Enterprise multi-service system centralized to-do processing method and system based on message driving

Abstract

The invention discloses a message-driven enterprise multi-service system centralized to-do processing method and system, and the method comprises the steps: S1, obtaining resource investment and task processing data, and obtaining a resource consumption and task state cluster through clustering; s2, using a regression algorithm to analyze correlation between the two, and determining a resource utilization rate change trend; s3, when the threshold value is exceeded, calculating a task completion rate statistical characteristic to obtain a resource optimization index; s4, constructing an interactive model by using a decision tree in combination with historical data, and judging an input strategy adjustment point; s5, rendering visual resource distribution; s6, abnormal tasks are screened, and an optimization area is determined; and S7, updating an instrument panel to obtain a decision view. Through data clustering, correlation analysis and dynamic adjustment, the resource and task matching degree is improved, and decision scientificity and efficiency are enhanced.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of business systems, and particularly relates to a centralized to-do processing method and system for enterprise multi-business systems based on message driving. BACKGROUND

[0002] In the operation and management of enterprise multi-business systems, the existing technology usually processes resource allocation and task execution through independent resource management modules and task scheduling modules. Specifically, the resource management module extracts resource input data such as the central processor usage rate and memory occupancy of the server from the database, and performs simple resource adjustment through a preset fixed threshold (such as increasing resource allocation when the central processor usage rate exceeds 80%). The task scheduling module records task processing data such as task completion time and success rate according to task submission time or priority, and triggers a retry mechanism when the task fails or times out.

[0003] However, the existing technology has obvious limitations: on the one hand, the correlation analysis of resource allocation and task processing is insufficient, and only single resource threshold or task attribute is used for decision-making, without classifying resource consumption patterns and task states by clustering algorithm, making it difficult to identify the matching rules of resources and tasks in different clusters, resulting in frequent problems of resource bottlenecks (such as low completion rate of high-load tasks due to insufficient resources) or resource redundancy (such as excessive resource occupation of low-load tasks); on the other hand, there is a lack of intelligent decision support based on historical data, which cannot mine the correlation and potential adjustment points of resource usage rate and task completion rate through regression analysis, decision tree model and other algorithms, and the resource usage state and task execution situation are mostly presented in the form of logs, with low visualization degree, making it difficult for management personnel to quickly locate the optimization direction, ultimately affecting the overall operation efficiency of multi-business systems.

[0004] Therefore, there is an urgent need for a method that can achieve precise matching of resource input and task processing through data clustering, correlation analysis, dynamic adjustment and visualization, to solve the problems of low matching degree of resources and tasks, lack of data-driven decision-making and insufficient visualization support in the existing technology. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a centralized to-do processing method and system for enterprise multi-business systems based on message driving, which improves the matching degree of resources and tasks and enhances the scientificity and efficiency of decision-making through data clustering, correlation analysis and dynamic adjustment.

[0006] To solve the above technical problems, the technical solution adopted by the present application is: A centralized to-do processing method and system for enterprise multi-business systems based on message driving, comprising the following steps: S1: Obtain resource input data and task processing data from the database, group the resource input data and task processing data through a clustering algorithm, and obtain classified resource consumption clusters and task state clusters; S2: According to the classified resource consumption clusters and task state clusters, use a regression algorithm to analyze the correlation between resource usage rate and task completion rate, and determine the resource usage rate trend; S3: If the resource usage rate trend exceeds the preset threshold, calculate the average and variance of the task completion rate through a data aggregation function, and obtain a resource optimization index set; S4: For the resource optimization index set, obtain historical decision data as reference input, use a decision tree algorithm to build an interactive model, and determine potential input strategy adjustment points; S5: Through the adjustment points generated by the interactive model, render intuitive charts in the dashboard framework, including bar charts and heat maps, to obtain visual resource usage rate distribution; S6: Extract real-time monitoring parameters from the visual resource usage rate distribution, use a filtering function to filter abnormal task states, and determine the resource input area that needs to be optimized; S7: According to the determined resource input area, update the dynamic components of the dashboard to reflect the optimized task completion rate, and obtain the final decision support view. Specifically, S1 sub-steps are: S1.1 Data acquisition: Extract resource input data and task processing data from the database through structured query language, and filter records that meet the time range or data format. Resource input data includes central processor usage, memory usage, and disk read / write speed; task processing data includes task completion time, task success rate, and task failure times. The database, such as an enterprise's internal resource management system, stores the computing resource usage, storage space occupation, and task execution log records of the server.

[0007] S1.2 Data preprocessing: Clean the resource input data and task processing data obtained from the database, identify and remove outliers (outliers may be caused by system failure) through a threshold method based on standard deviation, fill in missing values (missing values may be caused by sensor failure or log interruption) through an interpolation method (such as using the average of the previous and subsequent data), and standardize the cleaned data set to scale the data to a unified range.

[0008] S1.3 Feature extraction: Calculate statistical features from resource input data and task processing data. For resource input data, extract the mean, peak, and variance of central processor usage. For task processing data, extract the distribution of task completion rate and the average of task response time. Combine task status (including "success", "failure", "timeout") to generate a multi-dimensional feature vector, for example, generate a vector containing 10 dimensions of features, including central processor usage mean, task completion rate, task timeout rate, etc. These feature vectors constitute the feature set for clustering algorithm.

[0009] S1.4 Clustering processing: Use k-means clustering algorithm to group the feature set, algorithm parameters include the number of clusters k and the number of iterations. Evaluate the clustering effect by the silhouette coefficient, adjust the k value to optimize the classification. After clustering, generate clustering grouping results and cluster labels; analyze the relationship between cluster labels, resource consumption patterns and task status, for example, high consumption cluster may show that central processor usage is consistently higher than 80% and task completion rate is low, low consumption cluster may show that resource usage is lower than 30% and task timeout rate is high; based on the analysis, generate cluster distribution statistics to determine the final resource consumption cluster (such as high load, medium load, low load) and task status cluster (such as efficient completion, inefficient completion, failure).

[0010] Specifically, S2 sub-steps are: S2.1 Data extraction and cleaning: Obtain classified resource consumption cluster data from the database through structured query language, including server central processor usage, memory occupancy rate and other resource usage rate data, and task success completion rate, i.e. task completion rate data. Clean these data, eliminate abnormal high usage records caused by system restart, fill in missing task completion rate data through linear interpolation; standardize the cleaned data set.

[0011] S2.2 Correlation calculation: Calculate the statistical features of resource usage and task completion rate using Pearson correlation coefficient method. Pearson correlation coefficient is used to measure the linear correlation between two variables, its value ranges from -1 to 1. Set the threshold of correlation coefficient to 0.6, if the correlation coefficient exceeds the threshold, it is considered that there is significant correlation between resource usage and task completion rate, and it enters the regression analysis stage.

[0012] S2.3 Regression analysis: Linear regression algorithm is used to fit the relationship between resource usage and task completion rate, taking resource usage as the independent variable and task completion rate as the dependent variable. The least squares method is used to fit the curve, generating trend parameters such as the slope and intercept of the fitted curve (a steeper slope indicates that small changes in resource usage will result in significant changes in task completion rate). Combined with the distribution of task types (including batch processing tasks and real-time processing tasks), statistical data of resource usage and task completion rate are generated, for example, batch processing tasks may show a strong positive correlation between resource usage and task completion rate, while real-time processing tasks may have weaker correlation.

[0013] Specifically, S3 sub-steps are: S3.1 Trend judgment: Obtain resource usage data from time series data and judge whether its trend exceeds the preset threshold. The preset threshold is that the resource usage exceeds 85% for three consecutive days; by analyzing the time series data, if the optimization condition is triggered, the subsequent steps are entered.

[0014] S3.2 Data aggregation: Task completion rate data is processed by a data aggregation function to calculate the average value (to represent task execution efficiency) and variance (to represent task execution stability) of task completion rate within a period of time. Set the time window to 24 hours, calculate the average value and variance of task completion rate per hour based on the time window, and generate a set of statistical features (reflecting the overall trend and fluctuation of task execution, for example, high variance may correspond to resource usage peak period, and low variance may correspond to resource usage trough period).

[0015] S3.3 Dynamic adjustment: According to the statistical feature set, adjust the resource allocation ratio using the preset logical rules. If the average value of task completion rate is less than 80% and the variance is large, increase the allocation ratio of central processor or memory; if the task completion rate is high and stable, maintain the current allocation ratio. Through the dynamic adjustment mechanism, the resource allocation scheme is iteratively updated according to the updated resource usage data in the time series, for example, the allocation ratio is recalculated every hour according to the latest resource usage data. Finally, a set of resource optimization indicators are generated, including the adjusted resource allocation ratio, task completion rate statistical features, etc.

[0016] Specifically, S4 sub-steps are: S4.1 Data acquisition and cleaning: Obtain historical decision data from the database, including resource allocation ratio of server cluster, task completion rate and corresponding business scenario information (such as task load during peak or trough period). Clean these data to remove outliers (wrong records caused by system failure) and missing values (missing due to log interruption). Through feature selection criteria, extract key features related to business target weight (including central processor allocation ratio, average value and variance of task completion rate, task priority), generate model training set for training decision tree algorithm.

[0017] S4.2 Model training: Train the model training set using the decision tree algorithm, and construct a tree structure based on the feature conditions (for judging whether the resource allocation needs to be adjusted) by recursively partitioning the feature space. For example, the root node of the decision tree may be based on whether the central processor usage rate exceeds 80%, and the branch conditions include whether the task completion rate is less than 90%, whether the task priority is high priority, etc. If the prediction accuracy of the trained decision tree model is higher than the preset threshold (85%), an interactive model is generated; the interactive model can output potential input strategy adjustment points (such as suggesting to increase the allocation proportion of computing nodes when the central processor usage rate is high and the task completion rate is low) according to the input resource optimization indicators.

[0018] S4.3 Scene combination analysis: The analysis process of the interactive model combines specific business scenarios, and provides personalized adjustment suggestions for different business scenarios based on the patterns in historical data. For example, in the order processing system of an e-commerce platform, the model will analyze the high-load data caused by the surge in order volume during promotional activities and identify the adjustment points that require increased computing resources.

[0019] Specifically, S5 sub-steps are: S5.1 Data acquisition: Obtain real-time resource usage data through real-time data stream acquisition modules (such as monitoring tools), including server central processor usage rate, memory occupancy rate, etc.; use dynamic adjustment algorithm to generate point sets from these data, generate smoothed data points (reduce short-term fluctuation interference); the generated point set is used to render visual charts.

[0020] S5.2 Chart rendering: Generate resource occupancy column chart through column chart rendering tool, show central processor usage rate distribution of different server nodes, each column corresponds to the average usage rate of a node, the height of the column chart reflects the resource usage intensity, and the color distinguishes different node types (including computing nodes, storage nodes). Obtain density distribution data from the point set, generate heat map using color mapping method, show resource usage density of transaction processing server, color from blue to red represents resource usage rate from low to high, heat map area division is based on physical location or logical grouping of servers (such as partitioning by data center).

[0021] S5.3 Interaction design: The dashboard framework dynamically updates chart content through the front-end framework, ensuring real-time viewing of resource usage. Interactive functions include clicking a column in the column chart to display detailed resource usage data of the node, and selecting an area in the heat map to view the running status of a specific server.

[0022] Specifically, the S6 sub-step is: S6.1 Parameter extraction: acquire resource usage and task status data every second through monitoring tools, and the monitoring data stream includes central processor usage, task completion status, and task response time. Extract these real-time monitoring parameters from the visual resource usage distribution; use a preset threshold to determine whether the resource usage exceeds the normal range, and if the central processor usage exceeds 90%, it is marked as abnormal; based on this, generate a preliminary list of abnormal tasks, including tasks that are overdue and tasks that have failed.

[0023] S6.2 Abnormality screening: use a filtering function to screen the preliminary list of abnormal tasks, and the filtering function determines the specific range of abnormal tasks based on the threshold of task status (including task response time exceeding 2 seconds or task failure rate higher than 5%); the screened abnormal task list is used for further analysis of resource usage.

[0024] S6.3 Region determination: generate a visual distribution diagram of resource usage from monitoring parameters through a distribution extraction algorithm, which helps identify resource bottleneck areas (e.g., the central processor usage of a certain server node is consistently higher than 90% and the proportion of abnormal tasks is high). Combine task priority and scheduling strategy (such as allocating more computing resources to high-priority tasks first), dynamically adjust the resource scheduling scheme; consider the periodic characteristics of system load and business scenarios (such as the surge in order volume during the promotion of an e-commerce platform), and determine the resource investment area that needs to be optimized.

[0025] Specifically, the S7 sub-step is: S7.1 Data processing: obtain resource investment area data and task execution status data through stream processing tools, and verify the integrity of the data source; if the data loss rate is less than 5%, it is determined that the data source integrity meets the preset threshold, and a first data set containing resource investment areas and task execution status is generated, including the central processor usage of a certain server node, task completion rate, and task priority information.

[0026] S7.2 Calculation and adjustment: use distributed computing tools to calculate task completion rate and resource utilization, and process resource usage data of multiple nodes in parallel to calculate the average task completion rate and resource utilization of each node. If the task completion rate is lower than the preset threshold (85%), adjust the resource allocation weight (such as increasing the central processor allocation proportion of high-load nodes), and generate a second data set containing the optimized task completion rate and resource utilization.

[0027] S7.3 View generation: Map the task completion rate and resource utilization in the second dataset to the dashboard dynamic components, including the optimized task completion rate curve and resource utilization bar chart, through the data visualization tool. The view refresh period control mechanism ensures that the dynamic components are updated every minute, reflecting the latest resource allocation effect. Combined with business priority (such as the completion rate of high-priority tasks in a financial transaction system) and multi-dimensional data analysis (including resource utilization, task completion rate, and system load distribution), the final decision support view is generated.

[0028] A message-driven enterprise multi-service system centralized to-do processing system adopts the message-driven enterprise multi-service system centralized to-do processing method, and the system comprises: A data acquisition and clustering analysis module: acquires resource input data and task processing data from a database, and performs grouping processing on the resource input data and task processing data through a clustering algorithm to obtain classified resource consumption clusters and task state clusters; A correlation analysis module: according to the classified resource consumption clusters and task state clusters, a regression algorithm is used to analyze the correlation between resource utilization and task completion rate, and a resource utilization trend is determined; A resource optimization index generation module: if the resource utilization trend exceeds a preset threshold, the average value and variance of the task completion rate are calculated through a data aggregation function to obtain a resource optimization index set; An interactive model construction module: for the resource optimization index set, historical decision data is acquired as reference input, and a decision tree algorithm is used to construct an interactive model to determine potential input strategy adjustment points; A visual rendering module: through the adjustment points generated by the interactive model, intuitive charts including bar charts and heat maps are rendered in a dashboard framework to obtain a visual resource utilization distribution; An abnormality screening and region determination module: real-time monitoring parameters are extracted from the visual resource utilization distribution, and a filtering function is used to screen abnormal task states to determine resource input regions that need to be optimized; A decision support view generation module: according to the determined resource input regions, the dynamic components of the dashboard are updated to reflect the optimized task completion rate, and the final decision support view is obtained.

[0029] A computer device comprising: One or more processors, the processor storing one or more executable programs, When the one or more executable programs are executed by the one or more processors, the one or more executable programs are used to implement the message-driven enterprise multi-service system centralized to-do processing method.

[0030] The present application can achieve the following beneficial effects: The present application groups the resource input data and task processing data by clustering algorithm to obtain resource consumption clusters and task state clusters, analyzes the correlation between resource utilization rate and task completion rate by regression algorithm, accurately identifies the matching rules of resources and tasks under different clusters, avoids resource bottlenecks (such as low completion rate of high-load tasks due to insufficient resources) or resource redundancy (such as too many resources occupied by low-load tasks), and realizes efficient use of resources.

[0031] By presetting the threshold value to determine the change trend of resource utilization rate, the statistical characteristics of task completion rate are calculated by using data aggregation function when triggered, and the resource allocation scheme is iteratively updated by combining dynamic adjustment mechanism, so that the resource allocation can respond to the changes of task demand in real time; at the same time, based on the historical decision data, a decision tree interactive model is constructed, which can automatically identify potential input strategy adjustment points, reduce the subjectivity of manual decision, and improve the scientificity and timeliness of resource optimization.

[0032] By rendering column chart, heat map and other visual charts through the dashboard framework, the distribution of resource utilization rate can be intuitively presented, and the management personnel can quickly identify high-load nodes, resource bottleneck areas and abnormal task states; combined with the dynamically updated decision support view, the changes of task completion rate after optimization can be mastered in real time, and the decision efficiency and accuracy are significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0033] The present application will be further described below in conjunction with the drawings and examples: Fig. 1 The present application is a method flowchart. Fig. 2 The present application is a system structure diagram. DETAILED DESCRIPTION

[0034] The preferred scheme is as shown in Figs. 1-2 A message-driven enterprise multi-service system centralized to-be-done processing method, which can specifically include: Step S1, obtaining resource input data and task processing data from the database, grouping these data by clustering algorithm to obtain classified resource consumption clusters and task state clusters.

[0035] Specifically, when extracting resource input and task processing data from the database, structured query language is usually used. For example, the database can be an internal resource management system of an enterprise, which stores the server's computing resource usage, storage space occupation, and task execution log records. Resource input data includes, but is not limited to, central processor usage, memory usage, disk read / write speed, etc.; task processing data includes task completion time, task success rate, task failure times, etc. Structured query language can filter records that meet the time range or data format according to the preset field conditions. For example, in this embodiment, it is assumed that the database stores server running data within a month, and the query statement extracts the central processor usage and task completion status fields for a specific time period to ensure data integrity.

[0036] After obtaining the data, the data needs to be preprocessed to clean up missing values and outliers. For example, missing values can be caused by sensor failure or log interruption, and outliers can be caused by extreme values due to system failure. The cleaning process can identify outliers through statistical methods, such as using a threshold method based on standard deviation to mark and remove data points that exceed three times the standard deviation. Missing values can be filled by interpolation methods, such as using the average of the previous and subsequent data to fill in. The cleaned dataset is standardized, such as scaling the data to a uniform range through normalization, to ensure the accuracy of subsequent analysis.

[0037] Feature extraction of the standardized dataset is a key step. Feature extraction methods calculate statistical features based on resource input and task processing. For example, for resource input, the mean, peak, and variance of central processor usage can be extracted; for task processing, the distribution of task completion rate and the average of task response time can be extracted. These statistical features generate multi-dimensional feature vectors combined with task status. For example, in addition, task status includes three categories: "success", "failure", and "timeout", combined with statistical features of resource input, a 10-dimensional feature vector is generated, including central processor usage mean, task completion rate, and task timeout rate. These feature vectors constitute the feature set for the clustering algorithm.

[0038] In this embodiment, the k-means clustering algorithm is used to group the feature set. The k-means clustering algorithm assigns data points to different clusters through iterative optimization, ensuring that data points within each cluster have similar features. Algorithm parameters include the number of clusters k and the number of iterations. For example, suppose we are analyzing resource usage in a data center, initially set k to 3, indicating that the data is divided into three clusters of high consumption, medium consumption and low consumption. Preferably, the clustering effect is evaluated by the silhouette coefficient, and the k value is adjusted to optimize the classification of resource consumption patterns and task status. After clustering, the clustering grouping result and cluster label are generated, for example, the high consumption cluster may correspond to the high load task scenario, and the low consumption cluster corresponds to the light load task scenario.

[0039] Through the clustering grouping result, the relationship between the cluster label and the resource consumption pattern and the task status is analyzed. For example, the high consumption cluster may show that the central processor usage rate is consistently higher than 80%, while the task completion rate is low, indicating a resource bottleneck; the low consumption cluster may show that the resource usage rate is lower than 30%, but the task timeout rate is high, indicating insufficient resource allocation. Based on this analysis, cluster distribution statistics are generated to determine the final resource consumption cluster and task status cluster. For example, in the data center scenario, the resource consumption cluster can be divided into three categories: high load, medium load, and low load, and the task status cluster can be divided into three categories: efficient completion, inefficient completion, and failure. These clusters provide a data basis for subsequent resource optimization.

[0040] Step S2, according to the classified resource consumption cluster, a regression algorithm is used to analyze the correlation between resource usage rate and task completion rate, and the trend of resource usage rate is determined.

[0041] Specifically, when extracting the classified resource consumption cluster data from the database, the structured query language is also used to obtain the historical records of resource usage rate and task completion rate. For example, in an enterprise resource management system, resource usage rate data includes server central processor usage rate, memory occupancy rate, etc., and task completion rate data includes the proportion of successful task completion. The data cleaning process is similar to step S1, which removes missing values and outliers to ensure the standardization of the data set. For example, in this embodiment, for the running log of the server cluster, the cleaning process will eliminate abnormal high usage rate records caused by system restart, and fill in the missing task completion rate data through linear interpolation.

[0042] The Pearson correlation coefficient method is used to calculate the statistical characteristics of resource usage and task completion rate. The Pearson correlation coefficient is used to measure the linear correlation between two variables, and its value ranges from -1 to 1. For example, the correlation coefficient between CPU usage and task completion rate is calculated. If the result is close to 1, it indicates that the two are positively correlated, which may mean that increasing resource investment can significantly improve task completion rate; if it is close to 0, it indicates that the correlation between the two is weak. Preferably, the threshold of the correlation coefficient is set to 0.6, and if the correlation coefficient exceeds the threshold, it is considered that there is a significant correlation, and enters the regression analysis stage.

[0043] The linear regression algorithm is used to fit the relationship between resource usage and task completion rate. For example, in the data center scenario, assuming that resource usage is the independent variable and task completion rate is the dependent variable, the regression algorithm fits a curve through the least squares method, generating trend parameters such as the slope and intercept of the fitted curve. These parameters reflect the impact of changes in resource usage on task completion rate. For example, a steep slope indicates that a small change in resource usage will result in a significant change in task completion rate. In combination with the task type distribution, such as batch processing tasks and real-time processing tasks, distribution statistics of resource usage and task completion rate are generated. For example, in this embodiment, batch processing tasks may show a strong positive correlation between resource usage and task completion rate, while real-time processing tasks may have weaker correlation, indicating that different resource optimization strategies are needed for different task types.

[0044] Step S3, if the resource usage change trend exceeds the preset threshold, the average and variance of the task completion rate are calculated through the data aggregation function to obtain a set of resource optimization indicators.

[0045] In this embodiment, resource usage data is obtained from time series data to determine whether the change trend exceeds the preset threshold. For example, assuming that the preset threshold is that the resource usage exceeds 85% for three consecutive days, by analyzing the time series data, it is determined whether the optimization condition is triggered. If the condition is met, the task completion rate data is processed through the data aggregation function. For example, the data aggregation function can calculate the average and variance of the task completion rate within a period of time to represent the stability and efficiency of task execution. For example, in the data center scenario, if the average task completion rate is 90% and the variance is small, it indicates that the task execution is stable; if the variance is large, it indicates that the task completion rate fluctuates significantly, which may indicate a problem of uneven resource allocation.

[0046] Specifically, the data aggregation function can calculate the task completion rate data based on a time window. For example, in addition, set the time window to 24 hours, calculate the average and variance of the task completion rate per hour, and generate a set of statistical features. These feature sets reflect the overall trend and fluctuation of task execution. For example, high variance may correspond to a resource usage peak, and low variance may correspond to a resource usage trough. Based on the set of statistical features, the preset logical rules are used to adjust the resource allocation ratio. For example, if the average task completion rate is less than 80% and the variance is large, the allocation ratio of central processing unit or memory is increased; if the task completion rate is high and stable, the current allocation ratio is maintained.

[0047] Preferably, through a dynamic adjustment mechanism, the resource allocation scheme is iteratively updated according to the updated resource usage rate data in the time series. For example, in an enterprise resource management system, the dynamic adjustment mechanism can recalculate the allocation ratio every hour according to the latest resource usage rate data, ensuring that the resource allocation matches the task demand. The final generated resource optimization index set includes the adjusted resource allocation ratio, task completion rate statistical features, etc. For example, in this embodiment, the resource optimization index set may include a 10% increase in central processing unit allocation ratio, a 5% decrease in memory allocation ratio, and a target value of 95% for the task completion rate. These indicators provide a data basis for subsequent decision support.

[0048] The generation of the resource optimization index set can be combined with specific business scenarios. For example, in an order processing system of an e-commerce platform, resource usage rate may be related to order volume peak. Assuming that the increase in order volume during the peak period causes the central processing unit usage rate to exceed 90%, the task completion rate drops to 70%. Through the data aggregation function, the average and variance of the order processing task completion rate are calculated, and it is found that the completion rate fluctuates greatly. Based on this, the resource allocation of the server cluster is adjusted, for example, the proportion of computing nodes is increased and the proportion of storage nodes is reduced, and finally the optimization index set is generated. This scenario-based adjustment can effectively respond to the resource demand during the business peak period.

[0049] The implementation of the dynamic adjustment mechanism depends on the support of real-time data streams. For example, in a data center scenario, real-time data streams can collect resource usage rate and task completion rate data every minute through monitoring tools. The dynamic adjustment mechanism will update the resource allocation scheme regularly based on these data. For example, if it is monitored that the resource usage rate of a certain server node continuously exceeds the threshold, the dynamic adjustment mechanism will preferentially allocate tasks to other low-load nodes to ensure the overall efficiency of the system. This dynamic adjustment can significantly improve resource utilization efficiency and reduce the probability of task failure.

[0050] The generation of the resource optimization indicator set can be combined with task priority. For example, in a financial transaction system, high-priority tasks such as real-time transaction processing require higher resource guarantees. By analyzing the statistical characteristics of task completion rate, more computing resources are allocated to high-priority tasks. For example, if the completion rate of high-priority tasks is lower than expected, the dynamic adjustment mechanism will increase the number of central processing unit cores allocated, while reducing the resource occupation of low-priority tasks. This priority-based adjustment can ensure the stable operation of critical business.

[0051] Through the above steps, the complete process from data acquisition, feature extraction to resource optimization indicator generation is completed. These steps realize the effective matching of resource use and task completion through clustering analysis, regression analysis and dynamic adjustment, providing a solid foundation for subsequent decision support and visualization. The implementation of steps S4 to S7 will be described in the second generation.

[0052] In step S4, for the resource optimization indicator set, historical decision data is obtained as reference input, and a decision tree algorithm is used to build an interactive model to determine potential input strategy adjustment points.

[0053] Specifically, when obtaining resource optimization indicators from historical decision data, resource allocation records and task execution results in the past period are usually obtained through database queries. For example, in an enterprise resource management system, historical decision data includes resource allocation proportion of server clusters, task completion rate, and corresponding business scenario information such as task load during peak or off-peak periods. The data cleaning process is similar to the previous steps, removing outliers and missing values. For example, outliers may be due to system failure, and missing values may be due to log interruption. The cleaned data set is extracted through feature selection criteria to extract key features related to business target weights. For example, in this embodiment, key features include central processor allocation proportion, average and variance of task completion rate, task priority, etc., generating a model training set for training the decision tree algorithm.

[0054] The model training set is trained using a decision tree algorithm. The decision tree algorithm recursively partitions the feature space to build a tree structure based on feature conditions, which is used to determine whether resource allocation needs to be adjusted. For example, in a data center scenario, the root node of the decision tree may be based on whether the central processor usage rate exceeds 80%, and the branch conditions include whether the task completion rate is less than 90% and whether the task priority is high. If the prediction accuracy of the trained decision tree model is higher than the preset threshold, for example 85%, an interactive model is generated. The interactive model can output potential input strategy adjustment points according to the input resource optimization indicators. For example, the model may suggest increasing the allocation proportion of computing nodes when the central processor usage rate is high and the task completion rate is low.

[0055] Preferably, the analysis process of the interactive model will be combined with specific business scenarios. For example, in the order processing system of an e-commerce platform, historical decision-making data may show that during promotional activities, the order volume surged, causing the central processor usage rate to exceed 90% and the task completion rate to drop to 70%. The decision tree model will analyze these data and identify adjustment points that require increased computing resources. For example, in this embodiment, the model suggests increasing the central processor allocation ratio from 50% to 70% while reducing the resource occupation of storage nodes. This adjustment point is based on patterns in historical data and can effectively cope with similar high-load scenarios.

[0056] The construction of the interactive model needs to ensure the rationality of feature selection. For example, in a financial transaction system, key features may include transaction response time, transaction failure rate, and resource usage rate. The decision tree model will generate different branch paths based on these features. For example, if the transaction response time exceeds 1 second and the resource usage rate is higher than 85%, the model may suggest increasing computing resources; if the transaction failure rate is high but the resource usage rate is low, the model may suggest optimizing task scheduling strategies. In this way, the interactive model can provide personalized adjustment suggestions for different business scenarios.

[0057] In addition, the interactive model can analyze in combination with the time dimension. For example, in an enterprise resource management system, historical decision-making data may show that the resource usage rate and task completion rate patterns on weekdays are different from those on weekends. The decision tree model will generate different adjustment points based on time features such as weekdays or weekends. For example, weekdays may require increased computing resources to cope with high loads, while weekends may require reduced resources to reduce costs. This analysis based on the time dimension can further improve the accuracy of resource allocation.

[0058] Step S5, through the adjustment points generated by the interactive model, intuitive charts such as column charts and heat maps are rendered in the dashboard framework to obtain visual resource usage rate distribution.

[0059] Specifically, when obtaining real-time resource usage rate data from the data interaction interface, it is usually obtained through a real-time data stream collection module. For example, in a data center scenario, the data interaction interface collects server central processor usage rate, memory occupancy rate, etc. data every minute. Dynamic adjustment algorithm is used to generate points from these data. For example, the dynamic adjustment algorithm can generate smoothed data points according to the fluctuation of resource usage rate, reducing the interference of short-term fluctuations. The generated point data set is used to render visual charts.

[0060] In this embodiment, the resource occupation column chart is generated by the column chart rendering tool. For example, in the resource management system of an e-commerce platform, the column chart can show the central processor usage distribution of different server nodes, and each column corresponds to the average usage rate of a node. The height of the column chart reflects the resource usage intensity, and the color distinguishes different node types (such as computing nodes, storage nodes). This intuitive visualization can help administrators quickly identify high-load nodes. For example, if the column chart height of a certain node is significantly higher than that of other nodes, it indicates that the node may have a resource bottleneck and needs to adjust resource allocation.

[0061] Preferably, the density distribution data is obtained from the point data set, and the heat map is generated by using the color mapping method. For example, in a financial transaction system, the heat map can show the resource usage density of the transaction processing server, and the color from blue to red represents the resource usage rate from low to high. The area division of the heat map is based on the physical location or logical grouping of the server, such as displaying by data center partition. The administrator can quickly identify high-density areas of resource usage through the heat map, such as the computing resource usage rate of a certain data center is generally high, indicating that the resource investment in this area needs to be increased.

[0062] In addition, the resource occupation column chart and the resource density heat map are used in combination to analyze the resource allocation data through the performance monitoring tool. For example, the performance monitoring tool can compare the data consistency of the column chart and the heat map to verify the accuracy of the resource usage distribution. If it is found that the column chart of a certain node shows high usage rate, while the heat map shows that the overall resource usage of the area is low, it may indicate that the node has an abnormality and needs to be further checked. This multi-dimensional visualization analysis can improve the transparency and accuracy of resource allocation.

[0063] The rendering process of the dashboard framework needs to consider real-time and interactivity. For example, in an enterprise resource management system, the dashboard framework can dynamically update the chart content through the front-end framework to ensure that the administrator can view the resource usage in real time. The interactive functions include clicking a certain column in the column chart to display the detailed resource usage data of the node, or selecting a certain area in the heat map to view the running status of the specific server. This interactive design can significantly improve the decision-making efficiency of the administrator.

[0064] Step S6, extract real-time monitoring parameters from the visual resource usage distribution, filter abnormal task states by using the filtering function, and determine the resource investment area that needs to be optimized.

[0065] In this embodiment, when acquiring real-time monitoring data streams from the data acquisition module, the resource usage rate and task status data are usually collected by the monitoring tool every second. For example, in a data center scenario, the monitoring data stream includes central processor usage rate, task completion status, task response time, etc. A preset threshold is used to determine whether the resource usage rate exceeds the normal range. For example, if the central processor usage rate exceeds 90%, it is marked as abnormal. Based on this, a preliminary list of abnormal tasks is generated, such as tasks with timeout, task failure, etc.

[0066] Specifically, a filtering function is used to screen the preliminary list of abnormal tasks. For example, the filtering function can determine the specific range of abnormal tasks based on the threshold of task status, such as task response time exceeding 2 seconds or task failure rate higher than 5%. For example, in an order processing system of an e-commerce platform, the filtering function may identify that the order payment task is abnormal due to excessive response time. The screened abnormal task list is used for further analysis of resource usage.

[0067] In addition, a distribution extraction algorithm is used to generate a visual distribution diagram of resource usage from the monitoring parameters. For example, the distribution diagram can show the resource usage distribution of the server when the abnormal task occurs, helping administrators identify resource bottleneck areas. For example, if the distribution diagram shows that the central processor usage rate of a certain server node is consistently higher than 90% and the proportion of abnormal tasks is high, it indicates that this node is a resource investment area that needs to be optimized. Based on this, the resource scheduling scheme is dynamically adjusted in combination with task priority and scheduling strategy. For example, more computing resources are allocated to high-priority tasks to reduce resource occupation of low-priority tasks.

[0068] Preferably, the determination of the optimization area needs to consider the system load comprehensively. For example, in a financial transaction system, high load scenarios may cause the response time of transaction processing tasks to be prolonged. Through the distribution extraction algorithm, the server node of the high load area is identified, and the resource scheduling scheme is adjusted, such as increasing the number of computing nodes or improving the bandwidth. This dynamic adjustment can effectively reduce the occurrence of abnormal tasks and improve the overall stability of the system.

[0069] In addition, the optimization of the resource investment area can be combined with the periodic characteristics of the business scenario. For example, in an e-commerce platform, the surge in order volume during promotional activities may lead to an increase in abnormal tasks. By analyzing real-time monitoring data streams, high load nodes during promotional activities are identified, and the resource scheduling scheme is dynamically adjusted. For example, more computing resources are allocated to order processing nodes, while resource occupation of non-critical tasks is reduced. This optimization based on business cycles can significantly improve task completion efficiency.

[0070] Step S7, update the dashboard dynamic component according to the determined resource input area to reflect the optimized task completion rate, and obtain the final decision support view.

[0071] Specifically, when acquiring the resource input area data and task execution status data from the real-time data stream, the data source integrity is usually verified by a stream processing tool. For example, in a data center scenario, the stream processing tool can check whether the data stream contains complete resource usage and task completion rate records. If the data source integrity meets a preset threshold, for example, the data missing rate is less than 5%, a first data set containing the resource input area and the task execution status is generated. For example, the first data set can include the central processor usage, the task completion rate, and the task priority information of a certain server node.

[0072] In this embodiment, a distributed computing tool is used to calculate the task completion rate and the resource utilization rate. For example, in an enterprise resource management system, the distributed computing tool can process the resource usage data of multiple nodes in parallel to calculate the average task completion rate and the resource utilization rate of each node. If the task completion rate is lower than a preset threshold, for example, 85%, the resource allocation weight is adjusted. For example, the central processor allocation proportion of a high-load node is increased, and a second data set containing the optimized task completion rate and the resource utilization rate is generated. This data set reflects the effect of the adjusted resource allocation.

[0073] Preferably, the task completion rate and the resource utilization rate in the second data set are mapped to the dashboard dynamic component by a data visualization tool. For example, in a resource management system of an e-commerce platform, the dashboard dynamic component can display the optimized task completion rate curve and the resource utilization rate bar chart. The view refresh period control mechanism ensures that the dynamic component is updated every minute to reflect the latest resource allocation effect. For example, if the optimized task completion rate is increased from 80% to 95%, the dynamic component will update the curve in real time, and the administrator can visually see the optimization effect.

[0074] In addition, the update of the dashboard dynamic component needs to consider the view interactivity. For example, the administrator can view the detailed resource allocation and task execution data of a certain node in the dynamic component by clicking the node. An interactive front-end framework supports such dynamic interaction, for example, by adjusting the resource allocation proportion through a drag operation to view the change of the task completion rate in real time. Such interactive design can significantly improve the decision-making efficiency.

[0075] The generation of the final decision support view requires the combination of business priorities. For example, in a financial transaction system, the completion rate of high-priority tasks is a key indicator. The interactive front-end framework will prioritize the display of resource allocation and completion rate data for high-priority tasks, ensuring that administrators can quickly make decisions. For example, if the view shows that the completion rate of high-priority tasks is 98%, indicating that the resource optimization solution is effective, the administrator can further promote the solution.

[0076] In addition, the decision support view can incorporate multi-dimensional data analysis. For example, in a data center scenario, the view can simultaneously display resource utilization, task completion rate, and system load distribution. Administrators can switch between different dimensions of the view through interactive operations, such as switching from a resource utilization view to a task completion rate view, to gain a comprehensive understanding of system operation. This multi-dimensional view design can provide comprehensive decision support for complex business scenarios.

[0077] Through the above steps, the complete process from resource optimization indicator analysis to final decision support view generation is completed. These steps achieve precise optimization of resource allocation and significant improvement in task completion efficiency through interactive models, visual charts, and dynamic adjustments, providing an efficient decision support tool for enterprise resource management.

[0078] The above embodiments are only preferred technical solutions of the present application, and should not be considered as limiting the present application. The protection scope of the present application should be based on the technical solutions recited in the claims, including equivalent replacement solutions of the technical features recited in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present application.

Claims

1. A message-driven based centralized to-do processing method for enterprise multi-business systems, characterized in that The method comprises the following steps: S1: obtaining resource input data and task processing data from the database, grouping the resource input data and the task processing data through a clustering algorithm to obtain classified resource consumption clusters and task state clusters; S2: according to the classified resource consumption clusters and the task state clusters, analyzing the correlation between the resource usage rate and the task completion rate by using a regression algorithm to determine the change trend of the resource usage rate; S3: if the change trend of the resource usage rate exceeds a preset threshold, calculating the average value and the variance of the task completion rate by using a data aggregation function to obtain a resource optimization index set; S4: for the resource optimization index set, obtaining historical decision data as a reference input, constructing an interactive model by using a decision tree algorithm to determine a potential input strategy adjustment point; S5: rendering intuitive charts including a column chart and a heat map in a dashboard framework through the adjustment point generated by the interactive model to obtain a visual resource usage rate distribution; S6: extracting real-time monitoring parameters from the visual resource usage rate distribution, screening abnormal task states by using a filtering function to determine a resource input area that needs to be optimized; S7: according to the determined resource input area, updating a dynamic component of the dashboard to reflect the optimized task completion rate to obtain a final decision support view.

2. The centralized to-do processing method for enterprise multi-business systems based on message driving according to claim 1, characterized in that: The substep S1 comprises: S1.1: extracting resource input data and task processing data from the database by using a structured query language, and screening records meeting a time range or a data format; The resource input data includes a central processing unit usage rate, a memory occupancy, and a disk read-write speed; The task processing data includes a task completion time, a task success rate, and a task failure number; The database includes an enterprise internal resource management system, and stores a computing resource usage of a server, a storage space occupancy, and a task execution log record; S1.2: cleaning the resource input data and the task processing data obtained from the database, identifying and removing outliers by using a threshold method based on a standard deviation, and filling in missing values by using an interpolation method; and normalizing the cleaned data set to scale the data to a unified range; S1.3: calculating statistical features according to the resource input data and the task processing data; extracting a mean value, a peak value, and a variance of the central processing unit usage rate for the resource input data; extracting a distribution of the task completion rate and an average value of a task response time for the task processing data; and combining the task states to generate multi-dimensional feature vectors, which constitute a feature set for the clustering algorithm; S1.4: grouping the feature set by using a k-means clustering algorithm, and the algorithm parameters include a clustering number k and an iteration number; adjusting the k value to optimize the classification by evaluating the clustering effect through a silhouette coefficient; generating clustering grouping results and cluster labels after clustering; and analyzing the relationship between the cluster labels and resource consumption modes and task states; Based on the analysis, generate cluster distribution statistical data to determine the final resource consumption clusters and the task state clusters.

3. The centralized to-do processing method for message-driven enterprise multi-business system according to claim 1, characterized in that: The substep S2 comprises: S2.1 Obtain the classified resource consumption cluster data from the database through the structured query language, including the resource usage data of the central processing unit usage rate and memory occupancy rate of the server, and the task completion rate data of the proportion of successful task completion; clean the data, remove the abnormally high usage rate records caused by system restart, fill in the missing task completion rate data through linear interpolation; standardize the cleaned data set; S2.2 Calculate the statistical characteristics of resource usage and task completion rate by Pearson correlation coefficient method, Pearson correlation coefficient is used to measure the linear correlation between two variables, and its value range is between-1 and 1; set the threshold of correlation coefficient as 0.6, if the correlation coefficient exceeds the threshold, it is considered that there is significant correlation between resource usage and task completion rate, and it enters the regression analysis stage; S2.3 Fit the relationship between resource usage and task completion rate by linear regression algorithm, take resource usage as independent variable and task completion rate as dependent variable, fit the curve by least square method, generate trend parameters including slope and intercept of the fitted curve; combine the task type distribution to generate distribution statistical data of resource usage and task completion rate.

4. The centralized to-do processing method for message-driven enterprise multi-business system according to claim 1, characterized in that: The sub-steps of S3 are: S3.1 Obtain the resource usage data from the time series data, and judge whether the change trend exceeds the preset threshold; The preset threshold is that the resource usage exceeds 85% for three consecutive days; by analyzing the time series data, if the optimization condition is triggered, the subsequent steps are entered; S3.2 Process the task completion rate data by data aggregation function, calculate the average value and variance of task completion rate in a period of time; set the time window as 24 hours, calculate the average value and variance of task completion rate per hour based on the time window, generate a set of statistical characteristics; S3.3 Adjust the resource allocation proportion according to the set of statistical characteristics by using the preset logical rule; if the average value of task completion rate is lower than 80% and the variance is larger, increase the allocation proportion of central processing unit or memory; if the task completion rate is high and stable, maintain the current allocation proportion; through dynamic adjustment mechanism, update the resource allocation scheme iteratively according to the updated resource usage data in time series; finally generate a set of resource optimization indicators, including the adjusted resource allocation proportion and task completion rate statistical characteristics.

5. The centralized to-do processing method for message-driven enterprise multi-business system based on message driving according to claim 1, characterized in that: The sub-steps of S4 are: S4.1 Obtain historical decision data from the database, including resource allocation proportion of server cluster, task completion rate and corresponding business scenario information; clean the data to remove outliers and missing values; extract key features related to business target weight through feature selection standard, generate model training set for training decision tree algorithm; S4.2 Train the model training set by using decision tree algorithm, construct tree structure based on feature conditions by recursive partitioning feature space; if the prediction accuracy of the trained decision tree model is higher than the preset threshold, generate an interactive model; the interactive model can output potential input strategy adjustment points according to the input resource optimization indicators; S4.3 The analysis process of the interactive model combines specific business scenarios and provides personalized adjustment suggestions for different business scenarios based on the patterns in historical data.

6. The centralized to-do processing method for message-driven enterprise multi-business system based on message driving according to claim 1, characterized in that: S5 The sub-steps are: S5.1 Obtain real-time resource usage data, including server central processor usage and memory occupancy, through a real-time data stream acquisition module. Use a dynamic adjustment algorithm to generate points from these data, generating smoothed data points. The generated point data set is used to render a visual chart; S5.2 Generate a resource occupancy column chart through a column chart rendering tool to show the central processor usage distribution of different server nodes. Each column corresponds to the average usage of a node, and the height of the column chart reflects the resource usage intensity. Different node types are distinguished by color. Obtain density distribution data from the point data set and generate a heat map using a color mapping method to show the resource usage density of the transaction processing server. The color changes from blue to red to represent the resource usage rate from low to high. The heat map area is divided based on the physical location or logical grouping of the server; S5.3 The dashboard framework dynamically updates the chart content through the front-end framework to ensure real-time viewing of resource usage. Interactive functions include clicking on a column in the column chart to display detailed resource usage data for that node, and selecting an area in the heat map to view the running status of a specific server.

7. The centralized to-do processing method for message-driven enterprise multi-business system based on message driving according to claim 1, characterized in that: S6 The sub-steps are: S6.1 Collect resource usage and task status data every second through monitoring tools. The monitoring data stream includes central processor usage, task completion status, and task response time. Extract these real-time monitoring parameters from the visual resource usage distribution; Use a preset threshold to determine whether the resource usage exceeds the normal range. If the central processor usage exceeds 90%, it is marked as abnormal; Based on this, generate a preliminary list of abnormal tasks, including tasks that exceed the timeout and fail; S6.2 Use a filtering function to filter the preliminary list of abnormal tasks. The filtering function is based on the threshold of task status to determine the specific range of abnormal tasks. The filtered abnormal task list is used for further analysis of resource usage; S6.3 Generate a visual distribution chart of resource usage from monitoring parameters through a distribution extraction algorithm to help identify resource bottleneck areas. Combine task priority and scheduling strategy to dynamically adjust resource scheduling schemes. Consider the periodic characteristics of system load and business scenarios to determine the resource investment area that needs to be optimized.

8. The centralized to-do processing method for message-driven enterprise multi-business system based on message driving according to claim 1, characterized in that: S7 The sub-steps are: S7.1 Obtain resource investment area data and task execution status data through a stream processing tool to verify the integrity of the data source. If the data loss rate is less than 5%, the data source integrity meets the preset threshold, and a first data set containing resource investment areas and task execution status is generated. The first data set includes the central processor usage, task completion rate, and task priority information of a server node; S7.2 calculates the task completion rate and resource utilization rate using distributed computing tools, processes the resource usage data of multiple nodes in parallel, calculates the average task completion rate and resource utilization rate of each node, and adjusts the resource allocation weight if the task completion rate is lower than the preset threshold to generate a second data set containing the optimized task completion rate and resource utilization rate; S7.3 View generation: map the task completion rate and resource utilization rate in the second data set to the dashboard dynamic components through data visualization tools, including the optimized task completion rate curve and resource utilization rate bar chart; the view refresh cycle control mechanism ensures that the dynamic components are updated every minute to reflect the latest resource allocation effect; combined with business priority and multidimensional data analysis, the final decision support view is generated.

9. A message-driven based enterprise multi-business system centralized to-do processing system, characterized in that: A message-driven enterprise multi-service system centralized to-do processing method is adopted according to any one of claims 1-8, the system comprising: A data acquisition and clustering analysis module: acquires resource input data and task processing data from a database, and groups the resource input data and task processing data through a clustering algorithm to obtain classified resource consumption clusters and task state clusters; A correlation analysis module: according to the classified resource consumption clusters and task state clusters, a regression algorithm is used to analyze the correlation between resource usage rate and task completion rate to determine the resource usage rate trend; A resource optimization index generation module: if the resource usage rate trend exceeds the preset threshold, the average value and variance of the task completion rate are calculated through a data aggregation function to obtain a resource optimization index set; An interactive model construction module: for the resource optimization index set, historical decision data is obtained as reference input, and a decision tree algorithm is used to construct an interactive model to determine potential input strategy adjustment points; A visual rendering module: through the adjustment points generated by the interactive model, intuitive charts including bar charts and heat maps are rendered in the dashboard framework to obtain visual resource usage rate distribution; An abnormality screening and region determination module: real-time monitoring parameters are extracted from the visual resource usage rate distribution, and a filtering function is used to screen abnormal task states to determine the resource input area that needs to be optimized; A decision support view generation module: according to the determined resource input area, the dynamic components of the dashboard are updated to reflect the optimized task completion rate to obtain the final decision support view.

10. A computer device, comprising: It comprises: One or more processors, said processor having one or more executable programs stored thereon, When the one or more executable programs are executed by the one or more processors, a message-driven enterprise multi-service system centralized to-do processing method according to any one of claims 1-8 is implemented.

Images