A multi-task intelligent scheduling method and system for workshop equipment operation and maintenance

Abstract

The application provides a multi-task intelligent scheduling method and system for workshop equipment operation and maintenance, applied to the technical field of intelligent scheduling, by real-time sensing of personnel on-duty state, skills and load, a dual correction model of benchmark working hours and personnel efficiency is constructed, realizing accurate matching of tasks. Through the introduction of team load coefficient and personal efficiency coefficient, the method can automatically optimize and balance the work burden of each department, effectively solving the problem that scheduling allocation and actual resource matching are disconnected in the prior art. At the same time, by using the overtime recovery mechanism based on working hour prediction and the overload delay scheduling logic, the full life cycle dynamic closed loop control of operation and maintenance tasks is realized, and the task processing efficiency of equipment operation and maintenance and the overall resource utilization of the system are significantly improved.

Description

Technical Field

[0001] This application relates to the field of intelligent scheduling technology, and in particular to a multi-task intelligent scheduling method and system for workshop equipment operation and maintenance. Background Technology

[0002] In modern workshop operation and maintenance management, equipment maintenance tasks are diverse, and the skill levels, on-duty status, and current workload of different maintenance personnel vary greatly. Existing maintenance task scheduling methods mostly rely on manual scheduling or simple polling allocation. This extensive management approach has significant drawbacks: when the workshop faces various sources of sudden fault repairs, routine maintenance, and equipment modification tasks, the lack of precise dynamic matching of personnel's real-time workload and professional skills leads to delayed task scheduling, and even situations where some maintenance personnel are overworked while others are idle. Furthermore, when maintenance personnel cannot complete tasks on time, there is often a lack of automated recovery and rescheduling mechanisms, resulting in a backlog of maintenance tasks and severely impacting the workshop's production continuity and equipment operating efficiency. In summary, current systems suffer from inefficient technical bottlenecks in handling dynamic resource allocation and closed-loop task progress management in multi-task concurrent environments.

[0003] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0004] In view of the shortcomings of the prior art, this application provides a multi-task intelligent scheduling method and system for workshop equipment operation and maintenance, which aims to solve the technical problems of low resource utilization and delayed task processing caused by the lack of dynamic and accurate matching of personnel load and skills, and the lack of automated task recycling and redistribution mechanism in the existing workshop operation and maintenance task scheduling.

[0005] Firstly, a multi-task intelligent scheduling method for workshop equipment operation and maintenance is provided, the method comprising the following steps: S1: Real-time acquisition of the on-duty status, skill tags, and real-time workload of each operation and maintenance personnel's assigned tasks; S2: When the target task triggers scheduling, the baseline working time is determined according to the type of the target task, and the baseline working time is corrected by combining the historical operation and maintenance personnel's work duration data to determine the estimated execution time of the target task; S3: Based on the on-duty status, skill tags, and real-time workload, match the target task to the target maintenance personnel; S4: Set an overtime unprocessed threshold based on the estimated execution time, monitor the execution progress of the assigned target task, and when the execution progress meets the preset overtime unprocessed threshold, automatically reclaim the target task and re-trigger scheduling allocation; S5: When it is detected that the real-time workload of all tasks undertaken by the operation and maintenance personnel has reached the preset saturation threshold, the non-urgent target tasks are subject to delayed scheduling.

[0006] Furthermore, prior to step S1, the following is included: S01: Collect workshop operation and maintenance tasks from multiple sources in the workshop and store them in a unified scheduling pool. The target tasks are from the scheduling pool. The workshop operation and maintenance tasks include at least fault repair tasks, equipment maintenance tasks, and equipment modification tasks. In step S1, the on-duty status includes being on duty and not being on duty; The skill tags include at least electrical work, mechanical repair, and hydraulics; The real-time workload is obtained by summing the estimated remaining workload of all tasks that the corresponding maintenance personnel have undertaken but not yet completed.

[0007] Furthermore, step S2 includes: S21: When the target task triggers scheduling, obtain the type of the target task; S22: Using the task type identifier of the target task as an index, read the benchmark working hours corresponding to the type from the preset benchmark working hour library; S23: Obtain the identity identifier of the maintenance personnel, and use the identity identifier as an index to read the historical operation duration data of the maintenance personnel from the personnel management module; S24: Calculate the individual efficiency coefficient of the maintenance personnel based on the ratio of the historical operation duration data to the sum of the historical operation duration data of all the maintenance personnel. S25: Multiply the baseline working hours by the individual efficiency coefficient to obtain the estimated execution time of the target task.

[0008] Furthermore, step S3 includes: S31: Using the device location information associated with the target task as an index, determine the maintenance team corresponding to the device location information. A maintenance team includes multiple maintenance personnel. S32: Within the maintenance team, select maintenance personnel whose on-duty status is "already on duty" and whose skill tags match the skill requirements of the target task; S33: Calculate the total matching degree between each of the selected maintenance personnel and the target task based on their real-time workload. S34: Assign the target task to the maintenance personnel with the highest total matching degree.

[0009] Furthermore, step S32 includes: S321: Based on the on-duty status of each maintenance personnel Use these as filtering criteria to filter out... Maintenance personnel, retain The maintenance personnel, among whom, This indicates that the individual was not at their post. This indicates that the individual is already at their post. S322: Read the skill tags of each operations and maintenance personnel, and compare the skill tags of each operations and maintenance personnel with the skill requirements of the target task as the matching condition, and calculate the skill matching degree of each operations and maintenance personnel. The skill matching degree The ratio of the number of intersection elements between the skill tag and the skill requirement to the total number of skill requirement elements; (This is a partial translation of the original text, which is not directly related to the previous sentence.) Maintenance personnel, screening Maintenance personnel; Where i is the serial number of the maintenance personnel, i=[1,N], and N is the total number of maintenance personnel.

[0010] Furthermore, step S33 includes: S331: Obtain the real-time workload of the tasks already undertaken by the i-th operations and maintenance personnel. And obtain the maximum real-time workload of all maintenance personnel in the maintenance team j to which the i-th maintenance personnel belongs. ;in, , where j represents the sequence number of the maintenance team, j=[1,M], and M is the total number of maintenance teams; S332: According to the formula Calculate the idle matching degree of the i-th maintenance personnel. ; S333: Calculate the total matching degree of the i-th maintenance personnel using the formula. : ;in, For idle matching degree The weight, For skill matching The weights; Status: On duty; Let represent the load coefficient of the j-th maintenance team; where, TA represents the maximum real-time workload for all maintenance teams. The maximum value in, .

[0011] Furthermore, step S333 includes the following: S334: When there is no maintenance personnel in the maintenance team whose on-duty status is "on duty" and whose skill tags match the skill requirements of the target task, calculate the total matching degree of all maintenance personnel in the entire plant and assign the target task to the maintenance personnel with the highest total matching degree. The formula for calculating the overall matching degree of all maintenance personnel across the entire plant is as follows: , where λ represents the cross-team matching coefficient.

[0012] Furthermore, step S4 includes: S41: Calculate the timeout threshold of the target task by multiplying the estimated execution time by a preset timeout multiplier. S42: Monitor the execution progress of the assigned target task, wherein the execution progress is the duration of the target task from the time of assignment to the current time; S43: When the duration exceeds the timeout threshold, the target task is automatically reclaimed and steps S2 to S3 are re-executed for scheduling and allocation.

[0013] Furthermore, step S5 includes: S51: Compare the real-time workload of each maintenance worker with the preset saturation threshold; S52: When the real-time workload of all maintenance personnel reaches or exceeds the saturation threshold, the system as a whole is determined to be in a state of load saturation. S53: Obtain the priority identifier of the target task and determine whether the target task is a non-urgent task; S54: If the target task is a non-urgent task, then the target task will be subject to delayed scheduling and scheduling allocation will not be triggered for the time being. S55: If the target task is an urgent task, a manual intervention channel is reserved, and the administrator can perform cross-team support or temporary reinforcement.

[0014] Secondly, a multi-task intelligent scheduling system for workshop equipment operation and maintenance is provided to implement the steps of any of the methods described above, the system comprising: Acquisition module: Acquires real-time on-duty status, skill tags, and real-time workload of each operations and maintenance personnel's assigned tasks; Determining Module: When a target task triggers scheduling, a baseline working time is determined based on the type of the target task. The baseline working time is then corrected by combining the historical work duration data of the operation and maintenance personnel to determine the estimated execution time of the target task. Matching module: Based on the on-duty status, skill tags, and real-time workload, the target task is matched to the target maintenance personnel; Rescheduling module: Sets an overtime threshold based on the estimated execution time, monitors the execution progress of the assigned target task, and automatically reclaims the target task and re-triggers scheduling allocation when the execution progress meets the preset overtime threshold; Delayed processing module: When it is detected that the real-time workload of all tasks undertaken by the operation and maintenance personnel has reached the preset saturation threshold, the non-urgent target tasks are subject to delayed scheduling processing.

[0015] Beneficial Effects: This application proposes a multi-task intelligent scheduling method and system for workshop equipment operation and maintenance. By real-time sensing of personnel on-duty status, skills, and workload, a dual correction model based on baseline working hours and personnel efficiency is constructed, achieving precise task matching. This method, by introducing team load coefficients and individual efficiency coefficients, can automatically optimize and balance the workload of each department, effectively solving the problem of disconnect between scheduling allocation and actual resource matching in existing technologies. Simultaneously, by utilizing a timeout recovery mechanism based on working hour prediction and overload delay scheduling logic, dynamic closed-loop control of the entire lifecycle of operation and maintenance tasks is achieved, significantly improving the task processing efficiency of equipment operation and maintenance and the overall resource utilization rate of the system. Attached Figure Description

[0016] Figure 1 This is a flowchart of a multi-task intelligent scheduling method for workshop equipment operation and maintenance proposed in this application.

[0017] Figure 2 This is a schematic diagram of a multi-task intelligent scheduling system for workshop equipment operation and maintenance proposed in this application.

[0018] Labeling Explanation: 201. Acquisition Module; 202. Determination Module; 203. Matching Module; 204. Rescheduling Module; 205. Delay Processing Module. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. The components of the embodiments of this application described and marked in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] In modern manufacturing workshops, the variety of equipment and the complexity of their operating environments directly impact the overall production efficiency of the workshop. With the expansion of production scale, workshops generate numerous equipment maintenance needs daily, including unexpected equipment failures, regular maintenance, and planned equipment upgrades. Traditional workshop maintenance management often relies on manual dispatching or simple queuing mechanisms. The dispatch center cannot monitor the real-time working status, skill level, and workload of each maintenance worker. This information lag and simplistic dispatching methods easily lead to extremely uneven task allocation, with some personnel overworked while others are idle. Furthermore, critical emergency repair tasks may be delayed because they are assigned to personnel lacking the necessary skills or who are already busy, resulting in prolonged equipment downtime and severely impacting the overall workshop productivity.

[0022] For this, please refer to Figure 1 This application proposes a multi-task intelligent scheduling method for workshop equipment operation and maintenance, the method including the following steps: S1: Real-time acquisition of the on-duty status, skill tags, and real-time workload of each operation and maintenance personnel's assigned tasks; S2: When the target task triggers scheduling, the baseline working time is determined according to the type of the target task, and the baseline working time is corrected by combining the historical operation and maintenance personnel's work duration data to determine the estimated execution time of the target task; S3: Match target tasks to target maintenance personnel based on on-duty status, skill tags, and real-time workload; S4: Set an overtime threshold based on the estimated execution time, monitor the execution progress of the assigned target tasks, and automatically reclaim the target tasks and re-trigger scheduling when the execution progress meets the preset overtime threshold. S5: When it is detected that the real-time workload of all tasks undertaken by all maintenance personnel has reached the preset saturation threshold, the non-urgent target tasks will be delayed.

[0023] Maintenance and operation (O&M) personnel refer to workers in the workshop responsible for equipment repair, maintenance, and inspection to ensure the normal operation of equipment. O&M personnel can be either members of a fixed maintenance team or specialized support personnel categorized by profession, such as electricians handling electrical faults, mechanics handling mechanical disassembly and adjustment, and hydraulic repair personnel troubleshooting hydraulic circuits. Different O&M personnel typically differ in their skill set, scope of tasks they can handle, and work efficiency. Therefore, when scheduling, they should not be treated simply as a uniform number of execution units, but rather as scheduling objects with status and capability attributes.

[0024] Real-time workload refers to the total time required for all tasks undertaken but not yet completed by a specific individual at any given moment, used to measure the individual's current workload. On one hand, real-time workload needs to be dynamically updated based on events such as task acceptance, commencement, pause, reassignment, and completion; on the other hand, only incomplete tasks are counted, excluding closed and archived tasks. For example, if the same individual is currently undertaking a troubleshooting task, a maintenance and re-inspection task, and an pending modification coordination task, the remaining estimated execution time for each of these three tasks should be aggregated to reflect their actual workload.

[0025] Baseline working time refers to the standard reference time required to complete a specific type of task under standard working conditions and routine operating procedures. Baseline working time is not an arbitrarily assigned empirical value, but rather a fundamental reference for subsequent estimated execution time calculations. It can be formed by: categorizing and statistically analyzing historical work orders in the workshop and extracting typical time consumption at intermediate levels; or by having workshop process, equipment, and maintenance management personnel jointly confirm the baseline working time in conjunction with standard operating procedures, and then adjusting it according to the working time consumed for each identical task using an existing algorithm. For example, for tasks such as replacing standard-specification sensors, routine lubrication maintenance, and ordinary mechanical tightening and resetting, corresponding baseline working time records can be established. The time consumed by maintenance personnel for each task will be adjusted according to the corresponding baseline working time using the algorithm. It should be noted that baseline working time reflects the general time consumption level of the task type itself, not the actual time consumed by a specific individual.

[0026] Regarding the technical means of obtaining personnel information in step S1, in a specific embodiment, this application no longer relies solely on paper registration by the team leader and verbal inquiries via walkie-talkie. Instead, it obtains information through collaboration between the existing information system in the workshop and on-site terminals. Specifically, this can include the following implementation process: When each maintenance personnel enters their shift, they sign in via an attendance terminal, work badge swiping terminal, or mobile work terminal. The sign-in result is written into the personnel status table to establish a basic status of being on duty or not on duty. Skill tags can be pre-maintained in the personnel management module. The entry method can be that after personnel are hired, their positions are adjusted, or their skills are certified, the management end writes the corresponding job skills into the personnel file, for example, marking a person as "electrician + mechanic" or "hydraulics". For obtaining real-time workload, the system does not roughly estimate by verbally asking how many tasks are currently being performed. Instead, it reads all task records under the personnel's name that are assigned but not completed, in progress, or paused from the task management module, extracts the estimated execution time or remaining estimated execution time for each of these tasks, and then adds them up to obtain the personnel's current real-time workload.

[0027] To avoid load distortion caused by delayed task status updates, a more complete implementation can include a status verification mechanism: if a task is not initiated by maintenance personnel at the designated location within a preset reminder period using the NFC function of the personnel's terminal to swipe the designated task card, a task reminder is sent to the personnel's terminal. If the task remains unattended after the reminder period, or if the personnel's on-duty status changes to off-duty, the task is returned to the scheduling pool and reassigned to other on-duty personnel according to predetermined rules. Through this method, personnel on-duty status, skill tags, and real-time workload all have clear data sources, update trigger conditions, and verification paths, thus supporting the stable operation of subsequent scheduling steps.

[0028] Regarding the technical means of determining the estimated execution time in step S2, in order to ensure that this step can be truly implemented, the following example can be used as a reference: the scheduling system pre-establishes two data units: a baseline time database and a historical job duration database.

[0029] The baseline work time database is used to store the correspondence between task types and standard reference times. It can be established by first classifying historical work orders in the workshop, and then having equipment managers merge similar tasks to form a unified task type identifier, such as motor replacement, sensor calibration, hydraulic leak troubleshooting, periodic lubrication, etc. Subsequently, sample work orders with normal operating conditions, complete records, and no abnormal interruptions are selected from historical completed work orders, their actual time consumption is calculated, and a baseline work time is confirmed in combination with the standard operating procedure. Finally, it is written into the database in the form of task type identifier - baseline work time. After that, the algorithm is adjusted according to the time consumed by each maintenance personnel task.

[0030] The historical job duration database records the historical completion time of different maintenance personnel for different task types. The records include at least the personnel identification, task type identification, task start time, task completion time, whether there were external waits, whether spare parts were missing, and the final confirmed effective job duration. This setup aims to minimize time deviations caused by non-human capability factors such as waiting for spare parts, waiting for downtime windows, and waiting for process confirmation during subsequent efficiency adjustments.

[0031] In the specific processing, when a target task triggers scheduling, the system first reads the task type identifier. This task type identifier can originate from the task creation stage: for example, an equipment abnormality alarm is mapped to a certain fault type after rule parsing; maintenance work orders generated by MES directly carry maintenance category codes; and modification work orders issued by the process department specify the modification category using the work order template. The system uses this task type identifier as an index to read the corresponding baseline working hours from the baseline working hours database. If multiple sub-levels exist, the system prioritizes reading the task type with the finest granularity and established baseline working hours; if the finest granularity type has not yet been recorded, it falls back to the baseline working hours corresponding to the previous level task category to ensure that the system can still output usable results even when task type coverage is incomplete.

[0032] Subsequently, the system retrieves the historical job duration data for each candidate operations and maintenance personnel. Specifically, the historical job duration data preferably uses a set of effective completion times for similar tasks, rather than simply retrieving the total time for all tasks. This is because different task types have significantly different skill requirements and work paces; mixing unrelated tasks in the statistics would weaken the targeted nature of efficiency adjustments.

[0033] Specifically, the system first filters historical task duration records from the database based on the type of the target task, identifying similar tasks completed by the individual within a past statistical period. Then, it removes obviously abnormal records, such as those caused by prolonged equipment downtime, interruptions due to undelivered spare parts, or distorted records resulting from multiple people collaborating without splitting the work hours. For the remaining records, the system can use representative durations as the individual's historical task duration data for that type of task, such as a typical duration at an intermediate level or the average of multiple valid records. Simultaneously, the system also reads the historical task duration data for that type of task from all maintenance personnel possessing the same skill and with similar task processing records, forming a horizontal comparison benchmark.

[0034] Based on this, the system calculates the individual efficiency coefficient of each maintenance worker by comparing the historical job duration data with the sum of the historical job duration data of all maintenance workers. To ensure that this coefficient conforms to the scheduling logic, in actual implementation, this ratio can be further converted into a directional coefficient that indicates shorter execution time, higher efficiency, and a smaller estimated execution time after correction.

[0035] In other words, if an individual's historical effective working time for similar tasks is generally shorter than the overall level of the same skill group, then their individual efficiency coefficient should reflect a superior level; if an individual's historical effective working time is generally longer than the overall level, then their individual efficiency coefficient should reflect a relatively conservative level. After this processing, the baseline working time is multiplied by the individual efficiency coefficient to obtain the estimated execution time of the target task for the candidate.

[0036] To avoid excessive fluctuations in results due to insufficient historical samples, a more robust implementation approach involves preventing the system from directly applying strong corrections to an individual's historical data when their number of valid historical records for a particular task falls below a preset sample threshold. Instead, a weaker correction strategy is employed, such as referencing the average level of the skill group. The individualized correction is gradually increased as sufficient records are accumulated. This method preserves the standard timeframe for the task type while also reflecting the actual performance differences among individuals on similar tasks.

[0037] Regarding the matching and allocation process in step S3, the system does not use a polling method based on the order of personnel lists, or a method where the nearest personnel responds automatically via workshop broadcast. Instead, it executes according to a clearly defined screening order and allocation rules. Specifically, this process may include the following steps: First, the system filters out personnel whose on-duty status is "on duty" from all maintenance personnel, excluding those who are not on duty. Second, it matches the remaining personnel according to the skill tags required for the target task. For example, if the target task is electrical fault diagnosis, only personnel with the electrician skill tag are retained. If the task requires multiple skills, personnel with multiple corresponding skill tags are retained, or a set of personnel who meet the primary skill requirements and can collaborate with other personnel are retained. Third, the system reads the current real-time workload of these candidate personnel and, combined with the estimated execution time of the task at each candidate personnel obtained in step S2, determines whether the workload exceeds an acceptable range after the task is allocated.

[0038] If multiple candidates meet the criteria, the candidate with the lower total workload after assignment will be given priority. If the total workloads are similar, their efficiency in handling similar tasks, their current task priority, and the presence of significant geographical advantages will be further compared. Geographical location can be used as a secondary factor. For example, within the same workshop area, if two candidates have similar skills and workloads, the candidate closer to the target equipment will be given priority to reduce unnecessary travel time. However, this factor should not be more important than skill matching and load balancing.

[0039] Once a task is assigned to a specific person, the system pushes task information to the corresponding maintenance personnel's terminal. This information includes at least the task number, device identifier, task type, fault or job description, priority, required arrival time, and safety precautions. If the maintenance personnel do not confirm acceptance within the preset response time period, the system can mark the task as not validly received and automatically transfer it to the next candidate or re-trigger the scheduling process. Through this method, task allocation no longer relies on chance or subjective bidding, but is based on a comprehensive judgment of three core information categories: status, capability, and load.

[0040] Regarding timeout monitoring in step S4, the system does not use a fixed countdown timer for all tasks. Instead, it sets a timeout threshold based on the estimated execution time. This is because different tasks vary significantly in complexity, and a uniform threshold can lead to two problems: for simple tasks, an excessively long threshold may result in undetected abnormal delays; for complex tasks, an excessively short threshold may cause frequent false alarms. Therefore, in practice, after task allocation, the system generates a timeout threshold based on the estimated execution time for the personnel assigned to that task. This threshold can be understood as a combined control rule for allowed processing time and progress feedback requirements.

[0041] Specifically, the system monitors at least the following types of execution progress events: whether the task enters the work-in state after being accepted, whether there is continuous progress feedback after work starts, whether the task is completed within the reasonable range corresponding to the estimated execution time, and whether it remains in a paused state for a long time without resuming. If a task does not start for a long time after being accepted, it means that although the task has been assigned, it has not actually entered the execution phase; if a task has started but there are no progress updates for a long time, it means that there may be processing stagnation, personnel forgetting, terminal omission, or on-site obstruction; if the task continues to exceed the threshold corresponding to the estimated execution time and is still not completed, it means that the task may have timed out and not been processed.

[0042] Once the system determines that the preset timeout threshold has been met, it does not necessarily immediately and simply withdraw the process. Instead, it can first perform a verification process, such as sending a status confirmation request to the maintenance personnel's terminal, requiring them to choose a status such as continuing processing, waiting for external conditions, completed and awaiting submission, or unable to continue processing.

[0043] If the returned result shows that the task cannot proceed due to objective reasons such as waiting for spare parts or waiting for downtime permission, the system can put the task into an obstructed state and pause the timeout countdown; if the returned result shows that the personnel cannot continue to process or the terminal is unresponsive, the system will automatically reclaim the target task and re-trigger the scheduling and allocation.

[0044] During the recovery process, the system should simultaneously release the corresponding real-time workload of the original assigned personnel and retain information such as the current processing record, completed steps, and discovered faults in the task file. This allows the reassigned personnel to quickly take over and avoids redundant troubleshooting. Through this dynamic monitoring method based on estimated execution time, the timeout recovery mechanism can more accurately distinguish between normal long processing times and abnormally stalled or unprocessed situations.

[0045] Regarding the saturation handling in step S5, the system does not determine saturation by simply counting the number of tasks assigned to each employee, but rather by using real-time workload as the primary criterion. This is because the same number of tasks does not necessarily mean the same workload; a complex renovation task may be far more demanding than multiple simple inspection tasks.

[0046] In practice, the system pre-sets a saturation threshold for each maintenance worker. This threshold represents the maximum workload that worker can handle under the current shift, job responsibilities, and safety requirements. This threshold can be pre-configured by workshop management rules and can be differentiated based on shift type, job category, or special periods. For example, when there are fewer night shift workers, the threshold strategy can be more conservative; during high-risk operations, the threshold can be appropriately tightened to avoid safety issues caused by continuous high-intensity work. When the system detects that the real-time workload of all maintenance workers has reached their respective pre-set saturation thresholds, it determines that the current workshop has entered a saturation state.

[0047] At this point, for newly entered target tasks in the scheduling pool, the system first determines their urgency. If it is an urgent task, such as one that causes mainline downtime, poses a safety hazard, or may cause further equipment damage, it will still enter the forced scheduling process and may trigger an alarm on the management end, prompting that manual intervention, cross-shift support, or temporary resource adjustments are required. If it is a non-urgent task, such as routine inspections, general maintenance, or planned modification tasks for non-critical equipment that can be postponed, it will be subject to delayed scheduling.

[0048] The specific method for delayed scheduling can be as follows: The task is written to a deferred queue, recording the time it entered the deferred queue, its priority, the latest allowed release time, and the release conditions. When the real-time workload of any operations and maintenance personnel drops below its saturation threshold, or when the system detects the addition of new available operations and maintenance resources, the task is released from the deferred queue according to priority and re-enters scheduling. To prevent non-urgent tasks from being indefinitely postponed, the system can also set escalation rules for tasks in the deferred queue. That is, when a task remains in the deferred queue for a preset duration, or when the status of its associated equipment deteriorates, its priority is increased and it is re-evaluated. Through this workload-based saturation processing method, the system can prioritize critical tasks when resources are scarce, while avoiding misjudgments caused by simply judging based on the number of tasks.

[0049] Through the aforementioned overall technical solution, the system transforms the static task dispatching model into a dynamic closed-loop scheduling model. By acquiring real-time personnel status and load, it avoids assigning tasks to off-duty or severely overloaded employees, ensuring scheduling feasibility. Adjusting the baseline working hours makes task time assessments more aligned with actual conditions. Combined with an overtime recovery mechanism, it completely solves the deadlock problem of task stagnation caused by personnel stalling after task allocation in the traditional model. Finally, the saturation delay handling mechanism effectively prevents the scheduling system from crashing and personnel from becoming overly fatigued when the workshop faces sudden large-scale failures. This application significantly improves the flow efficiency and timeliness of workshop equipment maintenance tasks.

[0050] Furthermore, since a data link is formed between real-time workload, estimated execution time, timeout threshold, and saturation threshold, the system can make judgments based on the same set of statuses during each task allocation, recycling, and redistribution. Therefore, it not only improves the accuracy of a single task dispatch but also enhances the stability and consistency of continuous multi-task scheduling throughout the entire shift.

[0051] Furthermore, prior to step S1, the following is included: S01: Collect workshop operation and maintenance tasks from multiple sources in the workshop and store them in a unified scheduling pool. The target tasks are from the scheduling pool. Workshop operation and maintenance tasks include at least fault repair tasks, equipment maintenance tasks, and equipment modification tasks. In step S1, the on-duty status includes being on duty and not being on duty; Skill tags should include at least electrical work, mechanical repair, and hydraulics; Real-time workload is calculated by summing the estimated remaining hours of all tasks that the corresponding maintenance personnel have undertaken but not yet completed.

[0052] More specifically, in step S01, the system constructs a unified digital scheduling pool. Equipment anomaly alarms (fault repair tasks) automatically reported by IoT sensors in the workshop, lubrication / tightening instructions (equipment maintenance tasks) automatically generated periodically by the Manufacturing Execution System (MES), and production line upgrade work orders (equipment modification tasks) issued by the process department are all aggregated into this scheduling pool through a standard interface.

[0053] To ensure that tasks from different sources can enter the same scheduling process, the system standardizes task data upon access, forming at least the following fields: task number, task source, device identifier, device location, task type identifier, task description, creation time, urgency level, required skill tags, whether delay is allowed, and attachment information.

[0054] Taking equipment malfunction alarms as an example, the original sensor report may only contain the equipment number, alarm code, and alarm time. After the system is connected, it can map the fault category based on the alarm code and supplement the production line to which the equipment belongs, the equipment location, and suggested skill tags. Taking maintenance tasks generated by MES as an example, the system can directly read the work items and planned time windows in the maintenance template and write them into the scheduling pool. Taking modification work orders issued by the process department as an example, the system can read the modification object, planned downtime window, and collaboration requirements through the work order interface.

[0055] After the above standardization, although tasks from different sources have different generation mechanisms, they all appear as scheduled objects with a consistent structure at the scheduling level, which facilitates subsequent unified screening and sorting.

[0056] In step S1, the on-duty status of personnel is divided into two categories: on-duty and off-duty. The system can set clear status switching rules: after successful check-in, personnel initially enter the on-duty status; when personnel check out, request leave, leave their post, or their shift ends, they switch to off-duty.

[0057] Skill tags are further subdivided into professional categories such as electrician, mechanic, and hydraulics based on the actual job types in the workshop. Skill tags can be pre-maintained through the personnel management module, and the maintenance basis can include job appointment information, training qualification records, skill certification results, and historical authorization records. For example, after a person completes specialized hydraulic system training and passes the assessment, a hydraulics tag can be added to their file; if a person only possesses mechanical disassembly and assembly skills but lacks the qualification for live-line work, then only the mechanic tag will be retained, and no electrician tag will be assigned.

[0058] When calculating real-time workload, the system no longer simply counts, but instead iterates through all incomplete tasks under the employee's name, extracting the estimated remaining time for each task and accumulating it. Furthermore, in a more granular implementation, if a task has been partially completed, the system can prioritize using the estimated remaining time for accumulation, making the workload more closely reflect the current reality. For example, if an employee has a major overhaul task and an inspection task, if the major overhaul task has not yet started, its full estimated remaining time is counted; if the inspection task has been largely completed, only its estimated remaining time is counted. With this processing, the real-time workload reflects the amount of time still needed in the future, rather than how many tasks have been taken on historically.

[0059] The estimated remaining working hours are calculated and stored in the system during task allocation. Its essential meaning is the same as the estimated execution time in step S2, and the calculation method is the same.

[0060] Further, step S2 includes: S21: When the target task triggers scheduling, obtain the type of the target task; S22: Using the task type identifier of the target task as an index, read the corresponding baseline working hours from the preset baseline working hours library; S23: Obtain the identity identifier of the operation and maintenance personnel, and use the identity identifier as an index to read the historical operation and maintenance personnel's work duration data from the personnel management module; S24: Calculate the individual efficiency coefficient of the operations and maintenance personnel based on the ratio of historical operation time data to the sum of historical operation time data of all operations and maintenance personnel; S25: Multiply the baseline working hours by the individual efficiency coefficient to obtain the estimated execution time of the target task.

[0061] When assessing task time, relying solely on uniform reference standards or subjective experience often overlooks the individual skill differences among employees in the workshop. Experienced workers and newly hired apprentices require vastly different times to handle the same fault; a one-size-fits-all time estimate can lead to significant deviations in subsequent timeout monitoring and load calculations. Therefore, this specific implementation plan introduces a personalized efficiency correction mechanism based on historical data.

[0062] In practice, the system pre-builds a relational database as a baseline work time library, which records the standard time consumption for various tasks. This baseline work time library is not manually entered on an ad-hoc basis, but is formed through a relatively stable database creation process. Specifically, this includes: first, exporting completed task records within a statistical period from the historical work order system; then, classifying tasks according to equipment type, fault type, work content, and work method to form a unified task type identifier; subsequently, selecting sample tasks with complete records, normal working conditions, and no long external waiting times, and calculating the actual effective working time of these sample tasks; finally, the equipment management end reviews the statistical results to confirm that they represent the typical time consumption under standard working environment and routine operating procedures, and then writes them into the baseline work time library. Subsequently, the corresponding baseline work time is adjusted by algorithm based on the time consumption of future maintenance personnel's tasks. The baseline work time obtained in this way has a clear source and is traceable. Taking the task type of replacing a servo motor as an example, the system does not require a fixed value to be written, but rather saves the standard reference time confirmed after statistics and review as the baseline work time record for this type.

[0063] When a scheduling event is triggered, the system first extracts the type identifier of the current task. This type identifier can come from a template field when the task is created, or from an alarm code-task category mapping table. To ensure the stable use of the preset baseline work hour database, the system can also pre-establish a "task type identifier - baseline work hour" mapping table. This mapping table can be established by the management end first compiling a list of common maintenance tasks in the workshop, assigning a unique type identifier to each type of task, and then associating and saving each type identifier with the corresponding baseline work hour record.

[0064] For example, one type of electrical replacement task corresponds to one type identifier, and another type of hydraulic troubleshooting task corresponds to another type identifier. During scheduling, the system locates the corresponding record in the mapping table based on the type identifier carried by the current task, and then reads the baseline working hours. If the current task belongs to a newly emerging sub-type and no dedicated record has been established yet, the system can first roll back to use the baseline working hours of its parent category and mark the task as a sample to be added to the database for subsequent improvement of the baseline working hours database.

[0065] Subsequently, the system enters the personnel management module, using the employee's employee ID (identity identifier) ​​to query and extract the employee's historical work duration data for similar tasks within a past statistical period. The identity identifier preferably uses a unique and stable employee ID within the company to avoid data confusion caused by duplicate names, job changes, etc. The process of obtaining historical work duration data may include: first, locating the employee's historical work order records based on the employee ID; then, filtering similar tasks based on the target task type identifier; then, removing abnormal samples from these similar tasks, such as tasks suspended for extended periods due to production downtime, collaborative tasks without splitting individual work hours, or tasks with missing start or end times; finally, extracting representative durations from the remaining valid samples as the employee's historical work duration data for that type of task. Simultaneously, the system also aggregates the historical work duration data from all maintenance personnel with corresponding skill tags and valid records of similar tasks to form an overall comparison benchmark.

[0066] The system divides the employee's historical work time data by the sum of the historical work time data of all employees in the workshop with the same skill, resulting in a ratio as the individual efficiency coefficient. To ensure this coefficient accurately reflects efficiency, the system applies directional corrections and boundary constraints to this ratio in practical applications. That is, if an employee's historical work time is relatively short, their individual efficiency coefficient should relatively reduce the final estimated work time; conversely, if an employee's historical work time is relatively long, their individual efficiency coefficient should relatively increase the final estimated work time. Furthermore, to avoid distortion of the correction range due to extreme cases, the system can set a reasonable range for the individual efficiency coefficient. When an employee's historical sample size is too small or the fluctuation is too large, a more moderate correction result is prioritized. Finally, the system multiplies the baseline work time by the individual efficiency coefficient. Through this data processing logic, the system can generate an estimated work time with individual differences for each specific "task-person" combination, rather than outputting a uniform value that is the same for all employees.

[0067] Furthermore, step S3 includes: S31: Using the equipment location information associated with the target task as an index, determine the maintenance team corresponding to the equipment location information. A maintenance team includes multiple maintenance personnel. S32: Within the maintenance team, select maintenance personnel who are currently on duty, available to take on tasks, and whose skill tags match the skill requirements of the target task; S33: Calculate the total matching degree between each operation and maintenance personnel and the target task based on the real-time workload of each selected operation and maintenance personnel; S34: Assign the target task to the operations and maintenance personnel with the highest overall matching degree.

[0068] This specific implementation plan first simplifies the spatial and organizational structure by using equipment location information. Equipment location information is typically pre-maintained in the equipment ledger, representing the physical area where the equipment is located or the production line it belongs to. This equipment ledger can be an equipment master data table in an enterprise asset management system, equipment management system, or manufacturing execution system. The table should at least record fields such as equipment number, equipment name, installation location, production line, equipment location information, and responsible work team. Equipment location information can be maintained using text fields, such as stamping area, welding area, and final assembly area 1, or using coded fields, such as AREA-01, AREA-02, etc., establishing a one-to-one or one-to-many correspondence between the location code and the maintenance team code in the workshop organization table.

[0069] Maintenance teams are grassroots organizations within the workshop, divided according to region or specialty for ease of management. By establishing a mapping relationship between "equipment location information and maintenance team," the system can quickly locate the dedicated team responsible for that area. This mapping relationship can be obtained through initial configuration. For example, when the system goes live, equipment managers can enter the information into the configuration table one by one, based on the existing work group assignments, such as "stamping area - stamping maintenance team," "welding area - welding maintenance team," and "final assembly area - final assembly maintenance team." Subsequently, when the workshop layout is adjusted, equipment is relocated, or work groups are reorganized, the administrator updates the corresponding relationship in the backend interface.

[0070] Subsequently, a dual filtering process based on status and skills is implemented within the selected teams to eliminate personnel who are unable to accept orders or whose capabilities do not match, thus forming a pool of candidates. Finally, the overall matching degree is calculated based on real-time workload to select the best among the best. This processing path is adopted because equipment location information directly reflects the location where the task occurs, and the maintenance team reflects the actual jurisdictional responsibility boundary. By first narrowing the scope by location and then comparing personnel status and capabilities within the team, invalid traversal can be significantly reduced, and a large number of cross-regional work assignments can be avoided.

[0071] In practice, the workshop is assumed to be divided into a stamping area, a welding area, and a final assembly area, corresponding to the stamping maintenance team, the welding maintenance team, and the final assembly maintenance team, respectively. When a press in the stamping area malfunctions and triggers a task, the system reads the location information of the press as the stamping area, and then identifies the stamping maintenance team as the target maintenance team.

[0072] This reading process can be achieved by calling the equipment ledger interface through the task system. Specifically, when a press fault signal is uploaded by the PLC, equipment monitoring terminal, or manual repair terminal, the system queries the equipment master data table based on the equipment number carried in the task, returning the equipment's location field and shift responsibility information. Next, the system iterates through the 10 maintenance personnel in the stamping maintenance team, finding that 2 are on leave (not on duty), while the remaining 8 are on duty.

[0073] The on-duty status here can be obtained through the attendance system, scheduling system and work order execution system. For example, the leave status comes from the human resources scheduling module, and the on-site check-in status comes from the mobile terminal check-in record.

[0074] The system further compared the skill tags of these eight individuals with the mechanical and hydraulic skills required for press machine malfunctions, selecting three qualified candidates. Skill tags can originate from one or more of the following: personnel qualification files, training completion records, skill level certificates, and statistical results of historical maintenance records. For example, after completing hydraulic system maintenance training and assessment, the training management module automatically adds a hydraulic tag to an individual's file; if an individual independently completes multiple press machine hydraulic circuit malfunction repairs consecutively, the supervisor can review and add the corresponding skill tag.

[0075] Finally, the system calculates the real-time workload of each of the three individuals, determines their total matching degree, and assigns the task to the individual with the highest total matching degree (usually meaning the one who is currently the most available and best matched). The assignment can be sent simultaneously via mobile maintenance terminals, SMS, enterprise instant messaging tools, or workshop dashboards. The task can also include the equipment number, fault description, location navigation, safety tips, and arrival time limit.

[0076] The working principle of this specific implementation plan lies in transforming the originally complex global search into local optimization through a progressive logic of "location-team locking-initial status and skill screening-precise load matching." This approach effectively solves the computational redundancy and cross-regional scheduling problems caused by blind matching across the entire plant. By prioritizing task allocation within dedicated maintenance teams, not only is the physical distance for personnel to reach the site significantly shortened, but the experience advantage of team members being more familiar with the equipment in their area is also fully utilized, thereby significantly improving task response speed and first-time repair rate. Furthermore, team members are usually familiar with the installation location, common failure modes, spare parts storage locations, and safety isolation procedures of the equipment in their area, so they can arrive at the scene and enter the handling state more quickly after receiving an order; the system's initial restriction to dedicated teams also ensures that subsequent skill comparisons and load calculations are only performed within a smaller scope, reducing the processing pressure on the scheduling service in high-concurrency repair scenarios.

[0077] Further, step S32 includes: S321: Based on the on-duty status of each maintenance personnel Use these as filtering criteria to filter out... Maintenance personnel, retain The maintenance personnel, among whom, This indicates that the individual was not at their post. This indicates that the individual is already at their post. S322: Read the skill tag of each operation and maintenance personnel, take the skill requirement of the target task as the matching condition, compare the skill tag of each operation and maintenance personnel with the skill requirement of the target task, and calculate the skill matching degree of each operation and maintenance personnel , the skill matching degree is the ratio of the number of intersection elements between the skill tag and the skill requirement to the total number of elements of the skill requirement; retain the operation and maintenance personnel that , and screen out the operation and maintenance personnel that ; where i is the serial number of the operation and maintenance personnel, i=[1, N], and N is the total number of operation and maintenance personnel.

[0078] In this specific embodiment, by introducing a Boolean variable to binarize the on-duty status, the status screening can directly participate in the subsequent mathematical operations. At the same time, the skill matching is upgraded from qualitative judgment to quantitative skill matching degree calculation.

[0079] To describe this processing process more clearly and completely, in a practicable example, the on-duty status is not directly given by a single field, but is written into the scheduling cache after comprehensive judgment of multiple types of service information. Specifically, it can be executed in the following order: first, read the on-duty schedule of the personnel; if the current time is not within the scheduled shift period, the personnel is directly marked as off-duty; if it is within the scheduled shift period, then read the statuses such as leave, business trip, training, and standby for work suspension; if any status that makes the personnel unable to attend work exists, the personnel is marked as off-duty; if none of the above statuses exist, the personnel is marked as on-duty.

[0080] After such processing, the value of has a clear service meaning, which avoids the misjudgment that the attendance shows online but the personnel actually cannot receive the order. At the same time, the skill matching is upgraded from qualitative judgment to quantitative skill matching degree calculation. The calculation of the skill matching degree can be realized based on the coincidence ratio or weight score of the skill tags. For example, when the task requires multiple skills, the system can calculate according to the proportion of the number of skill items possessed by the personnel in the total number of required items; or assign different weights to different skills, with high weight for core skills and low weight for auxiliary skills, so as to calculate the weighted matching degree. For example: calculate the number of intersection elements between the skill tag set and the skill requirement set, divide it by the total number of elements in the skill requirement set, and take the obtained ratio as the value of . For example, if the target task requires two skills of electrician and hydraulics, and the skill tags of a certain operation and maintenance personnel are electrician and fitter, then his skill matching degree This approach allows for the flexible selection of personnel with some auxiliary skills, based on a comprehensive evaluation of subsequent factors such as workload, provided that core skills are met. It is suitable for operation and maintenance scenarios where skills can be partially replaced or can be used collaboratively.

[0081] In another optional embodiment, the skill tag set of the i-th maintenance personnel can be compared with the skill requirement set required by the target task. If the skill tag set completely covers the skill requirement set, then a skill match is determined, and... Otherwise, the skill mismatch is determined, and... This approach ensures that tasks are assigned only to personnel with all the necessary skills, making it suitable for operations and maintenance scenarios where skill thresholds are clearly defined and skill substitutions are not permitted.

[0082] The skill tags can be a set of tags from a standardized dictionary, rather than freely entered text. To ensure stable matching results, companies can pre-build a skill tag dictionary, including standard items such as advanced electrician, basic mechanical repair, hydraulics, PLC debugging, and bearing replacement, and assign a unified code to each tag to prevent semantically similar but textually inconsistent results such as hydraulic repair, hydraulic overhaul, and hydraulics.

[0083] The skill requirements for the target task can be obtained through a fault type-skill requirement mapping table. For example, a hydraulic system leak in a press can be mapped to mechanical repair and hydraulics, while a servo drive alarm can be mapped to advanced electrician and PLC debugging. This mapping table can also be compiled and entered into the system during the system implementation phase by the equipment supervisor, process personnel, and maintenance foreman to form preset rules for sustainable maintenance.

[0084] Furthermore, step S33 includes: S331: Obtain the real-time workload of the tasks already undertaken by the i-th operations and maintenance personnel. And obtain the maximum real-time workload of all maintenance personnel in the maintenance team j to which the i-th maintenance personnel belongs. ;in, , where j represents the sequence number of the maintenance team, j=[1,M], and M is the total number of maintenance teams; S332: According to the formula Calculate the idle matching degree of the i-th maintenance personnel. ; S333: Calculate the total matching degree of the i-th maintenance personnel using the formula. : ;in, For idle matching degree The weight, For skill matching The weights; Status: On duty; Let represent the load coefficient of the j-th maintenance team; where, TA represents the maximum real-time workload for all maintenance teams. The maximum value in, .

[0085] This specific implementation plan introduces the concept of relative idle time, by measuring the individual real-time workload. With the maximum load within this team Compare and calculate the idle matching degree. The real-time workload here This can be understood as the total estimated working hours occupied by a maintenance personnel's currently accepted but not yet completed tasks. It can come from the standard working hours in the work order, or from the remaining working hours dynamically adjusted based on the maintenance progress after the work order is dispatched. One feasible method for obtaining this is: the system first reads all work orders under the personnel's name that are in the pending, processing, or pending review stages, and then adds up the estimated remaining working hours for each work order to obtain the personnel's real-time working load.

[0086] If a work order has been completed or reassigned, it will not be included in the current load. Maximum load within the team. This is obtained by iterating through the real-time workload of all team members and represents the upper bound of the workload of the busiest person in the team. If no one in the team is currently performing any tasks, this value is zero. The calculation logic of idle matching degree can be understood as follows: the lower a person's current workload, the greater the gap with the busiest person in the team, and the higher their idle matching degree; conversely, if a person's workload is close to the highest level in the team, their idle matching degree will decrease significantly. The formula uses a square term to apply a non-linear penalty to high-load personnel, that is, when someone's workload approaches the team's maximum, their idle matching degree will drop sharply, thereby greatly reducing the probability of them being assigned new tasks.

[0087] To further explain, when the load difference between two people is small, the squared term has a more significant suppressive effect on the side with the high load. This is beneficial for prioritizing new tasks to people who are truly less busy, rather than continuing to concentrate them on people who are already close to saturation.

[0088] At the same time, the system also calculated the team load factor. This coefficient reflects the current maximum load on the team. Its relative position within the maximum load TA of all teams in the entire plant. The smaller the value, the busier the team is overall; The higher the score, the more relaxed the team is. In practical terms, when a team is generally under high pressure, even if one member is relatively less busy than others, the system will still reflect the team's overall workload in the total matching score, preventing the system from continuously adding new tasks to the team.

[0089] Final overall match Integrating idle matching degree Skill matching degree On-duty status and team load factor And through weights and To achieve a balance.

[0090] Weights can be configured based on workshop management objectives. For example, for critical equipment maintenance with high downtime losses and high skill thresholds, the importance of skill matching can be appropriately increased; for a large number of repetitive inspections, spot checks, or simple fault handling tasks, the importance of idle time matching can be appropriately increased. Weight parameters can be set by the administrator in the backend and adjusted based on the results of a period of operation. For example, it can be observed whether there are phenomena such as high skill levels but continuous overload, or high idle time but unstable processing quality, and then the weights can be adjusted to make work assignment more realistic.

[0091] Furthermore, step S333 includes the following: S334: When there are no maintenance personnel in the maintenance team whose on-duty status is on duty and whose skill tags match the skill requirements of the target task, calculate the total matching degree of all maintenance personnel in the entire plant and assign the target task to the maintenance personnel with the highest total matching degree. The formula for calculating the overall matching degree of all maintenance personnel across the entire plant is as follows: , where λ represents the cross-team matching coefficient.

[0092] This specific implementation plan establishes a cross-team overflow scheduling mechanism. When the system searches within the dedicated maintenance team to which the target equipment belongs and finds the candidate pool empty, it automatically triggers a plant-wide global search. An empty candidate pool can include the following executable decision scenarios: First, all personnel within the team... Both are 0, meaning there are no orders available; secondly, although there are... Personnel with a skill level of 1, but whose skill matching degree does not meet the preset matching conditions, are still considered unassignable.

[0093] In the global search, the system is no longer limited to a single team, but includes maintenance personnel from all other teams across the entire plant in the candidate pool. To reflect the cost of cross-team scheduling, such as long physical distances and unfamiliarity with remote equipment, a cross-team matching coefficient λ is introduced when calculating the overall matching degree. λ is typically a penalty factor less than 1, used to appropriately reduce the matching score when providing cross-team support, ensuring that cross-regional dispatch is only triggered when support personnel have a clear advantage. This cross-team matching coefficient can be obtained through management configuration, for example, set by the workshop based on plant layout, cross-regional access difficulty, equipment differences, and safety training requirements.

[0094] If two areas are close to each other, have similar equipment types, and personnel frequently support each other, the coefficient can be set relatively high; if the areas are far apart, have significant process differences, and require additional security authorization for cross-area entry, the coefficient can be set lower. Different coefficients can also be maintained based on team-to-team relationships, for example, workshop one to workshop two as one group, and workshop one to workshop three as another group, to more accurately reflect support costs. For ease of implementation, a support team relationship table can be created in the system, recording fields such as source team, target team, cross-team matching coefficient, effective time, and approval status. Administrator modifications will take effect immediately.

[0095] In practice, suppose the dedicated maintenance team in Workshop 1 is fully staffed due to handling an emergency, and then an urgent task requiring advanced hydraulic skills is added to Workshop 1. After failing to find a suitable team, the system initiates a plant-wide matching process. The system finds a maintenance worker (D) in Workshop 2 with advanced hydraulic skills who is currently completely available, and a maintenance worker (E) in Workshop 3 with basic hydraulic skills who is partially available.

[0096] The system sets the cross-team matching coefficient λ=0.8. When calculating the overall matching degree, personnel D has a high availability matching degree and skill matching degree. Even after multiplying by λ=0.8, their overall matching degree is still higher than that of personnel E. Finally, the system cross-teams the task from Workshop 1 to personnel D in Workshop 2, and specifically marks the task as a cross-regional support task in the assignment information, reminding personnel D to check the equipment location and safety regulations in Workshop 1.

[0097] To ensure the smooth execution of this cross-regional support mission, the dispatch message can also include a map of the target equipment's location, safety instructions for entering the area, standard operating procedures for the relevant equipment, and on-site contact information. If the company has cross-regional operation approval requirements, the system can automatically trigger an approval workflow before the mission is officially issued, for example, sending it to the target area's foreman or safety officer for confirmation before pushing it to the support personnel. This retains flexible scheduling capabilities while also taking into account on-site safety and organizational coordination.

[0098] Furthermore, step S4 includes: S41: Calculate the timeout threshold of the target task by multiplying the estimated execution time by the preset timeout multiplier. S42: Monitor the execution progress of the assigned target tasks. The execution progress is the duration of the target task from the time of assignment to the current time. S43: When the duration exceeds the timeout threshold, the target task is automatically reclaimed and steps S2 to S3 are re-executed for scheduling and allocation.

[0099] This specific implementation plan clarifies the quantitative calculation method for the overdue unprocessed threshold and the specific monitoring and recovery actions. In step S41, the system no longer uses a fixed absolute time (such as a uniform requirement to complete within 2 hours) as the evaluation standard, but dynamically generates the threshold based on the estimated execution time that has been individually modified in step S2. Specifically, the system pre-sets a timeout multiplier in the configuration table, which is usually greater than 1 (e.g., set to 1.5 or 2.0) to provide maintenance personnel with reasonable buffer space. The timeout multiplier can also be configured differently according to the priority or complexity of the task. For example, for high-priority urgent tasks, the timeout multiplier can be set to 1.2 to require a tighter execution rhythm; for routine inspection tasks, the timeout multiplier can be set to 2.0. The system multiplies the estimated execution time by the corresponding timeout multiplier to obtain the overdue unprocessed threshold specific to that task.

[0100] In steps S42 and S43, the system background starts a timing service to calculate the duration in real time from the moment the task is successfully assigned to the target maintenance personnel's terminal. This duration is not simply the passage of natural time; in more sophisticated implementation scenarios, the system can also deduct reasonable pause times confirmed by the system, such as waiting for spare parts or downtime permission, to reflect the actual execution delays. When the duration exceeds the calculated timeout threshold and the task has not yet been submitted as completed, the system determines that the task is in an abnormally stalled state. At this point, the system automatically triggers a recycling mechanism, revoking the original maintenance personnel's processing authority for the task, simultaneously deducting their corresponding real-time workload, and re-adding the task to the scheduling pool. Subsequently, the system re-executes steps S2 to S3, reassessing the time consumption and searching for a more suitable replacement from the latest personnel status pool.

[0101] Furthermore, step S5 includes: S51: Compare the real-time workload of each maintenance worker with the preset saturation threshold; S52: When the real-time workload of all maintenance personnel reaches or exceeds the saturation threshold, the system as a whole is determined to be in a state of load saturation. S53: Obtain the priority identifier of the target task and determine whether the target task is a non-urgent task; S54: If the target task is a non-urgent task, then the target task will be delayed in scheduling and scheduling allocation will not be triggered for the time being. S55: If the target task is an urgent task, the manual intervention channel is retained, and the administrator can perform cross-team support or temporary reinforcement.

[0102] This specific implementation plan constructs a system-level protection mechanism based on real load data. In steps S51 and S52, the system continuously monitors the real-time workload of all available maintenance personnel throughout the plant or a specific area. The preset saturation threshold represents the maximum workload that a single employee can bear while ensuring operational safety and quality. When the system finds that the real-time workload of all currently on-duty personnel with the ability to accept orders has reached or exceeded their respective saturation thresholds, the system no longer blindly searches for the relatively least idle personnel, but decisively determines that the current scheduling environment has entered a state of load saturation.

[0103] Once the system reaches saturation, it initiates task routing logic in steps S53 to S55. When a new target task enters the scheduling pool, the system first reads the priority identifier carried by the task. This identifier can be automatically mapped by the alarm level when the task is generated, or manually specified by the reporting personnel. If the task is identified as a non-urgent task (such as routine maintenance, troubleshooting minor abnormal noises in non-critical equipment, etc.), the system performs delayed scheduling and suspends it in the deferred queue. Tasks in the deferred queue will not be pushed to any terminal until the system detects that maintenance personnel have completed their work and released the workload, causing the overall environment to exit the saturation state, and then reactivates scheduling according to priority. Conversely, if the task is identified as an urgent task (such as main production line shutdown failure, leakage with serious safety hazards, etc.), since such tasks cannot be delayed, the system will skip the regular automatic allocation logic and directly trigger a high-level alarm (such as sending an emergency SMS to the workshop supervisor or a pop-up window on the scheduling screen), retaining the manual intervention channel. Once the administrator intervenes, they can, based on the actual situation on site, force cross-team deployment, coordinate external supplier support, or arrange temporary reinforcements for staff on leave through the system backend, so as to concentrate superior resources to prioritize resolving the core crisis.

[0104] By employing the aforementioned technical solutions, in extreme operating conditions with severely limited resources, the system effectively prevents deadlocks caused by continuous task dispatching by temporarily suspending non-urgent tasks. This also avoids safety accidents caused by excessive fatigue among on-site personnel. Meanwhile, the manual intervention channel reserved for urgent tasks ensures that critical faults receive the highest level of resource allocation. This approach significantly enhances the resilience and reliability of the workshop operation and maintenance management system in responding to sudden high-pressure scenarios.

[0105] Please refer to Figure 2This application also proposes a multi-task intelligent scheduling system for workshop equipment operation and maintenance, used to implement the steps of any of the above methods. The system includes: Module 201: Real-time acquisition of the on-duty status, skill tags, and real-time workload of each maintenance personnel's assigned tasks; Determine module 202: When the target task triggers scheduling, determine the baseline working time based on the type of the target task, and correct the baseline working time by combining the historical operation and maintenance personnel's work duration data to determine the estimated execution time of the target task; Matching module 203: Matches target tasks to target maintenance personnel based on on-duty status, skill tags, and real-time workload; Rescheduling module 204: Sets an overtime threshold based on the estimated execution time, monitors the execution progress of the assigned target tasks, and automatically reclaims the target tasks and re-triggers scheduling allocation when the execution progress meets the preset overtime threshold. Delayed processing module 205: When it is detected that the real-time workload of all tasks undertaken by maintenance personnel has reached the preset saturation threshold, delayed scheduling processing is performed on non-urgent target tasks.

[0106] Among them, the acquisition module 201 refers to the real-time acquisition of the on-duty status, skill tags and real-time work hours of each maintenance personnel. Specifically, it can be achieved by using sensor networks, RFID technology, mobile application check-in, or integration with enterprise resource planning (ERP) systems and manufacturing execution systems (MES). These technologies can collect personnel's location, work status, skill certification information and work hours data of the tasks currently being processed in real time.

[0107] The determination module 202 refers to determining the baseline working time based on the type of the target task when the target task is triggered for scheduling. The baseline working time is then corrected by combining the historical operation time data of the operation and maintenance personnel to determine the estimated execution time of the target task. Specifically, a preset working time database can be used, which stores the standard working time for different task types. The baseline working time can be adjusted in a personalized way through historical operation data. For example, it can be corrected based on factors such as the complexity of the task, the equipment model, and the experience level of the operation and maintenance personnel.

[0108] Among them, the matching module 203 refers to matching the target task to the target maintenance personnel based on the on-duty status, skill tags and real-time working hours and load, and weighting the various indicators of the maintenance personnel to select the most suitable personnel to perform the task.

[0109] The rescheduling module 204 refers to setting an overdue threshold based on the estimated execution time, monitoring the execution progress of the assigned target tasks, and automatically reclaiming the target tasks and re-triggering the scheduling allocation when the execution progress meets the preset overdue threshold. Specifically, it can be implemented using a timer mechanism and a task status monitoring system. After the task is assigned, the system starts a timer related to the estimated execution time. Once the timer reaches the set threshold and the task status has not been updated to complete, the reclamation and rescheduling process is automatically triggered.

[0110] The delayed processing module 205 refers to the process of delaying the scheduling of non-urgent target tasks when the real-time workload of all tasks undertaken by maintenance personnel reaches the preset saturation threshold. Specifically, it can be implemented using a system load balancing algorithm. That is, when the system detects that the overall resources are saturated, non-urgent tasks will be placed in a waiting queue and sorted according to preset priority rules or waiting time. They will then re-enter the scheduling process after the system load decreases.

[0111] Through the above technical solution, this application addresses the problems in modern workshop operation and maintenance management, such as the diverse types of equipment maintenance tasks, significant differences in the skill levels, on-duty status, and current load of different maintenance personnel, leading to delayed task scheduling, and even situations where some maintenance personnel are overworked while others are idle. Furthermore, the lack of an automated recovery and rescheduling mechanism when maintenance personnel cannot complete tasks on time results in a backlog of maintenance tasks, severely impacting workshop production continuity and equipment operating efficiency. This system achieves intelligent and efficient management of maintenance tasks by real-time acquisition of maintenance personnel status, intelligent estimation of task working hours, dynamic matching of tasks and personnel, and automated rescheduling and delayed processing mechanisms. This improves task scheduling efficiency, optimizes resource allocation, and ensures timely processing of maintenance tasks, thereby guaranteeing workshop production continuity and equipment operating efficiency.

[0112] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A multi-task intelligent scheduling method for workshop equipment operation and maintenance, characterized in that, The method includes the following steps: S1: Real-time acquisition of the on-duty status, skill tags, and real-time workload of each operation and maintenance personnel's assigned tasks; S2: When the target task triggers scheduling, the baseline working time is determined according to the type of the target task, and the baseline working time is corrected by combining the historical operation and maintenance personnel's work duration data to determine the estimated execution time of the target task; S3: Based on the on-duty status, skill tags, and real-time workload, match the target task to the target maintenance personnel; Step S3 includes: S31: Using the device location information associated with the target task as an index, determine the maintenance team corresponding to the device location information. A maintenance team includes multiple maintenance personnel. S32: Within the maintenance team, select maintenance personnel whose on-duty status is "already on duty" and whose skill tags match the skill requirements of the target task; S33: Calculate the total matching degree between each of the selected maintenance personnel and the target task based on their real-time workload. S34: Assign the target task to the maintenance personnel with the highest total matching degree; Step S33 includes: S331: Obtain the real-time workload of the tasks already undertaken by the i-th operations and maintenance personnel. And obtain the maximum real-time workload of all maintenance personnel in the maintenance team j to which the i-th maintenance personnel belongs. ;in, , where j represents the sequence number of the maintenance team, j=[1,M], and M is the total number of maintenance teams; S332: According to the formula Calculate the idle matching degree of the i-th maintenance personnel. ; S333: Calculate the total matching degree of the i-th maintenance personnel using the formula. : ;in, For idle matching degree The weight, For skill matching The weights; Status: On duty; Let represent the load coefficient of the j-th maintenance team; where, TA represents the maximum real-time workload for all maintenance teams. The maximum value in, ; S4: Set an overtime unprocessed threshold based on the estimated execution time, monitor the execution progress of the assigned target task, and when the execution progress meets the preset overtime unprocessed threshold, automatically reclaim the target task and re-trigger scheduling allocation; S5: When it is detected that the real-time workload of all tasks undertaken by the operation and maintenance personnel has reached the preset saturation threshold, the non-urgent target tasks are subject to delayed scheduling.

2. The multi-task intelligent scheduling method for workshop equipment operation and maintenance according to claim 1, characterized in that, Before step S1, the following are included: S01: Collect workshop operation and maintenance tasks from multiple sources in the workshop and store them in a unified scheduling pool. The target tasks are from the scheduling pool. The workshop operation and maintenance tasks include at least fault repair tasks, equipment maintenance tasks, and equipment modification tasks. In step S1, the on-duty status includes being on duty and not being on duty; The skill tags include at least electrical work, mechanical repair, and hydraulics; The real-time workload is obtained by summing the estimated remaining workload of all tasks that the corresponding maintenance personnel have undertaken but not yet completed.

3. The multi-task intelligent scheduling method for workshop equipment operation and maintenance according to claim 1, characterized in that, Step S2 includes: S21: When the target task triggers scheduling, obtain the type of the target task; S22: Using the task type identifier of the target task as an index, read the benchmark working hours corresponding to the type from the preset benchmark working hour library; S23: Obtain the identity identifier of the maintenance personnel, and use the identity identifier as an index to read the historical operation duration data of the maintenance personnel from the personnel management module; S24: Calculate the individual efficiency coefficient of the maintenance personnel based on the ratio of the historical operation duration data to the sum of the historical operation duration data of all the maintenance personnel. S25: Multiply the baseline working hours by the individual efficiency coefficient to obtain the estimated execution time of the target task.

4. The multi-task intelligent scheduling method for workshop equipment operation and maintenance according to claim 1, characterized in that, Step S32 includes: S321: Based on the on-duty status of each maintenance personnel Use these as filtering criteria to filter out... Maintenance personnel, retain The maintenance personnel, among whom, This indicates that the individual was not at their post. This indicates that the individual is already at their post. S322: Read the skill tags of each operations and maintenance personnel, and compare the skill tags of each operations and maintenance personnel with the skill requirements of the target task as the matching condition, and calculate the skill matching degree of each operations and maintenance personnel. The skill matching degree The ratio of the number of intersection elements between the skill tag and the skill requirement to the total number of skill requirement elements; (This is a partial translation of the original text, which is not directly related to the previous sentence.) Maintenance personnel, screening Maintenance personnel; Where i is the serial number of the maintenance personnel, i=[1,N], and N is the total number of maintenance personnel.

5. The multi-task intelligent scheduling method for workshop equipment operation and maintenance according to claim 1, characterized in that, Step S333 is followed by: S334: When there is no maintenance personnel in the maintenance team whose on-duty status is "on duty" and whose skill tags match the skill requirements of the target task, calculate the total matching degree of all maintenance personnel in the entire plant and assign the target task to the maintenance personnel with the highest total matching degree. The formula for calculating the overall matching degree of all maintenance personnel across the entire plant is as follows: , where λ represents the cross-team matching coefficient.

6. The multi-task intelligent scheduling method for workshop equipment operation and maintenance according to claim 1, characterized in that, Step S4 includes: S41: Calculate the timeout threshold of the target task by multiplying the estimated execution time by a preset timeout multiplier. S42: Monitor the execution progress of the assigned target task, wherein the execution progress is the duration of the target task from the time of assignment to the current time; S43: When the duration exceeds the timeout threshold, the target task is automatically reclaimed and steps S2 to S3 are re-executed for scheduling and allocation.

7. A multi-task intelligent scheduling method for workshop equipment operation and maintenance according to claim 1, characterized in that, Step S5 includes: S51: Compare the real-time workload of each maintenance worker with the preset saturation threshold; S52: When the real-time working hours of all maintenance personnel reach or exceed the saturation threshold, the system as a whole is determined to be in a state of load saturation. S53: Obtain the priority identifier of the target task and determine whether the target task is a non-urgent task; S54: If the target task is a non-urgent task, then the target task will be subject to delayed scheduling and scheduling allocation will not be triggered for the time being. S55: If the target task is an urgent task, a manual intervention channel is reserved, and the administrator can perform cross-team support or temporary reinforcement.

8. A multi-task intelligent scheduling system for workshop equipment operation and maintenance, characterized in that, The system, used to implement the method according to any one of claims 1-7, comprises: Acquisition module: Acquires real-time on-duty status, skill tags, and real-time workload of each operations and maintenance personnel's assigned tasks; Determining Module: When a target task triggers scheduling, a baseline working time is determined based on the type of the target task. The baseline working time is then corrected by combining the historical work duration data of the operation and maintenance personnel to determine the estimated execution time of the target task. Matching module: Based on the on-duty status, skill tags, and real-time workload, the target task is matched to the target maintenance personnel; The matching module is also used to determine the maintenance team corresponding to the equipment location information using the equipment location information associated with the target task as an index, wherein a maintenance team includes multiple maintenance personnel. Within the maintenance team, select maintenance personnel whose on-duty status is "already on duty" and whose skill tags match the skill requirements of the target task; Calculate the total matching degree between each of the selected maintenance personnel and the target task based on their real-time workload. The target task is assigned to the maintenance personnel with the highest total matching degree; The matching module is also used to obtain the real-time workload of the tasks already undertaken by the i-th maintenance personnel. And obtain the maximum real-time workload of all maintenance personnel in the maintenance team j to which the i-th maintenance personnel belongs. ;in, , where j represents the sequence number of the maintenance team, j=[1,M], and M is the total number of maintenance teams; According to the formula Calculate the idle matching degree of the i-th maintenance personnel. ; The total matching degree of the i-th maintenance personnel is calculated using the formula. : ;in, For idle matching degree The weight, For skill matching The weights; Status: On duty; Let represent the load coefficient of the j-th maintenance team; where, TA represents the maximum real-time workload for all maintenance teams. The maximum value in, ; Rescheduling module: Sets an overtime threshold based on the estimated execution time, monitors the execution progress of the assigned target task, and automatically reclaims the target task and re-triggers scheduling allocation when the execution progress meets the preset overtime threshold; Delayed processing module: When it is detected that the real-time workload of all tasks undertaken by the operation and maintenance personnel has reached the preset saturation threshold, the non-urgent target tasks are subject to delayed scheduling processing.

Images