Intelligent manufacturing system scheduling method driven by artificial intelligence

Abstract

The invention relates to the technical field of scheduling, in particular to an artificial intelligence-driven intelligent manufacturing system scheduling method, which comprises the following steps of monitoring order state parameters of a manufacturing execution system in real time, identifying equipment failure shutdown signals or order insertion signals, extracting associated equipment codes and current work order numbers, and establishing core disturbance information pairs. According to the method, the order state parameters of the manufacturing execution system are monitored in real time, the equipment fault shutdown signal and the order insertion signal are dynamically identified, and the equipment code and the current work order number are extracted, so that the core disturbance information pair highly associated with the disturbance source can be constructed, and the pertinence and accuracy of the scheduling process response are ensured. And in combination with the original plan completion time of the affected process, evaluating local delay caused by disturbance, gradually deducing to form the delay amount of the whole plan, describing the diffusion trend of the disturbance in the production progress, and defining a scheduling adjustment time window.

Description

Technical Field

[0001] The present invention relates to the technical field of scheduling, and in particular to a scheduling method for an intelligent manufacturing system driven by artificial intelligence. Background Art

[0002] The technical field of scheduling is a key component in the intelligent manufacturing and industrial automation systems, mainly focusing on how to reasonably arrange and optimize the job sequence, resource allocation, and execution timing in the production process under limited time, resource, and task constraints.

[0003] The existing technologies mainly rely on preset job sequences and static resource allocation logics to execute scheduling tasks. In scenarios with frequent disturbance events or complex task structures, due to the lack of in-depth analysis of real-time data and dynamic understanding of resource status, the scheduling response is often lagged. The sudden equipment failure signals are often confused with ordinary process pause states, and the scheduling intervention after order insertion lacks accuracy, resulting in repeated scheduling of some processes or aggravated resource conflicts. At the same time, the traditional scheduling process often ignores the impact of human resource status on scheduling stability, and the continuous operation load of operators is not reasonably dispersed, thus leading to problems such as process errors and output quality fluctuations. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of the present invention is to solve the drawbacks existing in the prior art, and to propose a scheduling method for an intelligent manufacturing system driven by artificial intelligence.

[0005] To achieve the above purpose, the present invention adopts the following technical solution, a scheduling method for an intelligent manufacturing system driven by artificial intelligence, including the following steps: Real-time monitor the order status parameters of the manufacturing execution system, identify equipment failure shutdown signals or order insertion signals, extract the associated equipment codes and the current work order numbers, and establish a pair of core disturbance information; Based on the pair of core disturbance information, evaluate the estimated completion time of the affected processes in the current production activities of the intelligent manufacturing system by the disturbance, obtain the estimated delay value of a single process, and based on the estimated delay value of the single process, combine with the original planned completion time of all affected processes, calculate the delay amount of the overall plan, and determine the scheduling adjustment time window; According to the scheduling adjustment time window and the resource changes reflected by the pair of core disturbance information, retrieve the processing capabilities of alternative equipment and the material supply cycles associated with the affected processes, obtain a set of available alternative resource parameters, and based on the set of available alternative resource parameters, select a resource combination for replacement according to the production efficiency index to form a preliminary scheduling operation instruction for coping with the disturbance; For the preliminary scheduling work instructions for coping with the disturbance, the continuous working time and skill proficiency of the target process operators are obtained, and the personnel's real-time status evaluation value is obtained. Based on the personnel's real-time status evaluation value, personnel allocation optimization is performed to generate a personnel scheduling execution plan.

[0006] Preferably, the steps of obtaining the core disturbance information pair are: Collect order status parameters in real time, analyze and judge the equipment failure shutdown signal and order insertion signal in the order status parameters respectively, and extract the equipment failure shutdown signal and order insertion signal; Based on the equipment failure shutdown signal and the order insertion signal, the corresponding equipment code and the current work order number are retrieved, and the association relationship between the equipment failure shutdown signal and the equipment code, and the association relationship between the order insertion signal and the current work order number are respectively established to form association mapping information; Based on the association mapping information, the device code and the current work order number are combined and matched, and the code data and the work order data are bound in the form of a key-value pair to form a core disturbance information pair.

[0007] Preferably, the steps for obtaining the estimated delay value of a single process are: Based on the core disturbance information pair, locate the affected process number in the current production activity, extract the planned completion time, unit standard working hours, resource start time and resource recovery time of each affected process, and form a set of production scheduling parameters for the affected processes; According to the affected process scheduling parameter set, the difference between the resource recovery available time and the actual resource start time is calculated to obtain the resource available delay length, and the affected process progress offset data group is generated; Based on the affected process progress deviation data group, a single process estimated delay value is calculated.

[0008] Preferably, the steps of obtaining the scheduling adjustment time window are: Based on the estimated delay value of the single process, the delay values of all affected processes are aggregated in the order of process numbers, and each delay value is matched with the original planned completion time of the corresponding process one by one to generate a comparison table of affected process plans; According to the affected process plan comparison table, the affected processes are sorted according to their original planned completion time, the delay value corresponding to the process with the latest planned completion time is extracted, and the delay value is added to the original planned completion time to obtain the maximum delay time point of the overall plan; Based on the maximum delay time point of the overall plan, the time difference between the end time of the current global scheduling cycle of the intelligent manufacturing system and the maximum delay time point of the overall plan is compared to determine whether the scheduling adjustment demand duration exceeds the acceptable range and obtain the scheduling adjustment time window.

[0009] Preferably, the step of obtaining the available alternative resource parameter set is as follows: Based on the scheduling adjustment time window and the equipment code and current work order number identified in the core perturbation information pair, query the associated equipment resource database and order resource mapping table in the intelligent manufacturing system, extract the equipment configuration records and original material call records associated with the affected processes, and generate an affected process resource binding information set; According to the affected process resource binding information set, retrieve the equipment items in the equipment capacity library that have the same process characteristics as the bound equipment, screen the equipment processing capacity parameters with available idle periods within the scheduling adjustment time window, and simultaneously retrieve the supply time intervals of the corresponding materials in the material supply schedule to generate a list of candidate resource capabilities and material supply cycle combinations; Based on the list of candidate resource capabilities and material supply cycle combinations, identify the combination items with processing capacity equivalence, time cycle matching, and scheduling compatibility, and screen the equipment processing capacity and material supply cycle that meet the replaceable conditions within the time window to obtain the available alternative resource parameter set.

[0010] Preferably, the step of obtaining the preliminary scheduling operation instruction for coping with perturbations is as follows: Based on the available alternative resource parameter set, extract the unit processing output, unit operation time, scheduling adjustment time window length, and material supply cycle duration corresponding to each resource combination, construct a standardized resource combination index item, and form a resource combination performance index set; According to the resource combination performance index set, calculate the efficiency matching value of the resource combination; Based on the efficiency matching value of the resource combination, sort the resource combinations from high to low according to the efficiency matching value of the resource combination, and select the resource combinations with an efficiency matching value greater than the average production efficiency threshold to form the preliminary scheduling operation instruction for coping with perturbations.

[0011] Preferably, the step of obtaining the personnel instant status evaluation value is as follows: Based on the preliminary scheduling operation instruction for coping with perturbations, extract the target process information corresponding to each scheduling operation instruction, match the operator number bound to the process, query the operator task record database, and obtain the continuous operation time period of each operator in the current cycle to generate a target process operator operation duration information set; According to the target process operator operation duration information set, combined with the operator training record file and performance evaluation form, extract the skill level, completion efficiency, and error rate of each operator under the corresponding process category to generate the personnel instant status evaluation value.

[0012] Preferably, the step of obtaining the personnel scheduling execution plan is as follows: Based on the instant status evaluation value of the personnel, extract the skill proficiency level, total continuous operation duration, and current assigned task quantity of each operator to form a personnel operation ability status set; Calculate the personnel fitness score value according to the personnel operation ability status set; Based on the personnel fitness score value, sort all candidate operators in descending order, and sequentially select the operator with the highest personnel fitness score value in the order of process matching priority for scheduling and allocation to generate a personnel scheduling execution plan.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are as follows: By real-time monitoring the order status parameters of the manufacturing execution system, dynamically identifying the equipment failure shutdown signal and order insertion signal, and extracting the equipment code and current work order number, the present invention can construct a core disturbance information pair highly correlated with the disturbance source, ensuring the pertinence and accuracy of the response in the scheduling process. Combining with the original planned completion time of the affected process, evaluate the local delay caused by the disturbance, and gradually deduce the delay amount of the overall plan to depict the diffusion trend of the disturbance in the production schedule and define the scheduling adjustment time window. Around the resource change characteristics and the scheduling adjustment time window, retrieve the processing capabilities of alternative equipment and the material supply cycle, screen and form a set of available alternative resource parameters, and quantitatively select through the efficiency matching value to improve the rationality of resource replacement and the operability of scheduling execution. At the personnel allocation level, combine the continuous operation duration and skill proficiency of the operators of the target process, calculate the instant status evaluation value of the personnel, and then perform personnel allocation optimization operations accordingly, making the scheduling result more in line with the actual state of human resources and avoiding the increase in error rate caused by high-intensity operations. In summary, it realizes the rapid response to abnormal states, the scientific screening of resource utilization, and the adaptation of personnel scheduling, thereby improving the recovery efficiency, resource coordination, and task fulfillment rate of the production system in the face of disturbances. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a step schematic diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0016] Please refer to Figure 1 , the present invention provides a technical solution, an artificial intelligence-driven intelligent manufacturing system scheduling method, including the following steps: Real-time monitor the order status parameters of the manufacturing execution system, identify equipment failure shutdown signals or order insertion signals, extract the associated equipment codes and current work order numbers, and establish core disturbance information pairs; Based on the core disturbance information pairs, evaluate the estimated completion time of the affected processes in the current production activities of the intelligent manufacturing system, obtain the estimated delay value of a single process, and based on the estimated delay value of a single process, combine with the original planned completion time of all affected processes to calculate the delay of the overall plan and determine the scheduling adjustment time window; According to the scheduling adjustment time window and the resource changes reflected by the core disturbance information pairs, retrieve the processing capabilities of alternative equipment and the material supply cycle associated with the affected processes, obtain the available alternative resource parameter set, and based on the available alternative resource parameter set, select a resource combination for replacement in accordance with the production efficiency indicators to form a preliminary scheduling operation instruction for coping with disturbances; For the preliminary scheduling operation instruction for coping with disturbances, obtain the continuous operation duration and skill proficiency of the operators of the target process to obtain the personnel immediate status evaluation value, and based on the personnel immediate status evaluation value, optimize the personnel allocation to generate a personnel scheduling execution plan.

[0017] The steps for obtaining the core disturbance information pairs are as follows: Real-time collect the order status parameters, respectively analyze and judge the equipment failure shutdown signals and order insertion signals in the order status parameters, and extract the equipment failure shutdown signals and order insertion signals; Based on the equipment failure shutdown signals and order insertion signals, retrieve the corresponding equipment codes and current work order numbers, and respectively establish the association relationship between the equipment failure shutdown signals and the equipment codes, and the association relationship between the order insertion signals and the current work order numbers to form association mapping information; Based on the association mapping information, combine and match the equipment codes and the current work order numbers, and bind the coded data and work order data in the form of key-value pairs to form core disturbance information pairs.

[0018] Specifically, the order status parameters are collected in real time. The specific operation is to continuously receive the original information flow containing timestamps, event sources, event types, and detailed data from various monitoring points or data interfaces of the Manufacturing Execution System (MES), such as device controller logs, database table change streams of the order management module, and manual input terminals. First, these original information flows are structurally parsed. For example, if the data is in JSON format, the values of fields with key names such as "eventType", "sourceID", "payload", etc. are extracted. If it is PLC raw data, it is converted according to the preset byte offset and data type rules. Subsequently, for the device failure shutdown signal in the parsed order status parameters, this judgment is based on a preset "device failure feature library". This library is jointly compiled by equipment maintenance engineers based on equipment manuals, historical failure records, and daily operation and maintenance experience, and is audited and updated at least once every quarter. It details the specific failure codes, status indicators, or abnormal operation parameter patterns of various devices. For example, for a Model A machining center, when "Error_Code: 503" or "Operational_Status: Emergency_Halt" appears in its status parameters, or its spindle load exceeds 150% of the rated load continuously for 10 seconds (this 150% threshold is obtained by statistically analyzing the operation data of this type of device in the 5 minutes before a typical failure, taking the sum of 1.5 times the historical average overload peak and one standard deviation. For example, if the historical average overload peak is 130% and the standard deviation is 10%, then the threshold is , here is an example. The actual setting of 150% is based on the balance consideration of safety redundancy and early warning, and the specific value is 150%), then it is determined as a device failure shutdown signal. For the judgment of the order insertion signal, it is based on an "order change rule set". This rule set is defined by the production planning department at the initial stage of system deployment according to order processing priorities and emergency response processes. For example, when the order status parameters indicate that a new order is created and its priority field is marked as "urgent" or "VIP", or the planned start time of an existing order is advanced by more than 24 hours (this 24-hour threshold is an empirical value set by the production planning department based on the average order processing cycle and the material preparation lead time to ensure there is enough time for resource coordination), then it is determined as an order insertion signal. After the above judgments are completed, the system classifies and aggregates the specific device failure shutdown signal instances (such as "Spindle overload failure occurred on device number CNC001") and order insertion signal instances (such as "Order number PO20250523001 is marked for emergency insertion") that are identified, to obtain the extracted device failure shutdown signals and order insertion signals.

[0019] Based on the equipment failure shutdown signal and order insertion signal extracted in the previous step, the system performs subsequent retrieval and association operations. For each identified equipment failure shutdown signal, such as the signal content being "Spindle overload failure occurred on equipment number CNC001 at 10:30:15 on May 23, 2025", the system will parse out the key equipment identification information from this signal, which is "CNC001". Then, using "CNC001" as the query key, it conducts a search in the pre-established and maintained "Equipment Basic Information Table", which details the static attributes of all equipment in the workshop, including equipment unique coding, model, location, affiliated production line, etc. For example, if the search result confirms that the standard equipment coding corresponding to "CNC001" is "EQP-MC-001", then a direct association relationship is established between this equipment failure shutdown signal and the equipment coding "EQP-MC-001", and it is recorded as paired information such as (Failure Signal ID: FS001, Equipment Coding: "EQP-MC-001"). For each identified order insertion signal, such as the signal content being "Order number PO20250523001 is marked as urgently inserted, product code PD007", the system also parses out the core order identification "PO20250523001" from it, and this identification is regarded as the current work order number. Subsequently, a direct association is established between this order insertion signal and the current work order number "PO20250523001", and it is recorded as paired information such as (Insertion Signal ID: IS001, Current Work Order Number: "PO20250523001"). These separately established association relationships are aggregated and stored to form a temporary set of associated mapping information. Each item in this set clearly indicates a specific perturbation signal and its directly corresponding equipment coding or current work order number, providing structured input for subsequent perturbation impact analysis and obtaining the associated mapping information.

[0020] Based on the set of associated mapping information formed in the previous stage, which includes the pairing of equipment failure shutdown signals and equipment codes, as well as the pairing of order insertion signals and current work order numbers, the system further performs a combined matching operation, aiming to clarify the impact on the equipment dimension and the impact on the order dimension for each type of perturbation. Specifically, if the entry in the associated mapping information is an equipment failure type, for example, (failure signal ID: FS001, equipment code: "EQP-MC-001"), the system will query the "Work Order Execution Status Table" in the real-time production database of the Manufacturing Execution System (MES) based on "EQP-MC-001" and the timestamp of the failure occurrence (associated by the failure signal ID: FS001). This table records the work order information currently being processed or most recently processed by each equipment. The query logic is to filter out the work order numbers with the status of "In Execution" or "Allocated but not started" on the equipment "EQP-MC-001" at the time of the failure. If there are multiple work order numbers, the most directly affected one is selected according to the work order priority and the planned start time. For example, if the current work order number obtained from the query is "WO202505007", then "EQP-MC-001" and "WO202505007" are bound. If the entry in the associated mapping information is an order insertion type, for example, (insertion signal ID: IS001, current work order number: "PO20250523001"), the system needs to match an initial target equipment code or a set of target equipment codes for this newly inserted work order number "PO20250523001". This matching process first consults the "Master Data Table of Product Process Routes", which is maintained by process engineers and defines the standard processing procedures for each product (the product code is obtained by associative query based on the current work order number) and the eligible equipment types and optional equipment lists for each procedure. For example, the product of the work order "PO20250523001" needs to go through the "milling" process, and the eligible equipment type is the "DX-500 CNC Milling Machine", and the optional equipment includes "EQP-ML-003" and "EQP-ML-004". The system then combines the "Equipment Real-Time Status Table" to query the current load, maintenance plan, and estimated idle time of these optional equipment, and selects an optimal equipment code. For example, if "EQP-ML-003", which is expected to be available earliest, is selected, then "EQP-ML-003" and "PO20250523001" are bound. After completing all combined matching operations, the system finally binds these pairings consisting of equipment codes and current work order numbers in the form of key-value pairs, such as {"equipment code": "EQP-MC-001", "current work order number": "WO202505007", "perturbation type": "equipment failure"} or {"equipment code": "EQP-ML-003", "current work order number": "PO20250523001", "perturbation type": "order insertion"}, thus forming the core perturbation information pairs.

[0021] The steps to obtain the estimated delay value of a single process are as follows: Based on the core perturbation information pair, locate the numbers of the affected processes in the current production activities, extract the planned completion time, unit standard working hours, resource start time, and resource recovery available time of each affected process, and form a set of production scheduling parameters for the affected processes; According to the set of production scheduling parameters for the affected processes, calculate the difference between the resource recovery available time and the actual resource start time to obtain the resource available delay length, and generate a set of progress deviation data for the affected processes; Based on the set of progress deviation data for the affected processes, calculate the estimated delay value of a single process. The calculation formula is: ; Wherein, is the estimated delay value of a single process for the c-th affected process, is the estimated completion time of the c-th affected process, is the originally planned completion time of the c-th affected process, is the unit standard working hours of the c-th affected process, is the resource available delay length of the c-th affected process, is the minimum time constant to prevent the denominator from being zero.

[0022] Specifically, based on the core perturbation information pair obtained in the foregoing steps, the system first analyzes the information pair to determine whether the perturbation source is a device failure or an order insertion, as well as the associated device code and the current work order number, and then locates the specific process numbers directly or indirectly affected by the perturbation in the current overall production plan. For example, if the core perturbation information pair indicates that the device "EQP-001" has failed and the process "OP-123B" of the work order "WO-123" is being processed at that time, then "OP-123B" is the primary affected process. At the same time, the corresponding processes of all other work orders that originally planned to use this device before the repair of "EQP-001" (such as the process "OP-124A" of the work order "WO-124") and all subsequent processes of the work order "WO-123" on "EQP-001" are identified as affected processes. If the core perturbation information pair indicates the insertion of a new work order "WO-789" into the device "EQP-002", then all processes that originally planned to use "EQP-002" during and after the time period occupied by "WO-789" (such as the process "OP-456C" of the work order "WO-456") are identified as affected processes. After completing the identification of all affected process numbers, the system extracts the original planned completion time of each identified affected process number from the production plan database and the master process path data. This time is the precise time point when the process is expected to end in the absence of perturbations. The unit standard man-hour is extracted. This man-hour is the standard operation time required to complete one unit of the product for this process, which is preset according to historical data statistics or man-hour quota standards. The original resource start time of this process in the original plan is extracted, as well as a key resource recovery available time. For device failure perturbations, this resource recovery available time refers to the time point when the failed device is expected to be able to be put back into production. This time point is usually estimated by the maintenance department based on the fault diagnosis results and the maintenance resource situation. For example, by consulting the device maintenance record system, it is known that the estimated repair time of the device "EQP-001" is 14:00 on May 26, 2025. For order insertion perturbations, the resource recovery available time refers to the time point when the originally occupied resource can actually start providing services to the subsequent affected processes due to the processing of the inserted order. For example, if the inserted work order "WO-789" is expected to take 4 hours on the device "EQP-002" and its start time squeezes the planned start time of the original process "OP-456C", then for "OP-456C", the resource recovery available time of the device "EQP-002" is the expected completion time of "WO-789" on "EQP-002". Through the above extraction, the system integrates the process number, original planned completion time, unit standard man-hour, original resource start time, and the calculated or obtained resource recovery available time of each affected process to form a set of scheduling parameters for the affected processes.

[0023] Based on the set of production scheduling parameters for affected processes formed in the previous step, which includes the original planned start time of resource usage and the updated resource recovery available time for each affected process, the system then calculates the resource availability delay length for each affected process. This calculation is achieved by comparing the resource recovery available time of the corresponding process with the actual resource start time of the original plan for that process. The specific calculation method is to subtract the actual resource start time of the process in the original plan from the resource recovery available time of a specific affected process. The resulting difference is preliminarily determined as the delay in resource availability. For example, in the original production plan of an affected process "OP-X", the planned start time of the required resource "RES-Y" is 08:00 on May 26, 2025. Due to a disturbance event (such as a failure of "RES-Y" or handling an emergency inserted order), the resource recovery available time of "RES-Y" is updated to 11:30 on May 26, 2025. Then the time difference between the two is 3 hours and 30 minutes. Considering that the resource recovery available time may be earlier than or equal to the resource start time of the original plan (such as the failure being repaired in advance), a negative delay length should not be generated at this time. Therefore, the final resource availability delay length takes the larger value between this calculated difference and zero. That is, if the calculated difference is negative or zero, the resource availability delay length of this process is recorded as zero, indicating that there is no substantial delay in resource availability. If the difference is positive, this positive value is the resource availability delay length of this process. For example, for the above process "OP-X", its resource availability delay length is 3 hours and 30 minutes. The system traverses each process in the set of production scheduling parameters for affected processes, performs this calculation for each process, and pairs and records each process number with its corresponding resource availability delay length to generate a data set of progress offsets for affected processes.

[0024] Formula: , the benefit of the formula is that it comprehensively evaluates the expected delay degree of a single process after a disturbance. It does not simply look at the absolute delay duration, but relativizes and normalizes both the deviation degree of the planned completion time and the delay degree of resource availability, and combines the two parts of the influence in a way similar to the Euclidean distance, so as to more accurately reflect the relative severity of the impact of disturbances on processes with different natures and different calibrated working hours, enabling managers to identify the bottleneck processes that are most "destructive" to the overall plan, so as to prioritize resource allocation or take countermeasures. The first squared term focuses on the deviation of the execution efficiency of the process itself from the plan, and the second squared term focuses on the restriction of the upstream resource supply on the process execution. The combination of the two can comprehensively describe the source and degree of the delay; Parameter The obtaining steps are as follows. It represents the estimated completion time of the c-th affected process. This time is not directly extracted but recalculated based on the disturbance impact. First, obtain the original planned start time of the c-th affected process and the total standard man-hours (if the unit standard man-hours refer to per piece, it needs to be multiplied by the number of work orders). At the same time, obtain the available resource delay length of this process from the previous step , then the actual available start time of the resources for this process is the original planned resource start time plus , the actual earliest start time of the process , and the final estimated completion time . All times need to be converted to a unified unit. For example, the original planned start time of process OP001 is the 100th hour of the production cycle, its total standard man-hours is 8 hours, its available resource delay length is 4 hours, and its original planned resource start time is also the 100th hour. Then the actual available start time of the resources is hours, the actual earliest start time hours. Therefore, the estimated completion time hours; The obtaining steps of parameter are as follows. It represents the original planned completion time of the c-th affected process. This parameter is directly extracted from the original production plan data stored in the Manufacturing Execution System (MES), corresponding to the end time point estimated for process c according to the initial schedule when no disturbance occurs. For example, for process OP001, its planned completion time recorded in the original production plan is the 108th hour of the production cycle, then hours; The obtaining steps of parameter are as follows. It represents the unit standard man-hours of the c-th affected process, or more precisely, the total standard man-hours required to complete the task of this process. This data is sourced from the enterprise's established process database or labor quota standard table, which details the standard time required for each process to process a specific product or component on a specific device. This standard time is usually set and regularly maintained by the industrial engineering department through methods such as time study and historical data averaging. For example, the total standard processing time defined for process OP001 in the process database is 8 hours (this value has considered the product of the processing batch of this process in this work order and the unit standard man-hours per piece, or is directly defined as the standard total time-consuming of this task package), then hours; The obtaining steps of parameter are as follows. It represents the available resource delay length of the c-th affected process. This parameter is calculated from the previous step "Generate the progress offset data group of affected processes", that is , which quantifies the delay in the time when the resources originally planned to be allocated to process c can actually start service due to reasons such as equipment failure repair or priority processing of inserted orders. For example, for process OP001, the resource availability delay length calculated in the previous step is 4 hours, then Hour; parameter The steps to obtain is, which represents a minimum time constant to prevent the denominator from being zero, and select Hour.

[0025] Calculation process: Taking process OP001 as an example, substitute the obtained parameter values for calculation: Hour, Hour, Hour, Hour, Hours. Substituting the parameters into the formula, we get 0.5013.

[0026] The result shows that the single process delay value of process OP001 is expected to be It is approximately 0.5013, which is a dimensionless relative delay indicator. This value comprehensively reflects the degree of deviation of the planned completion time of process OP001 compared with its own standard working hours, as well as the proportion of its resource availability delay relative to its new estimated completion time. The size of this value can be used to compare the delay severity of different affected processes horizontally.

[0027] The steps to obtain the scheduling adjustment time window are: Based on the estimated delay value of a single process, the delay values of all affected processes are aggregated in the order of process numbers, and each delay value is matched with the original planned completion time of the corresponding process one by one to generate a comparison table of affected process plans; According to the affected process plan comparison table, the affected processes are sorted according to their original planned completion time, the delay value corresponding to the process with the latest planned completion time is extracted, and the delay value is added to the original planned completion time to obtain the maximum delay time point of the overall plan; Based on the maximum delay time point of the overall plan, the time difference between the end time of the current global scheduling cycle of the intelligent manufacturing system and the maximum delay time point of the overall plan is compared to determine whether the scheduling adjustment demand duration exceeds the acceptable range and obtain the scheduling adjustment time window.

[0028] Specifically, the estimated delay value of each affected process calculated in the previous step , and in the calculation The original planned completion time of each step involved in the process and estimated completion time The system first collects this information in a structured manner. Specifically, the system traverses all processes identified as affected by the previous steps and creates a record for each affected process. The record contains at least the unique number of the process, the estimated delay value of the single process calculated for the process, and the , the planned completion time of the process set in the original production plan , and the estimated completion time after reassessment due to the disturbance , then the system preliminarily sorts or sorts these records according to the numbers of each process (for example, according to the alphanumeric order of the process numbers or their natural order in the process route) to ensure that all affected processes are included in the processing scope without omission. On this basis, the core operation is to calculate the single process estimated delay value for each process. The original planned completion time and estimated completion time Strict one-to-one matching is performed to ensure the consistency and accuracy of the data. For example, if the estimated delay value of a single process "OPN-102B" is calculated to be 0.45, and its original planned completion time is 16:00 on the 5th day, and the estimated completion time is 18:00 on the 5th day, then this information is accurately associated with the record "OPN-102B". This process does not involve new threshold judgments or complex calculations, but mainly involves data organization and verification. After completing the data matching of all affected processes, a structured data list or table is formed, namely the affected process plan comparison table.

[0029] According to the affected process plan comparison table generated in the previous stage, the table contains the process number and original planned completion time of each affected process. , and a key parameter, the estimated completion time of the process The system first compares all the items in the affected process plan comparison table according to the original planned completion time of each process. Sort in ascending order. After sorting, the last record in the table corresponds to the affected process with the latest planned completion time in the original plan. The system then extracts its original planned completion time from this record. and estimated completion time Then calculate the actual delay time of the process with the latest original planned completion time, and calculate the actual delay time as the estimated completion time of the process Subtract its original planned completion time ,get , and then the actual delay time calculated is Added back to the original planned completion time for the operation The specific operation is , the result is actually the estimated completion time of the most delayed affected process in this original plan after considering the disturbance. , for example, after sorting, if the process with the latest completion time in the original plan is "OPN-205C", its original planned completion time is 17:00 on the 8th day, and its estimated completion time is 20:00 on the 8th day, then its delay duration is 3 hours. Add these 3 hours to 17:00 on the 8th day, and the result is 20:00 on the 8th day. This time point is determined as the maximum delay time point of the overall plan.

[0030] Based on the maximum delay time point of the overall plan obtained in the previous step, for example, determined to be the 120th hour of the current production cycle, the system will then compare this time point with the end time of the preset current global scheduling cycle in the intelligent manufacturing system. The end time of this global scheduling cycle, for example, the 128th hour of the current cycle, is a fixed time point set by the production planning department based on the overall production order delivery commitment, resource availability, and reserved regular buffer time (for example, the planned production per week is 40 hours, but the end time of the global scheduling cycle may be set to the 48th hour, including 8 hours of fixed buffer). The time difference between the two is calculated by subtracting the maximum delay time point of the overall plan from the end time of the global scheduling cycle, that is hours. This difference represents the remaining surplus or the exceeded amount of the entire plan from exceeding the end time of the global scheduling cycle after the disturbance occurs. Subsequently, the system determines whether the required duration of the scheduling adjustment exceeds the acceptable range. The "acceptable range" here is not a single fixed value but is reflected by comparing the compliance of the adjusted plan with the global scheduling cycle. The specific judgment logic is: if the calculated time difference is positive or zero (such as the above 8 hours), it means that the maximum delay time point of the overall plan is still within the end time of the global scheduling cycle, and at this time, the scheduling adjustment requirement is time - acceptable. If the time difference is negative (for example, if the maximum delay time point of the overall plan is the 130th hour, then the difference is hours), it indicates that the overall plan has exceeded the preset end time of the global scheduling cycle, and the exceeded duration is 2 hours. At this time, the required duration of the scheduling adjustment has exceeded the acceptable zero - surplus range. Further, the system will refer to a "maximum allowable exceedance threshold". For example, the production department stipulates that the global scheduling cycle is allowed to exceed by a maximum of 2%. If the current cycle is 128 hours, then the allowed exceedance is For hours, if the calculated excess duration above is less than 2.56 hours and more than 2 hours, it is considered still within the "urgent tolerance range"; otherwise, it is completely outside the tolerance range. Based on this judgment result and the actual surplus or excess amount, the system finally determines and outputs a scheduling adjustment time window, which usually refers to the time period from the current moment to the maximum delay time point (the 120th hour) of the overall plan, along with a status flag indicating whether it is within the tolerance range, within the urgent tolerance range, or outside the range, thus obtaining the scheduling adjustment time window.

[0031] The steps for obtaining the available alternative resource parameter set are as follows: Based on the scheduling adjustment time window, the equipment codes and the current work order numbers identified in the core perturbation information pair, query the associated equipment resource database and the order resource mapping table in the intelligent manufacturing system, extract the equipment configuration records and the original material call records associated with the affected processes, and generate the affected process resource binding information set; According to the affected process resource binding information set, retrieve the equipment items in the equipment capacity library that have the same process characteristics as the bound equipment, screen the equipment processing capacity parameters with available idle periods within the scheduling adjustment time window, and simultaneously retrieve the supply time intervals of the corresponding materials in the material supply plan to generate a candidate resource capacity and material supply cycle combination list; Based on the candidate resource capacity and material supply cycle combination list, identify the combination items with processing capacity equivalence, time cycle matching, and scheduling compatibility, and screen the equipment processing capacity and material supply cycle that meet the replaceable conditions within the time window to obtain the available alternative resource parameter set.

[0032] Specifically, based on the scheduling adjustment time window determined in the previous step and the key equipment codes and current work order numbers identified in the core disturbance information obtained in an earlier step, the system first analyzes the core disturbance information pair to clarify the specific disturbance type (e.g., failure of equipment "EQP-X" or insertion of new order "WO-Y"), and then re-identifies all the process numbers directly or indirectly affected by this disturbance. For each identified affected process, the system sends a query request to the central database of the intelligent manufacturing system, specifically accessing the "equipment resource database" to obtain the detailed configuration records of the equipment originally planned to be bound to this process (e.g., equipment "EQP-X"). These records include the model of the equipment, the information of the specific tool group currently installed, the version of the control program running, and key performance parameters such as the rated processing speed and accuracy level. This database is jointly maintained by the equipment management department and process engineers, and the data comes from equipment files and regular parameter calibration. At the same time, the system also queries the "order resource mapping table" and the associated material requirements planning (MRP) system data to extract the detailed material call records of this affected process in the original plan. This includes the specific material codes, planned usage amounts, planned arrival times, and designated warehouse or supplier information of various raw materials and semi-finished products required. For example, for the affected process "PROC-101", the query result may show that it was originally planned to be processed on equipment "CNC-05", the equipment configuration requires the use of tool group "TG-A12", the control program is "NC-P101_V2", the required material "MAT-AL-003" has a planned usage amount of 10 pieces, and should be supplied by warehouse "WH-B" 2 hours before the start of the scheduling adjustment time window. The system integrates all these equipment configuration details and original material requirement information extracted for each affected process to form a structured set of resource binding information for affected processes.

[0033] Based on the set of affected process resource binding information generated in the previous step, which details the originally bound equipment, its configuration, and required materials for each affected process, and in combination with the determined scheduling adjustment time window, the system executes an alternative resource retrieval process for each affected process. First, according to the process characteristics of the originally bound equipment recorded in the affected process resource binding information set (for example, querying through the equipment code to find that the processing type marked in the "Equipment Capacity Library" is "five-axis precision milling", the workbench size range is "500mm - 800mm", and the machinable material is "titanium alloy"), a search is conducted in the "Equipment Capacity Library" to screen out other candidate equipment items with the same or highly similar process characteristics. The determination criteria for "the same process characteristics" are predefined by the process department. For example, for "five-axis precision milling", the search results are allowed to include equipment marked as "five-axis general milling" whose historical processing accuracy statistical data (for example, the average tolerance of similar part processing in the past six months meets more than 80% of the original equipment, and this 80% is the acceptable lower limit set by engineering experience) can meet the quality requirements of the current process. Next, for each candidate equipment item screened out, the system will query its current real-time status and future pre-scheduling plan to determine whether there are sufficient and continuous available idle periods within the scheduling adjustment time window (for example, within the next 24 hours from the current moment). This idle period must be able to accommodate the entire processing duration of the affected process (including the estimated equipment adjustment and preparation time), and record the specific processing capacity parameters of these candidate equipment (such as the actual speed range, feed rate, expected processing efficiency, etc.). At the same time, based on the original material call records in the affected process resource binding information set, the system queries the "Material Supply Scheduling Table" to obtain the current inventory level, in-transit quantity, and the supplier's promised supply time interval or the exact time point of the next arrival batch of the corresponding materials. For example, for material "MAT-AL-003", the current inventory is 5 pieces, 10 pieces are in transit and are expected to arrive in 2 hours, and the supplier's standard supply cycle is 24 hours. Through the above operations, a list of candidate resource capabilities and material supply cycle combinations is generated for each affected process, including candidate alternative equipment, their available periods, processing parameters, and the current and near-term availability of supporting materials.

[0034] Based on the list of candidate resource capabilities and material supply cycle combinations formed in the previous stage, which provides several potential (alternative equipment, material supply) combination options for each affected process, the system begins to carefully screen and evaluate these combination items. First, it conducts an identification of processing capacity equivalence. This involves not only comparing the process types but also contrasting the specific technical parameters of the candidate equipment (such as maximum cutting force, maximum spindle speed, tool magazine capacity) with the processing requirements of the affected process (for example, process "PROC-101" requires a cutting force of no less than 500N and a speed of 8000rpm). Only when the key parameters of the candidate equipment all meet or exceed the process requirements (as specified in the process design document) is it considered to have preliminary processing capacity equivalence. For example, if the maximum cutting force of a candidate equipment is 480N, it does not meet the equivalence. Secondly, it conducts an identification of time cycle matching. The system needs to verify that the starting point of the available idle period of the candidate equipment is not earlier than the expected available supply time point of the required material, and the length of this idle period is sufficient to complete the entire process (including the preparation time for changeover, which is statistically based on historical data, such as the average changeover time of similar equipment being 0.5 hours). At the same time, the expected completion time of the process on the alternative equipment must fall within the previously determined scheduling adjustment time window. Furthermore, it conducts an identification of scheduling compatibility. The system will evaluate whether selecting a certain combination item will have an adverse impact on other processes or the overall production goal. For example, whether it will overly occupy a certain scarce auxiliary resource (such as a group of skilled technicians with specific skills), or whether it violates the shop-level scheduling rules (such as "successive processes of the same order should be arranged in equipment groups with similar geographical locations", which is an empirical rule set by the production supervisor based on the shop layout and logistics efficiency). Combination items that do not meet the compatibility requirements will be excluded. For example, although a certain alternative equipment meets the first two items, its operation requires the only currently idle senior technician, who has been reserved for another higher-priority urgent task, so this combination is incompatible. Through the above multi-dimensional and multi-condition item-by-item identification and screening, finally, those equipment processing capacity and material supply cycle combinations that simultaneously meet processing capacity equivalence, time cycle matching, and scheduling compatibility, and can complete the replacement operation within the scheduling adjustment time window are selected to form the available alternative resource parameter set.

[0035] The steps to obtain the preliminary scheduling operation instructions for coping with disturbances are as follows: Based on the available alternative resource parameter set, extract the unit processing output, unit operation time, scheduling adjustment time window length, and material supply cycle duration corresponding to each resource combination, construct standardized resource combination index items, and form a resource combination performance index set; According to the resource combination performance index set, calculate the efficiency matching value of the resource combination. The calculation formula is: ; Among them, is the efficiency matching value of the k-th group of resource combinations, is the unit processing output of the k-th group of resource combinations, is the unit operation time of the k-th group of resource combinations, is the scheduling adjustment time window length of the k-th group of resource combinations, is the material supply cycle duration of the k-th group of resource combinations, is the minimum time constant to prevent the denominator from being zero; Based on the efficiency matching values of the resource combinations, sort the resource combinations in descending order of their efficiency matching values, and select the resource combinations with efficiency matching values greater than the average production efficiency threshold to form a preliminary scheduling operation instruction for coping with disturbances.

[0036] Specifically, based on the available alternative resource parameter set obtained from the previous step, this parameter set provides several sets of feasible alternative equipment and their associated material supply information for each affected process. Next, the system extracts and structures a series of key performance indicators for each group of resource combinations. Specifically, for the k-th group of resource combinations (for example, allocating the affected process X to the alternative equipment M and using the material B from the supplier S), the system first extracts its "unit processing output" , which is defined as the output quantity of a standard production batch (the quantity of which is specified by the work order information of process X, such as 100 pieces) when processing process X on the alternative equipment M. This value usually equals the standard quantity of this batch. Then, it extracts the "unit operation time" , which refers to the total effective operation time required for the alternative equipment M to complete the above-mentioned one standard production batch (i.e., the output of the quantity), including the necessary setup adjustment time, actual processing running time, and cleaning or turnover time after the process is completed on equipment M. These time data are obtained from the historical statistics or standard working hour tables of the corresponding equipment M in the equipment capacity library for processing similar processes. Furthermore, it extracts the "scheduling adjustment time window length" , and this length is calculated in the previous step and represents the total duration from the current decision moment to the point when the entire plan affected by the disturbance must be adjusted and reach a new stable state (i.e., the maximum delay time point of the overall plan). For the evaluation of the current batch of resource combinations, this value is a unified constraint condition. Finally, it extracts the "material supply cycle duration" , which is the shortest waiting time from the current moment until all the materials required for process X (such as material B) can be fully prepared and available for the alternative equipment M. This information is obtained from the material supply cycle recorded in the available alternative resource parameter set. All these extracted , , and The numerical values are jointly constructed into a standardized resource portfolio metric item. The system generates such a metric item for each available alternative resource portfolio, and they are jointly aggregated to form a set of resource portfolio performance metrics.

[0037] Formula: , and the benefit of the formula is that it provides a method for quantitatively evaluating the comprehensive effectiveness of alternative resource portfolios. It not only considers the inherent processing efficiency of the resource portfolio itself (i.e., the output capacity per unit time ), but also introduces a penalty factor that is associated with the timeliness of material supply ( ) and the length of the available scheduling adjustment time window ( ). This penalty factor will dynamically adjust the basic efficiency score. If the material supply is delayed too long, approaching or even exceeding the entire adjustment window, the efficiency matching value of this portfolio will decrease or even become zero, thus automatically excluding those unrealistic options caused by material problems during decision-making, even if the equipment processing speed itself is very fast. This design enables the system to select the overall optimal resources under the current material and time constraints, rather than simply pursuing the theoretical highest speed, thereby improving the practical feasibility and robustness of the scheduling plan; Parameter is obtained as follows. It represents the unit processing output of the k-th group of resource portfolios for processing a certain affected process. This refers to the standardized batch or quantity to be completed by this process in a complete operation cycle. This numerical value directly comes from the production order of the work order to which the affected process belongs or the planned production quantity in the bill of materials (BOM). For a specific affected process, its value is fixed and does not change with the selected k-th group of resource portfolios. For example, if the currently evaluated process is "OP-A35" and its work order "WO-2025007" requires the production of 100 pieces of products, then for all resource portfolios k considered for "OP-A35", their corresponding are all 100 pieces; Parameter is obtained as follows. It represents the k-th group of resource portfolios to complete the above The total unit operation time required for unit processing output, which includes all the time required for the process "OP-A35" on the specific alternative equipment, covering the preparation time for equipment adjustment and program loading, the running time of actual cutting or processing, and the necessary cleaning or conversion time after the process ends. This data is extracted from the "equipment capacity library" or historical production data of the manufacturing execution system (MES), and queried for the alternative equipment (the core of resource combination k) and specific process types. For example, for resource combination 1 (selecting equipment M1 for processing process "OP-A35"), the total operation time for completing 100 pieces is queried to be 5.0 hours. Therefore, hours. For resource combination 2 (selecting equipment M2), the total operation time is queried to be 4.5 hours. Therefore, hours; Parameter The acquisition steps are as follows. It represents the length of the scheduling adjustment time window faced by the kth group of resource combinations. This length is determined after a series of evaluations (such as calculating the maximum delay time point of the overall plan) after a disturbance occurs. From the current decision-making moment ( ), to the time point ( ) when the overall plan is expected to be completed at the latest after all affected processes are adjusted. The total duration between these two time points, this value is the same for all resource combinations k evaluated under the same disturbance event. It reflects the overall time constraint for the system to make effective adjustments. For example, if the current is the 100th hour of the production cycle and the maximum delay time point of the overall plan is calculated to be the 132nd hour, then hours; Parameter The acquisition steps are as follows. It represents the material supply cycle duration required for the materials in the kth group of resource combinations to be fully prepared and available for its core alternative equipment to start processing the target process. This duration is calculated from the current decision-making moment ( ) to the latest time point when all necessary materials (determined according to the BOM of the process) arrive at the designated alternative equipment or its buffer and complete preprocessing steps such as inspection. This information is from the material supply information associated with each group of resource combinations recorded in the "available alternative resource parameter set" generated in the previous step. For example, for resource combination 1 (equipment M1), all required materials are in stock and can be immediately called, then hours. For resource combination 2 (equipment M2), one key special material needs to be urgently allocated from an external supplier and is expected to arrive and be processed in 3 hours. Then hours; Parameter The acquisition steps are as follows. It represents a value to prevent the denominator from accidentally being zero when calculating , for example An extremely small or negative abnormal situation, although logically should be positive), a very small positive time constant is set, and the setting is hours.

[0038] Calculation process: Let be hours, . is hours, is hours.

[0039] ; ; ; This result shows that the efficiency matching value of resource combination 1 is 25, and the efficiency matching value of resource combination 2 is approximately 24.750. Although the theoretical equipment processing speed of resource combination 2 (28.571 units / hour) is higher than that of resource combination 1 (25 units / hour), due to its longer material supply cycle duration being larger (8 hours), its efficiency matching value has been more significantly reduced and finally lower than that of resource combination 1 with timely material supply. These efficiency matching values will be used for subsequent sorting and screening.

[0040] Based on the efficiency matching values calculated for each group of available alternative resource combinations in the previous step, the system first sorts these resource combinations together with their corresponding efficiency matching values in descending order of the efficiency matching value to form an ordered candidate list. For example, if the calculated efficiency matching values of three groups of resource combinations are , , , then the sorted list is A, B, C. Next, the system needs to screen this ordered list according to an "average production efficiency threshold". The setting method of this threshold is: First, refer to the standard production efficiency that can be achieved by the equipment allocated for this affected process in the original production plan (i.e., when there is no disturbance). This standard efficiency is obtained from the process database or MES historical data. For example, the standard production efficiency of process "OP - A35" on the original equipment is 22 units per hour. Then, combined with the current production urgency and the acceptable performance degradation range, a percentage coefficient is set. For example, for general disturbances, the acceptable lower limit of efficiency is 80% of the original standard efficiency. For very urgent orders, this coefficient may be reduced to 70%. This percentage coefficient is dynamically determined by the production planning department according to the emergency response level regulations in the "Production Operation Management Manual". If the current is a general disturbance, the average production efficiency threshold is calculated as , the system then traverses the sorted candidate list and selects all resource combinations with efficiency matching values greater than 17.6 units per hour. In the above example, , and are all greater than 17.6. Therefore, these three groups of resource combinations are all selected. The set of these selected resource combinations arranged in order of superiority and inferiority constitutes the preliminary scheduling job instructions for coping with the current disturbance.

[0041] The steps for obtaining the real-time status evaluation value of personnel are as follows: Based on the preliminary scheduling job instructions for coping with disturbances, extract the target process information corresponding to each scheduling job instruction, match the operator numbers bound to the processes, query the operator task record database, obtain the continuous working time periods of each operator in the current cycle, and generate the information set of the working hours of the operators for the target processes; According to the information set of the working hours of the operators for the target processes, combined with the operator training record files and performance evaluation forms, extract the skill levels, completion efficiencies, and error rates of each operator under the corresponding process categories, and generate the real-time status evaluation values of the personnel.

[0042] Specifically, based on the preliminary scheduling job instructions for coping with disturbances obtained after screening and sorting in the previous step, this instruction set contains alternative resource combinations (mainly alternative equipment) recommended for each affected process and optimized by efficiency. For each specific instruction in this instruction set (i.e., each pair of "affected process - selected alternative equipment"), the system first accurately extracts the target process number and its detailed information corresponding to this instruction, such as the process "OPN-102B" and its processing requirements, the product it belongs to, etc. Then, in order to allocate appropriate operators later, the system needs to query the "Process - Skill and Qualification Requirements Library", which is jointly maintained by the process department and the human resources department, and clarifies the minimum skill level, special operation certifications (such as high-precision welding certificates), and safety training requirements for each process category or for operating specific equipment (such as the selected alternative equipment in the instruction). Based on these requirements, the system preliminarily matches and screens out a list of operator numbers of all operators who have the qualifications to execute this target process in the "Total Operator Information Database". After obtaining this list of qualified operators, the system further queries the "Operator Task Record Database", which records the start and end times of each operator's daily work tasks, task types, and rest records during the process in real-time or near real-time. For each qualified operator in the list, the system retrieves all their job records during the "current period", where the "current period" is defined here as from the start time of this scheduling shift (for example, 8:00 am today) to the current query time (for example, currently 2:30 pm). The system analyzes these records, calculates the cumulative duration of all actual work performed by each operator during this current period, and specifically identifies the continuous operation time period that was not interrupted by an effective rest (effective rest is defined as continuous uninterrupted rest for more than 30 minutes) for the most recent time. The data such as the extracted target process information, the numbers of each qualified operator matched with this process, and the cumulative operation duration and the most recent continuous operation duration of each qualified operator during the current period are gathered to generate the target process operator operation duration information set.

[0043] Based on the target process operator operation duration information set generated in the previous step, which already contains the target process and the numbers of each potential qualified operator and their recent operation duration data, the system then makes a further refined assessment of the professional capabilities and current status of each candidate operator. This process is completed by integrating data sources from different management systems. First, the system queries the "Operator Training Record Archive" according to the number of the candidate operator and the category of the target process (such as "Precision CNC Milling Five-Axis Operation"). This archive is maintained by the human resources department and details the content of skills training participated by employees in the past, the assessment levels passed, the qualification certifications obtained and their validity periods. The official certification "skill level" of this operator for the current target process category is extracted from it. For example, it is divided into four levels: junior (L1), intermediate (L2), senior (L3), and technician (L4). This level division standard is formulated according to the "Post Skill Evaluation Standard" within the enterprise. This standard is reviewed and updated by the technical committee every year to ensure that the skills of operators are synchronized with the equipment process requirements. Subsequently, the system accesses the "Production Performance Evaluation System" to query the historical "completion efficiency" of this operator when performing the same or similar processes as the current target process category in the past evaluation period (for example, the last three months). This efficiency is usually recorded in the form of the reciprocal or percentage of the ratio of the actual time used to complete the task to the standard working hours. For example, if the average efficiency score of an operator in this type of process in the past three months is 95 points (on a 100-point scale, 100 points means exactly meeting the standard working hours), and the historical "error rate", which is recorded as the percentage of the number of defective products produced by the operator when performing this type of process in the total output, or the percentage of the number of processes that need to be reworked in the total number of completed processes, such as 0.5%. These performance data are all from the statistics of actual production data collected by the MES system and quality inspection results. The system combines the skill level, historical completion efficiency, and historical error rate extracted for each candidate operator, together with the existing operation duration information, to form a multi-dimensional set of personnel instant status evaluation values. Each set of evaluation values completely describes the current comprehensive status of a candidate operator in a specific process task.

[0044] The steps to obtain the personnel scheduling execution plan are as follows: Based on the personnel instant status evaluation values, extract the skill proficiency level, total continuous operation duration, and current assigned task quantity of each operator to form a personnel operation ability status set; According to the personnel operation ability status set, calculate the personnel fitness score value. The calculation formula is: ; Among them, is the fitness score value of the i-th operator, is the skill proficiency level corresponding to the i-th operator in the target process. is the total continuous operation duration of the i-th operator in the current cycle. is the cumulative number of error processes of the i-th operator in the current cycle. is the number of assigned tasks of the i-th operator in the current scheduling cycle; Based on the personnel fitness score values, all candidate operators are sorted in descending order, and the operator with the highest personnel fitness score value is selected in turn according to the process matching priority order for scheduling and assignment, generating a personnel scheduling execution plan.

[0045] Specifically, based on the "personnel instant status evaluation value" set obtained in the previous step, which includes the skill level, historical performance (completion efficiency and error rate), and recent operation duration of each candidate operator for a specific target process category, the system first extracts or converts specific parameters required for subsequent fitness calculation for each operator. The first item is the "skill proficiency level" , this value is directly taken from the "skill level" recorded in the "personnel instant status evaluation value" for the current target process category, and is converted from the original descriptive level (such as L1, L2, L3, L4) to a preset numerical score. This numerical system is jointly formulated by the human resources department and the production technology department. For example, L1 corresponds to 10 points, L2 corresponds to 20 points, L3 corresponds to 30 points, and L4 corresponds to 40 points. The score interval reflects the weight of the ability difference between each level. The second item is the "total continuous operation duration" , this data is also obtained from the "cumulative operation duration" included in the "personnel instant status evaluation value" or the summary analysis of the "recent continuous operation time period", accurately recording the total actual working duration of the operator in the current shift or a preset fatigue monitoring cycle (such as the recent 4 hours) without effective rest (the effective rest standard is defined as more than 30 consecutive minutes before), in hours. The third item is the "current number of assigned tasks" , this data does not directly exist in the "personnel instant status evaluation value", but the system needs to query the real-time task scheduling dashboard or database of the manufacturing execution system (MES) according to the operator number in real time, and count the total number of process tasks that have been clearly assigned to this operator and are in the status of "pending start" or "in progress" as of the current moment. The "current scheduling cycle" here usually refers to the remaining time of the current shift or a fixed short future planning time domain (such as the next 4 hours). Through the above extraction and query, a record containing the above three core data is generated for each candidate operator, jointly forming a personnel operation ability status set.

[0046] Formula: , The advantage of the formula is that it constructs a quantitative model for comprehensively evaluating the suitability of operators for specific process tasks. This model not only positively motivates high skill levels ( ), but also reasonably considers the immediate state of operators by introducing a non-linear penalty term, specifically reflected in: using the natural logarithm function to process the continuous operation duration, so that the negative impact of fatigue on the fitness increases with the growth of working hours, but the growth rate gradually slows down, which is more in line with the physiological characteristics of actual fatigue accumulation, avoiding the problems of being overly sensitive to short-term overtime or insufficient punishment for extremely long operations that may be caused by linear penalties. At the same time, for historical errors ( ) and the current task load ( ), through the square root function and as the denominator, it realizes the smooth processing of the impact of errors and the appropriate adjustment of the impact of individual errors during multi-task parallelism, making the scoring more robust and practically guiding. Finally, it helps the system make the optimal personnel assignment decision among many qualified operators based on four dimensions: skills, fatigue, historical performance, and current load; The acquisition step of parameter is that it represents the quantitative value of the skill proficiency level corresponding to the i-th operator on the target process. This value is derived from the quantified skill levels recorded for each operator in the previous step "forming the set of personnel operation ability states". The original skill levels (such as L1 to L4) are numerically converted according to the internal enterprise post skill and salary grade correspondence table, which is jointly formulated by the Human Resources Department and the Production Engineering Department. For example, L1 (junior operator) is counted as 10 points, L2 (intermediate operator, capable of working independently) is counted as 20 points, L3 (senior operator, capable of handling complex problems and guiding others) is counted as 30 points, and L4 (technician level, capable of process improvement and fault diagnosis) is counted as 40 points. This scoring system is adjusted every two years with the update of the enterprise skill certification system. For example, if operator Zhang San's certified skill level for the current target process "precision grinding 03B" is L3, then his corresponding ; The acquisition step of parameter is that it represents the total continuous operation duration of the i-th operator during the current evaluation period (for example, from the start of this shift to the current moment), in hours. This data is also extracted from the continuous working hours of the operator recorded in the previous step "forming the set of personnel operation ability states". This duration is calculated by analyzing the task check-in and check-out records of the operator in the MES system and the defined effective rest duration (for example, a rest of more than 30 minutes is considered an interruption of continuous operation). For example, after operator Zhang San goes to work, he has worked for a total of 3.5 hours, and the longest continuous rest period during this period is 15 minutes (not reaching the effective rest standard), then his hours; Parameter The obtaining steps are as follows. It represents the cumulative number of error processes directly caused by the operation of the ith operator in the past specified evaluation period (for example, in the past four consecutive complete working weeks, this period is set by the quality management department according to the product complexity and production rhythm to ensure that there is a sufficient sample size to reflect the true error level), that is, the number of processes that cause product scrapping or require significant rework. This data is obtained by querying the "Operator-related NCR (Non-Conformance Report) Record Database" of the quality management system and summarized and statistically analyzed according to the operator number. For example, in the past four weeks, among the processes responsible by operator Zhang San, there was 1 process that caused a quality problem requiring scrapping due to his operation error, then ; Parameter The obtaining steps are as follows. It represents the number of tasks that have been assigned and not yet completed by the ith operator when the current scheduling instruction is issued. These tasks include the tasks being executed and the tasks waiting to be executed in his current shift work queue. This data is obtained by real-time querying the MES system. For example, when evaluating whether to assign a new task to operator Zhang San currently, it is found that there are 2 tasks under his name in the "in progress" or "pending" state, then .

[0047] Calculation process: Taking operator Zhang San (code i = 1) as an example, substitute the obtained parameter values for calculation: Given , hours, , . Substitute the parameters into the formula to calculate and get 11.403.

[0048] This result indicates that the personnel fitness score value of operator Zhang San (i = 1) for the current target process to be assigned is approximately 11.403. This is a comprehensive dimensionless score. The higher this score, the better the comprehensive fitness of the operator for the target process at the current moment. This score value of 11.403 specifically reflects that Zhang San's relatively high skill proficiency (30 points) is reduced under the fatigue effect of a certain continuous operation duration (3.5 hours), and then deducts the negative impacts brought by his recent cumulative errors (1 time) and current task load (2 items). This score will be used as the core basis for horizontal comparison and selection among multiple candidate operators. For example, if the score of another operator Li Si is 10.5, then without considering other priority rules, Zhang San will be considered for the assignment of this task prior to Li Si.

[0049] Based on the personnel fitness score values for specific target processes calculated for each candidate operator in the previous step , the system first organizes all these candidate operators and their corresponding scores into a list, and strictly sorts the operators in the list in descending order according to the score values from high to low to form a priority queue. For example, if there are three candidate operators A, B, and C, and their scores are , , , then the sorted queue may be A, C, B (if the scores are the same, they can be sorted by secondary keys such as employee numbers, or the original relative order can be maintained). Then, the system refers to a preset "process matching priority order table" to determine the order of scheduling and allocation. This table is comprehensively formulated by the production planning department according to factors such as the urgency of the order, the importance of the customer, the scarcity of materials, or the dependency relationship of subsequent processes, and is updated regularly. For example, it is stipulated in the table that the "urgent order process" has the highest priority, followed by the "bottleneck equipment process", and then the "ordinary order process". The system processes the target processes that need to allocate operators in this priority order. For the target process with the highest priority being processed currently, the system selects the operator at the top of the queue (i.e., with the highest score) and whose current status is "available" (i.e., not selected by a previous higher-priority process or having reached the daily working hour limit) from the candidate operator queue sorted in descending order according to the fitness score, and binds it to the target process and its selected alternative equipment to form a specific personnel allocation instruction. For example, if the current process being processed is P1 and operator A has the highest score and is available, then A is allocated to P1. Subsequently, when the system processes the next-priority process P2, it will select from the remaining available candidates with the highest fitness score. This selection and allocation process will continue until all the target processes that need to be scheduled have been successfully matched with operators, or there are no suitable available operators. All these successful matches (process - equipment - operator) converge together to generate the final personnel scheduling execution plan.

[0050] The above is only a preferred embodiment of the present invention, and it does not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A scheduling method for an artificial intelligence-driven intelligent manufacturing system, characterized in that, The following steps are involved: Monitor the order status parameters of the manufacturing execution system in real time, identify equipment failure shutdown signals or order insertion signals, extract the associated equipment code and current work order number, and establish the core disturbance information pair; Based on the core disturbance information pair, the expected completion time of the affected process in the current production activity of the intelligent manufacturing system is evaluated to obtain the expected delay value of a single process, and based on the expected delay value of the single process and the original planned completion time of all affected processes, the delay amount of the overall plan is calculated to determine the scheduling adjustment time window; According to the resource changes reflected by the scheduling adjustment time window and the core disturbance information, the processing capacity and material supply cycle of the alternative equipment associated with the affected process are retrieved to obtain an available alternative resource parameter set, and based on the available alternative resource parameter set and in comparison with the production efficiency index, a resource combination is selected for replacement to form a preliminary scheduling operation instruction to cope with the disturbance; For the preliminary scheduling work instructions for coping with the disturbance, the continuous working time and skill proficiency of the target process operators are obtained, and the personnel's real-time status evaluation value is obtained. Based on the personnel's real-time status evaluation value, personnel allocation optimization is performed to generate a personnel scheduling execution plan.

2. The scheduling method of the artificial intelligence-driven intelligent manufacturing system according to claim 1, wherein, The steps of obtaining the core disturbance information pair are: Collect order status parameters in real time, analyze and judge the equipment failure shutdown signal and order insertion signal in the order status parameters respectively, and extract the equipment failure shutdown signal and order insertion signal; Based on the equipment failure shutdown signal and the order insertion signal, the corresponding equipment code and the current work order number are retrieved, and the association relationship between the equipment failure shutdown signal and the equipment code, and the association relationship between the order insertion signal and the current work order number are respectively established to form association mapping information; Based on the association mapping information, the device code and the current work order number are combined and matched, and the code data and the work order data are bound in the form of a key-value pair to form a core disturbance information pair.

3. The scheduling method of the artificial intelligence-driven intelligent manufacturing system according to claim 1, wherein, The steps for obtaining the estimated delay value of a single process are as follows: Based on the core disturbance information pair, locate the affected process number in the current production activity, extract the planned completion time, unit standard working hours, resource start time and resource recovery time of each affected process, and form a set of production scheduling parameters for the affected processes; According to the affected process scheduling parameter set, the difference between the resource recovery available time and the actual resource start time is calculated to obtain the resource available delay length, and the affected process progress offset data group is generated; Based on the affected process progress deviation data group, a single process estimated delay value is calculated.

4. The scheduling method of the artificial intelligence-driven intelligent manufacturing system according to claim 1, wherein The steps for obtaining the scheduling adjustment time window are: Based on the estimated delay value of the single process, the delay values of all affected processes are aggregated in the order of process numbers, and each delay value is matched with the original planned completion time of the corresponding process one by one to generate a comparison table of affected process plans; According to the affected process plan comparison table, the affected processes are sorted according to their original planned completion time, the delay value corresponding to the process with the latest planned completion time is extracted, and the delay value is added to the original planned completion time to obtain the maximum delay time point of the overall plan; Based on the maximum delay time point of the overall plan, the time difference between the end time of the current global scheduling cycle of the intelligent manufacturing system and the maximum delay time point of the overall plan is compared to determine whether the scheduling adjustment demand duration exceeds the acceptable range and obtain the scheduling adjustment time window.

5. The scheduling method of the artificial intelligence-driven intelligent manufacturing system according to claim 1, characterized in that, The steps for obtaining the available alternative resource parameter set are: Based on the equipment code and the current work order number identified in the scheduling adjustment time window and the core disturbance information, query the associated equipment resource database and order resource mapping table in the intelligent manufacturing system, extract the equipment configuration records and original material call records associated with the affected process, and generate the affected process resource binding information set; According to the affected process resource binding information set, retrieve the equipment items with the same process characteristics as the bound equipment in the equipment capability library, select the equipment processing capability parameters with available idle time periods within the scheduling adjustment time window, and simultaneously retrieve the supply time intervals of the corresponding materials in the material supply plan table to generate a list of candidate resource capabilities and material supply cycle combinations; Based on the list of candidate resource capabilities and material supply cycle combinations, identify combinations that have processing capability equivalence, time cycle matching, and scheduling compatibility, screen equipment processing capabilities and material supply cycles that meet replaceable conditions within the time window, and obtain a set of available alternative resource parameters.

6. The scheduling method of the artificial intelligence-driven intelligent manufacturing system according to claim 1, characterized in that, The steps for obtaining the preliminary scheduling operation instructions for coping with disturbances are: Based on the available alternative resource parameter set, the unit processing output, unit operation time, scheduling adjustment time window length and material supply cycle length corresponding to each resource combination are extracted, and standardized resource combination index items are constructed to form a resource combination performance index set; Calculating the efficiency matching value of the resource combination according to the resource combination performance indicator set; Based on the efficiency matching values of the resource combinations, the resource combinations are sorted from high to low according to their efficiency matching values, and resource combinations whose efficiency matching values are greater than the average production efficiency threshold are selected to form preliminary scheduling job instructions to cope with disturbances.

7. The scheduling method of the artificial intelligence-driven intelligent manufacturing system according to claim 1, characterized in that The steps for obtaining the personnel's instant status evaluation value are as follows: Based on the preliminary scheduling operation instructions for coping with disturbances, the target process information corresponding to each scheduling operation instruction is extracted, the operator number bound to the process is matched, the operator task record database is queried, the continuous operation time period of each operator in the current cycle is obtained, and the target process operator operation time information set is generated; According to the target process operator's working time information set, combined with the operator's training record file and performance evaluation form, the skill level, completion efficiency and error rate of each operator in the corresponding process category are extracted to generate the operator's real-time status evaluation value.

8. The scheduling method of the artificial intelligence-driven intelligent manufacturing system according to claim 1, wherein The steps for obtaining the personnel scheduling execution plan are as follows: Based on the instant status evaluation value of the personnel, extract the skill proficiency level, total continuous operation duration, and current number of assigned tasks of each operator to form a personnel operation ability status set; Calculate the personnel fitness score value according to the personnel operation ability status set; Based on the personnel fitness score value, sort all candidate operators in descending order, and sequentially select the operator with the highest personnel fitness score value in the order of process matching priority for scheduling and allocation to generate a personnel scheduling execution plan.

Images