Task allocation method based on seat state and related equipment

Abstract

The embodiment of the invention belongs to the technical field of artificial intelligence, and relates to a task allocation method based on an agent state and related equipment, and the method comprises the steps: reading an agent operation management system if a first candidate agent is in a leaving state, acquiring on-duty seat information and task attribute data corresponding to the emergency processing task in a seat operation management system; inputting the task attribute data into a large language model for data analysis operation to obtain task type data; and performing task rearrangement operation on the emergency processing task according to the differentiated redistribution strategy and the task type data. The method can be used for carrying out related seat task allocation processing in business systems such as financial science and technology, medical health, old-age care and the like, and the task processing quality and efficiency can be improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a task allocation method and related equipment based on agent status. Background Technology

[0002] In existing task processing systems, especially in scenarios involving agents handling various business tasks, such as customer service centers and business consultation and processing centers, task allocation methods typically employ a first-come, first-served approach or preset priority rules. However, this approach has several limitations: 1. Lack of real-time monitoring and dynamic adjustment: The inability to monitor task execution time in real time and the failure to promptly identify tasks that may time out result in some urgent tasks not being handled in a timely manner, affecting customer satisfaction and business efficiency.

[0003] 2. Inflexible seat allocation: When preset agents are unable to handle urgent tasks due to reasons such as being away from their posts, the system has difficulty in quickly and intelligently reassigning tasks, which can easily lead to task backlog and reduce overall business processing capacity.

[0004] 3. Inaccurate task type matching: When reassigning tasks, the specific type of task and the skill matching of the agents are not fully considered, which may lead to unreasonable task allocation and further affect processing efficiency and quality.

[0005] This shows that existing task processing systems suffer from low timeliness and accuracy in task processing. Summary of the Invention

[0006] The purpose of this application is to propose a task allocation method and related equipment based on agent status, so as to solve the problem that the timeliness and accuracy of task processing in existing task processing systems are low.

[0007] To address the aforementioned technical problems, this application provides a task allocation method based on agent status, employing the following technical solution: Real-time monitoring of the execution time information of pending tasks; When the execution time information reaches the preset warning time threshold, the task to be processed is identified as an emergency processing task, and the first candidate seat corresponding to the emergency processing task is obtained. If the first candidate agent is on duty, a task processing reminder signal corresponding to the emergency handling task is sent to the first candidate agent. If the first candidate agent is off duty, the agent operation management system is read, and the on-duty agent information and the task attribute data corresponding to the emergency handling task are obtained from the agent operation management system. The task attribute data is input into a large language model for data analysis to obtain task type data. The emergency handling tasks are reordered based on the differentiated redistribution strategy and the task type data.

[0008] Furthermore, the step of reordering the emergency processing tasks according to the differentiated reallocation strategy and the task type data specifically includes the following steps: If the task type data is a regular task type, then a second candidate agent that matches the emergency handling task is obtained from the on-duty agent information according to a greedy algorithm. The emergency handling task is assigned to the second candidate agent. If the task type data is a complex task type, then obtain the historical task data corresponding to the emergency handling task; The historical task data and on-duty agent information are input into the LightGBM classification prediction model for prediction to obtain the third candidate agent. The emergency handling task is assigned to the third candidate agent.

[0009] Furthermore, the on-duty agent information includes the total number of agents, the historical processing efficiency corresponding to each agent, and the current task load. The greedy algorithm is specifically expressed as follows:

[0010] in, This indicates the total number of seats. Indicates the first The historical processing efficiency of each agent seat Indicates the first Current task load of each agent, Represents decision variables.

[0011] Furthermore, after the step of sending a task processing reminder signal corresponding to the emergency handling task to the first candidate agent if the first candidate agent is on duty, the following step is also included: Predict the predicted execution time of the emergency response task; Obtain the historical experience records corresponding to the first candidate agent; Obtain task characteristic information corresponding to the emergency handling task; The potential execution time range is determined by analyzing the correlation between the historical experience records and the task characteristic information using the support vector machine algorithm. If the predicted execution time does not exceed the potential execution time range, a task processing reminder signal corresponding to the emergency processing task is sent to the first candidate agent. If the predicted execution time exceeds the potential execution time range, then the similarity score between the historical experience record of each on-duty agent and the task characteristic information is calculated to obtain the first matching result for each on-duty agent. The on-duty agent with the highest matching degree in the first matching result is determined as the fourth candidate agent; The emergency handling task is assigned to the fourth candidate agent.

[0012] Furthermore, after the step of analyzing the correlation between the historical experience records and the task characteristic information using the support vector machine algorithm to determine the potential execution time range, the following step is also included: If the predicted execution time exceeds the potential execution time range, then calculate the skill matching degree between each on-duty agent and the task characteristic information, and the real-time availability of the agent status. Based on the skill matching degree and the real-time availability of the seat status, it was determined as the fifth candidate seat; The emergency handling task is assigned to the fifth candidate agent.

[0013] Furthermore, after the step of analyzing the correlation between the historical experience records and the task characteristic information using the support vector machine algorithm to determine the potential execution time range, the following step is also included: If the predicted execution time exceeds the potential execution time range, the resource utilization rate of each on-duty agent is calculated separately. The resource utilization rate of each on-duty agent is filtered according to the resource utilization rate threshold to obtain a subset of candidate agents; Calculate the operating cost of each on-duty agent in the candidate agent subset; The on-duty agent with the lowest operating cost was selected as the sixth candidate agent. The emergency handling task is assigned to the sixth candidate seat.

[0014] To address the aforementioned technical problems, this application also provides a task allocation device based on agent status, employing the following technical solution: The execution time detection module is used to monitor the execution time information of tasks to be processed in real time. The first candidate seat acquisition module is used to determine the task to be processed as an emergency processing task when the execution time information reaches a preset warning time threshold, and to acquire the first candidate seat corresponding to the emergency processing task. The first reminder signal sending module is used to send a task processing reminder signal corresponding to the emergency handling task to the first candidate agent if the first candidate agent is on duty. The task attribute acquisition module is used to read the seat operation management system if the first candidate seat is off duty, and to obtain on-duty seat information and task attribute data corresponding to the emergency handling task from the seat operation management system. The task type acquisition module is used to input the task attribute data into the large language model for data analysis to obtain task type data. The task reordering module is used to perform task reordering operations on the emergency processing tasks according to the differentiated redistribution strategy and the task type data.

[0015] Furthermore, the task reordering module includes: The second candidate agent acquisition submodule is used to acquire a second candidate agent that matches the emergency handling task from the on-duty agent information according to a greedy algorithm if the task type data is a regular task type. The second candidate seat allocation submodule is used to allocate the emergency handling task to the second candidate agent; The historical task data acquisition submodule is used to acquire historical task data corresponding to the emergency handling task if the task type data is a complex task type. The third candidate seat acquisition submodule is used to input the historical task data and on-duty seat information into the LightGBM classification prediction model to perform prediction operations and obtain the third candidate seat. The third candidate seat allocation submodule is used to allocate the emergency handling task to the third candidate seat.

[0016] To address the aforementioned technical problems, this application also provides a computer device that employs the following technical solution: It includes a memory and a processor, wherein the memory stores computer-readable instructions, and the processor executes the computer-readable instructions to implement the steps of the task allocation method based on agent status as described above.

[0017] To address the aforementioned technical problems, this application also provides a computer-readable storage medium, employing the technical solution described below: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the task allocation method based on agent status as described above.

[0018] This application provides a task allocation method based on agent status, comprising: real-time monitoring of the execution time information of pending tasks; when the execution time information reaches a preset warning time threshold, identifying the pending task as an emergency task and obtaining a first candidate agent corresponding to the emergency task; if the first candidate agent is on duty, sending a task processing reminder signal corresponding to the emergency task to the first candidate agent; if the first candidate agent is off duty, reading the agent operation management system and obtaining on-duty agent information and task attribute data corresponding to the emergency task from the agent operation management system; inputting the task attribute data into a large language model for data analysis to obtain task type data; and performing task rearrangement operation on the emergency task according to a differentiated reallocation strategy and the task type data. Compared with existing technologies, this application, by monitoring the execution time information of pending tasks in real time, can promptly identify urgent tasks that may time out and take corresponding measures to improve the timeliness of task processing and effectively avoid adverse effects on customers due to task delays. When the first candidate agent leaves their post, the system can automatically read the agent operation management system to obtain on-duty agent information and task attribute data, and use a large model for data analysis. Based on a differentiated redistribution strategy, tasks are intelligently redistributed, improving the flexibility and rationality of agent resource allocation, making full use of existing agent resources, and reducing task backlog. By using the QWQ-32B large model to analyze task attribute data, accurate task type data is obtained. Combined with the differentiated redistribution strategy, urgent processing tasks can be accurately assigned to on-duty agents with the corresponding skills and capabilities, improving the quality and efficiency of task processing and enhancing customer satisfaction. Attached Figure Description

[0019] To more clearly illustrate the solutions in this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is an exemplary system architecture diagram to which this application can be applied; Figure 2 This is a flowchart illustrating the implementation of the task allocation method based on agent status provided in this application embodiment; Figure 3 This is a schematic diagram of the structure of the task allocation device based on the seat status provided in the embodiments of this application; Figure 4 This is a schematic diagram of the structure of one embodiment of the computer device according to this application. Detailed Implementation

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application, are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects, not to describe a particular order.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0024] like Figure 1 As shown, system architecture 100 may include terminal device 101, network 102, and server 103. Terminal device 101 may be a laptop 1011, tablet 1012, or mobile phone 1013. Network 102 is used as a medium to provide a communication link between terminal device 101 and server 103. Network 102 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0025] Users can use terminal device 101 to interact with server 103 via network 102 to receive or send messages, etc. Various communication client applications can be installed on terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social media platform software, etc.

[0026] Terminal device 101 can be various electronic devices with a display screen and support web browsing. In addition to laptops 1011, tablets 1012, or mobile phones 1013, terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 player (Moving Picture Experts Group Audio Layer IV), a laptop computer, and a desktop computer, etc.

[0027] Server 103 can be a server that provides various services, such as a backend server that provides support for the pages displayed on terminal device 101.

[0028] It should be noted that the task allocation method based on agent status provided in this application embodiment is generally executed by a server / terminal device, and correspondingly, the task allocation device based on agent status is generally set in the server / terminal device.

[0029] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0030] Continue to refer to Figure 2 The diagram shows a flowchart of an embodiment of the task allocation method based on agent status according to this application. The task allocation method based on agent status includes steps S201, S202, S203, S204, S205, S206, and S207.

[0031] In step S201, the execution time information of the tasks to be processed is monitored in real time.

[0032] In this embodiment, the application can use a timer or timestamp to achieve real-time monitoring of task execution time. For example, when the task starts execution, the system time is recorded as the starting timestamp. During subsequent monitoring, the current system time is periodically obtained, and the difference between the current time and the starting timestamp is calculated to obtain the task execution time information.

[0033] In step S202, when the execution time information reaches the preset warning time threshold, the task to be processed is identified as an emergency processing task, and the first candidate seat corresponding to the emergency processing task is obtained.

[0034] In this embodiment, a preset warning time threshold is stored in the system's configuration file. The emergency task identification module reads this threshold from the configuration file and compares it with the received execution time information. When the execution time information is greater than or equal to the preset warning time threshold, the emergency task identification logic is triggered, and the corresponding task is marked as an emergency handling task.

[0035] In step S203, if the first candidate agent is on duty, a task processing reminder signal corresponding to the emergency handling task is sent to the first candidate agent.

[0036] In this embodiment, the agent status can be determined by the agent's login status information. For example, an agent is marked as "on duty" when logging into the system and as "off duty" when logging out. The agent status determination module periodically queries the agent's status information to determine the status of the first candidate agent.

[0037] In step S204, if the first candidate agent is off duty, the agent operation management system is read, and the on-duty agent information and the task attribute data corresponding to the emergency handling task are obtained from the agent operation management system.

[0038] In this embodiment, the application can use standard interface protocols (such as RESTful API) to interact with the agent operation management system. By calling the interfaces provided by the agent operation management system, it obtains information on on-duty agents and task attribute data. During implementation, data format conversion and validation are required to ensure the accuracy and integrity of the data.

[0039] In step S205, the task attribute data is input into the large language model for data analysis to obtain task type data.

[0040] In this embodiment, a large language model is integrated into the system. Task attribute data is passed as input parameters to the large model by calling its interface. The large model analyzes and processes the input data and returns task type data. The system parses and stores the returned data for subsequent task reordering operations. The large language model can be a QWQ-32B large model.

[0041] In step S206, the emergency processing tasks are rearranged according to the differentiated redistribution strategy and task type data.

[0042] In this embodiment, a task reordering algorithm is designed based on a differentiated reallocation strategy. The algorithm comprehensively considers factors such as the skill level, work experience, and current task load of on-duty agents, combined with task type data, to calculate the suitability score for each on-duty agent to handle urgent tasks. The on-duty agent with the highest suitability score is selected as the new task handling agent, and the urgent task is reassigned to that agent.

[0043] In this embodiment, a task allocation method based on agent status is provided, comprising: real-time monitoring of the execution time information of tasks to be processed; when the execution time information reaches a preset warning time threshold, identifying the task to be processed as an emergency task and obtaining a first candidate agent corresponding to the emergency task; if the first candidate agent is on duty, sending a task processing reminder signal corresponding to the emergency task to the first candidate agent; if the first candidate agent is off duty, reading the agent operation management system and obtaining on-duty agent information and task attribute data corresponding to the emergency task from the agent operation management system; inputting the task attribute data into a large language model for data analysis to obtain task type data; and performing task rearrangement operation on the emergency task according to a differentiated reallocation strategy and task type data. Compared with existing technologies, this application, by monitoring the execution time information of pending tasks in real time, can promptly identify urgent tasks that may time out and take corresponding measures to improve the timeliness of task processing and effectively avoid adverse effects on customers due to task delays. When the first candidate agent leaves their post, the system can automatically read the agent operation management system to obtain on-duty agent information and task attribute data, and use a large model for data analysis. Based on a differentiated redistribution strategy, tasks are intelligently redistributed, improving the flexibility and rationality of agent resource allocation, making full use of existing agent resources, and reducing task backlog. By using the QWQ-32B large model to analyze task attribute data, accurate task type data is obtained. Combined with the differentiated redistribution strategy, urgent processing tasks can be accurately assigned to on-duty agents with the corresponding skills and capabilities, improving the quality and efficiency of task processing and enhancing customer satisfaction.

[0044] In some optional implementations of the embodiments of this application, the step of reordering emergency processing tasks based on the differentiated reallocation strategy and task type data specifically includes the following steps: If the task type data is a regular task type, then a second candidate agent that matches the emergency handling task is obtained from the on-duty agent information according to the greedy algorithm. Assign emergency handling tasks to the second candidate operator; If the task type data is a complex task type, then retrieve the historical task data corresponding to the emergency handling task; Historical task data and on-duty agent information are input into the LightGBM classification prediction model for prediction to obtain the third candidate agent. Assign emergency handling tasks to the third candidate seat.

[0045] In some optional implementations of the embodiments of this application, the above greedy algorithm is specifically expressed as follows:

[0046] in, This indicates the total number of seats. Indicates the first The historical processing efficiency of each agent seat Indicates the first Current task load of each agent, Represents decision variables.

[0047] In this embodiment, the LightGBM classification prediction model is mainly used to achieve optimal allocation decisions. The classification prediction model is mainly implemented as follows: Feature engineering: mainly collects historical data on agent processing tasks and the results of processing (customer satisfaction rating).

[0048] Feature data mainly includes three categories: Agent: Agent skills, project, current number of pending tasks, average task processing efficiency, etc.

[0049] Task: urgency level, products involved, cumulative call time, cumulative number of calls, associated customer information, task processing results, etc.

[0050] Model training: Based on the above feature data, construct a dataset and train the model.

[0051] Real-time prediction: The system performs feature processing on the tasks to be assigned and the information on the available agents, and then uses the model to predict the task processing results.

[0052] Model update: Add data for new tasks to be processed by agents to the feature data and train and update the model.

[0053] In this embodiment of the application, the customer service center receives an urgent customer inquiry task. By analyzing the characteristics of the inquiry (such as the business modules involved in the problem, the complexity of the problem, etc.) and matching them with preset task type rules, the task type data of the task is determined.

[0054] In this embodiment of the application, the process by which the task type determination module analyzes and determines the task type data can be as follows: (1) If the task is classified as a routine task: For example, if the customer is inquiring about the routine process of returning or exchanging goods. The routine task allocation module queries the information of the on-duty agents in the agent information management module according to the greedy algorithm. Assume that agent A currently has 2 tasks, the average time to process routine return and exchange issues is 10 minutes, and the historical success rate is 95%; agent B currently has 3 tasks, the average processing time is 12 minutes, and the historical success rate is 90%. According to the scoring function, agent A has a higher score, so agent A is selected as the second candidate agent, and a task allocation notification is sent to agent A to assign the urgent consultation task to agent A for processing.

[0055] (2) If the task is classified as complex: For example, if the customer inquires about a complex issue related to a system malfunction causing an order anomaly, the complex task processing module retrieves historical task data similar to this complex issue from the historical task data storage module, including the handling of similar malfunctions in the past and the performance of the agents handling the issue. Simultaneously, it retrieves information on agents currently on duty from the agent information management module. This historical task data and agent information are then input into a pre-trained LightGBM classification prediction model. The model predicts that agent C is best suited to handle this complex and urgent inquiry task. Therefore, agent C is selected as the third candidate agent, and a task assignment notification is sent to agent C to assign the task to agent C for processing.

[0056] Compared with existing technologies, this application can effectively realize the intelligent allocation of emergency handling task seats based on task type, improve the rationality and efficiency of task allocation, and enhance business processing quality and customer satisfaction.

[0057] In some optional implementations of the embodiments of this application, after the step of sending a task processing reminder signal corresponding to the emergency handling task to the first candidate agent if the first candidate agent is on duty, the following step is further included: Predict the estimated execution time of emergency response tasks; Retrieve historical experience records corresponding to the first candidate agent; Obtain task characteristic information corresponding to emergency handling tasks; Based on the support vector machine algorithm, the correlation between historical experience records and task characteristic information is analyzed to determine the potential execution time range; If the predicted execution time does not exceed the potential execution time range, a task processing reminder signal corresponding to the emergency handling task will be sent to the first candidate agent. If the predicted execution time exceeds the potential execution time range, the similarity score between the historical experience records and task characteristic information of each on-duty agent is calculated to obtain the first matching result for each on-duty agent. The on-duty agent with the highest matching degree in the first matching result is determined as the fourth candidate agent; Assign emergency handling tasks to the fourth candidate seat.

[0058] In this embodiment of the application, an example is taken of an emergency loan consultation task processing scenario in a financial customer service center: In this embodiment, the application can collect processing time data of similar emergency loan consultation tasks from the past through a task execution time prediction module, as well as factors such as the current system load and task complexity (e.g., the number and difficulty of consultation questions). A trained LSTM model is then used to predict the execution time of a new emergency loan consultation task, resulting in a predicted execution time of 30 minutes.

[0059] In this embodiment of the application, the application can obtain historical experience records corresponding to the first candidate agent: the historical experience record acquisition module queries the database for relevant information on loan consultation tasks previously handled by the first candidate agent, and finds that the agent has handled 5 similar tasks, with an average processing time of 25 minutes and a high success rate.

[0060] In this embodiment of the application, the application can obtain task characteristic information corresponding to the emergency handling task: the task characteristic information acquisition module analyzes the task description of the emergency loan consultation task and extracts task characteristic information, such as the business type being loan consultation, the issue involving loan application process and interest rate calculation, and the required skills being familiarity with loan business knowledge and communication skills, etc.

[0061] In this embodiment of the application, the application can input the historical experience records of the first candidate agent and the task characteristic information of the emergency loan consultation task into the trained SVM model through the correlation analysis module. The model analysis shows that under the current task characteristics, the potential execution time range for the agent to handle the task is 20-30 minutes.

[0062] In this embodiment, since the predicted execution time of 30 minutes does not exceed the potential execution time range of 30 minutes (which can be set to include boundary values), the task allocation module sends a task processing reminder signal to the first candidate agent, notifying the agent to process the emergency loan consultation task in a timely manner.

[0063] In this embodiment of the application, it is assumed that in another case, the predicted execution time is 35 minutes, which exceeds the potential execution time range. In this case: (1) Calculate similarity score: The agent matching module calculates the similarity score between the historical experience records of other agents and the task characteristic information of the emergency loan consultation task. For example, the similarity score between agent A's historical experience records and task characteristic information is 0.8, and the similarity score between agent B and agent B is 0.6.

[0064] (2) Determine the fourth candidate seat: The seat with the highest matching degree, A, is determined as the fourth candidate seat.

[0065] (3) Assign emergency processing tasks: The task assignment module assigns an emergency loan consultation task to agent A and notifies him to start processing.

[0066] Compared with existing technologies, this application can effectively realize emergency handling task seat allocation based on time prediction and matching, thereby improving the efficiency and quality of task processing.

[0067] In some optional implementations of the embodiments of this application, step S** specifically includes the following steps: In some optional implementations of the embodiments of this application, after determining the potential execution time range by analyzing the correlation between historical experience records and task characteristic information using the support vector machine algorithm, the following steps are also included: If the predicted execution time exceeds the potential execution time range, the skill matching degree between each on-duty agent and the task characteristic information and the real-time availability of the agent status are calculated separately. The fifth candidate agent was determined based on skill matching and real-time availability of agent status. Assign emergency handling tasks to the fifth candidate seat.

[0068] In this embodiment, the specific implementation process is illustrated using an emergency after-sales complaint handling scenario in an e-commerce customer service center as an example: (1) Predict the predicted execution time of emergency response tasks: In this embodiment, the task prediction module collects processing time data from similar emergency after-sales complaint tasks in the past, as well as factors such as the current system load and the complexity of the task (e.g., the number and difficulty of complaint issues). A trained LSTM model is used to predict the execution time of a new emergency after-sales complaint task, resulting in a predicted execution time of 40 minutes.

[0069] (2) Obtain the potential execution time range: In this embodiment, the historical data management module queries the actual execution time data of similar emergency after-sales complaint tasks in the past, and combines the task characteristic information extracted by the task characteristic analysis module (such as the business area being after-sales complaints, the problem involving product quality and refunds, etc.) to determine the potential execution time range as 30-35 minutes using statistical analysis methods.

[0070] (3) Determine whether the predicted execution time exceeds the potential execution time range: In this embodiment of the application, since the predicted execution time of 40 minutes exceeds the potential execution time range of 35 minutes, subsequent steps are executed.

[0071] (4) Calculate the skill matching degree between each on-duty agent and the task characteristic information, and the real-time availability of the agent status: ① Skill Matching Calculation: The task characteristic analysis module extracts the key skills required for this urgent after-sales complaint task, including familiarity with product quality standards, mastery of the refund process, and good communication skills. The skill matching calculation module uses a weighted scoring method to calculate the skill matching score for each agent based on the on-duty agent skill information provided by the agent information management module. For example, Agent A possesses all key skills at a high skill level and receives a skill matching score of 90; Agent B lacks the skill of mastering the refund process and receives a skill matching score of 70.

[0072] ② Monitor agent status and availability in real time: The agent status monitoring module monitors the status of agents on duty in real time. It finds that agent A is idle with an availability score of 1; agent B is handling a routine consultation task with an availability score of 0.5.

[0073] (5) The fifth candidate agent is determined based on skill matching and real-time availability of agent status: In this embodiment, the comprehensive evaluation module calculates the comprehensive evaluation value for each on-duty agent according to the evaluation rules of a skill matching degree weight of 0.7 and a real-time availability weight of 0.3. The comprehensive evaluation value of agent A is 90×0.7 + 1×0.3×100 = 93; the comprehensive evaluation value of agent B is 70×0.7 + 0.5×0.3×100 = 64. Agent A, with the highest comprehensive evaluation value, is selected as the fifth candidate agent.

[0074] (6) Assign emergency handling tasks to the fifth candidate agent: In this embodiment of the application, the task allocation module sends relevant information about the urgent after-sales complaint task to agent A, including the task description (such as a customer complaining about product quality issues and requesting a refund), task priority (urgent), etc., and notifies agent A to handle the task in a timely manner.

[0075] Compared with existing technologies, this application can effectively realize dynamic allocation of emergency handling task seats based on multi-dimensional evaluation, thereby improving the efficiency and quality of task processing.

[0076] In some optional implementations of the embodiments of this application, step S** specifically includes the following steps: In some optional implementations of the embodiments of this application, after determining the potential execution time range by analyzing the correlation between historical experience records and task characteristic information using the support vector machine algorithm, the following steps are also included: If the predicted execution time exceeds the potential execution time range, the resource utilization rate of each on-duty agent is calculated separately. Based on the resource utilization threshold, the resource utilization of each on-duty agent is filtered to obtain a subset of candidate agents; Calculate the operating cost of each on-duty agent in the candidate agent subset; The on-duty agent with the lowest operating cost was selected as the sixth candidate agent. Assign emergency handling tasks to the sixth candidate seat.

[0077] In this embodiment, when an emergency handling task enters the system, the system predicts the execution time of the task based on factors such as task type, historical data, and agent skill level. Assuming a preset potential execution time range of 30-60 minutes, if the predicted execution time is 70 minutes, exceeding this range, the process proceeds to the next step.

[0078] In this embodiment, the system obtains information on all currently active agents, including the number of tasks being processed, the task difficulty coefficient (pre-set according to task type and complexity), the time already worked, and the estimated remaining working time. Taking agent A as an example, the current number of tasks being processed is 2, with task difficulty coefficients of 0.8 and 0.6 respectively, the working time already completed is 2 hours, and the estimated total working time is 4 hours. According to the resource utilization rate calculation formula: Resource Utilization Rate = (Current Number of Tasks Processed × Task Difficulty Coefficient + Working Time Completed / Estimated Total Working Time) × 100%, the resource utilization rate of agent A is (2 × 0.8 + 2 / 4) × 100% = (1.6 + 0.5) × 100% = 210% (here it is assumed that the formula calculation result can exceed 100% to more comprehensively reflect the workload; in actual applications, the formula can be adjusted according to needs). The resource utilization rates of other active agents are calculated using the same method.

[0079] In this embodiment, the preset resource utilization threshold is 150%-250%. The resource utilization rate of each on-duty agent calculated above is compared with this threshold, and agents with resource utilization rates in the range of 150%-250% are selected to form a candidate agent subset. Assume that after screening, the candidate agent subset includes agent A, agent B, and agent C.

[0080] In this embodiment, the system obtains relevant cost information for each agent in the candidate agent subset, including daily salary, equipment depreciation costs, and estimated task processing time. Taking agent A as an example, the daily salary is 200 yuan, the equipment depreciation cost is 10 yuan / day, and the estimated processing time for the emergency task is 1 hour. Assuming an average daily usage time of 8 hours, according to the operating cost calculation formula: Operating cost = daily salary of agent + equipment depreciation cost / average daily usage time × estimated task processing time + other related costs (assuming other related costs are 0), the operating cost of agent A can be obtained as 200 + 10 / 8 × 1 = 200 + 1.25 = 201.25 yuan. The operating costs of agents B and C are calculated using the same method, assuming the operating cost of agent B is 210 yuan and the operating cost of agent C is 195 yuan.

[0081] In this embodiment of the application, the operating costs of seat A, seat B and seat C are compared in the candidate seat subset. It is found that the operating cost of seat C is the lowest. Therefore, seat C is determined as the sixth candidate seat.

[0082] In this embodiment, the system sends emergency handling task information to agent C via the task allocation module. Agent C begins processing the task upon receiving it. Simultaneously, the system monitors the task processing in real time, recording processing time, results, and other information for subsequent performance evaluation and data analysis.

[0083] Compared with existing technologies, this application can effectively improve the efficiency of handling emergency tasks.

[0084] In some optional implementations of the embodiments of this application, after performing task rearrangement operations on emergency processing tasks according to the differentiated reallocation strategy and task type data, the following steps are further included: Update the task queue and agent status data, use dynamic programming algorithm to optimize the urgency ranking of the remaining tasks, and determine the priority adjustment scheme for the next allocation; If a decline in service quality metrics is detected in the next priority adjustment scheme, historical experience records are retrieved from the updated agent status data, and the execution time range is re-predicted using the support vector machine algorithm to correct the matching process. For the corrected matching process, the change data of all task characteristics are obtained, and the K-means clustering algorithm is used to regroup the change data to determine and obtain the optimized skill matching score sequence. The resource utilization assessment is updated based on the optimized skill matching score sequence. The allocation path with the highest service quality is extracted from it, and the data of all steps are integrated through dynamic programming algorithm to obtain the complete task allocation loop.

[0085] In this embodiment, the system updates the task queue and agent status data every minute. The task queue data is obtained from the customer consultation interface and includes information such as consultation type (e.g., product consultation, after-sales issues), customer waiting time, and urgency level of the issue (divided into high, medium, and low levels). The agent status data is obtained from the agent's work terminal and includes information such as the number of tasks currently being processed by the agent, the time already worked, the remaining working time, the skills possessed (e.g., product knowledge, after-sales processing skills), and the skill level (divided into beginner, intermediate, and advanced levels).

[0086] In this embodiment, a dynamic programming algorithm is used to prioritize the remaining tasks based on their urgency. For example, if there are 10 remaining tasks, the dynamic programming algorithm considers the dependencies between tasks (such as needing to know product information before addressing certain after-sales issues) and time constraints (such as the response time required by customers). Combining the urgency of the tasks with the estimated processing time, the algorithm reorders the tasks and determines the priority adjustment scheme for the next allocation as follows: first, process the 2 high-urgency product consultation tasks; then, process the 3 medium-urgency after-sales issue tasks; and finally, process the 5 low-urgency other consultation tasks.

[0087] In this application embodiment, the service quality indicators after the implementation of the above priority adjustment scheme were examined. It was found that the average customer waiting time increased from 2 minutes to 3 minutes, and the task processing success rate decreased from 90% to 85%, indicating a decline in service quality indicators.

[0088] In this embodiment of the application, historical experience records are extracted from the updated agent status data. For example, agent A has handled 50 product consultation tasks, of which 30 took 3-5 minutes to process and 20 took 5-8 minutes to process; agent B has handled 40 after-sales problem tasks, of which 25 took 5-10 minutes to process and 15 took 10-15 minutes to process, etc.

[0089] In this embodiment, a support vector machine algorithm is used to construct an execution time prediction model. The model is trained using task type, agent skill level, and historical processing time as output labels. Then, based on the characteristics of the current task (e.g., a product consultation task, agent A's skill level) and the prediction model, the execution time range for the task is re-predicted to be 3-6 minutes. This corrects the previously inaccurate execution time prediction, making task allocation more reasonable.

[0090] In this application embodiment, for the modified matching process, change data of all task characteristics are obtained. For example, the complexity of the product consultation task may vary depending on the type of product, and the skill requirements for after-sales problem tasks may involve different aspects such as refund process and exchange process.

[0091] In this embodiment, the K-means clustering algorithm is used to regroup the task characteristics. Assuming the tasks are divided into 3 clusters, after cluster analysis, the first cluster consists of simple product consultation tasks, mainly involving basic product information queries; the second cluster consists of complex product consultation tasks, involving in-depth questions such as product performance and usage methods; and the third cluster consists of after-sales problem tasks, including refunds, exchanges, and repairs.

[0092] In this embodiment, a skill matching score is calculated for each agent based on their skill level and the cluster to which the task belongs. For example, agent A is skilled at handling simple product consultation tasks, and their skill matching score for tasks in the first cluster is 90 points; for tasks in the second cluster, it is 70 points; and for tasks in the third cluster, it is 50 points. Obtaining a sequence of skill matching scores for all agents more accurately reflects the degree of matching between agents and tasks.

[0093] In this embodiment, the resource utilization assessment is updated based on the optimized skill matching score sequence. The resource utilization of each agent is assessed by comprehensively considering factors such as the number of tasks currently being processed, remaining working time, and skill matching score. For example, although Agent A is currently processing 2 tasks, their skill matching score is high, resulting in a resource utilization assessment of 75%; Agent B is currently processing 3 tasks, but their skill matching score is average, resulting in a resource utilization assessment of 85%.

[0094] In this embodiment, the allocation path with the highest service quality is extracted from the service quality index analysis of different allocation paths. For example, there are three allocation path schemes, and after comparison, it is found that scheme two has the highest service quality, with an average customer waiting time of 2.5 minutes and a task processing success rate of 88%.

[0095] In this embodiment, a dynamic programming algorithm is used to integrate data from all steps, considering task allocation over a future period, to form a complete task allocation loop. This loop can be continuously adjusted and optimized based on real-time data, ensuring the continuous and efficient operation of the task allocation process.

[0096] Compared with existing technologies, this application can effectively shorten the average customer waiting time, improve the task processing success rate, and enhance the workload balance of agents.

[0097] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0098] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0099] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware through computer-readable instructions. These computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0100] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0101] Further reference Figure 3 As a response to the above Figure 2 The implementation of the method shown in this application provides an embodiment of a task allocation device based on agent status, which is similar to... Figure 2 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0102] like Figure 3 As shown, the task allocation device 200 based on agent status in this application embodiment includes: The execution time detection module 210 is used to monitor the execution time information of the tasks to be processed in real time. The first candidate seat acquisition module 220 is used to determine the task to be processed as an emergency processing task when the execution time information reaches the preset warning time threshold, and to acquire the first candidate seat corresponding to the emergency processing task. The first reminder signal sending module 230 is used to send a task processing reminder signal corresponding to the emergency handling task to the first candidate agent if the first candidate agent is on duty. The task attribute acquisition module 240 is used to read the seat operation management system if the first candidate seat is off duty, and to obtain the on-duty seat information and the task attribute data corresponding to the emergency handling task from the seat operation management system. The task type acquisition module 250 is used to input task attribute data into the large language model for data analysis to obtain task type data. The task reordering module 260 is used to reorder emergency tasks based on differentiated redistribution strategies and task type data.

[0103] In this embodiment of the application, a task allocation device 200 based on agent status is provided, including: an execution time detection module 210, used to monitor the execution time information of tasks to be processed in real time; a first candidate agent acquisition module 220, used to determine the task to be processed as an emergency processing task when the execution time information reaches a preset warning time threshold, and acquire the first candidate agent corresponding to the emergency processing task; a first reminder signal sending module 230, used to send a task processing reminder signal corresponding to the emergency processing task to the first candidate agent if the first candidate agent is on duty; a task attribute acquisition module 240, used to read the agent operation management system and acquire on-duty agent information and task attribute data corresponding to the emergency processing task if the first candidate agent is off duty; a task type acquisition module 250, used to input the task attribute data into a large language model for data analysis to obtain task type data; and a task reordering module 260, used to perform task reordering operation on the emergency processing task according to the differentiated reassignment strategy and the task type data. Compared with existing technologies, this application, by monitoring the execution time information of pending tasks in real time, can promptly identify urgent tasks that may time out and take corresponding measures to improve the timeliness of task processing and effectively avoid adverse effects on customers due to task delays. When the first candidate agent leaves their post, the system can automatically read the agent operation management system to obtain on-duty agent information and task attribute data, and use a large model for data analysis. Based on a differentiated redistribution strategy, tasks are intelligently redistributed, improving the flexibility and rationality of agent resource allocation, making full use of existing agent resources, and reducing task backlog. By using the QWQ-32B large model to analyze task attribute data, accurate task type data is obtained. Combined with the differentiated redistribution strategy, urgent processing tasks can be accurately assigned to on-duty agents with the corresponding skills and capabilities, improving the quality and efficiency of task processing and enhancing customer satisfaction.

[0104] In some optional implementations of the embodiments of this application, the task reordering module includes: The second candidate agent acquisition submodule is used to obtain the second candidate agent that matches the emergency handling task from the on-duty agent information according to the greedy algorithm if the task type data is a regular task type. The second candidate seat allocation submodule is used to assign emergency handling tasks to the second candidate seats. The historical task data acquisition submodule is used to acquire historical task data corresponding to the emergency handling task if the task type data is a complex task type. The third candidate seat acquisition submodule is used to input historical task data and on-duty seat information into the LightGBM classification prediction model for prediction to obtain the third candidate seat. The third candidate seat allocation submodule is used to assign emergency handling tasks to the third candidate seats.

[0105] In some optional implementations of the embodiments of this application, the task allocation device 200 based on agent status further includes: The execution time prediction module is used to predict the execution time of urgent processing tasks. The historical experience record acquisition module is used to acquire the historical experience records corresponding to the first candidate agent. The task characteristic acquisition module is used to acquire task characteristic information corresponding to emergency handling tasks; The execution time range determination module is used to analyze the correlation between historical experience records and task characteristic information based on the support vector machine algorithm to determine the potential execution time range; The second reminder signal sending module is used to send a task processing reminder signal corresponding to the emergency processing task to the first candidate agent if the predicted execution time does not exceed the potential execution time range. The first matching result module is used to calculate the similarity score between the historical experience records and task characteristic information of each on-duty agent if the predicted execution time exceeds the potential execution time range, and obtain the first matching result for each on-duty agent. The fourth candidate agent acquisition module is used to determine the on-duty agent with the highest matching degree of the first matching result as the fourth candidate agent; The fourth candidate seat allocation module is used to assign emergency handling tasks to the fourth candidate seat.

[0106] In some optional implementations of the embodiments of this application, the task allocation device 200 based on agent status further includes: The availability calculation module is used to calculate the skill matching degree between each on-duty agent and the task characteristic information and the real-time availability of the agent status if the predicted execution time exceeds the potential execution time range. The fifth candidate agent acquisition module is used to determine the fifth candidate agent based on skill matching degree and real-time availability of agent status; The fifth candidate seat allocation module is used to assign emergency handling tasks to the fifth candidate seat.

[0107] In some optional implementations of the embodiments of this application, the task allocation device 200 based on agent status further includes: The resource utilization calculation module is used to calculate the resource utilization of each on-duty agent if the predicted execution time exceeds the potential execution time range. The filtering module is used to filter the resource utilization of each on-duty agent based on a resource utilization threshold to obtain a subset of candidate agents. The operating cost calculation module is used to calculate the operating cost of each on-duty agent in the candidate agent subset. The sixth candidate seat acquisition module is used to identify the on-duty seat with the lowest operating cost as the sixth candidate seat. The sixth candidate seat assignment module is used to assign emergency handling tasks to the sixth candidate seat.

[0108] To address the aforementioned technical problems, embodiments of this application also provide a computer device. Please refer to [link / reference needed]. Figure 4 , Figure 4 This is a basic structural block diagram of a computer device according to an embodiment of this application.

[0109] The computer device 300 includes a memory 310, a processor 320, and a network interface 330 that are interconnected via a system bus. It should be noted that only the computer device 300 with components 310-330 is shown in the figure; however, it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Those skilled in the art will understand that the computer device described here is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0110] The computer device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device can interact with the user via a keyboard, mouse, remote control, touchpad, or voice control.

[0111] The memory 310 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 310 may be an internal storage unit of the computer device 300, such as the hard disk or memory of the computer device 300. In other embodiments, the memory 310 may also be an external storage device of the computer device 300, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. Of course, the memory 310 may also include both internal storage units and external storage devices of the computer device 300. In this embodiment, the memory 310 is typically used to store the operating system and various application software installed on the computer device 300, such as computer-readable instructions for task allocation methods based on agent status. Furthermore, the memory 310 can also be used to temporarily store various types of data that have been output or will be output.

[0112] In some embodiments, the processor 320 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 320 is typically used to control the overall operation of the computer device 300. In this embodiment, the processor 320 is used to execute computer-readable instructions stored in the memory 310 or to process data, for example, to execute computer-readable instructions of the task allocation method based on agent status.

[0113] The network interface 330 may include a wireless network interface or a wired network interface, which is typically used to establish communication connections between the computer device 300 and other electronic devices.

[0114] The computer equipment provided in this application can promptly identify potentially time-out urgent tasks by monitoring the execution time information of pending tasks in real time, and quickly take corresponding measures to improve the timeliness of task processing and effectively avoid adverse effects on customers due to task delays. When the first candidate agent leaves their post, the system can automatically read the agent operation management system, obtain on-duty agent information and task attribute data, and use a large model for data analysis. Based on a differentiated redistribution strategy, it can intelligently redistribute tasks, improving the flexibility and rationality of agent resource allocation, making full use of existing agent resources, and reducing task backlog. By using the QWQ-32B large model to analyze task attribute data, accurate task type data can be obtained. Combined with the differentiated redistribution strategy, urgent processing tasks can be accurately assigned to on-duty agents with the corresponding skills and abilities, improving the quality and efficiency of task processing and enhancing customer satisfaction.

[0115] This application also provides another embodiment, namely, providing a computer-readable storage medium storing computer-readable instructions that can be executed by at least one processor to cause the at least one processor to perform the steps of the task allocation method based on agent status as described above.

[0116] The computer-readable storage medium provided in this application can promptly identify potentially timed-out urgent tasks by monitoring the execution time information of pending tasks in real time, and quickly take corresponding handling measures to improve the timeliness of task processing and effectively avoid adverse effects on customers due to task delays. When the first candidate agent leaves their post, the system can automatically read the agent operation management system to obtain on-duty agent information and task attribute data, and use a large model for data analysis to intelligently reallocate tasks according to a differentiated redistribution strategy. This improves the flexibility and rationality of agent resource allocation, makes full use of existing agent resources, and reduces task backlog. By using the QWQ-32B large model to analyze task attribute data, accurate task type data is obtained. Combined with the differentiated redistribution strategy, urgent processing tasks can be accurately assigned to on-duty agents with the corresponding skills and abilities, improving the quality and efficiency of task processing and enhancing customer satisfaction.

[0117] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0118] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. A task allocation method based on agent status, characterized in that, Includes the following steps: Real-time monitoring of the execution time information of pending tasks; When the execution time information reaches the preset warning time threshold, the task to be processed is identified as an emergency processing task, and the first candidate seat corresponding to the emergency processing task is obtained. If the first candidate agent is on duty, a task processing reminder signal corresponding to the emergency handling task is sent to the first candidate agent. If the first candidate agent is off duty, the agent operation management system is read, and the on-duty agent information and the task attribute data corresponding to the emergency handling task are obtained from the agent operation management system. The task attribute data is input into a large language model for data analysis to obtain task type data. The emergency handling tasks are reordered based on the differentiated redistribution strategy and the task type data.

2. The task allocation method based on agent status according to claim 1, characterized in that, The step of reordering the emergency processing tasks according to the differentiated reallocation strategy and the task type data specifically includes the following steps: If the task type data is a regular task type, then a second candidate agent that matches the emergency handling task is obtained from the on-duty agent information according to a greedy algorithm. The emergency handling task is assigned to the second candidate agent. If the task type data is a complex task type, then obtain the historical task data corresponding to the emergency handling task; The historical task data and on-duty agent information are input into the LightGBM classification prediction model for prediction to obtain the third candidate agent. The emergency handling task is assigned to the third candidate agent.

3. The task allocation method based on agent status according to claim 2, characterized in that, The information on available seats includes the total number of seats, the historical processing efficiency corresponding to each seat, and the current task load. The greedy algorithm is specifically expressed as follows: in, This indicates the total number of seats. Indicates the first The historical processing efficiency of each agent seat Indicates the first Current task load of each agent, Represents decision variables.

4. The task allocation method based on agent status according to claim 1, characterized in that, After the step of sending a task processing reminder signal corresponding to the emergency handling task to the first candidate agent if the first candidate agent is on duty, the following step is also included: Predict the predicted execution time of the emergency response task; Obtain the historical experience records corresponding to the first candidate agent; Obtain task characteristic information corresponding to the emergency handling task; The potential execution time range is determined by analyzing the correlation between the historical experience records and the task characteristic information using the support vector machine algorithm. If the predicted execution time does not exceed the potential execution time range, a task processing reminder signal corresponding to the emergency processing task is sent to the first candidate agent. If the predicted execution time exceeds the potential execution time range, then the similarity score between the historical experience record of each on-duty agent and the task characteristic information is calculated to obtain the first matching result for each on-duty agent. The on-duty agent with the highest matching degree in the first matching result is determined as the fourth candidate agent; The emergency handling task is assigned to the fourth candidate agent.

5. The task allocation method based on agent status according to claim 4, characterized in that, After the step of analyzing the correlation between the historical experience records and the task characteristic information using the support vector machine algorithm to determine the potential execution time range, the following step is also included: If the predicted execution time exceeds the potential execution time range, then calculate the skill matching degree between each on-duty agent and the task characteristic information, and the real-time availability of the agent status. Based on the skill matching degree and the real-time availability of the seat status, it was determined as the fifth candidate seat; The emergency handling task is assigned to the fifth candidate agent.

6. The task allocation method based on agent status according to claim 4, characterized in that, After the step of analyzing the correlation between the historical experience records and the task characteristic information using the support vector machine algorithm to determine the potential execution time range, the following step is also included: If the predicted execution time exceeds the potential execution time range, the resource utilization rate of each on-duty agent is calculated separately. The resource utilization rate of each on-duty agent is filtered according to the resource utilization rate threshold to obtain a subset of candidate agents; Calculate the operating cost of each on-duty agent in the candidate agent subset; The on-duty agent with the lowest operating cost was selected as the sixth candidate agent. The emergency handling task is assigned to the sixth candidate seat.

7. A task allocation device based on agent status, characterized in that, include: The execution time detection module is used to monitor the execution time information of tasks to be processed in real time. The first candidate seat acquisition module is used to determine the task to be processed as an emergency processing task when the execution time information reaches a preset warning time threshold, and to acquire the first candidate seat corresponding to the emergency processing task. The first reminder signal sending module is used to send a task processing reminder signal corresponding to the emergency handling task to the first candidate agent if the first candidate agent is on duty. The task attribute acquisition module is used to read the seat operation management system if the first candidate seat is off duty, and to obtain on-duty seat information and task attribute data corresponding to the emergency handling task from the seat operation management system. The task type acquisition module is used to input the task attribute data into the large language model for data analysis to obtain task type data. The task reordering module is used to perform task reordering operations on the emergency processing tasks according to the differentiated redistribution strategy and the task type data.

8. The task allocation device based on agent status according to claim 7, characterized in that, The task reordering module includes: The second candidate agent acquisition submodule is used to acquire a second candidate agent that matches the emergency handling task from the on-duty agent information according to a greedy algorithm if the task type data is a regular task type. The second candidate seat allocation submodule is used to allocate the emergency handling task to the second candidate agent; The historical task data acquisition submodule is used to acquire historical task data corresponding to the emergency handling task if the task type data is a complex task type. The third candidate seat acquisition submodule is used to input the historical task data and on-duty seat information into the LightGBM classification prediction model to perform prediction operations and obtain the third candidate seat. The third candidate seat allocation submodule is used to allocate the emergency handling task to the third candidate seat.

9. A computer device, comprising a memory and a processor, characterized in that, The memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, it implements the steps of the task allocation method based on seat status as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the task allocation method based on seat status as described in any one of claims 1 to 6.

Images