Hospital supervision task processing method, system and device and storage medium

Abstract

The invention discloses a hospital supervision task processing method, system and device and a storage medium. The method comprises the following steps of S10, obtaining complex task text information parameters input in a natural language form; s20, automatically splitting the complex task text information parameters to obtain a plurality of sub-task text information parameters; s30, identifying logic dependency relationship data among the plurality of sub-task text information parameters, and obtaining historical execution efficiency of each execution department; s40, obtaining a department efficiency coefficient according to the historical execution efficiency; s50, calculating a sub-task time parameter set of a plurality of sub-task text information parameters according to the logic dependency relationship data and the department efficiency coefficient; s60, predicting the future workload of the execution department in the sub-task time parameter set; and S70, matching execution unit identifiers for the plurality of sub-task text information parameters according to the future workload.

Description

Technical Field

[0001] This invention relates to the field of hospital supervision task optimization technology, specifically to a method, system, equipment, and storage medium for processing hospital supervision tasks. Background Technology

[0002] In the daily operation and management of modern hospitals, the efficiency and quality of managing tasks, especially complex tasks involving textual information parameters that require collaboration among multiple departments (such as upgrading the outpatient building's information system or hospital-wide continuous improvement projects for medical quality), directly impact the overall operational level of the hospital. To effectively track and manage these tasks, hospitals typically employ information-based office automation (OA) or project management systems.

[0003] However, existing technologies still have a series of profound technical shortcomings when handling such complex supervision tasks, mainly relying on a large amount of manual intervention and subjective judgment by supervision and management personnel, which are specifically reflected in the following aspects:

[0004] Manual breakdown is time-consuming, and the results are often inconsistent, error-prone, and inefficient. For complex task documents, relying entirely on supervisors' personal experience for manual breakdown leads to excessively long average breakdown times per task. Different supervisors have varying understandings of the same task, resulting in inconsistent numbers and levels of detail in the sub-tasks, lacking a unified standard. During manual breakdown, it's easy to overlook crucial execution steps due to inexperience or negligence (e.g., omitting the "multi-department collaborative testing" step), creating potential problems for the project. The execution order is prone to errors, and time adjustments are severely delayed.

[0005] After subtasks are broken down, the logical relationships between them need to be manually clarified, which easily leads to basic logical errors such as "hardware procurement starting before requirements gathering," resulting in chaotic task flow and violating the orderly requirements of project management. Currently, the allocation of time for subtasks is only a rough estimate based on the total cycle, without taking into account the dependencies between tasks and the historical execution efficiency of each department for scientific scheduling. When a preceding subtask becomes overdue, the existing system cannot automatically adjust the time plans of all subsequent related tasks, requiring manual communication and modification, which is slow and inefficient.

[0006] The disconnect between tasks and future departmental workloads leads to task conflicts and a high risk of delays. When assigning sub-tasks to departments, managers can only see the "current" workload of each department and cannot predict its actual workload during the future task execution period. Due to this lack of foresight, tasks are often assigned to departments that are already at full capacity in the future, causing tasks to fail to start on time and requiring frequent re-coordination and re-assignment, increasing management costs. Summary of the Invention

[0007] In order to overcome the technical defects of existing technologies that rely on manual splitting of complex task text information parameters, resulting in non-standard splitting results and low efficiency, this invention provides a method, system, device and storage medium for processing hospital supervision tasks.

[0008] To solve the above problems, the present invention is implemented according to the following technical solution:

[0009] The first aspect of this invention provides a method for processing hospital supervision tasks, comprising the following steps:

[0010] S10: Obtain complex task text information parameters input in natural language form;

[0011] S20: Automatically split the complex task text information parameters to obtain multiple sub-task text information parameters;

[0012] S30: Identify the logical dependency data between the text information parameters of the multiple sub-tasks and obtain the historical execution efficiency of each executing department;

[0013] S40: The department efficiency coefficient is obtained based on historical execution efficiency.

[0014] S50: Based on the logical dependency data and department efficiency coefficient, calculate the sub-task time parameter set of multiple sub-task text information parameters. The calculated sub-task time parameter set is used for subsequent load prediction.

[0015] S60: Predict the future workload of the executing department within the set of time parameters for the subtask;

[0016] S70: Match execution unit identifiers to the multiple subtask text information parameters based on the future workload. In conjunction with the first aspect, the present invention also provides a first specific implementation of the first aspect, wherein S10 specifically includes:

[0017] Step S11: Extract the number of associated departments, technical difficulty coefficient, and deliverable complexity coefficient from the text information parameter description of the complex task as complexity features;

[0018] Step S12: Perform a weighted summation on the complexity features to obtain a complexity quantification score.

[0019] In conjunction with the first aspect, the present invention also provides a second specific embodiment of the first aspect, wherein the automatic splitting step in S20 specifically includes:

[0020] S21: Use a pre-trained language model to perform semantic analysis on the text information parameters of the complex task to obtain semantic analysis data;

[0021] S22: Generate multiple sub-task text information parameters based on the complex quantification score and semantic analysis data;

[0022] The complex metric score has a higher score and a lower score. Based on the higher score in the complex metric score, a subtask text information parameter with a larger number of subtasks is generated, and based on the lower score in the complex metric score, a subtask text information parameter with a smaller number of subtasks is generated.

[0023] In conjunction with the first aspect, the present invention also provides a third specific embodiment of the first aspect, which specifically includes identifying the logical dependency relationship data as follows:

[0024] A natural language processing model is used to scan the text descriptions of the multiple subtask text information parameters and compare them with a predefined dependency keyword library in the supervision task to identify the logical dependency data, which includes preceding dependency data, parallel dependency data, optional dependency data, and conditional dependency data.

[0025] In conjunction with the first aspect, the present invention also provides a fourth specific embodiment of the first aspect, wherein the departmental efficiency coefficient is calculated using the following formula:

[0026]

[0027] Where CE is the department efficiency coefficient, C i W represents the completion rate of the i-th historical task. i T represents the difficulty weight of the i-th historical task. i is the time span weight of the i-th historical task, and n is the total number of historical tasks within the statistical period.

[0028] In conjunction with the first aspect, the present invention also provides a fifth specific embodiment of the first aspect, which further includes:

[0029] S80: When any subtask in the plurality of subtask text information parameters becomes overdue, the subtask time parameter set of one or more subtask text information parameters executed after the overdue subtask is automatically adjusted according to the logical dependency relationship data.

[0030] S81: Generate a list of compressible subtasks, wherein the compressible duration of each subtask in the list is calculated using the following formula:

[0031] D c =T f ×C ev

[0032] Among them, D c For compressible duration, T fC represents the total float time of the subtask. ev This is the department efficiency fluctuation coefficient.

[0033] In conjunction with the first aspect, the present invention also provides a sixth specific embodiment of the first aspect, which further includes:

[0034] Use time series forecasting models to predict base load saturation;

[0035] The load correction coefficient is calculated based on planned training information, equipment maintenance information, seasonal business peak information, and policy node information.

[0036] The future workload is obtained based on the baseline load saturation and the load correction coefficient.

[0037] A second aspect of the present invention provides a management system for hospital supervision tasks, the management system being configured to perform a management method for the hospital supervision tasks, the management system comprising:

[0038] The automatic task splitting module is used to obtain complex task text information parameters input in natural language form, and to use a pre-trained language model to perform semantic parsing on the complex task text information parameters, splitting the complex task text information parameters into multiple sub-task text information parameters.

[0039] A subtask dependency parsing module is used to identify the logical dependency relationship data between the multiple subtask text information parameters.

[0040] The subtask time dynamic allocation module is used to obtain the historical execution efficiency of each executing department, obtain the department efficiency coefficient, and calculate the subtask time parameter set of multiple subtask text information parameters based on the logical dependency data and the department efficiency coefficient.

[0041] The subtask future execution matching module predicts the future workload of the executing department within the subtask time parameter set based on a time series prediction model, and matches the subtask with an execution unit identifier based on the future workload.

[0042] A third aspect of the present invention provides an electronic device, characterized in that the electronic device comprises:

[0043] At least one processor; and a memory communicatively connected to said at least one processor;

[0044] The memory stores a computer program that can be executed by the at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform a method for processing hospital supervision tasks as described in any of the first aspects of the invention.

[0045] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program configured to cause a processor to execute a method for processing hospital supervision tasks as described in any one of the first aspects of the present invention.

[0046] Compared with the prior art, the beneficial effects of the present invention are:

[0047] This invention discloses a method, system, equipment, and storage medium for processing hospital supervision tasks. Through steps S10 to S20, it automatically decomposes complex task text information parameters, replacing the subjective judgment and error-prone process of manual decomposition. This unifies sub-tasks and completes key links, improving the standardization and efficiency of task decomposition. In step S30, the invention automatically identifies logical dependencies between sub-tasks. Then, through steps S40 and S50, it introduces departmental efficiency coefficients calculated based on historical data to plan the time parameter set for sub-tasks. This overcomes the shortcomings of existing technologies that rely on manual dependency analysis, which is prone to errors, and time allocation that is based on rough estimations. It ensures that task time planning is based on data and logic, improving the scientific basis and accuracy of planning. It solves the problem of delayed time adjustments; when a preceding task is overdue, subsequent plans can be automatically adjusted, reducing the complexity of management and coordination and the risk of slow response. In step S60, the invention predicts the future workload of the executing department, and in step S70, it matches the execution unit identifier based on the prediction result. By introducing the prediction of future workload, the frequency of task re-coordination and the overall risk of overdue tasks are reduced. The overall approach reduces subjective intervention and manual time spent by supervisory managers through automated processes, supports integration with existing OA or project management systems, improves the full-cycle management level of complex supervisory tasks in hospitals, promotes overall operational efficiency optimization, and reduces the risk of overall task overdue. Attached Figure Description

[0048] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings, wherein:

[0049] Figure 1 This is a schematic diagram of the user interface for the automatic task splitting results of this invention;

[0050] Figure 2 This is a flowchart of the steps of the present invention;

[0051] Figure 3This is a flowchart of the semantic analysis process of this invention;

[0052] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0053] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0054] like Figures 1-4 As shown,

[0055] The first aspect of this invention provides a method for processing hospital supervision tasks, comprising the following steps:

[0056] S10: Obtain complex task text information parameters input in natural language form;

[0057] S20: Automatically split the text information parameters of complex tasks to obtain multiple sub-task text information parameters;

[0058] S30: Identify the logical dependency data between text information parameters of multiple subtasks and obtain the historical execution efficiency of each department;

[0059] S40: The department efficiency coefficient is obtained based on historical execution efficiency.

[0060] S50: Based on logical dependency data and department efficiency coefficients, calculate the sub-task time parameter set of multiple sub-task text information parameters. The calculated sub-task time parameter set is used for subsequent load prediction.

[0061] S60: Predict the future workload of the executing department within the sub-task time parameter set;

[0062] S70: Match execution unit identifiers to text information parameters for multiple subtasks based on future workload.

[0063] Step S10: Obtain the complex task text information parameters input in natural language form;

[0064] In this embodiment, the supervisor from the hospital's supervision department entered the following complex task text information parameters in natural language form on the "Task Creation Page" of the digital supervision management platform: "The Information Technology Department is required to take the lead, in conjunction with all outpatient departments and the Procurement Department, to complete the upgrade and transformation of the outpatient electronic medical record system in the first quarter of 2025. The project must include the entire process of needs assessment, system procurement, deployment and implementation, personnel training, and online verification, with the goal of improving medical record entry efficiency and reducing system lag."

[0065] In a specific embodiment of the present invention, natural language form refers to a piece of unstructured or semi-structured descriptive text entered by supervisors in a single text input box, following human daily language habits and writing logic.

[0066] This form of input contrasts sharply with traditional, structured form input. Traditional structured input requires users to fill in information such as "task name," "deadline," "department involved," and "budget amount" in multiple separate fields. The "natural language form" processed in this example, however, integrates all this information into a single, complete sentence or paragraph.

[0067] Specifically, the input text has the following examples:

[0068] Example 1: "The Information Technology Department is required to take the lead, in conjunction with all outpatient departments and the Procurement Department, to complete the upgrade and transformation of the outpatient electronic medical record system in the first quarter of 2025. The project should include the entire process of needs assessment, system procurement, deployment and implementation, personnel training and online verification, with the goal of improving medical record entry efficiency and reducing system lag."

[0069] Example 2: "Complete hospital-wide training on DRG payment policy for medical insurance by November 15, 2024, linking the medical insurance department and all clinical departments, with ≥8 training sessions (2 sessions for internal medicine, 2 sessions for surgery, and 4 sessions for medical technology departments), a participation rate of ≥95%, and an assessment pass rate of ≥90%."

[0070] Example 3: "Complete the procurement of equipment for the new surgical ICU ward before January 31, 2025. This involves the Equipment Department, the Surgery Department, and the Procurement Department. The procured equipment includes 3 ventilators and 6 monitors. The budget is ≤ 5 million yuan. The equipment must meet the latest medical equipment standards of 2024."

[0071] In another embodiment of the present invention, after the system obtains a complex task text information parameter in step S10, it assigns a unique task identifier (task ID) to it. All subtask lists, dependency data, time parameter sets, etc. generated in subsequent steps will be associated with the main task through this unique task identifier and stored in the corresponding database table.

[0072] Step S20: After receiving the task description from Example 1 above, the system performs automatic task splitting. This step is further broken down into the following actions in this embodiment:

[0073] First, a task complexity metric score is performed. The system extracts key features from the text in Example 1 and calculates the score according to the defined scoring method, as follows:

[0074] Scope of association: Identify "Information Technology Department, Outpatient Departments, and Procurement Department". Calculated based on 3 departments, this item scores 3 × 10 points / department = 30 points.

[0075] Technical difficulty level: Identifying descriptions such as "system upgrade and transformation" and "deployment and implementation" is classified as "highly technical" and scores 30 points for this item.

[0076] Output requirements: Identify descriptions such as "online verification" and determine that the deliverable is "system + verification". This item is worth 25 points.

[0077] Final score: 30 points + 30 points + 25 points = 85 points.

[0078] Based on the score of 85, the system determined that the task was a "high complexity" task.

[0079] Secondly, semantic parsing and template matching are performed to generate subtasks. The system uses a pre-trained language model (such as the BERT model) fine-tuned with hospital supervision task text to perform semantic parsing on the task description, extracting key actions such as "demand survey," "system procurement," and "training." Since the complexity score in the previous step is 85 points (high complexity), the system matches a "fine-grained" template from the task structure template library that is suitable for high-complexity (score ≥ 70 points) information technology construction tasks. In a specific implementation example, based on the semantic parsing results, the system automatically generates the following structured subtask list:

[0080] Subtask 1: Requirements Research and Analysis (Weight 12%)

[0081] Subtask 2: System Selection and Solution Design (Weight 18%)

[0082] Subtask 3: Procurement Process and Contract Signing (Weight 20%)

[0083] Subtask 4: System Deployment and Environment Setup (Weight 22%)

[0084] Subtask 5: Training for healthcare workers (weight 15%)

[0085] Subtask 6: Online Trial Run and Effect Verification (Weight 13%)

[0086] Each subtask includes clear completion criteria. For example, the completion criteria for "Requirements Survey and Analysis" are: "1. Conduct requirement interviews covering 3 departments in the outpatient department; 2. Output the "...Requirements Report"...; 3. The report is confirmed by both signatures of...".

[0087] Step S30: Identify the logical dependency data between the text information parameters of the multiple sub-tasks, and obtain the historical execution efficiency of each executing department. The historical execution efficiency is obtained from the historical task database of the supervised tasks, including completion rate, difficulty weight and time span.

[0088] First, the system identifies logical dependency data. It scans the names and completion criteria of the six subtasks generated in step S20 and matches them against a predefined dependency keyword library.

[0089] In this embodiment, system analysis revealed that the completion criteria description for subtask 2, "System Selection and Solution Design," explicitly included the text "relies on 'Requirements Research and Analysis' for completion." The system identified the keyword "relies on...for completion," thus determining that subtask 2 has a direct dependency on subtask 1.

[0090] By performing the same scan and analysis on the text descriptions of all six subtasks, the system can automatically and completely construct a topological graph of their dependencies, for example:

[0091] "System deployment and environment setup" depend on the completion of "procurement process and contract signing".

[0092] The "online trial operation" depends on the completion of "medical staff training".

[0093] Secondly, historical execution efficiency is obtained. Based on the execution departments identified in the complexity feature extraction (such as the information department and procurement department extracted from semantic analysis), the system retrieves historical data on the major departments involved in this task from the historical database, including the planned duration, actual completion time, and difficulty level of each historical task, in preparation for the next step of calculating the efficiency coefficient.

[0094] Step S40: Obtain the department efficiency coefficient based on historical execution efficiency;

[0095] In this embodiment, the system uses a weighted model to calculate the efficiency coefficient for each department. Taking the Information Technology Department as an example, the calculation process is as follows:

[0096] The calculation formula is:

[0097]

[0098] Where CE is the department efficiency coefficient, C i W represents the completion rate of the i-th historical task. i T represents the difficulty weight of the i-th historical task. i is the time span weight of the i-th historical task, and n is the total number of historical tasks within the statistical period.

[0099] Ci : The completion rate of the i-th historical subtask, calculated as: planned duration / actual completion time (if completed ahead of schedule, it is counted as 100%).

[0100] W i : The difficulty weight of the i-th historical subtask, which is a level value of 1-5 (1 = lowest, 5 = highest).

[0101] T i : The time span weight of the i-th historical subtask, a floating-point value of 0.5-1.2 (e.g., task duration < 7 days = 0.5, 7-30 days = 1.0, > 30 days = 1.2).

[0102] n: The total number of sub-tasks completed by the department within the statistical period (e.g., the last 3 months).

[0103] In one specific implementation, assuming the Information Technology Department has completed three sub-tasks in the past three months, the system calculates its departmental efficiency coefficient based on its historical data. The specific calculation process is as follows:

[0104] Step 1: Obtain parameters for each historical subtask. The system extracts the completion rate (C) of each subtask from the historical database. i ), difficulty weight (W) i ) and time span weight (T) i ).

[0105] Subtask 1: Development of the Outpatient Registration System

[0106] Completion rate (C) i : 95% (planned 28 days, actual 30 days); Difficulty weight (W) i :5 (High Difficulty); Time Span Weight (T) i ):1.2 (duration greater than 30 days)

[0107] Subtask 2: Monitor Calibration

[0108] Completion rate (C) i :100% (Planned 7 days, actual 5 days, early completion counts as 100%); Difficulty weight (W) i ):2 (Medium to low difficulty); Time span weight (T) i :0.5 (duration less than 7 days)

[0109] Subtask 3: Information Technology Training for Nurses

[0110] Completion rate (C) i : 90% (planned 12.6 days, actual 14 days); Difficulty weight (W) i :3 (Medium difficulty); Time span weight (T) i ):1.0 (Duration 7-30 days)

[0111] Step 2: Substitute into the formula for calculation. The system substitutes the above parameters into the formula for calculating the department efficiency coefficient:

[0112] Calculate the numerator (sum of weighted completion rates): (0.95 × 5 × 1.2) + (1.0 × 2 × 0.5) + (0.9 × 3 × 1.0) = 5.7 + 1.0 + 2.7 = 9.4

[0113] Calculate the denominator (sum of weighted factors): (5 × 1.2) + (2 × 0.5) + (3 × 1.0) = 6 + 1 + 3 = 10

[0114] Calculate the final coefficient: CE = (9.4 / 10) × 100% = 94%

[0115] Through the detailed calculations described above, the system ultimately arrives at the information technology department's...

[0116] The department efficiency coefficient (CE) is 94%. This coefficient will be used in subsequent steps to more accurately calculate the sub-task time parameter set of the task.

[0117] Step S50: Based on the logical dependency data and department efficiency coefficient, calculate the subtask time parameter set of multiple subtask text information parameters. The calculated subtask time parameter set is used for subsequent load prediction.

[0118] Based on the total project cycle (90 days in this embodiment), the dependency topology graph generated in step S30, and the department efficiency coefficient calculated in step S40 (e.g., 94% for the Information Technology Department), the system calculates the time parameter set for each of the six sub-tasks of the "Outpatient Electronic Medical Record System Upgrade".

[0119] The calculation process specifically includes:

[0120] (1) Decompose the total cycle and allocate the baseline duration. The system uses the total cycle (90 days) as a constraint and combines it with the weight of each subtask to initially allocate the baseline duration. For example, for the subtask "Requirements Survey and Analysis" with a weight of 12%, its baseline duration is initially calculated as: 90 days × 12% = 10.8 days.

[0121] (2) Adjusting the duration based on the department's efficiency coefficient. When allocating the final duration, the system will refer to the department's efficiency coefficient calculated in step S40. For departments with a high efficiency coefficient (such as the Information Technology Department at 94%), the task duration allocated to them will be appropriately shortened from the baseline duration; conversely, for departments with a low efficiency coefficient, the duration will be appropriately extended.

[0122] (3) The Critical Path Method (CPM) is used to calculate the final time window. Based on the dependency topology graph between subtasks, the system calculates the "earliest start time (ES), earliest deadline (EF), latest start time (LS), and latest deadline (LF)" for each subtask.

[0123] Its core calculation formula is as follows:

[0124] Earliest Start Time (ES) = Maximum value of the preceding subtask EF.

[0125] For example, the earliest start time (ES) of subtask 2 "System Selection and Solution Design" is equal to the earliest deadline (EF) of its only immediate predecessor, subtask 1 "Requirements Investigation and Analysis".

[0126] Earliest deadline (EF) = ES + adjusted task duration.

[0127] Through the above steps, the system calculates a complete and scientific time window arrangement for the six sub-tasks in this embodiment. For example, for sub-task 4, "System Deployment and Environment Setup," the execution time window defined by the calculated sub-task time parameter set is the 5th to 7th week of the future. This specific time window will serve as the precise time input for load forecasting in the next step, S60.

[0128] Step S60: Predict the future workload of the executing department within the set of time parameters for the sub-tasks;

[0129] Before matching execution unit identifiers to subtasks, the system predicts the load of relevant departments within the subtask time parameter set calculated in step S50. Taking subtask 4 "System Deployment and Environment Setup" as an example, its execution unit identifier is the Information Department, and it is assumed that its subtask time parameter set is the 5th to 7th week of the future.

[0130] The prediction steps are broken down as follows:

[0131] First, a time series forecasting model is used to predict the baseline workload saturation. The system employs the ARIMA model, which in this embodiment is a key time series forecasting tool whose core task is to predict the future baseline workload of each department.

[0132] Core Functionality: The ARIMA model, used to predict base load, does not directly calculate the final future workload that takes all dynamic factors into account. Its role is to analyze historical data to derive a base load saturation level. This baseline value serves as the starting point for subsequent fine-tuning.

[0133] The input data used by this model comes from the system's "Department Capabilities" module, which mainly includes historical data of a specific department over a period of time (e.g., the last 3 months), specifically: the number of tasks being handled by the department and staff leave records.

[0134] Predict the target and output;

[0135] Objective: To predict the department's "load saturation" over the next 1-4 weeks.

[0136] Output: A specific percentage value. For example, in your implementation, the ARIMA model predicts "Pediatric basal load saturation in the next 12 weeks (winter) = 75%".

[0137] More specifically in this embodiment, based on historical data such as the number of tasks in progress and staff leave records of the Information Department over the past 3 months, the basic load saturation of the department in the next 5-7 weeks is predicted, and its basic load is predicted to be 75%.

[0138] Secondly, adjustments are made based on dynamic factors. The system further acquires and analyzes dynamic factors that may affect the workload of the IT department in the next 5-7 weeks.

[0139] In this embodiment, the system finds that the Information Technology Department has a one-day company-wide business training session during this time period. The system calculates the resulting load correction factor according to the following formula.

[0140] Load correction factor = 1 - (Percentage of trainees × Training duration / Number of working days)

[0141] Assuming the week has 5 working days, the participation rate is 100%, and the training duration is 1 day, then: Load correction factor = 1 - (100% × 1 / 5) = 1 - 0.2 = 0.8

[0142] Finally, the final workload is derived. The system sequentially multiplies the base load saturation by all relevant load correction factors to obtain the final future workload. Future workload = Base load saturation × Load correction factor.

[0143] In this embodiment, the predicted future workload is 75% × 0.8 = 60%.

[0144] Step S70: Match execution unit identifiers to the multiple subtask text information parameters based on the future workload.

[0145] In this embodiment, the specific execution unit identifier is the department ID of the department in charge.

[0146] The system compares the final future workload (60%) of the Information Technology Department predicted in step S60 in the next 5-7 weeks with a preset load threshold and combines it with other dimensions for final matching.

[0147] The matching process is as follows:

[0148] First, a load capacity check is performed. The system compares the predicted load with a preset threshold. In this embodiment, the threshold is set to 70%.

[0149] Since the predicted load is 60% < 70%, the system determines that the Information Department has the capacity to take on new tasks within this time window.

[0150] Secondly, a professional matching degree verification is performed. The system will also verify the professional matching degree between the execution unit identifier and the subtask, requiring the matching degree to be no less than a preset value (e.g., ≥80%).

[0151] In this embodiment, subtask 4, "System Deployment and Environment Setup," is highly relevant to the professional field of the Information Technology Department, and the matching degree verification is passed.

[0152] Finally, the department assignment was completed. Since the Information Technology Department met both conditions of "load saturation < 70%" and "professional matching degree ≥ 80%", the system finally confirmed that subtask 4 "system deployment and environment setup" would be assigned to the Information Technology Department and updated the task status.

[0153] In contrast, if the predicted load is higher than the threshold:

[0154] In another scenario, the load on the predictive information department is 85%, meaning 85% > 70% (exceeding the threshold).

[0155] The system will then issue a warning to the supervisor, advising against assigning new tasks to the information department within that time window.

[0156] At the same time, the system will automatically filter other departments that match the specialty and recommend a list of suitable departments to the supervisor, including the original suggested departments and alternative departments, for the supervisor to make a final decision.

[0157] In conjunction with the first aspect, the present invention also provides a first specific embodiment of the first aspect, wherein S10 specifically includes:

[0158] Step S11: Extract the number of related departments, technical difficulty coefficient, and deliverable complexity coefficient from the text information parameter description of the complex task as complexity features;

[0159] Step S12: Perform a weighted summation of the complexity features to obtain a quantitative score for complexity measurement.

[0160] In this embodiment, the system extracts "complexity features" from the input task description text and uses fuzzy comprehensive evaluation to quantify the task complexity (0-100 points). The dimensions and specific indicators of the complexity features are shown in the table below:

[0161] Complexity features Specific indicators Weight Related scope Number of related departments (1 = 10 points, ≥4 = 40 points) 40％ Technical difficulty level Professional skills required? (No professional skills required = 5 points, highly skilled professionals = 30 points) 30％ Output requirements Deliverable complexity coefficient (report only = 5 points, system + verification = 25 points) 30％

[0162] Task complexity metric is performed after step S10 and before step S20.

[0163] The system performs a weighted summation of the feature scores extracted in step S11 according to preset weights to obtain the final complex quantitative score.

[0164] Let's take a highly complex task as an example to illustrate:

[0165] Take the highly complex task of "outpatient information transformation" as an example. Its task description contains key information such as "linking three departments: outpatient, information and nursing, requiring system development, and delivering the system + training report".

[0166] Task Description: "Complete the IT upgrade of the outpatient building by December 31, 2024, linking the outpatient office, information technology department, nursing department, and finance department. Develop an outpatient registration system, payment system, and electronic medical record query system, delivering 3 sets of systems plus an operation manual. After the system goes live, the average patient consultation time will be reduced by 30%."

[0167] The system's calculation steps are as follows:

[0168] Related scope score: Three departments were identified. According to the rule of "10 points for each department", the score for this item is: 3 × 10 = 30 points;

[0169] Technical difficulty rating: The requirement of "system development" has been identified, which is classified as "highly technical". According to the rules, this item is worth 30 points.

[0170] Output requirement scoring: The requirement to deliver "system + training report" is identified. According to the rules, this item scores 25 points.

[0171] Final Score: The final complex quantification score is obtained by simply summing the scores of the three items above.

[0172] 30 points + 30 points + 25 points = 85 points

[0173] Based on the score of 85, the system determines that the task is a "high-complexity" task, which will directly guide the subsequent step S20 in selecting the task template of which granularity to split.

[0174] Based on the total score, the system classifies this task as a high-complexity task (e.g., score ≥ 70 points), which will directly guide the subsequent step S20 in selecting the task template of which granularity to split.

[0175] In conjunction with the first aspect, the present invention also provides a second specific embodiment of the first aspect, in which the automatic splitting step S20 specifically includes:

[0176] S21: Use a pre-trained language model to perform semantic analysis on the parameters of textual information in complex tasks to obtain semantic analysis data;

[0177] S22: Generate multiple sub-task text information parameters based on complex quantified scoring and semantic analysis data;

[0178] The complex metric score has a high score and a low score. The higher score in the complex metric score generates a larger number of subtask text information parameters, while the lower score generates a smaller number of subtask text information parameters.

[0179] In this embodiment, the system uses a medical-domain fine-tuned BERT model, which is fine-tuned using pre-training data containing 500,000 hospital supervision task texts, and semantically annotates the input complex task text information parameters.

[0180] The semantic analysis process specifically includes:

[0181] Extracting core objectives: For example, from the task description "Complete the information technology transformation of the outpatient building before December 2024...", the core objective is "outpatient information technology transformation".

[0182] Extract key actions: Extract key execution steps from the task description, such as "requirements research", "system procurement", "training", and "effect verification".

[0183] Filtering irrelevant semantics: The system will filter out modifiers such as "important" and "routine" to focus on the core aspects of the task.

[0184] The system matches the semantic analysis results in S21 with a task structure template library built based on multiple historical complex task text information parameter cases, and dynamically determines the splitting granularity based on the "complexity quantification score" calculated in the previous steps.

[0185] Complex metric ratings have both high and low scores, and their specific applications are as follows:

[0186] Based on the higher score, more sub-tasks are generated (i.e., matching fine-grained templates). In the "Outpatient Electronic Medical Record System Upgrade" task in this embodiment, its complexity score is 85 points, which belongs to the high complexity task (e.g., score ≥ 70 points).

[0187] Therefore, the system will match a "fine-grained" template, breaking down the task into 6 to 8 detailed subtasks.

[0188] If the semantic parsing result of S21 omits a key subtask (e.g., "effect verification" is not extracted), the system will automatically complete the subtask based on the template and mark the "basis for completion" (e.g., "refer to the 2023 outpatient system transformation case").

[0189] Finally, based on the semantic parsing results, the system automatically generated the following structured list of subtasks for this highly complex task:

[0190] Subtask 1: Requirements Research and Analysis (Weight 12%)

[0191] Subtask 2: System Selection and Solution Design (Weight 18%)

[0192] Subtask 3: Procurement Process and Contract Signing (Weight 20%)

[0193] Subtask 4: System Deployment and Environment Setup (Weight 22%)

[0194] Subtask 5: Training for healthcare workers (weight 15%)

[0195] Subtask 6: Online Trial Run and Effect Verification (Weight 13%)

[0196] Let's take a low-complexity task as an example:

[0197] Task Description: "Archive the September 2024 monthly meeting minutes of the Administrative Department. This task is only related to the Administrative Department. Submit both electronic and paper copies of the minutes by October 7th."

[0198] Complexity calculation:

[0199] Scope of association: 1 department, 10 points

[0200] Technical difficulty level: No professional skills required, score 5 points

[0201] Output requirements: Report / minutes only, 10 points (Note: The original base score has been adjusted to match the total score).

[0202] Total score = 10 + 5 + 10 = 25 points

[0203] The total score is 25 points. Since the complexity score of this task is far below the threshold (e.g., <40 points), the system will match a "coarse-grained" template and only break it down into 2 key sub-tasks.

[0204] Take a medium-complexity task as an example

[0205] "Complete emergency skills training for all nurses in the hospital by November 2024, involving the Nursing Department and Training Department, covering nurses from internal medicine, surgery, and emergency departments (a total of 150 people), deliver training materials and examination papers, achieve a pass rate of ≥90%, and complete the task by November 30th."

[0206] Related areas: 2 departments (Nursing Department, Training Department), totaling 20 points.

[0207] Technical difficulty level: Basic professional skills (first aid training), 20 points

[0208] Output requirements: Courseware + Exam Paper + Report, total 15 points

[0209] Total score = 55 points (50-69 points, broken down into four sub-tasks: "Training Program Development", "Courseware Development", "Training Implementation", and "Assessment and Evaluation")

[0210] In this embodiment, more specifically, the BERT model is the core technology engine for achieving natural language understanding. It is not a general, off-the-shelf model, but rather a "specialist" that has been deeply customized and specifically trained to understand the language of the hospital supervision domain.

[0211] Based on the technical solution document you provided, the specific application and explanation of the BERT model in this patent are as follows:

[0212] Customized training of the model;

[0213] The base model is a Chinese pre-trained BERT model.

[0214] Key innovation: Domain-specific fine-tuning: In order for the model to accurately understand the special terminology and writing style of hospital supervision work, it was fine-tuned using a professional dataset containing 500,000 real hospital supervision task texts.

[0215] Training data source: These fine-tuning data come from the platform's "Supervision Library," which contains a large number of real historical task samples, such as: "Complete the calibration of 5 monitors in the internal medicine ward before October 31, 2024, and link the equipment department, internal medicine department, etc."

[0216] Continuous learning mechanism: The system also features an incremental training mechanism that can periodically (e.g., weekly) extract user-recorded fine-tuning of the automatic splitting results for model retraining, thereby continuously optimizing the accuracy of its semantic parsing.

[0217] In this embodiment, it has the following functions:

[0218] Function 1: Perform semantic parsing in the "Automatic Task Splitting Module";

[0219] When a supervisor inputs a complex task description in natural language (e.g., "Complete the outpatient building's IT upgrade by December 2024..."), the BERT model is responsible for deeply interpreting the meaning of this text. It can accurately extract structured key information, specifically including:

[0220] Core objective: e.g., "Outpatient information system transformation". Key actions: e.g., "Needs assessment", "System procurement", "Training". Constraints: e.g., "Completion before December 2024".

[0221] The extracted information forms the basis for subsequent complexity scoring, task template matching, and subtask list generation.

[0222] Function 2: Identify dependency keywords in the "Subtask Dependency Resolution Module"

[0223] When the textual information parameters of a complex task are broken down into multiple subtask textual information parameters, the BERT model intervenes again. At this point, it scans the name and completion criteria text of each subtask, identifying dependency keywords that indicate the logical relationships between tasks. For example, when it identifies phrases like "based on...result" or "only after...is completed" in the description of a subtask, it can determine that there is a priori dependency relationship between that subtask and other tasks.

[0224] In conjunction with the first aspect, the present invention also provides a third specific embodiment of the first aspect, which specifically includes identifying logical dependency relationship data including:

[0225] A natural language processing model is used to scan the text descriptions of multiple subtask text information parameters and compare them with a predefined dependency keyword library in the supervision task to identify logical dependency data, which includes preceding dependency data, parallel dependency data, optional dependency data, and conditional dependency data.

[0226] In this embodiment, the system uses a natural language processing model (BERT model) that has been fine-tuned by the hospital supervision task text to scan the text descriptions of the six sub-tasks of "Outpatient Electronic Medical Record System Upgrade" generated in step S20, including the sub-task names and completion standards, and automatically identifies the logical dependency relationship data between them by referring to a predefined dependency keyword library.

[0227] The specific keyword libraries required are shown in the table below:

[0228]

[0229]

[0230] In conjunction with the first aspect, the present invention also provides a fifth specific embodiment of the first aspect, which further includes:

[0231] S80: When any subtask in the multiple subtask text information parameters becomes overdue, the subtask time parameter set of one or more subtask text information parameters executed after the overdue subtask is automatically adjusted according to the logical dependency relationship data.

[0232] S81: Generate a list of compressible subtasks, where the compressible duration of each subtask in the list is calculated using the following formula:

[0233] D c =T f ×C ev

[0234] Among them, D c For compressible duration, T f C represents the total float time of the subtask. ev This is the department efficiency fluctuation coefficient.

[0235] In this embodiment, the "Outpatient Electronic Medical Record System Upgrade" task is taken as an example. During the project execution, its first sub-task 1, "Requirements Survey and Analysis," became overdue.

[0236] S80: When any subtask in the text information parameters of multiple subtasks becomes overdue, the system automatically adjusts the subtask time parameter set of one or more subtasks executed after the overdue subtask based on the logical dependency data. According to the calculation in step S50, the original earliest planned deadline (EF) for subtask 1 "Requirements Investigation and Analysis" was day 9. However, in actual execution, this task was delayed by 3 days and was completed on day 12. After monitoring this overdue, the system immediately and automatically triggers the linkage adjustment mechanism. Based on the dependency topology diagram established in step S30, the system synchronously postpones the earliest start time (ES) and earliest deadline (EF) of all subsequent subtasks that depend on "Requirements Investigation and Analysis" (including "System Selection", "Procurement", "Deployment", etc.) by 3 days.

[0237] This change extended the originally planned 90-day project duration to 93 days, exceeding the original plan. Therefore, the system needs to initiate the next step to shorten the project timeline.

[0238] S81: Generate a compressible task list. The compressible duration of each subtask in the compressible task list is calculated using the following formula. To restore the total project cycle to 90 days, the system needs to compress the project duration by 3 days. The system will automatically calculate and generate a "compressible subtask list" as follows:

[0239] First, identify non-critical paths and calculate the total float time. System analysis relies on the topology graph to identify non-critical path subtasks where the total float time is greater than 0. Total float time (T...) f () is the time that a task can be delayed without affecting the overall project duration.

[0240] In this embodiment, the system identifies "hardware procurement" as a non-critical path subtask, with a total float time (T). f (5 days)

[0241] Secondly, the department efficiency fluctuation coefficient is calculated. The system calculates the efficiency fluctuation coefficient (C) based on the historical efficiency data of the execution unit identifier. ev The calculation formula is as follows:

[0242] Departmental efficiency fluctuation coefficient (C) ev = 1 - (Standard deviation of departmental efficiency / Mean of departmental efficiency)

[0243] Finally, the compressible duration is calculated and an inventory is generated. For each non-critical path subtask, the system calculates its compressible duration (D) using the following formula. c ):

[0244] D c =T f ×C evIn this embodiment, it is assumed that the efficiency of the execution unit identifier of the "hardware procurement" task is stable, and its efficiency fluctuation coefficient (C) is... ev The value is 0.97.

[0245] The compressible timeframe for "hardware procurement" (D) c = 5 days × 0.97 ≈ 4.85 days (rounded down to 4 days).

[0246] The system generates a list of all compressible subtasks and their compressible durations, and based on the current total time required for compression (3 days), pushes adjustment suggestions to the supervisor: "Due to the requirement survey being 3 days overdue, it is recommended to compress the hardware procurement time by 3 days (from 22.5 days to 19.5 days). The adjusted total cycle will be in line with the original plan." After the supervisor confirms, the system will automatically update the time windows for all relevant tasks.

[0247] In conjunction with the first aspect, the present invention also provides a sixth specific embodiment of the first aspect, which further includes:

[0248] Use time series forecasting models to predict base load saturation;

[0249] The load correction coefficient is calculated based on planned training information, equipment maintenance information, seasonal business peak information, and policy node information.

[0250] The future workload is obtained based on the baseline load saturation and the load correction factor.

[0251] In this embodiment, taking subtask 4 "System Deployment and Environment Setup" as an example, the execution unit identifier of this task is the Information Department, and the subtask time parameter set is the 5th to 7th week of the future.

[0252] Step 1: Predict base load saturation using a time series forecasting model

[0253] The system uses the ARIMA time series model to predict the basic load saturation of the Information Department in the next 5-7 weeks based on historical data such as the number of tasks in progress and staff leave records over the past 3 months.

[0254] In this embodiment, the predicted base load is 75%.

[0255] Step 2: Calculate the load correction coefficient based on planned training information, equipment maintenance information, seasonal business peak information, and policy event information.

[0256] The system further acquires and analyzes dynamic factors that may affect the workload of the IT department in the next 5-7 weeks, and calculates a load correction factor for each factor. The calculation logic table for these factors and their correction factors is as follows:

[0257]

[0258]

[0259] A second aspect of the present invention provides a management system for hospital supervision tasks, the management system being configured to perform a management method for hospital supervision tasks, the management system comprising:

[0260] The automatic task splitting module is used to obtain complex task text information parameters input in natural language form, and use a pre-trained language model to perform semantic parsing on the complex task text information parameters, splitting the complex task text information parameters into multiple sub-task text information parameters.

[0261] The subtask dependency resolution module is used to identify the logical dependency relationships between text information parameters of multiple subtasks.

[0262] The subtask time dynamic allocation module is used to obtain the historical execution efficiency of each executing department, obtain the department efficiency coefficient, and calculate the subtask time parameter set of multiple subtask text information parameters based on logical dependency data and department efficiency coefficient.

[0263] The subtask future execution matching module predicts the future workload of the executing department within the subtask time parameter set based on the time series prediction model, and matches the execution unit identifier for the subtask based on the future workload.

[0264] The hospital supervision task management system of the present invention includes an automatic task splitting module, a subtask dependency parsing module, a subtask time dynamic allocation module, and a subtask future execution matching module. Each module is integrated with the existing hospital supervision platform (such as an OA system or project management system) through a data storage module. The data storage module includes a task splitting record table, a subtask dependency relationship table, a subtask time window table, and a department future load prediction table, which are associated with the task table and department table of the existing platform.

[0265] 1. The automatic task decomposition module receives complex task text descriptions in natural language and automatically breaks them down into structured subtasks, assigning weights and completion criteria. The system first extracts features from the task description, such as the number of related departments, technical difficulty coefficient, and deliverable complexity coefficient, calculating a quantified complexity score. Using the BERT model, deep semantic parsing is performed on the task description to extract core objectives and key actions. Based on the complexity score, a matching template is selected from a pre-classified task structure template library based on task type and complexity. Combined with the semantic parsing results, a final list of subtasks is generated.

[0266] 2. The subtask dependency resolution module automatically identifies and determines the logical dependencies between subtasks (such as predecessor, parallel, and conditional dependencies) and generates a visual dependency topology graph. It uses a natural language processing model to scan the description text of subtasks and matches it against a predefined dependency keyword library (e.g., keywords for predecessor dependencies: "based on…result", "only after completion"). The system transforms the identified dependencies into a topology graph represented by nodes (subtasks) and connections (dependency types), and can generate SVG format visualizations using tools such as Graphviz.

[0267] 3. The subtask time dynamic allocation module calculates a scientific time window for each subtask based on the total project duration, task dependencies, and departmental historical efficiency, and automatically adjusts the time window when a task is overdue. The system calculates an objective "departmental efficiency coefficient" using a weighted formula based on data such as each department's historical task completion rate, task difficulty, and time span. Using the Critical Path Method (CPM), combined with the total project duration, dependency topology, and efficiency coefficient, the system calculates the earliest / latest start and end times (ES, EF, LS, LF) for each subtask. The system monitors the actual completion time of tasks; if a task is overdue, it automatically extends the time window of subsequent related tasks and generates a "compressible subtask list" for administrator adjustments.

[0268] 4. The subtask future execution matching module proactively predicts the department's workload within a specific future time window and matches the most suitable execution unit identifier for each subtask accordingly. Using the ARIMA time series model, it predicts the department's "basic workload saturation" based on historical task data. This prediction is further adjusted based on a "load correction factor" calculated from dynamic influencing factors such as planned training information, equipment maintenance information, and seasonal business peak information. The final predicted workload saturation is compared with a preset threshold (e.g., 80%) to filter and assign execution unit identifiers with workloads below the threshold and matching the department's expertise to the subtasks.

[0269] This embodiment also includes a data storage and system integration module. The system adds a task splitting record table, a subtask dependency table, a subtask time window table, and a department future load prediction table, and associates these with the existing platform's task table and department table to ensure data sharing and system compatibility. The system backend can be implemented using Python and MySQL, the model can call TensorFlow and statsmodels libraries, and users can interact through the existing supervision platform's web interface.

[0270] In a specific embodiment, taking the application scenario as an example: Information technology transformation of outpatient building

[0271] In a "renovation of outpatient building information technology" project with a total cycle of 90 days, this system can realize the following automated management processes:

[0272] start

[0273] Automatic task breakdown: The task description is automatically broken down into 6 core sub-tasks (requirements research, hardware procurement, system deployment, and multi-department testing).

[0274] Dependency resolution: Clarify the immediate dependencies of "system deployment" on "hardware procurement" and other related relationships.

[0275] Time allocation: Based on the efficiency coefficient of 0.94 of the Information Department, a base duration of 30 days is allocated to the "system deployment" task it is responsible for.

[0276] Department Matching: The load saturation of the Information Department during task execution was predicted to be 60% (below the threshold), and it was successfully matched as the execution unit identifier.

[0277] Dynamic adjustment: If the "hardware procurement" is overdue by 5 days, the system will automatically postpone the start time of the "system deployment" by 5 days and provide a solution to compress the "test task" by 3 days to ensure that the total cycle remains unchanged.

[0278] Finish

[0279] A third aspect of the present invention provides an electronic device, characterized in that the electronic device comprises:

[0280] At least one processor; and a memory communicatively connected to the at least one processor;

[0281] The memory stores a computer program that can be executed by at least one processor, which is then executed by the at least one processor to enable the at least one processor to perform a method for processing hospital supervision tasks according to any of the first aspects of the invention.

[0282] Figure 4 A schematic diagram of an electronic device 10 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0283] like Figure 4As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0284] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0285] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as a method for processing hospital supervision tasks.

[0286] In some embodiments, a method for processing hospital supervision tasks is provided. This method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the chauffeur service order matching method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured for a method of managing hospital supervision tasks by any other suitable means (e.g., by means of firmware).

[0287] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0288] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program configured to cause a processor to execute a method for processing hospital supervision tasks as described in any one of the first aspects of the present invention.

[0289] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0290] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0291] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0292] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0293] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0294] The working principle of the hospital supervision task processing method, system, equipment, and storage medium described in this invention is as follows:

[0295] Users input a complex task description text in natural language through the hospital supervision platform interface (e.g., "Requires the Information Department to lead..." in Example 1). The automatic task splitting module receives this text and first performs complexity quantification. By extracting features such as the number of related departments, technical difficulty coefficient, and deliverable complexity coefficient, a quantified complexity score is calculated (e.g., 85 points is considered high complexity). The module uses a BERT model based on historical task texts from the hospital to perform semantic parsing of the task description to extract core objectives and key actions (e.g., "demand survey" and "system procurement"). Based on the complexity score, the system selects a matching template (e.g., a fine-grained template) from the task structure template library and, combined with the results of semantic parsing, automatically splits the complex task text information parameters into a list containing multiple sub-task text information parameters (e.g., 6 sub-tasks). Each sub-task has a clear weight and completion standard (e.g., "demand survey and analysis" has a weight of 12%). The sub-task dependency parsing module scans the text content in the sub-task list generated in the previous step. The module compares data with a predefined dependency keyword library (e.g., the preceding dependency keyword "based on... result") to determine logical dependencies between subtasks, such as preceding, parallel, or conditional dependencies (e.g., "system selection" depends on "requirements survey"). The system generates a visual dependency topology graph, where nodes represent subtasks and lines represent dependency types. The subtask time dynamic allocation module first calculates an objective departmental efficiency coefficient (e.g., 94% for the Information Technology Department) using a specific weighting formula based on historical task data for each department (e.g., completion rate, difficulty, duration). The system uses the Critical Path Method (CPM), combining the total task duration, dependency topology graph, and departmental efficiency coefficient, to calculate a scientific time window for each subtask, including the earliest / latest start and deadline times (ES, EF, LS, LF). When a subtask is overdue (e.g., requirements survey overdue by 3 days), the module automatically extends the time window of subsequent related tasks and generates a compressible subtask list (e.g., hardware procurement can be compressed by 4.85 days) for administrators to adjust. The subtask future execution matching module uses an ARIMA time series model to predict the base load saturation (e.g., 75%) of the execution unit identifier within the future task subtask time parameter set. This prediction is further adjusted based on a load correction factor (e.g., 0.8) calculated from dynamic influencing factors such as planned training information, equipment maintenance information, and seasonal business peak information, resulting in a final load prediction (e.g., 75% × 0.8 = 60%). The system compares the final predicted load saturation with a preset threshold, filtering and assigning execution unit identifiers (department IDs of the responsible departments) with loads below the threshold and matching professional requirements to the subtasks. All data generated in the process, such as task splitting records, dependencies, time windows, and load prediction results, are stored in a specially added database table (e.g., the task splitting record table stores descriptive text and subtask lists).These new tables will be linked to the existing task and department tables on the hospital's supervision platform via task ID and the department ID of the responsible department, to ensure data sharing and deep system integration.

[0296] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Therefore, any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for handling hospital supervision tasks, characterized in that, The method comprises the following steps: S10: obtaining complex task text information parameters input in natural language form; S20: automatically splitting the complex task text information parameters to obtain a plurality of sub-task text information parameters; S30: identifying logical dependency relationship data between the plurality of sub-task text information parameters, and obtaining historical execution efficiency of each execution department; S40: obtaining department efficiency coefficients according to the historical execution efficiency; S50: calculating sub-task time parameter sets of the plurality of sub-task text information parameters according to the logical dependency relationship data and the department efficiency coefficients; S60: predicting future work load of the execution department within the sub-task time parameter sets; S70: matching execution unit identifiers for the plurality of sub-task text information parameters according to the future work load.

2. The method of claim 1, wherein, The S10 specifically comprises: Step S11: extracting the number of associated departments, technical difficulty coefficients and delivery complexity coefficients in the complex task text information parameter description as complexity characteristics; Step S12: performing weighted summation on the complexity characteristics to obtain a complexity quantitative score.

3. A method of processing a hospital follow-up task according to claim 1 or 2, characterized in that, The automatic splitting step in the S20 specifically comprises: S21: performing semantic analysis on the complex task text information parameters by using a pre-trained language model to obtain semantic analysis data; S22: generating a plurality of sub-task text information parameters according to the complexity quantitative score and the semantic analysis data; Wherein, the complexity quantitative score has a higher score and a lower score, according to the higher score in the complexity quantitative score, a larger number of sub-task text information parameters are generated, and according to the lower score in the complexity quantitative score, a smaller number of sub-task text information parameters are generated.

4. The method of claim 1, wherein, Identifying the logical dependency relationship data specifically comprises: Scanning the text description of the plurality of sub-task text information parameters by using a natural language processing model, and identifying the logical dependency relationship data by comparing with a pre-defined dependency keyword library in the supervision task, the logical dependency relationship data includes immediate dependency data, parallel dependency data, optional dependency data and conditional dependency data.

5. The method of claim 1, wherein, The department efficiency coefficient is calculated by the following formula: Wherein, CE is the department efficiency coefficient, C i is the completion rate of the i-th historical task, W i is the difficulty weight of the i-th historical task, T i is the time span weight of the i-th historical task, and n is the total number of historical tasks in the statistical period.

6. The method of claim 1, wherein, Further comprising: S80: when any sub-task in the plurality of sub-task text information parameters is overdue, automatically adjusting the sub-task time parameter sets of one or more sub-task text information parameters executed after the overdue sub-task according to the logical dependency relationship data; S81: generating a compressible sub-task list, wherein the compressible time length of each sub-task in the compressible sub-task list is calculated by the following formula: D c = T f x C ev Where D c is the compressible duration, T f is the total float time of the subtask, C ev is the department efficiency fluctuation coefficient.

7. The method of claim 1, wherein, The S60 specifically comprises: Using a time series prediction model to predict a basic load saturation degree; Calculating a load correction coefficient according to planned training information, equipment maintenance information, seasonal business peak information and policy node information; Obtaining the future work load according to the basic load saturation degree and the load correction coefficient.

8. A processing system for hospital oversight tasks, characterized by, The hospital supervision task management system is configured to execute the hospital supervision task management method, and the management system comprises: The task automatic splitting module is configured to acquire complex task text information parameters input in natural language, perform semantic analysis on the complex task text information parameters by using a pre-trained language model to obtain semantic analysis data, and split the complex task text information parameters into a plurality of sub-task text information parameters; The sub-task dependency analysis module is configured to identify logical dependency relationship data among the plurality of sub-task text information parameters; The sub-task time dynamic allocation module is configured to acquire historical execution efficiency of each execution department, obtain a department efficiency coefficient, and calculate a sub-task time parameter set of the plurality of sub-task text information parameters according to the logical dependency relationship data and the department efficiency coefficient; The sub-task future execution matching module is configured to predict future work load of the execution department within the sub-task time parameter set according to a time sequence prediction model, and match an execution unit identifier for the sub-task according to the future work load.

9. An electronic device, comprising: The electronic device comprises: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the processing method of the hospital supervisory task according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is used to enable the processor to perform the processing method of the hospital supervisory task according to any one of claims 1 to 7 when executed.

Images