The exploration value of various network capacity solutions in the network.

[0126] Step 7: Calculate the current network capacity plan: .

[0127] By maximizing the acquisition function, we can get the network capacity solution that needs to be selected in this round of iteration .

[0128] Steps 8-11: Perform Q independent simulations based on the service deployment strategy actually adopted by the ISP to evaluate the performance of the network capacity solution obtained in step 7.

[0129] Step 9: Distribution based on business requests Generate a set of requests .

[0130] Step 10: Simulate Deployment , get the network performance under the current network capacity solution .

[0131] Step 12: Evaluate the current network capacity solution Overall performance under .

[0132] In order to improve the evaluation accuracy, the mean of Q simulation results is considered at the same time and variance Determine the current network capacity plan Overall performance under .

[0133] Specifically expressed as a network capacity solution Performance under ,in is an adjustable weight parameter.

[0134] Step 13: Expand the evaluation point set: .

[0135] The evaluation points obtained in this round of iteration Added to the evaluation point set for updating the proxy model and acquisition function in the next iteration.

[0136] Step 15: After the iteration process is completed, the optimal network capacity solution is selected from the iteration history. .

[0137] When using Bayesian optimization for network capacity planning, a key issue is how to improve the convergence speed of the network capacity planning algorithm. This is because the service deployment process needs to be simulated multiple times in each round of iteration. Complex service deployment scenarios may cause each round of iteration to take a long time, generally in the order of minutes. Therefore, in order to improve the efficiency of Bayesian optimization, it is necessary to minimize the number of iterations required for the algorithm to converge. This disclosure provides two methods to accelerate convergence to optimize the network capacity planning algorithm: Reduced Search Space (RSS) and Transfer Learning (TL). Each method of accelerating convergence can make full use of various types of prior information to guide the convergence of Bayesian optimization.

[0138] Method 1: Narrowing the search space

[0139] Among them, in an optional embodiment, the search space of the resource capacity of each node and each link in the network is determined according to the business request distribution, including: allocating a corresponding initial resource capacity search space to each node and each link in the network topology; based on the business request distribution and the network topology, according to the target routing algorithm, determining the shortest path between the starting and ending points of each group of requests; determining each of the shortest paths as a shortest path set; determining the frequency of each link in the network topology in the shortest path set, and determining the corresponding frequency ratio; the frequency ratio represents: the frequency of the corresponding link being selected in the route used by the business request distribution; through the frequency ratio, reallocating the corresponding optimized resource capacity search space to each node and each link in the network topology; and determining the resource capacity search space of each node and each link in the network based on the optimized resource capacity search space.

[0140] In the initial stages of Bayesian optimization, you can first make a rough estimate of network resource requirements. Based on an understanding of traffic demand characteristics at different time periods and locations, you can narrow the search space. This can be done using the following methods.

[0141] An initial resource capacity search space is allocated for each link and node in the network topology. The initial resource capacity search space can be determined based on basic network design parameters and a preset capacity range, such as link bandwidth, latency, load, and other characteristics.

[0142] Based on the business request distribution and network topology, a target routing algorithm is selected to calculate the shortest path between the starting and ending points of each group of requests. The target routing algorithm can be a routing calculation algorithm such as the Dijkstra algorithm.

[0143] All the calculated shortest paths are aggregated into a shortest path set. The shortest path set contains all possible business traffic paths in the network under the business request distribution. The shortest path set can be expressed as .

[0144] The shortest path set is traversed, and the frequency of each link in the network topology appearing in the shortest path set is counted. The frequency ratio is determined by the ratio of the link frequency to the sum of the frequencies of all links. This frequency ratio reflects how often the link is selected in the routes used for service request distribution.

[0145] Analyze the statistical results to identify links with high and low frequencies. Links with high frequencies can be considered hot links. Hot links typically exhibit high traffic loads and are frequently used.

[0146] According to the frequencies at which different links are selected under the service request distribution, the search space of resource capacity is reallocated for each link and each node, so as to achieve the reduction of the search space.

[0147] The frequency ratio is used to reallocate the search space for the optimized resource capacity for each node and link in the network topology. For example, a larger search space may be allocated to a link with a higher frequency and its corresponding node, specifically expanding the capacity selection range of the link and possibly increasing the upper limit of the search space. A smaller search space may be allocated to a link with a lower frequency and its corresponding node, specifically narrowing the capacity selection range of the link and possibly lowering the upper limit of the search space.

[0148] Based on the initial search space and the optimized search space, a search space for resource capacity of each node and each link in the network corresponding to the network topology is determined. In a subsequent Bayesian optimization process, a network capacity solution is determined from the search space.

[0149] After analyzing and identifying hot and infrequent links, we can assign initial capacity values ​​to each link based on its frequency ratio. This initial capacity solution is then input into the first round of Bayesian optimization. Compared to the initial capacity solution determined by the initial search space, the initial capacity solution determined after narrowing the search space provides a better starting point for Bayesian optimization.

[0150] By using the embodiments of the present disclosure, by analyzing the service request distribution and setting a larger search space for frequently used links, it is possible to meet higher service demands under the service request distribution, while narrowing the search space for less frequently used links, effectively reducing unnecessary search and computational burdens. This allows for effective utilization of prior information about the network and service requirements, making the search more accurate and efficient. Because the search space has been optimized based on the service request distribution and network topology, Bayesian optimization can perform iterative searches within a smaller search space, reducing the number of iterations and improving the efficiency of each iteration.

[0151] Method 2: Transfer Learning

[0152] To speed up the convergence of Bayesian models, you can also use transfer learning (TL). The following is a specific example of using transfer learning to speed up the convergence of Bayesian models.

[0153] In an optional embodiment, it also includes: determining the historical business request distribution corresponding to the business request distribution based on the periodic changes in the request distribution; determining the historical business request distribution and optimizing the historical iteration data for the network capacity solution; and pre-training the proxy model based on the historical iteration data before the iteration begins through a transfer learning algorithm.

[0154] Since the distribution of business requests often changes periodically in actual business application scenarios, the optimization history can be used as prior information to improve the efficiency of Bayesian optimization.

[0155] Real-time monitoring of business request data, including key information such as request time, type, and quantity. The monitored business request data can be used to identify historical changes in the distribution of business requests using statistical methods such as time series analysis or machine learning algorithms.

[0156] Based on the cycle phase of the current request distribution, a historical business request distribution that matches the current cycle phase is selected from the historical data. The historical business request distribution can be verified to determine whether it is comparable to the request distribution currently being iterated.

[0157] The historical iteration data can be stored in a historical database, so that the historical iteration data of the network capacity solution corresponding to the selected historical service request distribution can be found in the historical database. The historical iteration data can include information such as the network capacity solution for each iteration, the corresponding performance indicators, and the agent model.

[0158] You can choose a suitable transfer learning algorithm based on the characteristics and goals of historical iterative data, such as feature-based transfer, instance-based transfer, or model-based transfer.

[0159] Data from historical optimization tasks is used as input to construct a training set. The selected transfer learning algorithm is used to train the proxy model to capture the relationship between the service request distribution and network capacity solutions in the historical optimization tasks. This allows the knowledge and experience from the historical iteration data to be transferred to the currently running Bayesian optimization-based capacity planning algorithm, and a proxy model is trained to match the current service request distribution.

[0160] In a specific implementation, for a business request distribution , assuming that there are K iterations of Bayesian optimization before, before building the proxy model for the business request distribution, that is, before running the network capacity planning algorithm, As input, use transfer learning algorithms such as RGPE (Relational Gaussian Process Expertise), SGPR (Sparse Gaussian Process Regression), and TransBO (Transfer Bayesian Optimization) to pre-train the proxy model. .

[0161] By adopting the embodiments of the present disclosure, by determining the proxy model corresponding to this round of iteration based on the evaluation point set and using the transfer learning algorithm to integrate the prior information in the historical data, the convergence speed of Bayesian optimization in dealing with periodically changing business request distribution can be significantly improved, which not only improves the efficiency of network capacity planning but also reduces the consumption of computing resources.

[0162] To verify the effectiveness of the two accelerated convergence methods provided by this disclosure, simulation experiments can be performed. The simulation experiments test the network capacity planning algorithm OSS without optimization using methods one and two, the network capacity planning algorithm RSS optimized only using method one, and the network capacity planning algorithm TL optimized only using method two.

[0163] Figure 2 FIG. 1 is a schematic diagram of a test network topology shown in an embodiment of the present disclosure. Figure 2 As shown, the test network topology This is the network topology used for simulation experiment verification. Test network topology There are 15 nodes and 30 links in total. Figure 2The larger nodes have computing power and can deploy services; the smaller nodes have no computing power and can only forward service data.

[0164] Based on the test network topology , now given a business request distribution , in business request distribution The possible starting and ending points of the request are Figure 2 The ISP plans to deploy 200 service requests in the network, of which 150 requests will start and end at randomly selected nodes in the rectangular boxes, and the other 50 requests will start and end at randomly selected points across the entire network.

[0165] Since network bandwidth resources are limited, the price of renting a unit of bandwidth resource on each link is 1, and the constraint condition is set as the total cost of renting network resources. .

[0166] The optimization goal is to minimize the request rejection rate, that is, the performance indicator The number of rejected requests out of 200 requests.

[0167] In order to verify the effectiveness of the method 1 for accelerating convergence provided by the present disclosure. Figure 2 Given a network topology and a given business request distribution, the network capacity planning algorithm is run using different search spaces. The initial search space and the optimized search space are shown in the following table:

[0168]

[0169] Table 2: Initial search space

[0170]

[0171] Table 3: Optimized search space

[0172] The initial bandwidths of each link in Table 2 represent the initial network capacity solution determined during the initial search space, while the initial bandwidths of each link in Table 3 represent the initial network capacity solution determined during the optimized search space. Compared to the initial network capacity solution determined during the initial search space, the initial network capacity solution determined during the optimized search space provides a better initial solution for Bayesian optimization.

[0173] In order to verify the effectiveness of the second method for accelerating convergence provided by the present disclosure, 10 Bayesian optimization iteration histories are used as input to pre-train the surrogate model, and the initial search space is used for Bayesian optimization.

[0174] OSS, RSS, and TL were used to conduct 30 independent experiments, each with 200 iterations. The convergence of Bayesian optimization is shown in Figure 2. Figure 3 shown.

[0175] Figure 3 This is a schematic diagram of a Bayesian optimization convergence performance comparison diagram shown in an embodiment of the present disclosure. Figure 3 As shown in the figure, the convergence curve composed of circles represents the average convergence of OSS over 30 independent experiments, the convergence curve composed of triangles represents the average convergence of RSS over 30 independent experiments, and the convergence curve composed of rectangles represents the average convergence of TL over 30 independent experiments. The corresponding colored range indicates the maximum fluctuation range of the convergence curve over 30 experiments.

[0176] Comparing the convergence of the OSS, RSS, and TL experiments, RSS and TL converged significantly faster than OSS, with significantly less convergence curve fluctuation than OSS. OSS failed to converge within 200 iterations. Comparing RSS and TL reveals that RSS converged faster than TL because it was able to identify a more effective initial network capacity solution by narrowing the search space.

[0177] based on Figure 3 The simulation results of OSS, RSS, and TL are shown. Method 1 and Method 2 provided by this disclosure to improve convergence speed can effectively accelerate the convergence of the network capacity planning algorithm, thereby improving the real-time performance of the algorithm and helping ISPs quickly plan new network capacity when business requests change.

[0178] Based on the same technical concept, the present disclosure provides a network capacity planning device based on Bayesian optimization. Figure 4 This is a block diagram of a network capacity planning device based on Bayesian optimization shown in an embodiment of the present disclosure. Figure 4 As shown, the device includes:

[0179] A search space setting module 410 is configured to determine a search space for resource capacity of each node and each link in the network according to the service request distribution; the search space represents a selection range of the resource capacity;

[0180] A solution determination module 420 is configured to determine a network capacity solution for each iteration from the search space based on a set of historical evaluation points, wherein the predicted network capacity solution represents: allocating a specific resource capacity to each node and each link in the network topology; the set of historical evaluation points includes: each network capacity solution determined in historical iterations and the performance indicators achieved;

[0181] A performance evaluation module 430 is configured to perform a performance evaluation on the network capacity solution to obtain a performance indicator;

[0182] The iteration ending module 440 is configured to determine, when an iteration ending condition is met, the network capacity solution corresponding to the optimal performance indicator in the historical iteration as the network capacity solution corresponding to the service request distribution.

[0183] Optionally, the range determination module is specifically configured to execute:

[0184] Allocating a corresponding initial search space of resource capacity to each node and each link in the network topology;

[0185] Based on the service request distribution and the network topology, determining the shortest path between the starting and ending points of each group of requests according to a target routing algorithm;

[0186] Determine each of the shortest paths as a shortest path set;

[0187] Determining the frequency of each link in the shortest path set in the network topology and determining a corresponding frequency ratio; the frequency ratio represents: the frequency with which the corresponding link is selected in the route used for the service request distribution;

[0188] reallocating a search space of corresponding optimized resource capacity for each node and each link in the network topology according to the frequency ratio;

[0189] The search space of resource capacity of each node and each link in the network is determined according to the optimized search space of resource capacity.

[0190] Optionally, the solution determination module is specifically configured to execute:

[0191] Determine a set of historical evaluation points; any evaluation point in the set of historical evaluation points represents a performance indicator under a certain network capacity solution;

[0192] Determining a proxy model corresponding to this iteration based on the historical evaluation point set; the proxy model represents an implicit functional relationship between the network capacity solution and the performance indicator;

[0193] Determining an acquisition function based on the proxy model; the acquisition function represents: the exploration value of each network capacity solution determined from the search space based on the proxy model;

[0194] We will explore the network capacity solution with the greatest value and determine it as the network capacity solution for this round of iteration.

[0195] Optionally, the device further includes:

[0196] A historical service request distribution determining module, configured to determine a historical service request distribution corresponding to the service request distribution based on historical changes in the request distribution;

[0197] A historical iteration data determination module, configured to determine historical iteration data of the historical service request distribution; the historical iteration data is historical iteration data optimized for a network capacity solution;

[0198] The pre-training module is used to pre-train the agent model before the iteration starts based on the historical iteration data by using a transfer learning algorithm.

[0199] Optionally, the solution determination module is specifically configured to execute:

[0200] Based on the service deployment history, an initial evaluation point set is obtained; the initial evaluation point set at least includes evaluation points consisting of an initial network capacity solution and corresponding performance indicators;

[0201] Combine the network capacity solutions and corresponding performance indicators obtained in the previous iterations to obtain various evaluation points.

[0202] The evaluation point is added to the initial evaluation point set to obtain the historical evaluation point set.

[0203] Optionally, the performance evaluation module is specifically configured to execute:

[0204] Perform performance evaluation on the network capacity solution to obtain performance indicators, including:

[0205] Perform a preset number of independent simulations based on the actual service deployment strategy to obtain corresponding performance indicators. A set of requests is generated for each simulation based on the business request distribution.

[0206] Determine, based on each of the performance indicators, a mean performance indicator and a variance performance indicator corresponding to the predetermined number of independent simulations;

[0207] A performance indicator of the network capacity solution is determined according to the mean performance indicator and the variance performance indicator.

[0208] Optionally, the solution determination module is specifically configured to execute:

[0209] Determine constraints based on actual needs, where the constraints represent constraints of the network capacity solution;

[0210] Based on the constraints, a network capacity solution that needs to be explored in each iteration is determined from the search space.

[0211] It should be noted that the device embodiment is similar to the method embodiment, so the description is relatively simple, and the relevant parts can be referred to the method embodiment.

[0212] The present disclosure also provides an electronic device, Figure 5 , Figure 5 Schematic diagram of an electronic device according to an embodiment of the present disclosure. Figure 5 As shown, the electronic device 500 includes: a memory 510 and a processor 520. The memory 510 and the processor 520 are connected via a bus communication. A computer program is stored in the memory 510. The computer program can be run on the processor 520 to implement the steps in the network capacity planning method based on Bayesian optimization disclosed in the embodiment of the present disclosure.

[0213] The embodiment of the present disclosure further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the network capacity planning method based on Bayesian optimization disclosed in the embodiment of the present disclosure are implemented.

[0214] The embodiment of the present disclosure further provides a computer program product, including a computer program. When the computer program is executed by a processor, the steps of the network capacity planning method based on Bayesian optimization disclosed in the embodiment of the present disclosure are implemented.

[0215] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0216] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, apparatuses, or computer program products. Thus, the embodiments of the present disclosure may take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, the embodiments of the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0217] The embodiments of the present disclosure are described with reference to the flowcharts and / or block diagrams of the methods, apparatuses, electronic devices, and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0218] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0219] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0220] Although some embodiments of the present disclosure have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiment and all changes and modifications that fall within the scope of the present disclosure.

[0221] The above is a detailed introduction to the network capacity planning method, device, equipment and medium based on Bayesian optimization provided by the present disclosure. Specific examples are used in this article to illustrate the principles and implementation methods of the present disclosure. The description of the above embodiments is only used to help understand the method and core ideas of the present disclosure. At the same time, for those skilled in the art, according to the ideas of the present disclosure, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as limiting the present disclosure.

Claims

1. A network capacity planning method based on Bayesian optimization, characterized in that: include: Determine the search space for resource capacity of each node and link in the network based on the distribution of service requests; The search space represents a selection range of the resource capacity; Determining a network capacity solution for each iteration from the search space based on a set of historical evaluation points; The network capacity solution characterization: allocating specific resource capacity to each node and each link in the network topology; the historical evaluation point set includes: each network capacity solution determined by historical iterations and the performance indicators achieved; Performing performance evaluation on the network capacity solution to obtain performance indicators; When the iteration end condition is met, the network capacity solution corresponding to the optimal performance indicator in the historical iteration is determined as the network capacity solution corresponding to the service request distribution; Based on the service request distribution, the search space for resource capacity of each node and link in the network is determined, including: Allocating a corresponding initial search space of resource capacity to each node and each link in the network topology; Based on the service request distribution and the network topology, determining the shortest path between the starting and ending points of each group of requests according to a target routing algorithm; Determine each of the shortest paths as a shortest path set; Determining the frequency of each link in the shortest path set in the network topology and determining a corresponding frequency ratio; the frequency ratio represents: the frequency with which the corresponding link is selected in the route used for the service request distribution; reallocating a search space of corresponding optimized resource capacity for each node and each link in the network topology according to the frequency ratio; Determining the search space for resource capacity of each node and each link in the network according to the optimized search space for resource capacity; Based on the historical evaluation point set, a network capacity solution for each iteration is determined from the search space, including: Determine a set of historical evaluation points; any evaluation point in the set of historical evaluation points represents a performance indicator under a certain network capacity solution; Determining a proxy model corresponding to this iteration based on the historical evaluation point set; the proxy model represents an implicit functional relationship between the network capacity solution and the performance indicator; Determining an acquisition function based on the proxy model; the acquisition function represents: the exploration value of each network capacity solution determined from the search space based on the proxy model; We will explore the network capacity solution with the greatest value and determine it as the network capacity solution for this round of iteration.

2. The method according to claim 1, characterized in that Also includes: Determining a historical service request distribution corresponding to the service request distribution based on historical changes in the request distribution; Determining historical iteration data of the historical service request distribution; the historical iteration data is historical iteration data optimized for a network capacity solution; Through the transfer learning algorithm, the agent model is pre-trained before the iteration begins based on the historical iteration data.

3. The method according to claim 1, characterized in that Determine the set of historical assessment points, including: Based on the service deployment history, an initial evaluation point set is obtained; the initial evaluation point set at least includes evaluation points consisting of an initial network capacity solution and corresponding performance indicators; Combine the network capacity solutions and corresponding performance indicators obtained in the previous iterations to obtain various evaluation points. The evaluation point is added to the initial evaluation point set to obtain the historical evaluation point set.

4. The method according to claim 1, wherein Perform performance evaluation on the network capacity solution to obtain performance indicators, including: Perform a preset number of independent simulations based on the actual service deployment strategy to obtain corresponding performance indicators. A set of requests is generated for each simulation based on the business request distribution. Determine, based on each of the performance indicators, a mean performance indicator and a variance performance indicator corresponding to the predetermined number of independent simulations; A performance indicator of the network capacity solution is determined according to the mean performance indicator and the variance performance indicator.

5. The method according to any one of claims 1 to 4, characterized in that: Based on the historical evaluation point set, a network capacity solution for each iteration is determined from the search space, including: Determine constraints based on actual needs, where the constraints represent constraints of the network capacity solution; Based on the constraints, a network capacity solution that needs to be explored in each iteration is determined from the search space.

6. A network capacity planning device based on Bayesian optimization, characterized in that: include: A search space setting module is used to determine the search space of resource capacity of each node and each link in the network according to the distribution of service requests; The search space represents a selection range of the resource capacity; A solution determination module, configured to determine a network capacity solution for each iteration from the search space based on a set of historical evaluation points; The network capacity solution characterization: allocating specific resource capacity to each node and each link in the network topology; the historical evaluation point set includes: each network capacity solution determined by historical iterations and the performance indicators achieved; A performance evaluation module, configured to perform a performance evaluation on the network capacity solution to obtain performance indicators; An iteration end module, configured to, when an iteration end condition is met, determine the network capacity solution corresponding to the optimal performance indicator in the historical iteration as the network capacity solution corresponding to the service request distribution; The search space setting module is specifically used to: Allocating a corresponding initial search space of resource capacity to each node and each link in the network topology; Based on the service request distribution and the network topology, determining the shortest path between the starting and ending points of each group of requests according to a target routing algorithm; Determine each of the shortest paths as a shortest path set; Determining the frequency of each link in the shortest path set in the network topology and determining a corresponding frequency ratio; the frequency ratio represents: the frequency with which the corresponding link is selected in the route used for the service request distribution; reallocating a search space of corresponding optimized resource capacity for each node and each link in the network topology according to the frequency ratio; Determining the search space for resource capacity of each node and each link in the network according to the optimized search space for resource capacity; The solution determination module is specifically used for: Determine a set of historical evaluation points; any evaluation point in the set of historical evaluation points represents a performance indicator under a certain network capacity solution; Determining a proxy model corresponding to this iteration based on the historical evaluation point set; the proxy model represents an implicit functional relationship between the network capacity solution and the performance indicator; Determining an acquisition function based on the proxy model; the acquisition function represents: the exploration value of each network capacity solution determined from the search space based on the proxy model; We will explore the network capacity solution with the greatest value and determine it as the network capacity solution for this round of iteration.

7. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of the network capacity planning method based on Bayesian optimization are implemented as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the network capacity planning method based on Bayesian optimization are implemented as described in any one of claims 1 to 5.

Images