Edge computing optimization method based on big data network
By collecting and preprocessing data in an edge computing environment, performing distributed storage and machine learning analysis, and using a fuzzy comprehensive evaluation model to evaluate equipment capabilities and dynamically schedule tasks, the coordination problem of edge computing resource allocation is solved, and low latency is achieved. Highly efficient data processing and resource utilization.
Patent Information
- Application Number
- CN202411012573.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-12-03
AI Technical Summary
The existing edge computing resource allocation methods lack consideration for the overall network resources and fail to effectively optimize the coordination between storage, computing and transmission resources, resulting in low resource utilization and high latency.
Data is collected and preprocessed in an edge computing environment, stored through distributed file systems and in-memory databases, data analysis is performed using machine learning algorithms, and edge device capabilities are evaluated using fuzzy comprehensive evaluation models to dynamically schedule tasks to match resource requirements.
It reduces data transmission latency, improves system performance and resource utilization, reduces the risk of data leakage, enhances data privacy and security, and realizes dynamic evaluation and task matching of edge device capabilities.
Smart Images

Figure CN119065831A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of edge computing technology, and in particular to an edge computing optimization method based on a big data network. Background Art
[0002] Nowadays, we use technologies to improve data processing efficiency and reduce latency by processing and analyzing data at the edge of the network. The main background technologies of this approach include big data processing, network communication technology, and edge computing. Big data processing technology is used to process and analyze large-scale data sets, network communication technology is used to transmit data between the cloud and edge devices, and edge computing technology is used to process and analyze data at the edge of the network.
[0003] At present, there are many resource allocation methods for edge computing, but they only consider the allocation of a few resources, such as limited storage resources or limited computing resources, and lack consideration of the overall network resources. The storage, computing and transmission resources in the wireless communication network should be analyzed as a whole, and the resources can also be "exchanged" with each other through the adaptation mechanism to achieve the purpose of overall optimization. Secondly, the resource allocation technology in the current research is only analyzed for one or several resource allocation technologies, such as how to store the most popular resources and how to manage wireless resources. However, the research on these technologies lacks relevance, which is not only related to the user's behavior, but also to the network's capabilities, so further optimization of edge computing is needed. Summary of the invention
[0004] The purpose of the present invention is to solve the shortcomings of the above-mentioned prior art and provide an edge computing optimization method based on a big data network, comprising the following steps:
[0005] S1: Collect network data in the edge computing environment and pre-process the collected data, including data cleaning, deduplication and format conversion;
[0006] S2: storing and analyzing the preprocessed data, wherein the data analysis includes data exploration, data mining and prediction model;
[0007] S3: Decompose tasks, schedule tasks according to data processing requirements and the assessed edge device capabilities, and reasonably distribute computing tasks to the edge devices;
[0008] S4: Collect feedback on task execution results and provide feedback to users, and adjust the edge computing environment based on the results of task completion.
[0009] Wherein, in step S1, the network data includes user behavior, device status data and environmental data, and the network data is collected by sensors, cameras and intelligent collection devices.
[0010] Furthermore, in step S1, the data cleaning includes removing noise and irrelevant data generated during the collection process, and deleting or filling the missing values;
[0011] The data deduplication monitors and deletes duplicate data by using a hash value comparison method to ensure the uniqueness of the data;
[0012] The format conversion converts data from different sources into a unified format to facilitate subsequent data processing and analysis, and at the same time standardizes the timestamp as needed and adjusts the data accuracy to meet analysis requirements.
[0013] Further, in step S2, the storage process includes:
[0014] S21: Select to use a distributed file system as a data storage medium, and import the pre-processed data into a designated storage system;
[0015] S22: performing version control, backup, recovery and security protection on the imported data;
[0016] S23: merging the pre-processed data from different sources and creating an index for the data;
[0017] S24: Use an in-memory database to cache frequently accessed data.
[0018] Further, in step S2, the data in the data analysis includes:
[0019] Explore the use of statistical analysis and visualization tools for preliminary data exploration, quickly obtain the basic statistical characteristics of the data set, display the data distribution through histograms, box plots, and scatter plots, identify patterns, anomalies, and data clustering trends in the data, and analyze the correlation between different variables through matrices and heat maps to avoid the impact of multicollinearity on subsequent models.
[0020] Further, in step S2, the data mining in the data analysis uses a machine learning algorithm to mine patterns in the data, and the machine learning algorithm used includes a decision tree, a support vector machine or a neural network algorithm, specifically including:
[0021] The decision tree algorithm constructs a tree structure by recursively dividing the data set into different subsets, each node represents a feature, each branch represents a decision rule for a feature, and each leaf node represents a prediction result;
[0022] The support vector machine maps the original data into a high-dimensional space by using a kernel function, and finds a hyperplane that maximizes the interval between two types of data;
[0023] The neural network algorithm introduces nonlinear characteristics through activation functions to increase the learning of complex patterns from data.
[0024] Further, in step S2, the predictive modeling in the data analysis includes:
[0025] Clarify the forecasting target, collect historical data, build a forecasting model based on historical data, use regression analysis or time series analysis methods, select an appropriate model, and use the fitted model to perform forecasting analysis on relevant data.
[0026] Furthermore, in step S3, the task scheduling includes the following steps:
[0027] S31: Identify the tasks that need to be performed and break them down into smaller subtasks;
[0028] S32: Evaluate the capabilities of the edge devices, assign the decomposed subtasks to the most suitable edge devices according to the evaluation results, and perform scheduling according to the shortest job first strategy;
[0029] S33: Monitor the execution status of the task on the edge device and make timely adjustments as needed.
[0030] Preferably, in step S32, the capabilities of the edge device include processing capability, storage space, and battery life, and the evaluation of the edge device capabilities adopts a fuzzy comprehensive evaluation model, including the following steps:
[0031] S321: Combine the hierarchical structure to be analyzed and establish a factor set Q = {c1, c2 ..., c m}, where each single factor subset c i It also includes several elements;
[0032] S322: Establishing an evaluation result set V = {V1, V2..., V K}, take the median of each interval as the grade parameter, then the corresponding evaluation grade parameter column vector is
[0033] S323: Evaluate the level of each indicator of the specific object, establish a fuzzy relationship matrix, evaluate each indicator, grade the actual situation of the qualitative indicator, and then count the frequency of comments at different levels for each indicator, that is, obtain the C of each subset. i The single factor evaluation matrix R i ;
[0034] S324: Perform single factor evaluation, and calculate the single factors c i , and each single factor ci The weight coefficient vector W i , and the single factor evaluation matrix R i , perform synthesis operation: B = W i o R i ;
[0035] S325: Repeat S324 to perform a comprehensive evaluation of multiple factors, and obtain a comprehensive evaluation matrix for each subset in Q: R = (B1, B2, ..., B m ) T , until the fuzzy comprehensive evaluation result of the criterion layer relative to the target is obtained: E = W c o R c =(e i ,e2,…,e n ), where E is a membership result vector of factor set Q, and the final evaluation result is the algebraic value obtained by multiplying E and the evaluation level parameter column vector V: Where Z represents the overall capability assessment result of the assessment object.
[0036] Compared with the prior art, the present invention has the following beneficial effects:
[0037] (1) The present invention can greatly reduce the distance and time of data transmission by processing data on edge devices near the data source, thereby reducing latency. This is of great significance for applications with high real-time requirements, such as autonomous driving and industrial automation. Edge computing can reduce the transmission of large amounts of data from data sources to cloud servers, thereby reducing bandwidth consumption. This is very beneficial for application scenarios with limited network bandwidth. Edge computing can distribute computing tasks across multiple edge devices, thereby improving system performance and throughput. At the same time, edge devices can take advantage of their proximity to data sources to perform faster and more accurate data processing. Edge computing can perform data processing near the data source, reducing data transmission in the network, thereby reducing the risk of data leakage. Edge devices can implement local storage and processing of data, enhancing data privacy and security.
[0038] (2) By decomposing tasks, the present invention can more easily assign them to different computing resources, ensuring that tasks match the capabilities of execution resources to optimize performance and energy efficiency. Task characteristics and device capabilities are used for dynamic matching to achieve the lowest response time and the highest resource utilization. By selecting an appropriate scheduling strategy, execution speed and resource utilization can be balanced in a multi-tasking environment. Real-time monitoring helps to discover problems and reallocate tasks in a timely manner to maintain the overall performance of the system. By reasonably allocating tasks, the load of all devices is close to the average value to avoid overloading some devices while others are idle. Computing tasks are assigned to devices as close to the data source as possible to reduce data transmission time and delay. Considering the energy consumption of the device, tasks are assigned to devices with higher energy efficiency to extend the service life of the device. The reliability and stability of the device are considered when assigning tasks to ensure that the system can still operate normally in the face of device failures.
[0039] (3) The present invention uses a fuzzy evaluation model to evaluate the capabilities of edge devices, allowing evaluation indicators and weights to be adjusted according to actual needs. It can adapt to the requirements for edge device capability evaluation in different scenarios, and can realize dynamic evaluation of edge device capabilities. As the performance of the device changes, the evaluation indicators and weights can be adjusted in time to reflect the capability level of the device in real time, and more accurately support the matching results between subtasks and edge devices. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.
[0041] In the attached picture:
[0042] Figure 1 This is a flow chart of an edge computing optimization method based on a big data network of the present invention;
[0043] Figure 2 A schematic diagram of the steps of an optimization method of an edge computing optimization method based on a big data network according to the present invention;
[0044] Figure 3 A schematic diagram of data storage steps of an edge computing optimization method based on a big data network according to the present invention;
[0045] Figure 4 This is a schematic diagram of the computing task scheduling process of an edge computing optimization method based on a big data network in the present invention. DETAILED DESCRIPTION
[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0047] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.
[0048] The edge computing optimization method based on big data network of the present invention is as follows Figure 2 As shown, it includes: data collection, collecting network data in the edge computing environment; data preprocessing, preprocessing the collected data; data allocation analysis, storing and analyzing the preprocessed data; computing task scheduling, reasonably allocating computing tasks to edge devices according to data processing requirements and device performance; result feedback adjustment, adjusting the edge computing environment according to the results of the edge devices completing the tasks.
[0049] The specific implementation of the present invention is described below with reference to the accompanying drawings and embodiments.
[0050] Example 1
[0051] like Figure 1 As shown, this embodiment provides an edge computing optimization method based on a big data network, comprising the following steps:
[0052] S1: Collect network data in the edge computing environment and pre-process the collected data, including data cleaning, deduplication and format conversion;
[0053] S2: storing and analyzing the preprocessed data, wherein the data analysis includes data exploration, data mining and prediction model;
[0054] S3: Decompose tasks, schedule tasks according to data processing requirements and the assessed edge device capabilities, and reasonably distribute computing tasks to the edge devices;
[0055] S4: Collect feedback on task execution results and provide feedback to users, and adjust the edge computing environment based on the results of task completion.
[0056] The network data includes user behavior, device status data and environmental data, and the network data is collected through sensors, cameras and intelligent collection devices.
[0057] Furthermore, in step S1, the data cleaning includes removing noise and irrelevant data generated during the collection process, and deleting or filling the missing values;
[0058] The data deduplication monitors and deletes duplicate data by using a hash value comparison method to ensure the uniqueness of the data;
[0059] The format conversion converts data from different sources into a unified format to facilitate subsequent data processing and analysis, and at the same time standardizes the timestamp as needed and adjusts the data accuracy to meet analysis requirements.
[0060] Preferably, in step S2, the storage process is as follows: Figure 3 As shown, the following steps are included:
[0061] S21: Select to use a distributed file system as a data storage medium, and import the pre-processed data into a designated storage system;
[0062] S22: performing version control, backup, recovery and security protection on the imported data;
[0063] S23: merging the pre-processed data from different sources and creating indexes for the data to improve query efficiency;
[0064] S24: Use an in-memory database to cache frequently accessed data.
[0065] Furthermore, the data in the data analysis includes:
[0066] Explore the use of statistical analysis and visualization tools for preliminary data exploration, quickly obtain the basic statistical characteristics of the data set, display the data distribution through histograms, box plots, and scatter plots, identify patterns, anomalies, and data clustering trends in the data, and analyze the correlation between different variables through matrices and heat maps to avoid the impact of multicollinearity on subsequent models.
[0067] Furthermore, the data mining in the data analysis uses a machine learning algorithm to mine patterns in the data, and the machine learning algorithm used includes a decision tree, a support vector machine or a neural network algorithm, specifically including:
[0068] The decision tree algorithm constructs a tree structure by recursively dividing the data set into different subsets, each node represents a feature, each branch represents a decision rule for a feature, and each leaf node represents a prediction result;
[0069] The support vector machine maps the original data into a high-dimensional space by using a kernel function, and finds a hyperplane that maximizes the interval between two types of data;
[0070] The neural network algorithm introduces nonlinear characteristics through activation functions to increase the learning of complex patterns from data.
[0071] Preferably, the predictive modeling in the data analysis includes:
[0072] Clarify the forecasting target, collect historical data, build a forecasting model based on historical data, use regression analysis or time series analysis methods, select an appropriate model, and use the fitted model to perform forecasting analysis on relevant data.
[0073] Further, in step S3, the task scheduling is as follows: Figure 4 As shown, the following steps are included:
[0074] S31: Identify the tasks that need to be performed and break them down into smaller subtasks;
[0075] S32: Evaluate the capabilities of the edge devices, assign the decomposed subtasks to the most suitable edge devices according to the evaluation results, and perform scheduling according to the shortest job first strategy;
[0076] S33: Monitor the execution status of the task on the edge device and make timely adjustments as needed.
[0077] Preferably, in step S32, the edge device includes a router, a sensor, and a smart phone, and the capabilities of the edge device include processing power, storage space, and battery life. The evaluation of the edge device capabilities adopts a fuzzy comprehensive evaluation model, including the following steps:
[0078] S321: Combine the hierarchical structure to be analyzed and establish a factor set Q = {c1, c2 ..., c m}, where each single factor subset ci includes several elements;
[0079] S322: Establishing an evaluation result set V = {V1, V2..., V K}, take the median of each interval as the grade parameter, then the corresponding evaluation grade parameter column vector is
[0080] S323: Evaluate the level of each indicator of the specific object, establish a fuzzy relationship matrix, evaluate each indicator, grade the actual situation of the qualitative indicator, and then count the frequency of comments at different levels for each indicator, that is, obtain the C of each subset. i The single factor evaluation matrix R i ;
[0081] S324: Perform single factor evaluation, and calculate the single factors c i , and each single factor c i The weight coefficient vector W i , and the single factor evaluation matrix R i , perform synthesis operation: B = W i o R i ;
[0082] S325: Repeat S324 to perform a comprehensive evaluation of multiple factors, and obtain a comprehensive evaluation matrix for each subset in Q: R = (B1, B2, ..., B m ) T , until the fuzzy comprehensive evaluation result of the criterion layer relative to the target is obtained: E = W c o R c =(e i ,e2,…,e n ), where E is a membership result vector of factor set Q, and the final evaluation result is the algebraic value obtained by multiplying E and the evaluation level parameter column vector V: Where Z represents the overall capability assessment result of the assessment object.
[0083] Fuzzy comprehensive evaluation is an evaluation method based on fuzzy set theory, which is suitable for the evaluation of complex systems or multi-criteria problems. This method divides the evaluation indicators into different levels and uses the membership function in fuzzy mathematics to describe the degree to which each evaluation indicator belongs to a certain level. Then, by synthesizing the membership functions of these evaluation indicators, the final evaluation value is obtained; it can effectively deal with the uncertainty and ambiguity in the evaluation process and obtain more reasonable evaluation results. Fuzzy comprehensive evaluation has strong adaptability and can be applied to evaluation problems of different types and scales; it can comprehensively consider multiple evaluation indicators to obtain comprehensive evaluation results. The methods and steps of fuzzy comprehensive evaluation are relatively simple and easy to understand and operate.
[0084] In summary, the present invention can greatly reduce the distance and time of data transmission by processing data on edge devices near the data source, thereby reducing latency. This is of great significance for applications with high real-time requirements, such as autonomous driving and industrial automation. Edge computing can reduce the transmission of large amounts of data from data sources to cloud servers, thereby reducing bandwidth consumption, which is very beneficial for application scenarios with limited network bandwidth. Edge computing can distribute computing tasks on multiple edge devices, thereby improving system performance and throughput. At the same time, edge devices can take advantage of their proximity to data sources to perform faster and more accurate data processing. Edge computing can perform data processing near the data source, reducing data transmission in the network, thereby reducing the risk of data leakage. Edge devices can implement local storage and processing of data, enhancing data privacy and security.
[0085] Secondly, by decomposing tasks, it is easier to assign them to different computing resources, ensuring that tasks match the capabilities of execution resources to optimize performance and energy efficiency. Task characteristics and device capabilities are used for dynamic matching to achieve the lowest response time and the highest resource utilization. By selecting an appropriate scheduling strategy, execution speed and resource utilization can be balanced in a multi-tasking environment. Real-time monitoring helps to detect problems and reallocate tasks in a timely manner to maintain the overall performance of the system. By reasonably allocating tasks, the load of all devices is close to the average value to avoid overloading some devices while others are idle. Computing tasks are assigned to devices as close to the data source as possible to reduce data transmission time and delay. Considering the energy consumption of the device, tasks are assigned to devices with higher energy efficiency to extend the service life of the device. Consider the reliability and stability of the device when assigning tasks to ensure that the system can still operate normally in the face of device failures.
[0086] The fuzzy evaluation model allows the evaluation indicators and weights to be adjusted according to actual needs. It can adapt to the requirements for edge device capability evaluation in different scenarios, and can realize dynamic evaluation of edge device capabilities. As the performance of the device changes, the evaluation indicators and weights can be adjusted in time to reflect the capability level of the device in real time, and more accurately support the matching results between subtasks and edge devices.
[0087] Finally, it should be noted that the above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.
[0088] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
Claims
1. An edge computing optimization method based on a big data network, characterized in that: The steps include: S1: Collect network data in the edge computing environment and pre-process the collected data, including data cleaning, deduplication and format conversion; S2: storing and analyzing the preprocessed data, wherein the data analysis includes data exploration, data mining and prediction model; S3: Decompose tasks, schedule tasks according to data processing requirements and the assessed edge device capabilities, and reasonably distribute computing tasks to the edge devices; S4: Collect feedback on task execution results and provide feedback to users, and adjust the edge computing environment based on the results of task completion.
2. The edge computing optimization method based on big data network according to claim 1 is characterized in that: In step S1, the network data includes user behavior, device status data and environmental data, and the network data is collected by sensors, cameras and intelligent collection devices.
3. The edge computing optimization method based on big data network according to claim 1 is characterized in that: In step S1, the data cleaning includes removing noise and irrelevant data generated during the collection process, and deleting or filling the missing values; The data deduplication monitors and deletes duplicate data by using a hash value comparison method to ensure the uniqueness of the data; The format conversion converts data from different sources into a unified format to facilitate subsequent data processing and analysis, and at the same time standardizes the timestamp as needed and adjusts the data accuracy to meet analysis requirements.
4. The edge computing optimization method based on big data network according to claim 1 is characterized in that: In step S2, the storage process includes: S21: Select to use a distributed file system as a data storage medium, and import the pre-processed data into a designated storage system; S22: performing version control, backup, recovery and security protection on the imported data; S23: merging the pre-processed data from different sources and creating an index for the data; S24: Use an in-memory database to cache frequently accessed data.
5. The edge computing optimization method based on big data network according to claim 1 is characterized in that: In step S2, the data in the data analysis includes: Explore the use of statistical analysis and visualization tools for preliminary data exploration, quickly obtain the basic statistical characteristics of the data set, display the data distribution through histograms, box plots, and scatter plots, identify patterns, anomalies, and data clustering trends in the data, and analyze the correlation between different variables through matrices and heat maps to avoid the impact of multicollinearity on subsequent models.
6. The edge computing optimization method based on big data network according to claim 1 is characterized in that: In step S2, the data mining in the data analysis uses a machine learning algorithm to mine patterns in the data. The machine learning algorithm used includes a decision tree, a support vector machine or a neural network algorithm, specifically including: The decision tree algorithm constructs a tree structure by recursively dividing the data set into different subsets, each node represents a feature, each branch represents a decision rule for a feature, and each leaf node represents a prediction result; The support vector machine maps the original data into a high-dimensional space by using a kernel function, and finds a hyperplane that maximizes the interval between two types of data; The neural network algorithm introduces nonlinear characteristics through activation functions to increase the learning of complex patterns from data.
7. The edge computing optimization method based on big data network according to claim 1 is characterized in that: In step S2, the predictive modeling in the data analysis includes: Clarify the forecasting target, collect historical data, build a forecasting model based on historical data, use regression analysis or time series analysis methods, select an appropriate model, and use the fitted model to perform forecasting analysis on relevant data.
8. The edge computing optimization method based on big data network according to claim 1 is characterized in that: In step S3, the task scheduling includes the following steps: S31: Identify the tasks that need to be performed and break them down into smaller subtasks; S32: Evaluate the capabilities of the edge devices, assign the decomposed subtasks to the most suitable edge devices according to the evaluation results, and perform scheduling according to the shortest job first strategy; S33: Monitor the execution status of the task on the edge device and make timely adjustments as needed.
9. The edge computing optimization method based on big data network according to claim 8 is characterized in that: In step S32, the capabilities of the edge device include processing power, storage space, and battery life. The evaluation of the edge device capabilities adopts a fuzzy comprehensive evaluation model, including the following steps: S321: Combine the hierarchical structure to be analyzed and establish a factor set Q = {c1, c2 ..., c m }, where each single factor subset ci includes several elements; S322: Establishing an evaluation result set V = {V1, V2..., V K }, take the median of each interval as the grade parameter, then the corresponding evaluation grade parameter column vector is S323: Evaluate the level of each indicator of the specific object, establish a fuzzy relationship matrix, evaluate each indicator, grade the actual situation of the qualitative indicator, and then count the frequency of comments at different levels for each indicator, that is, obtain the C of each subset. i The single factor evaluation matrix R i ; S324: Perform single factor evaluation, and calculate the single factors c i , and each single factor c i The weight coefficient vector W i , and the single factor evaluation matrix R i , perform synthesis operation: B = W i o R i ; S325: Repeat S324 to perform a comprehensive evaluation of multiple factors, and obtain a comprehensive evaluation matrix for each subset in Q: R = (B1, B2, ..., B m ) T , until the fuzzy comprehensive evaluation result of the criterion layer relative to the target is obtained: E = W c o R c =(e i ,e2,…,e n ), where E is a membership result vector of factor set Q, and the final evaluation result is the algebraic value obtained by multiplying E and the evaluation level parameter column vector V: Where Z represents the overall capability assessment result of the assessment object.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by the processor, the edge computing optimization method based on the big data network is implemented as described in any one of claims 1 to 9.
11. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, enables the one or more processors to implement the edge computing optimization method based on a big data network as described in any one of claims 1-9.
Citation Information
Cited By
Regional energy management optimization method based on edge calculation and multi-objective optimization
CN119295260A
