A FlexE-based method for managing power communication resources in bearer networks

By constructing the FlexE bearer network topology and employing deep learning and multi-swarm optimization algorithms, the complexity, compatibility, and security issues of MS-OTN technology were resolved, enabling flexible management and scheduling of power communication resources and improving the stability and availability of power communication.

CN117857338BActive Publication Date: 2026-11-10CHINA SOUTHERN POWER GRID COMPANY
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202311679824.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-07
Publication Date
2026-11-10
Estimated Expiration
2043-12-07

AI Technical Summary

Technical Problem

Existing MS-OTN technology suffers from high technical complexity, compatibility issues, high investment costs, and insufficient security, making it difficult to meet the demands of power communication networks for high-speed, high-efficiency, and multi-type service processing capabilities.

Method used

A FlexE bearer network topology is constructed, and power communication equipment is connected through the FlexE bearer network topology. A deep learning model and a multi-group optimization algorithm are used to optimize the network status detection model, so as to realize flexible management and scheduling of power communication resources.

Benefits of technology

It enables flexible management and scheduling of power communication resources, improves network stability and availability, optimizes network performance and security, and is applicable to power monitoring, power metering and fault diagnosis in power systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117857338B_ABST
    Figure CN117857338B_ABST
Patent Text Reader

Abstract

This invention discloses a method for managing power communication resources in a FlexE-based bearer network, relating to the field of power system automation technology. By constructing a FlexE bearer network topology, connecting power communication equipment to this topology, setting different communication paths and traffic allocation rules, monitoring network status and power communication quality in real time, flexibly managing and scheduling power communication resources, and optimizing network performance and security, this method enables flexible management and scheduling of power communication resources, improving the stability and availability of power communication. Furthermore, it can be widely applied in various fields within the power system, such as power monitoring, energy metering, and fault diagnosis.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of power system automation technology, and specifically to a method for managing power communication resources in a bearer network based on FlexE. Background Technology

[0002] With the development of power communication networks, the requirements for networks are becoming increasingly stringent. Networks need not only to provide high-speed, efficient transmission but also to handle multiple types of services. Currently, only MS-OTN (Multi-service OTN) unified line board technology has been developed to meet these needs. As an advanced OTN technology, unified line board technology has many advantages, such as multi-plane service processing capabilities, multi-granularity multi-service carrying capabilities, and arbitrary ratio mixed transmission capabilities. However, it may also have some disadvantages, specifically: Technical complexity: It integrates multiple technologies, including SDH (Synchronous Digital Hierarchy), OTN (Optical Transport Network), and PTN (Packet Transport Network), resulting in relatively high technical complexity. This may place higher demands on the configuration and training of technical personnel and may increase the difficulty and cost of network operation and maintenance. Compatibility issues: Despite having multi-plane service processing and multi-granularity multi-service carrying capabilities, certain compatibility issues may still exist in practical applications. Interoperability between different manufacturers and different equipment may be limited, increasing the complexity of network deployment and management. High investment costs: The use of advanced wavelength block services and wavelength division multiplexing (WDM) technology results in relatively high equipment costs. This can lead to high initial investment costs in network construction, potentially impacting its market adoption and application. Security issues: Although MS-OTN utilizes OTN's built-in switching modules to achieve physical hard isolation between leased lines, ensuring they do not interfere with each other and providing high security, certain security issues may still exist in practical applications, such as network attacks and data breaches. Summary of the Invention

[0003] The purpose of this invention is to provide a method for managing power communication resources in a bearer network based on FlexE, which solves the problems existing in the prior art.

[0004] This invention is achieved through the following technical solution:

[0005] A method for managing power communication resources in a bearer network based on FlexE, comprising:

[0006] Construct the bearer network topology of FlexE and connect the power communication equipment to the bearer network topology of FlexE so that the power communication equipment can access the data management system through the bearer network topology of FlexE.

[0007] Based on preset rules or rules input through human-computer interaction, corresponding communication paths and traffic are allocated to different power communication devices, thereby completing the initialization of the power communication devices, the bearer network topology of FlexE, and the data management system.

[0008] A network status detection model is deployed in the data management system, and the communication quality of power communication equipment when it accesses the data management system is detected by the network status detection model to obtain the network status detection results.

[0009] Based on the network status detection results, a preset network optimization strategy and power communication resource scheduling strategy are obtained through a table lookup matching method, and the network optimization strategy and power communication resource scheduling strategy are executed to complete the resource management of power communication.

[0010] In one possible implementation, a network status detection model is deployed in the data management system, and the communication quality of power communication equipment accessing the data management system is detected using the network status detection model to obtain network status detection results, including:

[0011] A network state detection model is constructed using a deep learning model, and a multi-group optimization algorithm is used to optimize the network state detection model to obtain a network state detection model with network state detection capability.

[0012] Deploy a network status detection model in the data management system, and collect real-time network traffic characteristics when the power communication equipment accesses the data management system;

[0013] By deploying a network status detection model with network status detection capabilities, real-time network traffic characteristics are monitored to obtain network status detection results.

[0014] In one possible implementation, a multi-group optimization algorithm is used to optimize the network state detection model to obtain a network state detection model with network state detection capabilities, including:

[0015] A deep learning model is used to construct a network state detection model, and a chaotic mapping strategy is used to initialize the model parameters of the network state detection model to obtain the initial population;

[0016] The initial population is divided into four subpopulations in a ratio of 6:6:7:11, resulting in the first subpopulation, the second subpopulation, the third subpopulation, and the fourth subpopulation; the number of individuals in the initial population must be a multiple of 30.

[0017] Based on the training data, obtain the fitness value of each individual in the first subpopulation, the second subpopulation, the third subpopulation, and the fourth subpopulation;

[0018] Based on the fitness values ​​of each individual in the first, second, third, and fourth subpopulations, record the globally worst individual, the globally best individual, and the locally best individual corresponding to the fourth subpopulation.

[0019] Based on the worst individual in the world, the first global walk strategy is used to update each individual in the first subpopulation to obtain the updated first subpopulation.

[0020] The second global walk strategy is used to update each individual in the second subpopulation to obtain the updated second subpopulation;

[0021] Based on the globally optimal individual, the best region search strategy is used to update each individual in the third subpopulation, resulting in the updated third subpopulation.

[0022] Based on the globally optimal individual and the locally optimal individual corresponding to the fourth subpopulation, a balanced guided search strategy is used to update each individual in the fourth subpopulation to obtain the updated fourth subpopulation.

[0023] The updated first subpopulation, the updated second subpopulation, the updated third subpopulation, and the updated fourth subpopulation are subjected to boundary crossing processing to obtain the boundary-crossed first subpopulation, boundary-crossed second subpopulation, boundary-crossed third subpopulation, and boundary-crossed fourth subpopulation.

[0024] Update the record of the globally optimal individual based on the first subpopulation after the boundary crossing process, the second subpopulation after the boundary crossing process, the third subpopulation after the boundary crossing process, and the fourth subpopulation after the boundary crossing process.

[0025] Determine if the current update count is greater than the maximum update count. If so, output the globally optimal individual; otherwise, return the partitioning steps for the first, second, third, and fourth subpopulations.

[0026] The model parameters in the output of the globally optimal individual are used as the final model parameters of the network state detection model, thus obtaining a network state detection model with network state detection capability.

[0027] In one possible implementation, based on the fitness value of each individual in the first, second, third, and fourth subpopulations, the globally worst individual, the globally best individual, and the locally best individual corresponding to the fourth subpopulation are recorded, including:

[0028] The individual with the highest fitness value in the first, second, third, and fourth subpopulations is designated as the globally optimal individual, and the individual with the lowest fitness value is designated as the globally worst individual.

[0029] The individual with the highest fitness value in the fourth subpopulation is taken as its corresponding local optimum.

[0030] In one possible implementation, based on the globally worst individual, a first global walk strategy is used to update each individual in the first subpopulation to obtain an updated first subpopulation, including:

[0031] A random decision probability is generated between (0,1), and it is determined whether the random decision probability is less than or equal to a preset threshold. If so, each individual in the first subpopulation is globally updated based on the worst individual in the world to obtain the updated first subpopulation. Otherwise, the individuals in the first subpopulation are updated based on their historical states to obtain the updated first subpopulation.

[0032] Specifically, based on the globally worst individual, a global update is performed on each individual in the first subpopulation as follows:

[0033]

[0034] In the formula, Let i represent the i-th first individual in the first subpopulation during the t-th update, where i = 1, 2, ..., I, and I represents the total number of first individuals in the first subpopulation. Indicates the updated α represents a constant term with a value of 1 or -1, and k represents the deflection coefficient, where k is a constant term between (0, 0.2). Let represent the i-th individual in the first subpopulation during the (t-1)-th update, levy(λ) represent Levy's flight path, and b represent a random number between (0,1). This represents the globally worst individual during the t-th update process;

[0035] The individuals in the first subpopulation are updated based on their historical states, resulting in the updated first subpopulation:

[0036]

[0037] Where θ represents the update direction control factor between [0, π].

[0038] In one possible implementation, the second global walk strategy is used to update each individual in the second subpopulation as follows:

[0039]

[0040] in, Let J represent the j-th second individual in the second subpopulation during the t-th update, where j = 1, 2, ..., J, and J represents the total number of second individuals in the second subpopulation. This represents the j-th second individual in the second subpopulation during the (t-1)-th update process. Indicates the updated

[0041] In one possible implementation, based on the globally optimal individual, an optimal region search strategy is used to update each individual in the third subpopulation, resulting in an updated third subpopulation, including:

[0042] The optimal search area is determined as follows:

[0043]

[0044]

[0045] Where lb' represents the lower bound of the optimal search region, and ub' represents the upper bound of the optimal search region. Let R represent the globally optimal individual in the t-th update process, R represent the variable parameters, and R = 1 - t / T, T represent the maximum number of updates, lb represent the lower bound of the model parameters of the automatic verification model, and ub represent the upper bound of the model parameters of the automatic verification model.

[0046] Based on the optimal search region, each individual in the third subpopulation is updated as follows:

[0047]

[0048] in, Let m represent the m-th third individual in the third subpopulation during the t-th update, where m = 1, 2, ..., M, and M represents the total number of third individuals in the third subpopulation. Indicates the updated c1 represents a random number between (0,1), and c2 represents a random number between (0,1).

[0049] In one possible implementation, based on the globally optimal individual and the locally optimal individual corresponding to the fourth subpopulation, a balance-guided search strategy is used to update each individual in the fourth subpopulation as follows:

[0050]

[0051] in, Let represent the nth fourth individual in the fourth subpopulation during the t-th update, where n = 1, 2, ..., N, and N represents the total number of third individuals in the third subpopulation. Indicates the updated S represents the coefficient of the constant term, g represents a random number following a uniform distribution (0,1), and w(t) represents the inertia weight. This represents the locally optimal individual corresponding to the fourth subpopulation; e represents the natural constant.

[0052] In one possible implementation, boundary crossing handling is performed on the updated first subpopulation, the updated second subpopulation, the updated third subpopulation, and the updated fourth subpopulation, including:

[0053] Get the target individuals that have crossed the boundary in the updated first subpopulation, the updated second subpopulation, the updated third subpopulation, and the updated fourth subpopulation;

[0054] Determine the out-of-bounds dimension d for each of the target out-of-bounds individuals, and update the out-of-bounds dimension d for each of the target out-of-bounds individuals as follows:

[0055]

[0056] in, This represents the out-of-bounds dimension d among the updated target out-of-bounds individuals. denoted as the parameter of the same dimension as the out-of-bounds dimension d in the globally optimal individual during the t-th update process, randn represents a random number that follows a standard normal distribution, and γ represents the scaling factor.

[0057] This invention provides a FlexE-based method for managing power communication resources in a bearer network. By constructing a FlexE bearer network topology, connecting power communication equipment to this topology, setting different communication paths and traffic allocation rules, monitoring network status and power communication quality in real time, flexibly managing and scheduling power communication resources, and optimizing network performance and security, this method enables flexible management and scheduling of power communication resources, improving the stability and availability of power communication. Furthermore, it can be widely applied in various fields of power systems, such as power monitoring, energy metering, and fault diagnosis. Attached Figure Description

[0058] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0059] Figure 1 A flowchart illustrating a method for managing power communication resources in a bearer network based on FlexE, provided as an embodiment of the present invention. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0061] With the development of smart grids, the scale and complexity of power communication networks are constantly increasing, placing higher demands on the management and maintenance of power communication network resources. Currently, the management of power communication network resources still faces some problems, such as low utilization rate of communication resources, unstable quality, and insecure data transmission. Therefore, this invention provides a FlexE-based method for managing power communication resources in bearer networks to address the problems existing in the prior art.

[0062] like Figure 1 As shown, a method for managing power communication resources in a bearer network based on FlexE includes:

[0063] S101. Construct the bearer network topology of FlexE and connect the power communication equipment to the bearer network topology of FlexE so that the power communication equipment can access the data management system through the bearer network topology of FlexE.

[0064] S102. Based on preset rules or rules input by human-computer interaction, allocate corresponding communication paths and traffic to different power communication devices, and complete the initialization of the power communication devices, the bearer network topology of FlexE, and the data management system.

[0065] Power communication equipment can be connected to the FlexE bearer network topology, such as power monitoring cameras and power monitoring equipment.

[0066] In this embodiment, multiple power monitoring cameras and power monitoring devices are connected to the FlexE bearer network. By configuring the port and IP (Internet Protocol) address of the network devices, the connection and data transmission between the power communication devices and the FlexE bearer network are realized.

[0067] Depending on the type of power communication equipment and its communication requirements, different communication paths and traffic allocation rules can be configured in the FlexE bearer network to ensure the security and availability of power communication resources. For example, for power monitoring cameras with high real-time requirements, higher-priority communication paths and traffic allocation rules are set to ensure the stability and real-time performance of their data transmission. Conversely, for power monitoring equipment with lower real-time requirements, lower-priority communication paths and traffic allocation rules are set to fully utilize network bandwidth resources.

[0068] The communication data of power communication equipment can be transmitted to the data management system via the FlexE bearer network. This system can monitor network status and power communication quality in real time, while also monitoring and managing power communication data traffic. Through the data management system, problems in the network can be detected promptly, ensuring the stability and security of power communication.

[0069] S103. Deploy a network status detection model in the data management system, and use the network status detection model to detect the communication quality when power communication equipment accesses the data management system, and obtain the network status detection results.

[0070] The system monitors the FlexE Ethernet network status and power communication quality in real time through a data management system, while also monitoring and managing power communication data traffic. If any anomalies or faults are detected in the network, the system can promptly issue alarms or take corresponding measures to ensure the stability and security of power communication.

[0071] S104. Based on the network status detection results, obtain the preset network optimization strategy and power communication resource scheduling strategy through a table lookup matching method, and execute the network optimization strategy and power communication resource scheduling strategy to complete the power communication resource management.

[0072] Based on actual needs, flexible management and scheduling of power communication resources can be achieved by configuring network devices and management systems. This includes scheduling and controlling power communication data traffic, and configuring and managing power communication equipment. For example, flexible configuration and management of power communication equipment can be achieved by configuring network device ports and IP addresses. Furthermore, flexible scheduling and control of power communication data traffic can be achieved by configuring the QoS (Quality of Service) function of network devices.

[0073] Adjustments are made to network performance and security based on actual operating conditions to ensure the stability and availability of power communication. For example, network performance and security can be optimized by adjusting network device parameters and configurations. Additionally, network performance and security can be improved by upgrading network devices and software versions.

[0074] In this embodiment, the network status detection result can be the result corresponding to a single power communication device or the result of multiple power communication devices. Different network status detection results correspond to different preset network optimization strategies and power communication resource scheduling strategies, thereby achieving optimization.

[0075] Optionally, to further optimize the process, this embodiment provides a method for determining network optimization strategies and power communication resource scheduling strategies. Specifically, it involves: acquiring different network status detection results for different power communication devices; constructing a data vector or data matrix from these network status detection results; assigning category labels for network optimization strategies to the data vector or data matrix using manually defined labels; and then using deep learning technology to learn the correlation between the data vector or data matrix and the category labels. This allows for the identification and execution of network status detection results during subsequent communication processes. The method for acquiring power communication resource scheduling strategies is similar and will not be described in detail in this embodiment.

[0076] This invention provides a FlexE-based method for managing power communication resources in a bearer network. By constructing a FlexE bearer network topology, connecting power communication equipment to this topology, setting different communication paths and traffic allocation rules, monitoring network status and power communication quality in real time, flexibly managing and scheduling power communication resources, and optimizing network performance and security, this method enables flexible management and scheduling of power communication resources, improving the stability and availability of power communication. Furthermore, it can be widely applied in various fields of power systems, such as power monitoring, energy metering, and fault diagnosis.

[0077] In one possible implementation, a network status detection model is deployed in the data management system, and the communication quality of power communication equipment accessing the data management system is detected using the network status detection model to obtain network status detection results, including:

[0078] A network state detection model is constructed using a deep learning model, and a multi-group optimization algorithm is used to optimize the network state detection model to obtain a network state detection model with network state detection capabilities.

[0079] A network status detection model is deployed in the data management system, and real-time network traffic characteristics are collected when power communication equipment accesses the data management system. The collection of network traffic characteristics is relatively conventional in existing technologies, and will not be described in detail in this embodiment.

[0080] By deploying a network status detection model with network status detection capabilities, real-time network traffic characteristics are monitored to obtain network status detection results. Obtaining these results effectively ensures network security and enables network optimization.

[0081] In one possible implementation, a multi-group optimization algorithm is used to optimize the network state detection model to obtain a network state detection model with network state detection capabilities, including:

[0082] A deep learning model is used to construct a network state detection model, and a chaotic mapping strategy is used to initialize the model parameters of the network state detection model to obtain the initial population.

[0083] Deep learning models can be configured as convolutional neural networks, backpropagation (BP) neural networks, or other neural network models. In existing technologies, to enable deep learning models to process and recognize data, parameter optimization is often required. Current techniques commonly employ optimization algorithms such as gradient descent, particle swarm optimization, genetic algorithms, and ant colony optimization. However, these algorithms suffer from slow convergence speed and poor convergence accuracy, ultimately leading to poor parameter optimization performance and data recognition capabilities in the deep learning model. Therefore, this invention proposes a multi-swarm optimization algorithm to improve the convergence speed and accuracy of deep learning models, thereby enhancing their data recognition capabilities.

[0084] Initializing the model parameters of the network state detection model using a chaotic mapping strategy can make the distribution of individual solutions in the solution space more uniform at the beginning of the algorithm, thereby improving the convergence speed.

[0085] Optionally, to ensure proper data identification, real-time network traffic characteristics need to be processed into the input form of the network status detection model, and non-numerical data needs to be processed into numerical data.

[0086] The initial population is divided into four subpopulations in a ratio of 6:6:7:11, resulting in subpopulations 1, 2, 3, and 4. The number of individuals in each initial population is a multiple of 30. By ensuring that the number of individuals in the initial population is a multiple of 30, each subpopulation receives an integer number of individuals during the subpopulation process.

[0087] Based on the training data, obtain the fitness value of each individual in the first subpopulation, the second subpopulation, the third subpopulation, and the fourth subpopulation.

[0088] Training data can include historical network traffic features and corresponding network state detection results. Historical network traffic features can be used as input, and the corresponding network state detection results can be used to construct the desired output, thus enabling the training of a deep learning model. The network state detection results can be set according to actual needs, or simplified to two cases: abnormal network state or normal network state.

[0089] Optionally, the fitness value can be the reciprocal of the individual error function value. To avoid the denominator being zero when taking the reciprocal, the individual error function value can be added to a constant before taking the reciprocal. Different neural networks can choose different error functions, which will not be described in detail in this embodiment.

[0090] Based on the fitness values ​​of each individual in the first, second, third, and fourth subpopulations, record the globally worst individual, the globally best individual, and the locally best individual corresponding to the fourth subpopulation.

[0091] Based on the worst individual in the global population, the first global walk strategy is used to update each individual in the first subpopulation, resulting in the updated first subpopulation.

[0092] The second global walk strategy is used to update each individual in the second subpopulation, resulting in the updated second subpopulation.

[0093] Based on the globally optimal individual, the best region search strategy is used to update each individual in the third subpopulation, resulting in the updated third subpopulation.

[0094] Based on the globally optimal individual and the locally optimal individual corresponding to the fourth subpopulation, a balanced guided search strategy is used to update each individual in the fourth subpopulation, resulting in the updated fourth subpopulation.

[0095] The updated first subpopulation, the updated second subpopulation, the updated third subpopulation, and the updated fourth subpopulation are subjected to boundary crossing processing to obtain the boundary-crossed first subpopulation, boundary-crossed second subpopulation, boundary-crossed third subpopulation, and boundary-crossed fourth subpopulation.

[0096] Update the record of the globally optimal individual based on the first, second, third, and fourth subpopulations after the boundary crossing process.

[0097] Determine if the current update count is greater than the maximum update count. If so, output the globally optimal individual; otherwise, return the partitioning steps for the first, second, third, and fourth subpopulations.

[0098] The model parameters in the output of the globally optimal individual are used as the final model parameters of the network state detection model, thus obtaining a network state detection model with network state detection capability.

[0099] Updating a single population can easily lead to getting stuck in local optima. This embodiment uses a multi-population update method, which has a certain probability of retaining some individuals that always execute the same update strategy. At the same time, the combination of global search, balanced search and local search terms is more conducive to finding the global optimum, which not only improves the convergence speed but also improves the convergence accuracy.

[0100] In one possible implementation, based on the fitness value of each individual in the first, second, third, and fourth subpopulations, the globally worst individual, the globally best individual, and the locally best individual corresponding to the fourth subpopulation are recorded, including:

[0101] The individual with the highest fitness value in the first, second, third, and fourth subpopulations is designated as the globally optimal individual, and the individual with the lowest fitness value is designated as the globally worst individual.

[0102] The individual with the highest fitness value in the fourth subpopulation is taken as its corresponding local optimum.

[0103] In one possible implementation, based on the globally worst individual, a first global walk strategy is used to update each individual in the first subpopulation to obtain an updated first subpopulation, including:

[0104] A random decision probability is generated between (0,1), and it is determined whether the random decision probability is less than or equal to a preset threshold (for example, it can be set to 0.9, so that a global search is performed with a higher probability and the search direction is changed with a lower probability). If so, each individual in the first subpopulation is globally updated according to the worst individual in the world to obtain the updated first subpopulation. Otherwise, the individuals are updated according to the historical state of the individuals in the first subpopulation to obtain the updated first subpopulation.

[0105] Specifically, based on the globally worst individual, a global update is performed on each individual in the first subpopulation as follows:

[0106]

[0107] In the formula, Let i represent the i-th first individual in the first subpopulation during the t-th update, where i = 1, 2, ..., I, and I represents the total number of first individuals in the first subpopulation. Indicates the updated α represents a constant term with a value of 1 or -1, and k represents the deflection coefficient, where k is a constant term between (0, 0.2). Let represent the i-th individual in the first subpopulation during the (t-1)-th update, levy(λ) represent Levy's flight path, and b represent a random number between (0,1). This represents the worst individual globally during the t-th update process.

[0108] The multidimensional parameter optimization problem is actually a multi-peaked optimization problem, which is prone to getting trapped in local optima. To avoid getting trapped in local optima, we use historical states as a reference and perturb through the global worst position and Lévy flight to finally update the first volume, thereby enabling effective global search.

[0109] The individuals in the first subpopulation are updated based on their historical states, resulting in the updated first subpopulation:

[0110]

[0111] Here, θ represents the update direction control factor between [0, π]. If the search continues in only one direction, the algorithm may stagnate. Therefore, this embodiment provides a method for updating based on historical states, which can effectively change the search direction of individuals and is more beneficial for global search.

[0112] In one possible implementation, the second global walk strategy is used to update each individual in the second subpopulation as follows:

[0113]

[0114] in, Let J represent the j-th second individual in the second subpopulation during the t-th update, where j = 1, 2, ..., J, and J represents the total number of second individuals in the second subpopulation. This represents the j-th second individual in the second subpopulation during the (t-1)-th update process. Indicates the updated Similarly, the second subpopulation updates individuals by changing their search direction in conjunction with Levi's flight, thus achieving a global search.

[0115] In one possible implementation, based on the globally optimal individual, an optimal region search strategy is used to update each individual in the third subpopulation, resulting in an updated third subpopulation, including:

[0116] The optimal search area is determined as follows:

[0117]

[0118]

[0119] Where lb' represents the lower bound of the optimal search region, and ub' represents the upper bound of the optimal search region. Let R represent the globally optimal individual in the t-th update process, R represent the variable parameters and R = 1 - t / T, T represent the maximum number of updates, lb represent the lower bound of the model parameters of the automatic verification model, and ub represent the upper bound of the model parameters of the automatic verification model.

[0120] Based on the optimal search region, each individual in the third subpopulation is updated as follows:

[0121]

[0122] in, Let m represent the m-th third individual in the third subpopulation during the t-th update, where m = 1, 2, ..., M, and M represents the total number of third individuals in the third subpopulation. Indicates the updated c1 represents a random number between (0,1), and c2 represents a random number between (0,1).

[0123] By setting an optimal search area and having individuals from the third subpopulation search within this optimal search area, the search effect and accuracy can be effectively guaranteed.

[0124] In one possible implementation, based on the globally optimal individual and the locally optimal individual corresponding to the fourth subpopulation, a balance-guided search strategy is used to update each individual in the fourth subpopulation as follows:

[0125]

[0126] in, Let represent the nth fourth individual in the fourth subpopulation during the t-th update, where n = 1, 2, ..., N, and N represents the total number of third individuals in the third subpopulation. Indicates the updated S represents the coefficient of the constant term, g represents a random number following a uniform distribution (0,1), and w(t) represents the inertia weight. This represents the locally optimal individual corresponding to the fourth subpopulation. e represents the natural constant.

[0127] This embodiment introduces adaptive inertia weights to improve solution quality. These adaptive inertia weights are exponential; larger weights result in stronger search performance and enhanced global search capabilities. As the number of updates increases, precise searches can be performed around the optimal solution, strengthening local search capabilities and effectively balancing search diversity with convergence accuracy.

[0128] In one possible implementation, boundary crossing handling is performed on the updated first subpopulation, the updated second subpopulation, the updated third subpopulation, and the updated fourth subpopulation, including:

[0129] Get the target outbound individuals that exhibit outbound behavior in the updated first subpopulation, updated second subpopulation, updated third subpopulation, and updated fourth subpopulation.

[0130] Determine the out-of-bounds dimension d for each of the target out-of-bounds individuals, and update the out-of-bounds dimension d for each of the target out-of-bounds individuals as follows:

[0131]

[0132] in, This represents the out-of-bounds dimension d among the updated target out-of-bounds individuals. denoted as the parameter of the same dimension as the out-of-bounds dimension d in the globally optimal individual during the t-th update process, randn represents a random number that follows a standard normal distribution, and γ represents the scaling factor.

[0133] The out-of-bounds handling method guided by the globally optimal individual provided in this embodiment of the invention reduces fluctuations during the tracking process and can effectively improve the convergence speed and accuracy of the algorithm.

[0134] To verify the effectiveness of the proposed FlexE-based power communication resource management method for bearer networks, a series of experiments were conducted. One hundred power monitoring cameras and 50 power monitoring devices were used as experimental subjects. These devices were connected to the FlexE bearer network and monitored and managed through a network monitoring and management system. During the experiments, indicators such as network status, power communication quality, and data traffic were monitored and recorded in real time. Simultaneously, communication paths and traffic allocation rules for different types of devices were set and optimized. Experimental results show that the method of this invention can effectively improve the stability and availability of power communication. Regarding network status, the method of this invention can achieve fast network connection and stable transmission. Regarding power communication quality, the method of this invention can achieve high-quality data transmission. Regarding data traffic, the method of this invention can achieve flexible data traffic scheduling and control.

[0135] In summary, the method of this invention can be widely applied to various fields in power systems, such as power monitoring, energy metering, and fault diagnosis. By constructing the FlexE bearer network topology, connecting power communication equipment to the FlexE bearer network topology, setting different communication paths and traffic allocation rules, real-time monitoring of network status and power communication quality, flexible management and scheduling of power communication resources, and optimization of network performance and security, flexible management and scheduling of power communication resources can be achieved, improving the stability and availability of power communication. Furthermore, the method of this invention also has high scalability and customizability, and can meet the actual needs of different power systems.

[0136] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0137] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0138] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0139] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0140] Those skilled in the art will understand that all or part of the steps in the above facts and methods can be implemented by a program instructing related hardware. The program or the program described therein can be stored in a computer-readable storage medium. When the program is executed, it includes the following steps: at this time, the corresponding method steps are introduced. The storage medium can be ROM / RAM, magnetic disk, optical disk, etc.

[0141] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for managing FlexE-based bearer network power communication resources, characterized in that, include: Construct the bearer network topology of FlexE and connect the power communication equipment to the bearer network topology of FlexE so that the power communication equipment can access the data management system through the bearer network topology of FlexE. Based on preset rules or rules input through human-computer interaction, corresponding communication paths and traffic are allocated to different power communication devices, thereby completing the initialization of the power communication devices, the bearer network topology of FlexE, and the data management system. A network status detection model is deployed in the data management system, and the communication quality of power communication equipment when it accesses the data management system is detected by the network status detection model to obtain the network status detection results. Based on the network status detection results, a preset network optimization strategy and power communication resource scheduling strategy are obtained through a table lookup matching method, and the network optimization strategy and power communication resource scheduling strategy are executed to complete the resource management of power communication. The process involves deploying a network status detection model within the data management system and using this model to detect the communication quality of power communication equipment accessing the data management system, thereby obtaining network status detection results, including: A network state detection model is constructed using a deep learning model, and a multi-group optimization algorithm is used to optimize the network state detection model to obtain a network state detection model with network state detection capability. Deploy a network status detection model in the data management system, and collect real-time network traffic characteristics when the power communication equipment accesses the data management system; By deploying a network status detection model with network status detection capabilities, real-time network traffic characteristics are monitored to obtain network status detection results. The optimization of the network state detection model using multiple swarm optimization algorithms to obtain a network state detection model with network state detection capabilities includes: A deep learning model is used to construct a network state detection model, and a chaotic mapping strategy is used to initialize the model parameters of the network state detection model to obtain the initial population; The initial population is divided into four subpopulations in a ratio of 6:6:7:11, resulting in the first subpopulation, the second subpopulation, the third subpopulation, and the fourth subpopulation; the number of individuals in the initial population must be a multiple of 30. Based on the training data, obtain the fitness value of each individual in the first subpopulation, the second subpopulation, the third subpopulation, and the fourth subpopulation; Based on the fitness values ​​of each individual in the first, second, third, and fourth subpopulations, record the globally worst individual, the globally best individual, and the locally best individual corresponding to the fourth subpopulation. Based on the worst individual in the world, the first global walk strategy is used to update each individual in the first subpopulation to obtain the updated first subpopulation. The second global walk strategy is used to update each individual in the second subpopulation to obtain the updated second subpopulation; Based on the globally optimal individual, the best region search strategy is used to update each individual in the third subpopulation, resulting in the updated third subpopulation. Based on the globally optimal individual and the locally optimal individual corresponding to the fourth subpopulation, a balanced guided search strategy is used to update each individual in the fourth subpopulation to obtain the updated fourth subpopulation. The updated first subpopulation, the updated second subpopulation, the updated third subpopulation, and the updated fourth subpopulation are subjected to boundary crossing processing to obtain the boundary-crossed first subpopulation, boundary-crossed second subpopulation, boundary-crossed third subpopulation, and boundary-crossed fourth subpopulation. Update the record of the globally optimal individual based on the first subpopulation after the boundary crossing process, the second subpopulation after the boundary crossing process, the third subpopulation after the boundary crossing process, and the fourth subpopulation after the boundary crossing process. Determine if the current update count is greater than the maximum update count. If so, output the globally optimal individual; otherwise, return the partitioning steps for the first, second, third, and fourth subpopulations. The model parameters in the output of the globally optimal individual are used as the final model parameters of the network state detection model, thus obtaining a network state detection model with network state detection capability.

2. The FlexE-based bearer network power communication resource management method according to claim 1, characterized in that, Based on the fitness values ​​of each individual in the first, second, third, and fourth subpopulations, record the globally worst individual, the globally best individual, and the locally best individual corresponding to the fourth subpopulation, including: The individual with the highest fitness value in the first, second, third, and fourth subpopulations is designated as the globally optimal individual, and the individual with the lowest fitness value is designated as the globally worst individual. The individual with the highest fitness value in the fourth subpopulation is taken as its corresponding local optimum.

3. The FlexE-based bearer network power communication resource management method according to claim 1, characterized in that, Based on the globally worst individual, the first global walk strategy is used to update each individual in the first subpopulation, resulting in the updated first subpopulation, including: A random decision probability is generated between (0,1), and it is determined whether the random decision probability is less than or equal to a preset threshold. If so, each individual in the first subpopulation is globally updated based on the worst individual in the world to obtain the updated first subpopulation. Otherwise, the individuals in the first subpopulation are updated based on their historical states to obtain the updated first subpopulation. Specifically, based on the globally worst individual, a global update is performed on each individual in the first subpopulation as follows: ; wherein, denotes the i-th first individual of the first subpopulation in the t-th update process, i = 1, 2, …, I, I denotes the total number of first individuals in the first subpopulation, denotes the updated , denotes a constant term with value 1 or -1, k denotes a deflection coefficient, and k is a constant term between (0, 0.2), denotes the i-th first individual of the first subpopulation in the t-1-th update process, denotes a Levy flight path, denotes a random number between (0, 1), denotes the global worst individual in the t-th update process; The individuals in the first subpopulation are updated based on their historical states, resulting in the updated first subpopulation: ; wherein denotes an update direction control factor between [0, π].

4. The FlexE-based bearer network power communication resource management method according to claim 3, characterized in that, The second global walk strategy is used to update each individual in the second subpopulation as follows: ; wherein, denotes the jth second individual of the second subpopulation in the tth update process, j = 1, 2, …, J, J denotes the total number of second individuals in the second subpopulation, denotes the jth second individual of the second subpopulation in the t-1th update process, denotes the updated .

5. The FlexE-based bearer network power communication resource management method according to claim 4, characterized in that, Based on the globally optimal individual, the best region search strategy is used to update each individual in the third subpopulation, resulting in the updated third subpopulation, which includes: The optimal search area is determined as follows: ; ; wherein, represents a lower limit of the optimal search region, represents an upper limit of the optimal search region, represents a globally optimal individual in the tth update process, represents a variable parameter, and = 1 - t / T, T represents a maximum number of updates, represents a lower limit of a model parameter of the automatic verification model, represents an upper limit of a model parameter of the automatic verification model; Based on the optimal search region, each individual in the third subpopulation is updated as follows: ; wherein, denotes the m-th third individual of the third subpopulation in the t-th update process, m = 1, 2, …, M, M denotes the total number of third individuals in the third subpopulation, denotes the updated , denotes a random number between (0, 1), denotes a random number between (0, 1).

6. The FlexE-based bearer network power communication resource management method according to claim 5, characterized in that, Based on the globally optimal individual and the locally optimal individual corresponding to the fourth subpopulation, a balance-guided search strategy is used to update each individual in the fourth subpopulation as follows: ; wherein, denotes the nth fourth individual of the fourth subpopulation in the tth update process, n = 1, 2, …, N, N denotes the total number of third individuals in the third subpopulation, denotes the updated , S denotes the coefficient of the constant term, denotes a random number subject to a (0, 1) uniform distribution, denotes an inertia weight, denotes the local optimal individual corresponding to the fourth subpopulation; , denotes a natural constant.

7. The FlexE-based bearer network power communication resource management method according to claim 6, characterized in that, Out-of-bounds handling is performed on the updated first subpopulation, the updated second subpopulation, the updated third subpopulation, and the updated fourth subpopulation, including: Get the target individuals that have crossed the boundary in the updated first subpopulation, the updated second subpopulation, the updated third subpopulation, and the updated fourth subpopulation; Determine the out-of-bounds dimension d for each of the target out-of-bounds individuals, and update the out-of-bounds dimension d for each of the target out-of-bounds individuals as follows: ; wherein, represents the out-of-bound dimension d in the updated target out-of-bound individual, represents the same-dimension parameter in the global optimal individual in the tth update process with the out-of-bound dimension d, represents a random number satisfying the standard normal distribution, represents a scale factor.

Citation Information

Patent Citations

  • Computing unloading and resource optimization method in ultra-dense network

    CN116155728A

  • Intelligent industrial gateway design method and device

    CN117118781A