Large-screen application shopping mall operation optimization system and method thereof

Abstract

The invention discloses a large-screen application shopping mall operation optimization system and a large-screen application shopping mall operation optimization method. The system comprises a basic configuration module, a module used for collecting user terminals and carrying out grouping data analysis on terminal users, an operation display module used for basic operation of a system mall and displaying customized content, and an upgrading strategy module used for selecting optimal upgrading time and mode. The method comprises visual operation layout, user group differentiation display and a multi-dimensional terminal application upgrading strategy. The operation efficiency of the screen application shopping mall can be remarkably improved, diversified requirements of different users are met, the user satisfaction is improved, the user experience is enhanced, the load of the server can be effectively dispersed, stable operation of the system is guaranteed, and the pressure of the server is relieved.

Description

Technical Field

[0001] The present invention relates to an operation optimization system and method for a large-screen application mall, belonging to the technical field of operation of Internet application malls. Background Art

[0002] With the rapid development of mobile Internet technology, as an important channel for software distribution, the large-screen application mall faces challenges such as how to operate efficiently, push accurately, and optimize the user experience while providing rich application resources. Most of the large-screen application malls on the market currently adopt a fixed layout mode and a unified upgrade strategy, lacking the ability to adjust in real time according to the differences in user groups. Operators usually formulate promotion strategies through background data analysis, but this process often lags behind market changes and is difficult to achieve refined operation. At the same time, traditional application malls often have problems such as low operation efficiency, single user experience, and high server load. Especially when dealing with a large-scale user group, how to perform differential display according to user characteristics and flexible application upgrade management has become a technical problem to be solved urgently. Summary of the Invention

[0003] The purpose of the present invention is to provide an operation optimization system and method for a large-screen application mall, which realizes efficient operation, personalized recommendation, and effective alleviation of server pressure through visual operation layout, differential display of user groups, and flexible multi-dimensional terminal application upgrade strategies, thereby improving the user experience.

[0004] To solve the above technical problems, the present invention adopts the following technical solutions: A large-screen application mall operation optimization system, which is connected to multiple user terminals. The system includes: A basic configuration module: used for system management to realize the basic function framework of the system; A data analysis module: used to collect the basic data of user terminals and the behavior data generated during use, and group terminal users. Specifically, through user portrait technology and machine learning algorithms, the K-means clustering algorithm is used to group terminal users; An operation display module: used for the basic operation of the system mall. At the same time, according to the interests and needs of different user groups, customized application recommendations and mall layouts can be displayed. According to the interests and needs of different user groups, customized application recommendations and mall layouts are displayed to achieve the application mall display effect of "one person, one face". And a drag-and-drop interface design is adopted, and a layout optimization algorithm and a real-time preview mechanism are configured. Specifically, the basic layout module and component library are initialized through the VUE framework, including various UI components and layout templates. The initialization process includes component registration and basic layout setting. The drag-and-drop and placement functions are implemented using D3.js. During the drag process, the position of the components is dynamically adjusted through a self-developed layout optimization algorithm to avoid overlap and layout imbalance; An upgrade strategy module: According to the basic data and behavior data of user terminals, configure application upgrade strategies and select the optimal upgrade time and method. The specific implementation method is to allow operation personnel to configure application upgrade strategies according to multiple dimensions such as device type, user tags, and geographical location, and automatically select the optimal upgrade time and method to reduce the server pressure and improve the user experience.

[0005] A large-screen application mall operation optimization method includes the following steps:

[0006] A. Visual operation layout: A drag-and-drop interface design is adopted on the user terminal interface, and a layout optimization algorithm and a real-time preview mechanism are set. The layout optimization algorithm can calculate and display the optimal layout effect in real time when the user adjusts the layout, and the real-time preview mechanism ensures that the user can immediately see the change results when adjusting the layout;

[0007] B. Differentiated display of user groups: Collect the behavior data during the use of the terminal, group terminal users, and display customized application recommendations and mall layouts according to the interests and needs of different user groups;

[0008] C. Multi-dimensional terminal application upgrade strategy: Operation personnel configure application upgrade strategies according to multiple dimensions such as device type, user tags, geographical location, and network environment, and automatically select the optimal upgrade time and method.

[0009] For the foregoing large-screen application mall operation optimization method, the specific implementation steps of step A are as follows:

[0010] A1. Initialize the basic layout module and component library through the VUE framework, including various UI components and layout templates. Among them, the initialization process includes component registration and basic layout setting;

[0011] A2. Implement the drag-and-drop function of components in the user interface using D3.js. Users can adjust the component positions in real time, and the component positions are dynamically adjusted through a layout optimization algorithm during the dragging process;

[0012] A3. Real-time preview and save the layout: Real-time preview is achieved through WebSocket connection to ensure that users can immediately see the change results when adjusting the layout. At the same time, local storage and server synchronization technologies are used to ensure the consistency of the layout in different sessions. The aforementioned method for optimizing the operation of a large-screen application mall, the layout optimization algorithm includes:

[0013] a11. Calculate the distance and layout balance between components in real time during the component dragging process. The formula is as follows: D_{ij}=\sqrt{(x_i - x_j)^2+(y_i - y_j)^2}

[0014] Where, (D_{ij}) is the distance between component (i) and (j), and (x_i,y_i) and (x_j,y_j) are the coordinates of component (i) and (j) respectively;

[0015] a12. Optimize the layout to minimize the total distance. The formula is as follows:

[0016] \min\sum_{i,j}D_{ij}

[0017] Where:

[0018] (\sum_{i,j}) represents the summation over all component pairs (i,j);

[0019] (D_{ij}) represents the distance between component i and component j.

[0020] The aforementioned method for optimizing the operation of a large-screen application mall, the specific implementation steps of step B are as follows:

[0021] B1. Collect user behavior data: Data collection is achieved through the client SDK and server logs to ensure the real-time and integrity of the data. Each user behavior data record contains the following fields: user ID, browsing time, click count, session duration;

[0022] B2. Data cleaning and feature extraction: Use data cleaning algorithms to clean the acquired user behavior data, and use feature extraction algorithms to extract features from the cleaned data;

[0023] B3. User Segmentation: Standardize the extracted feature data to ensure comparability between different features, and use the K-means clustering algorithm to segment users and generate user portraits.

[0024] B4. Dynamically Adjust the Mall Interface and Recommended Content: Dynamically adjust the mall interface and recommended content based on the user portrait and update the user portrait in real time.

[0025] For the aforementioned operation optimization method of a large-screen application mall, in step B2, the data cleaning algorithm includes the following steps: b11. Remove Missing Values: Detect and remove missing values in the data to ensure data integrity.

[0026] b12. Delete Duplicate Records: Detect and delete duplicate user behavior records to avoid data redundancy.

[0027] b13. Remove Outliers: Use statistical methods to detect and remove outliers to improve data quality.

[0028] For the aforementioned operation optimization method of a large-screen application mall, in step B2, the feature extraction algorithm includes the following steps: b21. Extract features from the cleaned data and extract key features of user behavior.

[0029] b22. Use the time window method to aggregate the user's behavior data at regular time intervals to capture the time features of user behavior.

[0030] For the aforementioned operation optimization method of a large-screen application mall, the specific steps of the K-means clustering algorithm in step B3 are as follows:

[0031] b31: Initialize the clustering centers: Randomly select K initial clustering centers.

[0032] b32: Calculate the distance from each data point to the clustering centers and assign the data points to the nearest clustering center.

[0033] b33: Recalculate the clustering centers: Take the mean of the data points in each cluster as the new clustering centers.

[0034] b34: Repeat the above steps until the clustering centers no longer change significantly.

[0035] For the aforementioned operation optimization method of a large-screen application mall, the specific implementation steps of step C are as follows:

[0036] C1. Collect Multidimensional Data: Data collection is achieved through the client SDK and server logs to ensure the real-time and integrity of the data, and the data records include information in multiple dimensions such as device type, user tags, geographical location, and network environment.

[0037] C2. Optimize and upgrade the strategy using a genetic algorithm: Based on the collected data, use a genetic algorithm to optimize the upgrade strategy and select the optimal upgrade time and method;

[0038] C3. Dynamically adjust the upgrade strategy: Based on real-time monitoring data, adjust the upgrade time and upgrade batches to ensure that the upgrade is carried out when the server load is low, and dynamically adjust the upgrade strategy according to real-time data to ensure server load balancing;

[0039] C4. Real-time monitor the upgrade process: Real-time monitor the upgrade process through a monitoring system, collect and display key indicators such as server load and upgrade progress to ensure the smooth progress of the upgrade..

[0040] For the aforementioned operation optimization method of a large-screen application mall, the genetic algorithm described in step C2 includes:

[0041] c11. Generate the initial population: Randomly generate a group of upgrade strategies as the initial population;

[0042] c12. Fitness function: Calculate the fitness value according to the execution effect of the upgrade strategy. The fitness value formula is as follows:

[0043] F = \alpha\cdot(1-\frac{L}{L_{max}})+\beta\cdot U

[0044] Where, (L) represents the current load, (L_{max}) represents the maximum load, (U) represents user satisfaction, and (\alpha) and (\beta) are weight coefficients;

[0045] c13. Selection operation: Select individuals with high fitness as parents;

[0046] c14. Crossover operation: Generate new individuals through the crossover operation;

[0047] c15. Mutation operation: Perform a mutation operation on the new individuals to increase population diversity.

[0048] Compared with the prior art, the present invention has at least the following beneficial effects:

[0049] (1) The present invention adopts a visual operation layout, enabling operation personnel to quickly respond to market changes, adjust strategies, and improve operation efficiency.

[0050] (2) The present invention designs and develops a user group differentiation display algorithm for user group differentiation display and personalized recommendation, meeting the diverse needs of different users, enhancing user satisfaction, and improving the user experience.

[0051] (3) The present invention effectively disperses the server load by using a flexible multi-dimensional upgrade strategy, ensures the stable operation of the system, and reduces the server pressure. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a schematic diagram of the system architecture of the present invention;

[0053] Figure 2 is the system flow chart of the present invention.

[0054] The present invention will be further described below in conjunction with the drawings and specific embodiments. SPECIFIC EMBODIMENTS

[0055] Embodiment 1: An operation optimization system for a large-screen application mall. The system is connected to multiple user terminals and includes: a basic configuration module for system management to realize the basic function framework of the system; a data analysis module for collecting the basic data of user terminals and the behavior data generated during use, clustering terminal users. Specifically, through user portrait technology and machine learning algorithms, the K-means clustering algorithm is used to cluster terminal users; an operation display module for the basic operation of the system mall. At the same time, according to the interests and needs of different user groups, customized application recommendations and mall layouts can be displayed, and customized application recommendations and mall layouts are displayed according to the interests and needs of different user groups to achieve the effect of a personalized application mall display for each user. And a drag-and-drop interface design is adopted, and a layout optimization algorithm and a real-time preview mechanism are configured. Specifically, the basic layout module and the component library are initialized through the VUE framework, including various UI components and layout templates. The initialization process includes the registration of components and the setting of the basic layout. The drag-and-drop and placement functions are implemented by D3.js. During the dragging process, the positions of components are dynamically adjusted through a self-developed layout optimization algorithm to avoid overlap and layout imbalance; an upgrade strategy module for configuring application upgrade strategies according to the basic data and behavior data of user terminals, and selecting the optimal upgrade time and method. The specific implementation method is to allow operation personnel to configure application upgrade strategies according to multiple dimensions such as device type, user tags, and geographical location, automatically select the optimal upgrade time and method, reduce the server pressure, and improve the user experience.

[0056] Embodiment 2: An operation optimization method for a large-screen application mall, including the following steps:

[0057] A. Visual operation layout: A drag-and-drop interface design is adopted on the user terminal interface, and a layout optimization algorithm and a real-time preview mechanism are set. The layout optimization algorithm can calculate and display the optimal layout effect in real time when the user adjusts the layout, and the real-time preview mechanism ensures that the user can immediately see the change results when adjusting the layout. The specific implementation steps are as follows:

[0058] A1. Initialize the basic layout module and component library through the VUE framework, including various UI components and layout templates. Among them, the initialization process includes component registration and basic layout setting;

[0059] A2. Use D3.js to implement the drag-and-drop function of components in the user interface. Users can adjust the component positions in real time, and during the dragging process, the component positions are dynamically adjusted through a layout optimization algorithm to avoid overlap and layout imbalance. The layout optimization algorithm is as follows:

[0060] a11. Calculate the distance and layout balance between components in real time during the component dragging process. The formula is as follows: D_{ij}=\sqrt{(x_i - x_j)^2+(y_i - y_j)^2}

[0061] Where:

[0062] (D_{ij}) is the distance between component (i) and (j), and (x_i,y_i) and (x_j,y_j) are the coordinates of component (i) and (j) respectively;

[0063] a12. Optimize the layout to minimize the total distance. The formula is as follows:

[0064] \min\sum_{i,j}D_{ij}

[0065] Where:

[0066] (\sum_{i,j}) represents the summation over all component pairs (i,j);

[0067] (D_{ij}) represents the distance between component i and component j.

[0068] a13. Real-time preview and save the layout: When the user adjusts the layout, the layout optimization algorithm calculates and displays the optimal layout effect in real time to avoid layout confusion. It specifically realizes real-time preview through a WebSocket connection to ensure that the user can immediately see the change results when adjusting the layout. At the same time, local storage and server synchronization technologies are used to ensure the consistency of the layout in different sessions. Among them, the specific implementation of data persistence includes real-time synchronization of layout data to the server through WebSocket and using a database for persistent storage.

[0069] B. Differentiated display for user groups: Collect behavioral data during the use of the terminal. Through user portrait technology and machine learning algorithms, divide terminal users into groups. According to the interests and needs of different user groups, display customized application recommendations and mall layouts to achieve a personalized application mall display effect. The specific implementation steps are as follows:

[0070] B1. Collect user behavior data: Collect data such as user browsing records, click records, and usage duration. The data collection is implemented through the client SDK and server logs to ensure the real-time and integrity of the data. Each user behavior data record contains the following fields: user ID, browsing time, click count, and session duration;

[0071] B2. Data cleaning and feature extraction: Use data cleaning algorithms to clean the obtained user behavior data, and use feature extraction algorithms to extract features from the cleaned data. The data cleaning algorithm includes the following steps:

[0072] b11. Remove missing values: Detect and remove missing values in the data to ensure data integrity;

[0073] b12. Delete duplicate records: Detect and delete duplicate user behavior records to avoid data redundancy;

[0074] b13. Remove outliers: Use statistical methods to detect and remove outliers to improve data quality;

[0075] Specifically, the code implementation example of the above data cleaning algorithm is as follows:

[0076] def clean_data(data):

[0077] # Remove missing values

[0078] data = data.dropna()

[0079] # Delete duplicate records

[0080] data = data.drop_duplicates()

[0081] # Remove outliers

[0082] q_low = data.quantile(0.01)

[0083] q_high = data.quantile(0.99)

[0084] data = data[(data >= q_low) & (data <= q_high)]

[0085] return data

[0086] The above feature extraction algorithm includes the following steps:

[0087] b21. Extract features from the cleaned data to extract key features of user behavior, such as browsing time, click count, session duration, etc.;

[0088] b22. Use the time window method to aggregate the user's behavior data at regular time intervals to capture the time characteristics of the user's behavior;

[0089] Specifically, the code implementation example of the above feature extraction algorithm is as follows:

[0090] def extract_features(data):

[0091] features = []

[0092] for record in data:

[0093] features.append(

[0094] record['browsing_time'],

[0095] record['click_count'],

[0096] record['session_duration'] )

[0098] return features

[0099] B3. User clustering: Standardize the extracted feature data to ensure the comparability between different features, and use the K-means clustering algorithm to cluster users and generate user portraits. The specific steps of the K-means clustering algorithm are as follows:

[0100] b31: Initialize the cluster centers: Randomly select K initial cluster centers;

[0101] b32: Calculate the distance from each data point to the cluster centers and assign the data points to the nearest cluster center;

[0102] b33: Recalculate the cluster centers: Take the mean of the data points in each cluster as the new cluster centers;

[0103] b34: Repeat the above steps until the cluster centers no longer change significantly.

[0104] B4. Dynamic adjustment of the mall interface and recommended content: Dynamically adjust the mall interface and recommended content based on the user profile to improve the user experience and the accuracy of recommendations. Update the user profile in real time to ensure that changes in user behavior can be promptly reflected in the recommended content. Specifically, users can be divided into different groups according to the user profile, and each group corresponds to a different mall interface and recommended content. At the same time, monitor user behavior data in real time and dynamically adjust the user profile to ensure the timeliness and accuracy of the recommended content. C. Multi-dimensional terminal application upgrade strategy: The operation staff configures the application upgrade strategy according to multiple dimensions such as device type, user tags, geographical location, and network environment, and automatically selects the optimal upgrade time and method. The specific implementation steps are as follows:

[0105] C1. Collect multi-dimensional data: Data collection is achieved through the client SDK and server logs to ensure the real-time and integrity of the data. The data records include information on multiple dimensions such as device type, user tags, geographical location, and network environment;

[0106] C2. Optimize the upgrade strategy using the genetic algorithm: Based on the collected data, use the genetic algorithm to optimize the upgrade strategy and select the optimal upgrade time and method. The genetic algorithm includes:

[0107] c11. Generation of the initial population: Randomly generate a set of upgrade strategies as the initial population;

[0108] c12. Fitness function: Calculate the fitness value according to the execution effect of the upgrade strategy. The formula for the fitness value is as follows:

[0109] F = \alpha\cdot(1-\frac{L}{L_{max}})+\beta\cdot U

[0110] Where:

[0111] (L) represents the current load, (L_{max}) represents the maximum load, (U) represents user satisfaction, and (\alpha) and (\beta) are weight coefficients;

[0112] c13. Selection operation: Select individuals with high fitness as parents;

[0113] c14. Crossover operation: Generate new individuals through the crossover operation;

[0114] c15. Mutation operation: Perform mutation operations on the new individuals to increase the diversity of the population;

[0115] C3. Dynamic adjustment and upgrade strategy: Based on real-time monitoring data, use scheduling algorithms for calculation, adjust the upgrade time and upgrade batches to ensure that the upgrade is carried out when the server load is low, and dynamically adjust the upgrade strategy according to real-time data to ensure server load balancing. The pseudo-code implementation example of the scheduling algorithm is as follows:

[0116]

[0117] C4. Real-time monitoring of the upgrade process: Real-time monitor the upgrade process through the monitoring system, collect and display key indicators such as server load and upgrade progress to ensure the smooth progress of the upgrade. The real-time monitoring system uses monitoring tools (Prometheus and Grafana) for real-time monitoring, collect and display key indicators such as server load and upgrade progress.

[0118] The pseudo-codes of the various algorithms described above in the present invention are all used to represent the functions implemented by the algorithms and their program execution processes, and are all one of the examples, without limiting their specific content.

[0119] The present invention provides an efficient operation optimization solution. The visual operation layout enables operation personnel to quickly respond to market changes and adjust strategies; the differentiated display and personalized recommendation of user groups meet the diverse needs of different users and improve user satisfaction; the flexible multi-dimensional upgrade strategy effectively disperses the server load and ensures the stable operation of the system; the present invention has been tested in multiple rounds in actual applications, including simulating user behaviors, collecting real user feedback, and monitoring server load, etc. The experimental results show that the present invention can significantly improve operation efficiency, enhance user experience, and effectively relieve server pressure.

Claims

1. A large-screen application mall operation optimization system, characterized in that: The system is connected to a plurality of user terminals, and the system comprises: Basic configuration module: used for system management and to implement the basic functional architecture of the system; Data analysis module: used to collect basic data of user terminals and behavioral data generated during use, and to group terminal users; Operation and display module: used for the basic operation of the system mall, and can display customized application recommendations and mall layout according to the interests and needs of different user groups; Upgrade strategy module: configure application upgrade strategy based on basic data and behavior data of user terminals, and select the optimal upgrade time and method.

2. A method for optimizing the operation of a large-screen application mall, using the system of claim 1, characterized in that: The following steps are involved: A. Visual operation layout: The user terminal interface adopts a drag-and-drop interface design, and is equipped with a layout optimization algorithm and a real-time preview mechanism. The layout optimization algorithm can calculate and display the optimal layout effect in real time when the user adjusts the layout. The real-time preview mechanism ensures that the user can see the change results immediately when adjusting the layout; B. Differentiated display by user groups: collect behavioral data during the use of terminals, group terminal users, and display customized application recommendations and mall layouts based on the interests and needs of different user groups; C. Multi-dimensional terminal application upgrade strategy: Operators configure application upgrade strategies based on multiple dimensions such as device type, user tag, geographic location, and network environment, and automatically select the optimal upgrade time and method.

3. A large-screen application mall operation optimization method according to claim 2, characterized in that: The specific implementation steps of step A are as follows: A1. Initialize the basic layout module and component library through the VUE framework, including various UI components and layout templates, wherein the initialization process includes component registration and basic layout setting; A2. Use D3.js to implement the drag and drop function of components in the user interface. Users can adjust the position of components in real time, and the position of components can be dynamically adjusted through the layout optimization algorithm during the dragging process; A3. Real-time preview and layout saving: Real-time preview is achieved through WebSocket connection to ensure that users can see the changes immediately when adjusting the layout. At the same time, local storage and server synchronization technology are used to ensure the consistency of the layout in different sessions.

4. A large-screen application mall operation optimization system and method according to claim 3, characterized in that: The layout optimization algorithm includes: a11. Calculate the distance between components and layout balance in real time during the component dragging process. The formula is as follows: D_{ij}=\sqrt{(x_i-x_j)^2+(y_i-y_j)^2} Where (D_{ij}) is the distance between components (i) and (j), (x_i, y_i) and (x_j, y_j) are the coordinates of components (i) and (j), respectively; a12. Optimize the layout to minimize the total distance. The formula is as follows: \min\sum_{i,j}D_{ij} in: (\sum_{i,j}) means summing all component pairs (i,j); (D_{ij}) represents the distance between component i and component j.

5. A large-screen application mall operation optimization method according to claim 2, characterized in that: The specific implementation steps of step B are as follows: B1. Collect user behavior data: Data collection is achieved through client SDK and server logs to ensure the real-time and integrity of data. Each user behavior data record contains the following fields: user ID, browsing time, number of clicks, and session duration; B2. Data cleaning and feature extraction: Use data cleaning algorithms to clean the acquired user behavior data, and use feature extraction algorithms to extract features from the cleaned data; B3. User grouping: Standardize the extracted feature data to ensure the comparability between different features, use the K-means clustering algorithm to group users and generate user portraits; B4. Dynamically adjust the mall interface and recommended content: Dynamically adjust the mall interface and recommended content based on user portraits, and update user portraits in real time.

6. A large-screen application mall operation optimization method according to claim 5, characterized in that: In step B2, the data cleaning algorithm includes the following steps: b11. Remove missing values: Detect and remove missing values ​​in the data to ensure data integrity; b12. Delete duplicate records: Detect and delete duplicate user behavior records to avoid data redundancy; b13. Remove outliers: Use statistical methods to detect and remove outliers to improve data quality.

7. A large-screen application mall operation optimization method according to claim 5, characterized in that: In step B2, the feature extraction algorithm includes the following steps: b21. Perform feature extraction on the cleaned data to extract key features of user behavior; b22. Use the time window method to aggregate user behavior data at certain time intervals to capture the time characteristics of user behavior.

8. A large-screen application mall operation optimization method according to claim 5, characterized in that: The specific steps of the K-means clustering algorithm in step B3 are as follows: b31: Initialize cluster centers: randomly select K initial cluster centers; b32: Calculate the distance from each data point to the cluster center and assign the data point to the nearest cluster center; b33: Recalculate the cluster center: take the mean of the data points in each cluster as the new cluster center; b34: Repeat the above steps until the cluster center no longer changes significantly.

9. A large-screen application mall operation optimization method according to claim 2, characterized in that: The specific implementation steps of step C are as follows: C1. Collect multi-dimensional data: Data collection is achieved through client SDK and server logs to ensure the real-time and integrity of data. The data records include information on device type, user tags, geographic location, and network environment in multiple dimensions; C2. Use genetic algorithms to optimize upgrade strategies: Based on the collected data, use genetic algorithms to optimize upgrade strategies and select the optimal upgrade time and method; C3. Dynamically adjust the upgrade strategy: Based on real-time monitoring data, adjust the upgrade time and upgrade batches to ensure that the upgrade is carried out when the server load is low. Dynamically adjust the upgrade strategy according to real-time data to ensure balanced server load. C4. Real-time monitoring of the upgrade process: The upgrade process is monitored in real time through the monitoring system, and key indicators such as server load and upgrade progress are collected and displayed to ensure the smooth progress of the upgrade.

10. A large-screen application mall operation optimization method according to claim 9, characterized in that: The genetic algorithm in step C2 includes: c11. Initial population generation: randomly generate a set of upgrade strategies as the initial population; c12. Fitness function: The fitness value is calculated according to the execution effect of the upgrade strategy. The fitness value formula is as follows: F=\alpha\cdot(1-\frac{L}{L_{max}})+\beta\cdot U Among them, (L) represents the current load, (L_{max}) represents the maximum load, (U) represents user satisfaction, (\alpha) and (\beta) are weight coefficients; c13, selection operation: select individuals with high fitness as parents; c14, Crossover operation: Generate new individuals through crossover operation; c15. Mutation operation: Perform mutation operations on new individuals to increase population diversity.

Images