Multi-key indicator fusion service resource scheduling method based on AI big model drive

Abstract

The present invention discloses a multi-key indicator fusion service resource scheduling method driven by an AI big model. The present invention relates to the field of intelligent resource scheduling technology. When a new service resource is generated, the scheduling process is automatically triggered to identify and collect the service task information of the service resource; collect the latest engineer-related data, and integrate the engineer data and task information to obtain a multi-dimensional task data set. The multi-key indicator fusion service resource scheduling method driven by an AI big model avoids excessive fatigue and unnecessary waste of travel time for engineers by comprehensively considering factors such as work saturation and distance, can respond to resources quickly, and reduce the average time of resource processing, thereby improving overall service efficiency. In addition, resources are allocated according to the engineer's skills and historical praise rate, ensuring that resources are handled by engineers with corresponding capabilities and high service quality, thereby improving the success rate of resource processing and customer satisfaction.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent resource scheduling, and specifically to a multi-key indicator fusion service resource scheduling method driven by an AI big model. Background Art

[0002] In modern service-oriented enterprises, resource scheduling is a key link to ensure service quality and efficiency. Traditional resource scheduling methods often have many limitations. For example, resources are allocated based on simple scheduling rules without fully considering the current work saturation of engineers. This may cause some engineers to be overloaded with tasks while others are idle, affecting the overall service efficiency.

[0003] Moreover, most existing scheduling systems do not fully integrate key information such as engineer skills, service areas, engineer levels, distance from resource locations, and urgency, which makes resource allocation unreasonable. Resources requiring specific skills may be allocated to engineers who do not have the corresponding capabilities, or engineers who are far away from the resource location may be arranged to provide services, increasing response time and cost.

[0004] With the development of artificial intelligence technology, especially the powerful capabilities of AI big models in data processing and analysis, the multi-key indicator fusion service resource scheduling method driven by it has become a new way to solve the above problems. Summary of the invention

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a multi-key indicator fusion service resource scheduling method driven by an AI big model, comprising the following steps:

[0006] Step 1: When a new service resource is generated, the scheduling process is automatically triggered to identify and collect the service task information of the service resource, where the service task information includes service content, service location, and resource urgency parameters;

[0007] Step 2: Collect the latest engineer-related data, including scheduling, work saturation, skills, service areas, and historical praise rate information, and integrate engineer data and task information to obtain a multi-dimensional task data set;

[0008] Step 3: Input the relevant data of the multi-dimensional task data set into the AI ​​big model, use the machine learning algorithm to perform feature analysis, and use the neural network architecture in deep learning to perform fusion analysis on multiple key indicator data;

[0009] Step 4: Use the AI ​​big model to process and analyze the input data and calculate the matching score of each engineer for the service task;

[0010] Step 5: Sort all available engineers according to the engineer matching scores output by the AI ​​big model, and analyze the matching degree between the engineers and the service tasks;

[0011] Step 6: After determining the engineer who will perform the task, the scheduling result will be notified to the engineer and the relevant service management system. After receiving the task notification, the engineer will start to perform the service task. During the task execution process, the engineer's work saturation information will be continuously collected and updated so that new resources that may appear in the future can be scheduled in a timely manner.

[0012] Preferably, in step 1, the process of obtaining service task information includes:

[0013] New service resources (maintenance requests, customer inquiries, orders, etc.) enter the service management system through the system interface, API interface, email, and SMS. The service management system monitors the entry of new service resources in real time through the event monitoring mechanism. When new service resources are detected, the trigger configuration mechanism is automatically activated to trigger the scheduling process. The service management system has multi-channel receiving capabilities to ensure that no new service resources are missed;

[0014] Analyze the service task content in the new service resources, identify specific maintenance types (such as electrical appliance maintenance, pipe unblocking, etc.) or consulting question categories (such as product usage consultation, after-sales service, etc.), and use natural language processing technology to convert the service content into a standardized format that the system can recognize;

[0015] Extract service location information from service resources, including customer address and service area, and set resource urgency parameters according to service resource description and customer requirements, including emergency, ordinary, and appointment levels;

[0016] Integrate the identified service content, service location, and urgency parameter information to form complete service task information, and store the integrated service task information in the database of the service management system for subsequent scheduling and analysis;

[0017] The service management system prepares to dispatch engineers based on the integrated service task information, and passes the service task information to the AI ​​big model for further matching and allocation.

[0018] Preferably, in step 2, the process of acquiring the multi-dimensional task data set includes:

[0019] Establish an API interface with the existing engineer scheduling system to regularly synchronize the latest scheduling information at a frequency of 20 minutes, including the engineers' available time periods and rest days. After receiving the scheduling data, verify it to ensure the integrity and accuracy of the data, and integrate the data into the scheduling information database of the dispatching system;

[0020] Establish a task progress tracking system to record each engineer's current task assignment, the progress of completed tasks, and the estimated remaining working time. Calculate the engineer's work saturation based on the engineer's task progress and estimated working time, and ensure that the saturation information can be updated in real time so that the engineer's current load can be considered during scheduling;

[0021] Establish a connection with the engineer's personal profile and skills database to regularly synchronize the latest professional skills, certification status and training experience. Automatically confirm the service area based on the engineer's work area and store it in the system. Set up a data update mechanism to ensure that the skills and service area information can be updated as the engineer's career develops;

[0022] Establish an interface with the customer feedback system, regularly collect historical information on engineers' favorable comments, including customer evaluation and satisfaction surveys, conduct data analysis on the collected customer feedback, calculate engineers' favorable comments and satisfaction, and establish a feedback mechanism to allow engineers to view their own customer evaluations to encourage improvement of service quality;

[0023] Integrate the collected engineer data (scheduling, work saturation, skills, service area, historical praise rate) with the service task information (task content, location, urgency), build a multi-dimensional task data set based on the integrated data, and store the constructed multi-dimensional task data set in the database of the service management system to ensure the security and accessibility of the data for subsequent analysis and scheduling.

[0024] Preferably, the calculation expression of the working saturation is:

[0025] in, is the work saturation, is the actual working time, is the work intensity coefficient (dimensionless), which indicates the impact of work difficulty and complexity on working time. Its value range is between 1 and 5. The larger the value, the higher the work intensity. is the amount of completed tasks, is the task quality coefficient (dimensionless), which indicates the impact of the quality of the completed task on the work saturation. Its value range is between 0 and 1. The closer the value is to 1, the higher the task quality. The maximum working capacity refers to the maximum working time that an engineer can undertake within a certain period of time. It is the standard working time, which refers to the standard working time of engineers;

[0026] The calculation expression of the engineer's satisfaction is: ;

[0027] in, It is the engineer's satisfaction rating. is the rating of the i-th customer (1-5 points, 5 points is the highest), is the expected value of the i-th customer, which is scored according to the degree of match between the customer's feedback expectations and the actual service (0-1, 1 is a perfect match), is the number of complaints from the ith customer (0 means no complaint), is the severity of the complaint of the i-th customer, which is scored based on the scope of the complaint and the difficulty of solving it (0-1, 1 is very serious), and the engineer's satisfaction The value range is from 0% to 100%. The higher the value, the higher the customer satisfaction the engineer obtains.

[0028] Preferably, in step 3, the feature analysis process includes:

[0029] Extract engineer data and service task information from the multi-dimensional task data set, clean the extracted data, and use a unique identifier (engineer ID) to identify and delete duplicate records to ensure data accuracy and consistency. The fields extracted from the engineer data include engineer ID, name, skills, service area, and historical task records.

[0030] Use machine learning algorithms to extract service task features from preprocessed data, including engineer skills, service areas, and historical favorable reviews. Standardize the features of different units. For engineer skills, encode the skills as numerical variables (skill proficiency scores). For service areas, use geographic coordinates to digitize the area codes, and directly extract the historical favorable reviews as numerical features.

[0031] Define the architecture of the AI ​​big model based on the convolutional neural network, input the feature data into the selected AI big model, use the back propagation training algorithm to train the model, use cross-validation, accuracy, recall, and F1 score indicators to evaluate the performance of the model, and tune the model according to the evaluation results;

[0032] The features of different units are integrated, and the nonlinear transformation ability of neural networks is used to analyze the relationship between the features. The importance of different features is weighted using the attention mechanism method.

[0033] Deploy the trained model to the production environment, process new data and output prediction results to achieve automated resource scheduling.

[0034] Preferably, in step 4, the calculation process of the matching score includes:

[0035] Extract the characteristic data of engineers and service tasks from the multi-dimensional task data set, including engineer ID, skills, service area, historical praise rate, as well as task type, location, urgency, and required skills. Pair the characteristic data of each engineer with the characteristic data of the service task and prepare to input it into the AI ​​big model.

[0036] The paired feature data is input into the AI ​​big model, which processes the input feature data, performs reasoning analysis on it, compares the engineer’s features with the service task’s features, and analyzes the degree of match between the two.

[0037] For each engineer, the matching degree between his / her characteristics and the service task is analyzed, and based on the matching degree analysis results, the matching score of each engineer for the current service task is calculated.

[0038] Preferably, the calculation expression of the matching score is: ;

[0039] in, is the matching score, is the engineer's skill proficiency score, is the engineer's historical favorable rating, is the urgency of the task, ranging from 0 to 10. is the distance between the engineer and the mission location, is the engineer's historical task completion rate, ranging from 0 to 1. is the benchmark task completion rate, ranging from 0 to 1, indicating the average task completion rate.

[0040] Preferably, in step 5, the process of analyzing the matching degree between the engineer and the service task includes:

[0041] Obtain the matching score of each engineer for the current service task from the AI ​​big model, organize the matching scores output by the AI ​​big model into a list, and associate each score with the corresponding engineer ID;

[0042] Sort all available engineers in descending order based on the matching scores, with engineers with higher scores at the front. If there are multiple engineers with similar matching scores, further analyze the engineers' historical service task response times;

[0043] According to the matching score and historical service task response time, the matching evaluation coefficient is calculated to analyze the matching degree between the engineer and the service task. The matching evaluation coefficient is then combined to rank the engineers with different matching evaluation coefficients.

[0044] Based on the value of the matching evaluation coefficient, the engineer with the highest matching evaluation coefficient is given priority to perform the current service task, and a response time threshold is preset. If the engineer with the highest matching evaluation coefficient does not respond, the service tasks are assigned in sequence until the engineer responds, thus completing the assignment of the service task.

[0045] Preferably, the calculation expression of the matching evaluation coefficient is: ;

[0046] in, is the matching evaluation coefficient, is the matching score, is the average historical service task response time of engineers, is the maximum allowed response time, is the average match score of all engineers.

[0047] Preferably, in step 6, the execution process of the service task includes:

[0048] Use SMS and the work management platform to notify the selected engineer of the scheduling results, including service task details, task priority, expected completion time, and relevant contact information;

[0049] Update the status of the service management system, mark the task as assigned, and record the assignment information of engineer ID, task ID, and assignment time;

[0050] After receiving the notification, the engineer confirms the task information and prepares to start the task. If necessary, the engineer can further communicate with the contact person related to the service task to obtain more task background or details. Then the engineer starts to perform the service task according to the task requirements;

[0051] During the task execution process, the system continuously collects the engineer's work saturation information, including the engineer's current number of tasks, task progress, estimated completion time, and available working time. The collected work saturation information is updated in real time to the service management system, which helps the system understand the engineer's work status for subsequent resource scheduling or task allocation.

[0052] The service management system continuously monitors the execution of tasks, including task progress and engineer work saturation. If problems are found in task execution or engineer work saturation is too high, measures will be taken to make adjustments, reallocate resources, assign additional engineers to assist in completing tasks, or adjust task priorities and assignments.

[0053] The present invention provides a multi-key indicator fusion service resource scheduling method based on AI big model drive. It has the following beneficial effects:

[0054] 1. This multi-key indicator fusion service resource scheduling method driven by the AI ​​big model avoids excessive fatigue of engineers and unnecessary waste of travel time by comprehensively considering factors such as work saturation and distance. It can respond to resources quickly and reduce the average time of resource processing, thereby improving overall service efficiency. It also allocates resources according to the engineer's skills and historical praise rate to ensure that resources are handled by engineers with corresponding capabilities and high service quality, thereby improving the success rate of resource processing and customer satisfaction.

[0055] 2. This multi-key indicator fusion service resource scheduling method driven by AI big model makes resource allocation more reasonable by making full use of engineers' skills and working time, avoids idleness and waste of human resources, and realizes the optimal configuration of service resources. Due to the use of AI big model, it can handle multi-dimensional and complex key indicator relationships, can adapt to the different types of resources and diverse engineering teams of enterprises, and shows good adaptability and robustness in complex business scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a flow chart of the multi-key indicator fusion service resource scheduling method driven by the AI ​​big model of the present invention;

[0057] Figure 2 A flowchart for obtaining a multi-dimensional task data set of the present invention;

[0058] Figure 3 A flow chart for calculating the matching score of the present invention;

[0059] Figure 4 The flowchart of the present invention is used to analyze the matching degree between the engineer and the service task. DETAILED DESCRIPTION

[0060] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention are provided for the purpose of illustration and description, and are not intended to be exhaustive or to limit the present invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention, and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses.

[0061] The first embodiment, as Figure 1 , Figure 2 As shown, the present invention provides a technical solution: a multi-key indicator fusion service resource scheduling method driven by an AI big model, comprising the following steps:

[0062] Step 1: When a new service resource is generated, the scheduling process is automatically triggered to identify and collect the service task information of the service resource. The service task information includes service content, service location, and resource urgency parameters. New service resources (maintenance requests, customer consultations, orders, etc.) enter the service management system through the system interface, API interface, email, and SMS. The service management system monitors the entry of new service resources in real time through the event monitoring mechanism. When a new service resource is detected, the trigger configuration mechanism is automatically activated to trigger the scheduling process. The service management system has multi-channel receiving capabilities to ensure that no new service resources are missed. The service task content in the new service resource is analyzed to identify the specific maintenance type (such as electrical appliance maintenance, pipe unblocking, etc.) or consultation problem category ( Such as product usage consultation, after-sales service, etc.), and use natural language processing technology to convert service content into a standardized format that the system can recognize, extract service location information from service resources, including customer address and service area, and set resource urgency parameters according to the description of service resources and customer requirements, which are emergency, ordinary, and appointment. The identified service content, service location, and urgency parameter information are integrated to form complete service task information, and the integrated service task information is stored in the database of the service management system for subsequent scheduling and analysis. The service management system prepares to schedule engineers based on the integrated service task information, and passes the service task information to the AI ​​big model for further matching and allocation;

[0063] Step 2: Collect the latest engineer-related data, including scheduling, work saturation, skills, service areas, and historical praise rate information, and integrate engineer data and task information to obtain a multi-dimensional task data set. Establish an API interface with the existing engineer scheduling system, and regularly synchronize the latest scheduling information at a frequency of 20 minutes, including the engineer's available time period and rest days. After receiving the scheduling data, verify it to ensure the integrity and accuracy of the data, and integrate the data into the scheduling information database of the scheduling system. Establish a task progress tracking system to record each engineer's current task allocation, the progress of completed tasks, and the estimated remaining working time. Calculate the engineer's work saturation based on the engineer's task progress and estimated working time, and ensure that the saturation information can be updated in real time so that the engineer's current load can be considered during scheduling. Establish a connection with the engineer's personal profile and skill database to regularly synchronize the latest professional skills, recognition, and other related information. Based on the engineer's certification status and training experience, the service area is automatically confirmed according to the engineer's work area and stored in the system. A data update mechanism is set up to ensure that the skills and service area information can be updated as the engineer's career develops. An interface is established with the customer feedback system to regularly collect the engineer's historical praise rate information, including customer evaluation and satisfaction surveys. Data analysis is performed on the collected customer feedback to calculate the engineer's praise satisfaction. A feedback mechanism is established to allow engineers to view their own customer evaluations to encourage improvement of service quality. The collected engineer data (scheduling, work saturation, skills, service area, historical praise rate) is integrated with the service task information (task content, location, urgency). Based on the integrated data, a multi-dimensional task data set is constructed, and the constructed multi-dimensional task data set is stored in the database of the service management system to ensure the security and accessibility of the data for subsequent analysis and scheduling.

[0064] Furthermore, the calculation expression of working saturation is:

[0065] in, is the work saturation, is the actual working time, is the work intensity coefficient (dimensionless), which indicates the impact of work difficulty and complexity on working time. Its value range is between 1 and 5. The larger the value, the higher the work intensity. is the amount of completed tasks, is the task quality coefficient (dimensionless), which indicates the impact of the quality of the completed task on the work saturation. Its value range is between 0 and 1. The closer the value is to 1, the higher the task quality. The maximum working capacity refers to the maximum working time that an engineer can undertake within a certain period of time. It is the standard working time, which refers to the standard working time of engineers. and The value range is determined according to the specific nature and requirements of the work. For high-intensity and high-difficulty work, will be higher; for tasks completed with high quality, will be close to 1, and The ratio reflects the ratio of the engineer's actual working time to his maximum working capacity. The higher the ratio, the closer the engineer's workload is to saturation. The value range of is between 0% and 100%, and a job saturation between 70% and 80% is considered ideal, more than 80% indicates oversaturation, and less than 60% may indicate undersaturation;

[0066] The calculation expression of engineer's satisfaction is: ;

[0067] in, It is the engineer's praise satisfaction. is the rating of the i-th customer (1-5 points, 5 points is the highest), is the expected value of the i-th customer, which is scored according to the degree of match between the customer's feedback expectations and the actual service (0-1, 1 is a perfect match), is the number of complaints from the ith customer (0 means no complaint), is the severity of the complaint of the i-th customer, which is scored based on the scope of the complaint and the difficulty of solving it (0-1, 1 is very serious), and the engineer's satisfaction The value range is from 0% to 100%. The higher the value, the higher the customer satisfaction the engineer gets. and The sum of the products of When the ratio of the sum is high, it means that the customer has a high evaluation of the engineer's service, and the engineer's satisfaction increases. and The sum of the products of When the ratio of the sum is low, it means that the impact of complaints is small and the satisfaction of engineers’ favorable comments increases;

[0068] Step 3: Input the relevant data of the multi-dimensional task data set into the AI ​​big model, use the machine learning algorithm to perform feature analysis, and use the neural network architecture in deep learning to integrate and analyze multiple key indicator data. In step 3, the feature analysis process includes extracting engineer data and service task information from the multi-dimensional task data set, cleaning the extracted data, and using a unique identifier (engineer ID) to identify and delete duplicate records to ensure data accuracy and consistency. The fields extracted from the engineer data include engineer ID, name, skills, service area, and historical task records. Use machine learning algorithms to extract service task features from the preprocessed data, including engineer skills, service area, and historical praise rate, and standardize the features of different units. For engineer skills, encode the skills as numerical variables (skill proficiency scores). For service areas, use geographic coordinates to digitize the area codes. The historical praise rates are directly extracted as numerical features. The architecture of the AI ​​big model is defined based on the convolutional neural network. The feature data is input into the selected AI big model. The model is trained using the back propagation training algorithm. The performance of the model is evaluated using cross-validation, accuracy, recall, and F1 score indicators. The model is tuned based on the evaluation results, and the features of different units are integrated. The nonlinear transformation capabilities of the neural network are used to analyze the relationship between the features. The importance of different features is weighted using the attention mechanism method. The trained model is deployed to the production environment, new data is processed, and the prediction results are output to achieve automated resource scheduling.

[0069] Step 4: Use the AI ​​big model to process and analyze the input data and calculate the matching score of each engineer for the service task;

[0070] Step 5: Sort all available engineers according to the engineer matching scores output by the AI ​​big model, and analyze the matching degree between the engineers and the service tasks;

[0071] Step 6: After determining the engineer who will perform the task, the scheduling result will be notified to the engineer and the relevant service management system. After receiving the task notification, the engineer will start to perform the service task. During the task execution process, the engineer's work saturation information will be continuously collected and updated so that new resources that may appear in the future can be scheduled in a timely manner.

[0072] The second embodiment is based on the first embodiment. Figure 3 , Figure 4 As shown, in step 4, the calculation process of the matching score includes:

[0073] Extract the characteristic data of engineers and service tasks from the multi-dimensional task data set, including engineer ID, skills, service area, historical praise rate, and task type, location, urgency, and required skills, and pair the characteristic data of each engineer with the characteristic data of the service task, and prepare to input them into the AI ​​big model. Input the paired characteristic data into the AI ​​big model, and the AI ​​big model processes the input characteristic data, performs reasoning analysis on it, compares the characteristics of the engineer with the characteristics of the service task, and analyzes the degree of match between the two. For each engineer, analyze the degree of match between his characteristics and the service task, and calculate the matching score of each engineer for the current service task based on the matching degree analysis results.

[0074] Furthermore, the calculation expression of the matching score is: ;

[0075] in, is the matching score, is the engineer's skill proficiency score, is the engineer's historical favorable rating, is the urgency of the task, ranging from 0 to 10. is the distance between the engineer and the mission location, is the engineer's historical task completion rate, ranging from 0 to 1. is the benchmark task completion rate, ranging from 0 to 1, indicating the average task completion rate. and When it is higher, Larger, thus improving the matching score, when When it is higher, Larger, thus improving the matching score, when When smaller, Larger, thus improving the matching score, when Greater than hour, Larger, thus improving the matching score;

[0076] In step 5, the process of analyzing the matching degree between the engineer and the service task includes:

[0077] Obtain the matching score of each engineer for the current service task from the AI ​​big model, organize the matching scores output by the AI ​​big model into a list, and associate each score with the corresponding engineer ID. Sort all available engineers in descending order according to the matching score, and arrange engineers with higher scores in front. If there are multiple engineers with similar matching scores, further analyze the historical service task response time of the engineer, calculate the matching evaluation coefficient according to the matching score and the historical service task response time, analyze the matching degree between the engineer and the service task, and sort the engineers with different matching evaluation coefficients according to the matching evaluation coefficient. Based on the value of the matching evaluation coefficient, give priority to the engineer with the highest matching evaluation coefficient to perform the current service task, and preset a response time threshold. If the engineer with the highest matching evaluation coefficient does not respond, the service tasks are assigned in order until the engineer responds, that is, the assignment of the service task is completed;

[0078] Furthermore, the calculation expression of the matching evaluation coefficient is: ;

[0079] in, is the matching evaluation coefficient, is the matching score, is the average historical service task response time of engineers, is the maximum allowed response time, is the average matching score of all engineers, Indicates that the matching score is adjusted using an exponential function. When the engineer's matching score is higher than the average score, the value of the exponential term is less than 1, thereby increasing the matching evaluation coefficient. Higher and When it is lower, The larger the value, the higher the matching degree between the engineer and the service task. Lower than When , the value of the exponential term is greater than 1, thus reducing the matching evaluation coefficient;

[0080] In step 6, the execution process of the service task includes:

[0081] The selected engineer is notified of the scheduling result using SMS and the work management platform. The notification content includes service task details, task priority, expected completion time, and relevant contact information. The status of the service management system is updated, the task is marked as assigned, and the allocation information of the engineer ID, task ID, and allocation time is recorded. After receiving the notification, the engineer confirms the task information and prepares to start the task. If necessary, the engineer can further communicate with the contact person related to the service task to obtain more task background or detailed information. The engineer then starts to perform the service task according to the task requirements. During the task execution process, the system continuously collects the engineer's work saturation information, where the work saturation information includes the engineer's current number of tasks, task progress, expected completion time, and available working time. The collected work saturation information is updated in real time to the service management system, which helps the system understand the engineer's work status for subsequent resource scheduling or task allocation. The service management system continuously monitors the task execution, including task progress and the engineer's work saturation. If problems are found in task execution or the engineer's work saturation is too high, measures are taken to adjust, reschedule resources, assign additional engineers to assist in completing the task, or adjust task priority and allocation.

[0082] Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field and related fields without creative work should fall within the scope of protection of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention are implemented according to the conventional means in the field unless otherwise specified and limited.

Claims

1. A multi-key indicator fusion service resource scheduling method driven by an AI big model, characterized in that: The following steps are involved: Step 1: When a new service resource is generated, the scheduling process is automatically triggered to identify and collect the service task information of the service resource. In step 1, the process of obtaining the service task information includes: New service resources enter the service management system through the system interface, API interface, email, and SMS. The service management system monitors the entry of new service resources in real time through the event monitoring mechanism. When new service resources are detected, the trigger configuration mechanism is automatically activated to trigger the scheduling process; Analyze the service task content in the new service resources, identify the specific maintenance type or consulting problem category, and use natural language processing technology to convert the service content into a standardized format that the system can recognize; Extract service location information from service resources, including customer address and service area, and set resource urgency parameters according to service resource description and customer requirements, including emergency, ordinary, and appointment levels; Integrate the identified service content, service location, and urgency parameter information to form complete service task information, and store the integrated service task information in the database of the service management system; The service management system prepares to dispatch engineers based on the integrated service task information and passes the service task information to the AI ​​big model for further matching and allocation; Step 2: Collect the latest engineer-related data, and integrate the engineer data and task information to obtain a multi-dimensional task data set. In step 2, the process of obtaining the multi-dimensional task data set includes: Establish an API interface with the existing engineer scheduling system to synchronize the latest scheduling information at a frequency of 20 minutes, including the engineers' available time periods and rest days. After receiving the scheduling data, verify it and integrate the data into the scheduling information database of the dispatching system. Establish a task progress tracking system to record each engineer's current task assignment, the progress of completed tasks, and the estimated remaining working time, and calculate the engineer's work saturation based on the engineer's task progress and estimated working time; Establish a connection with the engineer's personal profile and skills database to regularly synchronize the latest professional skills, certification status and training experience. Automatically confirm the service area based on the engineer's work area and store it in the system; Establish an interface with the customer feedback system to regularly collect historical favorable ratings of engineers, including customer evaluations and satisfaction surveys, conduct data analysis on collected customer feedback, and calculate engineer favorable ratings and satisfaction; Integrate the collected engineer data with the service task information, construct a multi-dimensional task data set based on the integrated data, and store the constructed multi-dimensional task data set in the database of the service management system; The calculation expression of the working saturation is: in, is the work saturation, is the actual working time, is the work intensity coefficient, is the amount of completed tasks, is the task quality coefficient, is the maximum working capacity, It is the standard working time, which refers to the standard working time of engineers; The calculation expression of the engineer's satisfaction is: ； in, It is the engineer's satisfaction rating. is the rating of the i-th customer, is the expected value of the ith customer, is the number of complaints from the i-th customer, is the severity of the complaint of the ith customer; Step 3: Input the relevant data of the multi-dimensional task data set into the AI ​​big model and use the machine learning algorithm to perform feature analysis; Step 4: Use the AI ​​big model to process and analyze the input data and calculate the matching score of each engineer for the service task; Step 5: Sort all available engineers according to the engineer matching scores output by the AI ​​big model, and analyze the matching degree between the engineers and the service tasks; Step 6: After the engineer who will perform the task is determined, the scheduling result will be notified to the engineer and the relevant service management system. During the task execution process, the engineer's work saturation information will be continuously collected and updated.

2. The method for scheduling multi-key indicator fusion service resources based on AI big model drive according to claim 1 is characterized by: In step 3, the feature analysis process includes: Extract engineer data and service task information from the multi-dimensional task data set, clean the extracted data, and use unique identifiers to identify and delete duplicate records. The fields extracted from the engineer data include engineer ID, name, skills, service area, and historical task records. Use machine learning algorithms to extract service task features from preprocessed data, including engineer skills, service areas, and historical praise rates. Standardize the features of different units. For engineer skills, encode the skills as numerical variables. For service areas, use geographic coordinates to digitize the area codes. The historical praise rates are directly extracted as numerical features. Define the architecture of the AI ​​big model based on the convolutional neural network, input the feature data into the selected AI big model, use the back propagation training algorithm to train the model, use cross-validation, accuracy, recall, and F1 score indicators to evaluate the performance of the model, and tune the model according to the evaluation results; The features of different units are integrated, and the nonlinear transformation ability of neural networks is used to analyze the relationship between the features. The importance of different features is weighted using the attention mechanism method. Deploy the trained model to the production environment, process new data and output prediction results to achieve automated resource scheduling.

3. The method for scheduling multi-key indicator fusion service resources based on AI big model drive according to claim 2 is characterized by: In step 4, the calculation process of the matching score includes: Extract the characteristic data of engineers and service tasks from the multi-dimensional task data set, including engineer ID, skills, service area, historical praise rate, as well as task type, location, urgency, and required skills. Pair the characteristic data of each engineer with the characteristic data of the service task and prepare to input it into the AI ​​big model. The paired feature data is input into the AI ​​big model, which processes the input feature data, performs reasoning analysis on it, compares the engineer’s features with the service task’s features, and analyzes the degree of match between the two. For each engineer, the matching degree between his / her characteristics and the service task is analyzed, and based on the matching degree analysis results, the matching score of each engineer for the current service task is calculated.

4. The method for scheduling multi-key indicator fusion service resources based on AI big model drive according to claim 3 is characterized by: The calculation expression of the matching score is: ; in, is the matching score, is the engineer's skill proficiency score, is the engineer's historical favorable rating, is the urgency of the task, is the distance between the engineer and the mission location, is the engineer's historical task completion rate, is the benchmark task completion rate, which represents the average task completion rate.

5. The method for scheduling multi-key indicator fusion service resources based on AI big model drive according to claim 4 is characterized by: In step 5, the process of analyzing the matching degree between the engineer and the service task includes: Obtain the matching score of each engineer for the current service task from the AI ​​big model, organize the matching scores output by the AI ​​big model into a list, and associate each score with the corresponding engineer ID; Sort all available engineers in descending order based on the matching scores, with engineers with higher scores at the front. If there are multiple engineers with similar matching scores, further analyze the engineers' historical service task response times; According to the matching score and historical service task response time, the matching evaluation coefficient is calculated to analyze the matching degree between the engineer and the service task. The matching evaluation coefficient is then combined to rank the engineers with different matching evaluation coefficients. Based on the value of the matching evaluation coefficient, the engineer with the highest matching evaluation coefficient is given priority to perform the current service task, and a response time threshold is preset. If the engineer with the highest matching evaluation coefficient does not respond, the service tasks are assigned in sequence until the engineer responds, thus completing the assignment of the service task.

6. The method for scheduling multi-key indicator fusion service resources based on AI big model drive according to claim 5 is characterized by: The calculation expression of the matching evaluation coefficient is: ; in, is the matching evaluation coefficient, is the matching score, is the average historical service task response time of engineers, is the maximum allowed response time, is the average match score of all engineers.

7. The method for scheduling multi-key indicator fusion service resources based on AI big model drive according to claim 6 is characterized by: In step 6, the execution process of the service task includes: Use SMS and the work management platform to notify the selected engineer of the scheduling results, including service task details, task priority, expected completion time, and relevant contact information; Update the status of the service management system, mark the task as assigned, and record the assignment information of engineer ID, task ID, and assignment time; After receiving the notification, the engineer confirms the task information and prepares to start the task. If necessary, the engineer can further communicate with the contact person related to the service task. The engineer can then start to perform the service task according to the task requirements; During the task execution process, the system continuously collects the work saturation information of engineers and updates the collected work saturation information to the service management system in real time; The service management system continuously monitors the execution of tasks, including task progress and engineer work saturation. If problems are found in task execution or engineer work saturation is too high, measures will be taken to make adjustments, reallocate resources, assign additional engineers to assist in completing tasks, or adjust task priorities and assignments.

Images