Production path optimization method and system in industrial internet

Abstract

The invention relates to the technical field of industrial production management, in particular to a production path optimization method and system in the industrial internet. The method comprises the steps of obtaining and standardizing order, production line, resource and process data, and constructing a production basic data set; generating a reference production scheduling scheme based on a multi-objective optimization algorithm; monitoring a production execution state and an external event in real time to capture a preset production disturbance event; after the disturbance event is detected, the type and the influence level of the disturbance event are analyzed and evaluated; selecting a rescheduling strategy matched with the disturbance influence from a preset strategy library according to an evaluation result; and executing the selected strategy to generate a corrected scheduling scheme, and issuing the corrected scheduling scheme to a related production unit. According to the method, the hierarchical disturbance response mechanism is introduced, so that the impact on the whole production plan is reduced to the greatest extent while the emergency situation is effectively dealt with, and the huge calculation burden caused by unnecessary global re-optimization is avoided.

Description

Technical Field

[0001] This invention relates to the field of industrial production management technology, specifically to a production path optimization method and system in the Industrial Internet. Background Technology

[0002] With the rapid development of industrial internet technology, the manufacturing industry is gradually transforming towards intelligence and collaboration. Production path optimization, as a core component of intelligent manufacturing, directly affects production efficiency, cost control, and order delivery capabilities. Its goal is to plan the optimal processing path and scheduling scheme for multi-variety, multi-process production tasks under complex production resource constraints.

[0003] Currently, traditional production path optimization methods largely rely on static planning, which generates fixed production scheduling schemes based on initial order information and resource status. This approach struggles to adapt to dynamically changing production environments. In actual production execution, various uncertainties and disturbances frequently occur, such as: sudden equipment failures causing processing interruptions, raw material supply delays leading to material shortages, urgent orders disrupting the original production rhythm, and component quality defects requiring rework. These disturbances undermine the feasibility of pre-set scheduling schemes, resulting in production delays, decreased resource utilization, and soaring production costs.

[0004] To address the above issues, some dynamic rescheduling methods have emerged in the existing technology, but they still have the following limitations:

[0005] Insufficient flexibility in disturbance response: Most methods employ a single rescheduling strategy (such as global re-optimization). For low-impact disturbances (such as short-term shutdowns of small equipment), excessive adjustments can lead to decreased production plan stability and excessively high rescheduling costs. For high-impact disturbances (such as long-term failures of critical equipment), simple local adjustments are insufficient to eliminate their cascading effects on overall production, which can easily cause scheduling schemes to fail.

[0006] Low integration of multi-objective optimization with actual constraints: Production path optimization often involves multiple objectives such as total production time, total cost, and resource load balancing. Existing methods often focus on single-objective optimization, or fail to fully consider the strict sequential constraints of production processes (such as components must be processed according to preset procedures) and resource uniqueness constraints (such as equipment and molds cannot be occupied by multiple tasks at the same time) in multi-objective solutions, making it difficult to implement the generated scheduling scheme in actual production. Summary of the Invention

[0007] To address the problems in existing technologies, this invention provides a production path optimization method and system in the Industrial Internet. Through multi-objective benchmark scheduling, real-time disturbance detection, and hierarchical rescheduling strategy selection, it achieves dynamic adaptation to complex production environments and improves the anti-interference capability and overall optimization level of the production system.

[0008] The production path optimization method in the Industrial Internet of Things proposed in this solution includes the following steps:

[0009] S1. Obtain production data including order information, capacity parameters of multiple production lines, material resource information, and component process flow information, and perform standardization processing to form a structured production basic dataset;

[0010] S2. Based on the aforementioned production base dataset, apply a multi-objective optimization algorithm to generate an initial baseline production scheduling scheme with the goal of minimizing total production time and total production cost.

[0011] S3. During production execution, continuously monitor the actual operating status data of the production line and external event data from the enterprise management system to detect predefined production disturbance events in real time.

[0012] S4. When the production disturbance event is detected, analyze and determine the type of the disturbance event and its potential impact level on the existing scheduling scheme;

[0013] S5. Based on the determination results of the type and impact level of the disturbance event, select and execute an appropriate rescheduling strategy from a preset rescheduling strategy set to generate a modified production scheduling scheme.

[0014] The preset rescheduling strategy set includes:

[0015] A low-complexity local adjustment strategy for handling low-impact disturbance events;

[0016] A highly complex global re-optimization strategy for handling high-impact perturbation events;

[0017] S6. Distribute the modified production scheduling scheme to the corresponding production line control unit and management system.

[0018] Preferably, before generating the baseline production scheduling scheme in step S2, a mathematical optimization model is first constructed. This model takes the process allocation and sequencing of components as decision variables and includes the following constraints: process sequence constraints to ensure that each component must be processed according to its preset process flow, and resource uniqueness constraints to ensure that each piece of equipment or each set of molds can only be used for one production task at any given time.

[0019] Preferably, the application of a multi-objective optimization algorithm, with the goal of minimizing total production time and total production cost, generates an initial baseline production scheduling scheme, specifically including:

[0020] Randomly generate an initial population consisting of multiple initial scheduling schemes;

[0021] The population is evolved iteratively by performing the following operations:

[0022] The scheduling scheme in the current population is evaluated and selected based on the non-dominated ordering and crowding distance to generate the parent generation;

[0023] Crossover and mutation operators are applied to the parent generation to generate the offspring population;

[0024] The parent and offspring populations are merged, and a new generation of populations is selected from them until the preset termination conditions are met.

[0025] After the algorithm terminates, it outputs a Pareto optimal solution set consisting of multiple non-dominated scheduling schemes;

[0026] Based on the Pareto optimal solution set, the final baseline production scheduling scheme is determined.

[0027] Preferably, determining the final baseline production scheduling scheme based on the Pareto optimal solution set specifically includes:

[0028] Based on the Pareto optimal solution set, the optimal and worst values ​​of each optimization objective are determined, thereby constructing positive ideal solutions and negative ideal solutions;

[0029] Calculate the geometric distance between each scheduling scheme in the solution set and the positive ideal solution and the negative ideal solution;

[0030] Based on the geometric distance, calculate the proximity of each scheduling scheme to the ideal solution;

[0031] The scheduling scheme with the highest degree of approximation will be selected as the final benchmark production scheduling scheme.

[0032] Preferably, the types of production disruption events in step S4 include: emergency order insertion, sudden production equipment failure, raw material supply delay, and component quality inspection failure.

[0033] Preferably, step S4 involves analyzing and determining the type of the disturbance event and its potential impact level on the existing scheduling scheme, specifically including:

[0034] The production disruption events are categorized into a predefined category, which includes resource-related disruptions, material-related disruptions, and order-related disruptions.

[0035] Based on the classification results, the chain reaction of the disturbance on the existing scheduling scheme is calculated and used as the impact index value of the disturbance;

[0036] The calculated impact index value is compared with a preset threshold using a rule-based inference engine to determine the impact level of the disturbance.

[0037] Preferably, the local adjustment strategy in step S5 includes:

[0038] The tasks affected by the disturbance event are identified, and candidate production lines capable of processing the tasks are determined.

[0039] Obtain the existing scheduling plans for each candidate production line and find the available time window that meets the task processing time and does not violate the task delivery deadline;

[0040] If at least one available time window is found, the optimal time window is selected to insert the task, and the planned time of subsequent tasks is updated.

[0041] If no available time window is found, then tasks with lower priority than the current task on the candidate production line are obtained as replacement tasks for the current task.

[0042] Select the option that has the least impact on the overall stability of the plan and execute the task exchange as the replacement task for the current task.

[0043] Preferably, the global re-optimization strategy in step S5 includes:

[0044] Obtain the real-time status of the production system, including a list of completed tasks, the progress of tasks in execution, and the availability of production resources;

[0045] The planned but not yet started production tasks are merged with new tasks generated by disturbance events to form a brand new set of tasks to be scheduled, while the resource constraints are updated according to the real-time status obtained.

[0046] The new set of tasks to be scheduled and the updated resource constraints are taken as input, and the multi-objective optimization algorithm is applied to solve them to generate a new Pareto optimal solution set.

[0047] Choose a final modified scheduling scheme from the new Pareto optimal solution set.

[0048] The production path optimization system in the Industrial Internet of this solution includes a data acquisition and processing module, a baseline scheduling module, a monitoring module, an analysis and selection module, a rescheduling execution module, and an instruction distribution module.

[0049] The data acquisition and processing module is used to acquire and standardize production-related data from multiple data sources;

[0050] The baseline scheduling module is connected to the data acquisition and processing module and is used to generate an initial baseline production scheduling scheme.

[0051] The monitoring module is used to continuously monitor the production status and external events during the production process in order to detect production disturbances;

[0052] The analysis and selection module is communicatively connected to the monitoring module and is used for:

[0053] Receive and classify production disturbance event data reported by the monitoring module;

[0054] Based on the pre-defined decision logic, and considering the classification results of disturbance events and their potential impact, the most suitable rescheduling strategy is selected from an available rescheduling strategy library.

[0055] The rescheduling execution module is connected to the analysis and selection module and is used to execute the corresponding algorithm according to the selected strategy to generate a modified production scheduling scheme.

[0056] The instruction distribution module is used to send the revised scheduling scheme to the production site.

[0057] Compared with the prior art, the advantages of this invention are:

[0058] For common, low-impact disturbances (such as small-batch order insertions), the system can quickly respond using a "local adjustment" strategy with extremely low computational costs, rather than waiting for global optimization calculations as required by existing technologies, thus greatly improving its responsiveness to changes in the production environment.

[0059] By adopting a localized adjustment strategy, this invention can limit the impact of disturbances to a minimum, avoiding the chaotic situation where the entire production plan has to be "started from scratch" due to minor changes, and ensuring the stability and predictability of most production tasks. Attached Figure Description

[0060] Figure 1 This is a flowchart of the production path optimization method in the Industrial Internet proposed in this invention. Detailed Implementation

[0061] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0062] This invention provides a production path optimization method in the Industrial Internet. This method aims to achieve dynamic optimal adjustment of production plans through data-driven approaches and intelligent algorithms, and includes the following steps:

[0063] S1. Acquiring and processing production data:

[0064] S101. Before the start of a production cycle or when rescheduling is required, collect the core data needed for production from multiple heterogeneous data sources within the enterprise.

[0065] The main data sources include: Enterprise Resource Planning (ERP) systems, which provide order information such as customer order numbers, product models, required quantities, expected delivery dates, and order priorities, as well as material resource information such as inventory levels of major raw materials and supplier information; Manufacturing Execution Systems (MES), which provide production execution-related data, including specific component process flow information, i.e., the sequence of processes required for each product or component, the standard processing time for each process, and the required resources (equipment, molds, tools); and Equipment Data Acquisition and Monitoring Control Systems and Programmable Logic Controllers, which acquire real-time capability parameters of multiple production lines, such as the current status of each piece of equipment (running, idle, faulty, under maintenance), processing speed, energy consumption rate, and equipment start-up and changeover preparation time.

[0066] The collected data specifically includes:

[0067] Order information: Structured representation as a set containing order ID, product ID, quantity, delivery deadline, and priority coefficient.

[0068] Production line capacity parameters: Create a file for each piece of equipment, including equipment ID, list of executable processes, processing time function for each process (may be related to workpiece material), unit time operating cost, and time matrix for switching between different tasks.

[0069] Material resource information: including raw material inventory list (material ID, current quantity, safety stock threshold) and tool / mold list (ID, availability status, remaining life).

[0070] Component process flow information: Represented in the form of a directed acyclic graph (DAG), where nodes represent processes and edges represent the sequential constraints between processes.

[0071] S102. The collected data is then standardized and integrated into a production base dataset. This includes:

[0072] Data cleaning: Identify and remove outliers or missing values, such as a negative processing time record or a non-existent device ID.

[0073] Data transformation: Convert all data into a unified format and unit. For example, unify all time units to seconds.

[0074] Data structuring: Cleaned and transformed data is integrated and loaded into a unified, structured production base dataset. This dataset can be a relational database or an in-memory data object model, providing a data source for subsequent scheduling calculations regarding factory status and production needs.

[0075] S2. Generate a baseline production scheduling plan:

[0076] Before production begins, an initial, globally optimal baseline production scheduling scheme is generated based on the structured production dataset created above.

[0077] S201. Abstract the production scheduling problem into a mathematical optimization model.

[0078] The decision variables include process allocation variables and process sequencing and time variables.

[0079] The process allocation variable can be defined as a three-dimensional binary variable X. ijk If process j of component i is assigned to production line (or equipment) k for processing, then X ijk =1, otherwise 0.

[0080] The process sequence and time variable can be defined as a continuous variable S. ij Let represent the start time of process i for component i, and let C be a continuous variable. ij Indicates the completion time.

[0081] The optimization objectives are set as minimizing production time and minimizing production costs.

[0082] Constraints include process sequence constraints and resource uniqueness constraints.

[0083] The process sequence constraint is used to ensure that all processes of each component must be processed in the order of its preset process flow. For any two adjacent processes j and j+1 of component i, the following must be satisfied:

[0084] S i,j+1 ≥C ij ;

[0085] That is, the start time of the subsequent process must not be earlier than the completion time of the preceding process.

[0086] The resource uniqueness constraint is used to ensure that at any given time, each piece of equipment or each set of molds can only be used for one production task. For any two different tasks assigned to the same equipment k (e.g., task A, i.e., operation (i,j), and task B, i.e., operation (i',j')), the following must be satisfied:

[0087] S ij ≥Ci'j' or S i'j' ≥C ij ;

[0088] Among them, S ij and C ij These are the start and finish times for task A, S. i'j' and C i'j' These are the start and finish times of task B, respectively. This constraint ensures that the execution of these two tasks on device k is serial, not parallel.

[0089] S202. Solve the model.

[0090] Since the above model has multiple optimization objectives, a heuristic algorithm—the Non-Dominated Sorting Genetic Algorithm (NSGA-II)—is used to solve it, specifically including:

[0091] Step 1: Encode each "individual" (i.e., a potential scheduling scheme) using a two-segment chromosome. The first segment is the process permutation string, which contains a global permutation of all processes for all orders; the second segment is the equipment allocation string, which assigns a feasible processing equipment to each process.

[0092] Step 2: Randomly generate an initial population consisting of multiple (e.g., 100) chromosomes. Each generated chromosome must be a valid scheduling scheme, i.e., satisfying all hard constraints.

[0093] Step 3: For each scheduling scheme in the current population, calculate its corresponding total production time and total production cost. Then, apply non-dominated sorting to stratify the population, with the first stratum being optimal, the second stratum being even better, and so on. Within the same stratum, evaluate the density of solutions around an individual by calculating crowding distance, prioritizing individuals in sparse regions to ensure solution diversity. In this way (prioritizing those with higher non-dominated levels and larger crowding distances), superior individuals are selected as parents.

[0094] Step 4: Apply genetic operators to the selected parents to generate offspring. Crossover operators (such as process-priority crossover) are used to exchange superior gene segments between two parent schemes, generating new process permutations. Mutation operators (such as exchange mutation or redistribution mutation) make small, random modifications to a single scheme, such as swapping the positions of two processes or reassigning a process to another available machine, to explore new solution spaces.

[0095] Step 5: Merge the parent population with the newly generated offspring population. Perform non-dominated sorting and crowding calculations on this merged large population again, and select the optimal individuals with the same number as the original population to form a new generation population.

[0096] Step 6: Repeat steps 3-5 above until a preset termination condition is met, such as reaching the maximum number of iterations (e.g., 500 generations) or the Pareto front showing no significant change over multiple generations. After the algorithm terminates, it outputs a Pareto optimal solution set consisting of multiple non-dominated scheduling schemes. Each scheme in this solution set is an "optimal" solution that cannot improve another objective (e.g., time) without sacrificing one objective (e.g., cost).

[0097] S203. Determine the baseline scheduling scheme based on the above Pareto optimal solution set.

[0098] A multi-criteria decision-making method, namely the approximation of ideal solution ranking method, is used to select a final solution from the Pareto optimal solution set. Specifically, this includes:

[0099] Step 1: In the Pareto solution set, find the minimum total production time and the minimum total production cost among all solutions. These two optimal values ​​constitute a theoretically possible, but usually unattainable, "perfect" solution, i.e., the positive ideal solution.

[0100] Step 2: Similarly, find the maximum value of total production time and the maximum value of total production cost in the solution set to form the worst solution, i.e., the negative ideal solution.

[0101] Step 3: For each scheduling scheme in the Pareto solution set, calculate its Euclidean distance to the positive ideal solution and the negative ideal solution.

[0102] Step 4: Based on the two distances mentioned above, calculate the closeness of each solution to the positive ideal solution. A good solution should be as close to the positive ideal solution as possible, while being as far away from the negative ideal solution as possible.

[0103]

[0104] In the formula, C i Let represent the closeness of the i-th solution to the ideal solution. This closeness value is between 0 and 1; the closer the value is to 1, the better the overall performance of the solution. Let be the Euclidean distance between the i-th solution and the negative ideal solution. Let be the Euclidean distance between the i-th solution and the ideal solution.

[0105] Step 5: Select the scheduling scheme with the highest proximity as the final benchmark production scheduling scheme.

[0106] S3, Monitor production status:

[0107] During production execution, real-time operational status data of the production line is captured through continuous communication with MES, SCADA, and IoT devices. This includes the real-time status of each device (processing, idle, faulty), the progress of the current processing task, reports of completed tasks, and material consumption. Simultaneously, external event data from the enterprise management system is monitored. For example, requests for urgent order insertions can be obtained instantly through API polling or message queue subscription.

[0108] Based on a predefined rules engine, when the monitored data stream meets specific conditions, it is determined that a "production disturbance event" has occurred. For example:

[0109] Rule 1: If a device changes its status from "running" to "faulty" and this status continues for more than a preset time (e.g., 2 minutes), a "device fault" event is triggered.

[0110] Rule 2: If a new order with an "urgent" flag is passed into the ERP system interface, the "urgent order insertion" event is triggered.

[0111] Rule 3: If the MES reports that a component is "unqualified" in the quality inspection process, a "quality non-conformity" event is triggered.

[0112] Once a disturbance event is detected, information such as the event type, occurrence time, and related objects (e.g., faulty equipment ID, non-conforming component ID) is immediately packaged for further analysis.

[0113] S4. Disturbance event analysis and judgment.

[0114] S401. Upon receiving a disturbance event, the event is first classified.

[0115] The main categories of events include:

[0116] Resource-related disturbances: Events that directly affect the availability of production resources, such as sudden equipment failures or tool damage.

[0117] Material-related disturbances: Events that affect the supply of materials, such as delays in the supply of raw materials or batch material quality issues.

[0118] Order-related disturbances: Events stemming from changes in the order itself, such as the insertion of urgent orders, order cancellations, or modifications.

[0119] Quality disturbances: Events caused by production quality problems, such as components failing quality inspection.

[0120] S402. Then assess the potential impact level of the disturbance on existing scheduling schemes. Specifically:

[0121] First, simulate the "chain reaction" of the disturbance event on the current scheduling plan. For example, for a single equipment failure, quickly calculate: how many tasks scheduled to execute on that equipment will be affected? How many subsequent tasks (on other equipment) will be delayed due to the delays in these tasks? What is the total value or priority of the affected orders? Sum these quantitative indicators in a weighted manner to obtain a comprehensive "impact index value".

[0122] Then, the calculated impact index value is compared with the preset threshold to determine its impact level.

[0123] S403. Based on the determination of the type and impact level of the disturbance event, select and execute an appropriate rescheduling strategy from a preset set of rescheduling strategies. For example, for low impact levels, select a local adjustment strategy; for high impact levels, select a global re-optimization strategy.

[0124] S5. Select and execute the rescheduling policy.

[0125] Based on the analysis and judgment results of step S4, a rescheduling algorithm is executed to generate a revised production scheduling scheme. This step includes two core strategies.

[0126] Local adjustment strategy: This strategy is executed when the instruction to "perform local adjustment" is received. The specific process is as follows:

[0127] Step 1: First, pinpoint the task directly affected by the disturbance (e.g., a task originally planned to be performed on a faulty machine) and identify all candidate production lines capable of processing this task from the production base dataset.

[0128] Step 2: For each candidate production line, check its current scheduling plan and find a sufficiently long, idle "time window" whose length must meet the processing time of the task and that inserting the task into the window will not cause it to violate the final delivery deadline.

[0129] Step 3: If at least one available time window is found, select the optimal one according to preset rules (such as selecting the earliest window or the window with the least impact on subsequent tasks), insert the affected task, and postpone the planned start time of all subsequent tasks on the production line.

[0130] Step 4: If no available time window is found on any candidate production lines, initiate the task swapping logic. Identify tasks already scheduled on these production lines that have a lower priority than the currently affected task, designating them as "replaceable tasks." Then, assess the overall schedule volatility resulting from swapping the affected task with each replaceable task, and select the swapping scheme with the least impact on global stability.

[0131] Global re-optimization strategy: This strategy is executed when the instruction to "execute global re-optimization" is received. The specific process is as follows:

[0132] Step 1: Obtain the current status of the production site. This includes: a list of tasks that are 100% complete; the exact progress of tasks in progress (e.g., a task that takes 60 minutes has been completed for 20 minutes); and updated production resource availability (e.g., a piece of equipment is unavailable for the next 8 hours).

[0133] Step 2: Merge all production tasks that were originally planned but have not yet started with new tasks generated by the disturbance event (such as urgent orders or rework tasks) to form a completely new set of tasks to be scheduled.

[0134] Step 3: Take this new set of tasks to be scheduled, along with the resource constraints updated according to the real-time status (such as the device availability time window), as new inputs, and call the multi-objective optimization algorithm in step S2 again.

[0135] Step 4: Find a completely new Pareto optimal solution set, which reflects all possible new optimal solutions starting from the current time point.

[0136] Step 5: From this new solution set, use the multi-criteria decision-making method described above again to select a solution as the final revised production scheduling solution.

[0137] S6. Distribute the revised scheduling scheme.

[0138] After generating the revised production scheduling plan, it is converted into executable instructions and issued to the production site. Specifically, this can be done as follows:

[0139] The updated electronic work order is pushed to the workstation terminal of the MES system to inform the operator of the next task and related process parameters.

[0140] Send specific machining programs and start / stop commands to the production line control unit (such as a PLC or CNC controller).

[0141] Send new material delivery instructions to the material handling system (such as the AGV scheduling system) to ensure that the right material is delivered to the right workstation at the right time.

[0142] Update the estimated completion time of the order with the relevant management system (such as ERP).

[0143] This invention also provides a production path optimization system in the Industrial Internet. This system aims to achieve dynamic optimal adjustment of production plans through data-driven approaches and intelligent algorithms. The system includes a data acquisition and processing module, a baseline scheduling module, a monitoring module, an analysis and selection module, a rescheduling execution module, and an instruction distribution module.

[0144] The data acquisition and processing module is used to execute step S1 in the above method, to acquire and standardize production-related data from multiple data sources.

[0145] The baseline scheduling module is connected to the data acquisition and processing module and is used to execute step S2 in the above method to generate an initial baseline production scheduling scheme.

[0146] The monitoring module is used to execute step S3 in the above method, continuously monitoring the production status and external events during the production process to detect production disturbances.

[0147] The analysis and selection module is communicatively connected to the monitoring module and is used to execute step S4 in the above method, receive and classify production disturbance event data reported by the monitoring module, and select the most suitable rescheduling strategy according to the preset decision logic.

[0148] The rescheduling execution module is connected to the analysis and selection module and is used to execute step S5 in the above method. Based on the selected strategy, it executes the corresponding algorithm to generate a modified production scheduling scheme.

[0149] The instruction distribution module is used to execute step S6 in the above method, and to send the revised scheduling scheme to the production site.

[0150] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0151] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A production path optimization method in the Industrial Internet, characterized in that, Includes the following steps: S1. Obtain production data including order information, capacity parameters of multiple production lines, material resource information, and component process flow information, and perform standardization processing to form a structured production basic dataset; S2. Based on the aforementioned production base dataset, apply a multi-objective optimization algorithm to generate an initial baseline production scheduling scheme with the goal of minimizing total production time and total production cost. S3. During production execution, continuously monitor the actual operating status data of the production line and external event data from the enterprise management system to detect predefined production disturbance events in real time. S4. When the production disturbance event is detected, analyze and determine the type of the disturbance event and its potential impact level on the existing scheduling scheme; S5. Based on the determination results of the type and impact level of the disturbance event, select and execute an appropriate rescheduling strategy from a preset rescheduling strategy set to generate a modified production scheduling scheme. The preset rescheduling strategy set includes: A low-complexity local adjustment strategy for handling low-impact disturbance events; A highly complex global re-optimization strategy for handling high-impact perturbation events; S6. Distribute the modified production scheduling scheme to the corresponding production line control unit and management system.

2. The production path optimization method in the Industrial Internet according to claim 1, characterized in that, Before generating the baseline production scheduling scheme in step S2, a mathematical optimization model is first constructed. This model takes the process allocation and sequencing of components as decision variables and includes the following constraints: process sequence constraints to ensure that each component must be processed according to its preset process flow, and resource uniqueness constraints to ensure that each piece of equipment or each set of molds can only be used for one production task at any time.

3. The production path optimization method in the Industrial Internet according to claim 1, characterized in that, The application uses a multi-objective optimization algorithm to minimize total production time and total production cost, generating an initial baseline production scheduling scheme, specifically including: Randomly generate an initial population consisting of multiple initial scheduling schemes; The population is evolved iteratively by performing the following operations: The scheduling scheme in the current population is evaluated and selected based on the non-dominated ordering and crowding distance to generate the parent generation; Crossover and mutation operators are applied to the parent generation to generate the offspring population; The parent and offspring populations are merged, and a new generation of populations is selected from them until the preset termination conditions are met. After the algorithm terminates, it outputs a Pareto optimal solution set consisting of multiple non-dominated scheduling schemes; Based on the Pareto optimal solution set, the final baseline production scheduling scheme is determined.

4. The production path optimization method in the Industrial Internet according to claim 3, characterized in that, The determination of the final baseline production scheduling scheme based on the Pareto optimal solution set specifically includes: Based on the Pareto optimal solution set, the optimal and worst values ​​of each optimization objective are determined, thereby constructing positive ideal solutions and negative ideal solutions; Calculate the geometric distance between each scheduling scheme in the solution set and the positive ideal solution and the negative ideal solution; Based on the geometric distance, calculate the proximity of each scheduling scheme to the ideal solution; The scheduling scheme with the highest degree of approximation will be selected as the final benchmark production scheduling scheme.

5. The production path optimization method in the Industrial Internet according to claim 1, characterized in that, The types of production disruption events in step S4 include: emergency order insertion, sudden production equipment failure, raw material supply delay, and component quality inspection failure.

6. The production path optimization method in the Industrial Internet according to claim 1, characterized in that, Step S4 involves analyzing and determining the type of the disturbance event and its potential impact on the existing scheduling scheme, specifically including: The production disruption events are categorized into a predefined category, which includes resource-related disruptions, material-related disruptions, and order-related disruptions. Based on the classification results, the chain reaction of the disturbance on the existing scheduling scheme is calculated and used as the impact index value of the disturbance; The calculated impact index value is compared with a preset threshold using a rule-based inference engine to determine the impact level of the disturbance.

7. The production path optimization method in the Industrial Internet according to claim 1, characterized in that, The local adjustment strategy in step S5 includes: The tasks affected by the disturbance event are identified, and candidate production lines capable of processing the tasks are determined. Obtain the existing scheduling plans for each candidate production line and find the available time window that meets the task processing time and does not violate the task delivery deadline; If at least one available time window is found, the optimal time window is selected to insert the task, and the planned time of subsequent tasks is updated. If no available time window is found, then tasks with lower priority than the current task on the candidate production line are obtained as replacement tasks for the current task. Select the option that has the least impact on the overall stability of the plan and execute the task exchange as the replacement task for the current task.

8. The production path optimization method in the Industrial Internet according to claim 1, characterized in that, The global re-optimization strategy in step S5 includes: Obtain the real-time status of the production system, including a list of completed tasks, the progress of tasks in execution, and the availability of production resources; The planned but not yet started production tasks are merged with new tasks generated by disturbance events to form a brand new set of tasks to be scheduled, while the resource constraints are updated according to the real-time status obtained. The new set of tasks to be scheduled and the updated resource constraints are taken as input, and the multi-objective optimization algorithm is applied to solve them to generate a new Pareto optimal solution set. Choose a final modified scheduling scheme from the new Pareto optimal solution set.

9. A production path optimization system in the Industrial Internet, characterized in that, It includes a data acquisition and processing module, a baseline scheduling module, a monitoring module, an analysis and selection module, a rescheduling execution module, and an instruction distribution module; The data acquisition and processing module is used to acquire and standardize production-related data from multiple data sources; The baseline scheduling module is connected to the data acquisition and processing module and is used to generate an initial baseline production scheduling scheme. The monitoring module is used to continuously monitor the production status and external events during the production process in order to detect production disturbances; The analysis and selection module is communicatively connected to the monitoring module and is used for: Receive and classify production disturbance event data reported by the monitoring module; Based on the pre-defined decision logic, and considering the classification results of disturbance events and their potential impact, the most suitable rescheduling strategy is selected from an available rescheduling strategy library. The rescheduling execution module is connected to the analysis and selection module and is used to execute the corresponding algorithm according to the selected strategy to generate a modified production scheduling scheme. The instruction distribution module is used to send the revised scheduling scheme to the production site.

Images