Workshop production scheduling optimization method and system based on digital model

Abstract

The invention discloses a workshop production scheduling optimization method and system based on a digital model, and relates to the technical field of digital models, and the method comprises the following steps: S1, collecting data, and building an initial digital model; s2, collecting equipment data in real time; s3, generating a dynamically updated digital model based on the equipment data; s4, local optimization is carried out in combination with a collaborative particle swarm optimization algorithm; s5, performing global search by using a firefly algorithm; s6, performing simulation verification evaluation on the preliminary production scheduling scheme; s7, updating the dynamically updated digital model in real time; and S8, monitoring the execution effect and correcting the execution effect in real time. According to the invention, intelligent optimization and dynamic closed-loop control of workshop scheduling are realized, and the production efficiency and the resource utilization rate are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital models, and particularly to an optimization method and system for workshop production scheduling based on digital models. Background Art

[0002] With the development of intelligent manufacturing, industrial Internet, and workshop informatization technologies, modern manufacturing enterprises have put forward higher requirements for production scheduling efficiency and resource utilization rate. As the most core execution unit in the manufacturing system, the rationality of production scheduling in the workshop directly affects the product delivery cycle, equipment load balance, and energy consumption level. Therefore, how to achieve efficient scheduling optimization of the workshop production process has become a key link in improving the intelligent level of the manufacturing system.

[0003] Most traditional workshop production scheduling methods rely on static rule tables, manual operation plans, or scheduling engines based on heuristic strategies. Common scheduling methods include first-come-first-served, shortest processing time first, earliest due date first, etc. These methods have simple logic and low implementation costs, and are applicable to some production environments with stable rhythms. However, in the face of practical challenges such as diverse product types, complex process routes, frequent order interspersions, and dynamic changes in resource status, the limitations of traditional scheduling methods have become increasingly prominent, mainly manifested as problems such as lagging scheduling response, low resource utilization rate, serious energy consumption redundancy, and large plan execution deviation.

[0004] In recent years, with the wide application of data acquisition technologies, multi-source information such as equipment operation status, order processing progress, and material flow location can be obtained in real time at the manufacturing site. Although the information acquisition means have been continuously improved, these data are often not effectively transformed into dynamic models to support scheduling decisions, resulting in serious underestimation of data value. Especially in traditional scheduling systems, the models are often statically constructed and cannot reflect actual situations such as equipment failures, load fluctuations, and process offsets in real time, thus disconnecting scheduling decisions from on-site status.

[0005] To improve the scheduling optimization effect, some research has begun to introduce intelligent algorithms, such as genetic algorithms, particle swarm optimization algorithms, ant colony algorithms, etc., and attempts to find better solutions through intelligent search methods. However, the application of these methods in complex workshop systems still faces two challenges: one is the lack of dynamic update ability, that is, the scheduling algorithm still runs on static inputs and cannot perceive the evolution of the system state in real time; the other is that the optimization goal is too single, and most algorithms only focus on production efficiency indicators and ignore the coordinated optimization of composite indicators such as resource load balance and energy consumption control, resulting in the lack of comprehensiveness and feasibility of scheduling results.

[0006] In addition, there is still a certain degree of uncertainty in the actual implementation of the scheduling plan. For example, temporary equipment failures, logistics congestion, material delays and other problems often disrupt the original scheduling plan. In traditional scheduling systems, manual intervention is usually used for remedial adjustments. This processing method is not only slow to respond and relies on experience, but also cannot guarantee the optimality or suboptimality of the adjustment results. Therefore, the lack of closed-loop feedback capabilities and the inability to achieve real-time monitoring and correction during the scheduling execution process are the key shortcomings faced by the current scheduling system in actual operation.

[0007] Based on this background, there is an urgent need for a comprehensive scheduling optimization method and system that integrates modeling, data fusion, intelligent optimization and dynamic feedback to open up the full-chain intelligent closed loop of "modeling-optimization-execution-correction". The system should have the ability to build high-precision digital models, and be able to uniformly model multi-source data such as workshop equipment, process paths, material flow and energy consumption parameters; it should have real-time state perception and dynamic model update mechanisms, and be able to continuously reflect micro-changes in the production process; it should have multi-objective scheduling optimization capabilities, and achieve dynamic trade-offs between delivery time, load balancing and energy saving; at the same time, it should also have scheduling simulation and feedback control capabilities, which can verify the feasibility of scheduling and realize deviation correction in the execution stage, thereby improving the intelligence level and practical value of the scheduling system.

[0008] Therefore, how to provide a workshop production scheduling optimization method and system based on a digital model is a problem that technical personnel in this field urgently need to solve. Summary of the invention

[0009] One object of the present invention is to propose a workshop production scheduling optimization method and system based on a digital model. The present invention fully integrates digital modeling, real-time data perception, multi-objective intelligent optimization and simulation feedback correction technology, and describes in detail the process of scheduling optimization through collaborative particle swarm algorithm and improved firefly algorithm, realizing intelligent modeling, dynamic optimization and adaptive adjustment of the entire production process, and has the advantages of high scheduling efficiency, high resource utilization and lower energy consumption.

[0010] A workshop production scheduling optimization method based on a digital model according to an embodiment of the present invention comprises the following steps:

[0011] S1. Collect data from the workshop production system and establish an initial digital model of the workshop production system; S2. Deploy IoT sensors to collect device data from IoT sensors in real time; S3. Input device data into the initial digital model, generate a dynamically updated digital model, and set scheduling optimization goals;

[0012] S4. Based on the scheduling optimization objective, use the cooperative particle swarm optimization algorithm to perform local variable grouping and parallel optimization on the digitally updated model, and generate an optimal production scheduling sub-scheme;

[0013] S5. Map each optimal production scheduling sub-scheme to an initial population, and apply the firefly algorithm to further optimize in the global search space according to the brightness attraction mechanism to generate a preliminary production scheduling scheme; S6. Conduct simulation verification on the preliminary production scheduling scheme. If the simulation verification result meets the preset evaluation criteria, it is determined as the target production scheduling scheme. If not, return to step S5 to regenerate the preliminary production scheduling scheme; S7. Send the target production scheduling scheme to the workshop production system for execution, and update the digitally updated model in real time; S8. Based on the digitally updated model in real time, monitor the execution effect. If it is detected that the execution deviation exceeds the preset threshold, trigger a correction.

[0014] Optionally, S3 specifically includes:

[0015] S31. Locate each equipment node in the initial digital model, update the attributes of the processing capacity, equipment load parameter, and unit time energy consumption of the equipment node to the current real-time data. Based on the processing capacity and equipment load parameter, calculate the equipment processing capacity utilization rate, load change rate, and energy consumption change rate. According to the preset weighting coefficient, perform a weighted sum of the processing capacity utilization rate, load change rate, and energy consumption change rate to generate a dynamic status label for the equipment node;

[0016] S32. Calculate the remaining processing time of the process node in the initial digital model;

[0017] S33. Predict the overall path residence time of the material in the initial digital model as the dynamic attribute of the material flow path edge;

[0018] S34. Integrate and write the dynamic status label of the equipment node, the remaining processing time of the process node, and the dynamic attribute of the material flow path edge into the corresponding node and edge attributes respectively to form a digitally updated model after dynamic update;

[0019] S35. Set a multi-objective scheduling optimization objective function and associate the set multi-objective scheduling optimization objective function with the digitally updated model after dynamic update.

[0020] Optionally, S4 specifically includes:

[0021] S41. Based on the set multi-objective scheduling optimization objective function, extract the processing capacity of the equipment node, equipment load parameter, unit time energy consumption, standard processing time of the process node, remaining processing time, and predicted residence time of the material flow path edge in the digitally updated model after dynamic update as a set of decision variables for scheduling optimization, denoted as ;

[0022] S42. Group the decision variable set locally according to the attribute category to form three variable subsets. The variables related to the equipment nodes form the equipment group , the variables related to the process nodes form the process group , and the variables related to the material flow path form the logistics group ;

[0023] S43. For each variable subset , and , initialize the corresponding particle swarm populations respectively. Let the particle position represent a scheduling sub-scheme, and the particle velocity represent the change trend of the scheduling variables. Each particle swarm independently initializes the population size, position range and velocity range;

[0024] S44. Inside each variable subset, use the cooperative particle swarm optimization algorithm for local optimization. Specifically: Each particle adjusts its flight speed according to its own historical best position and the current global best position, and updates its position. The update rule is: ;

[0025] Among them, is the position of particle at time , is the velocity of particle at time , is the position of particle at time , is the inertia weight, and are the learning factors, and are random numbers, is the historical best position of particle , is the current global best position of the particle swarm;

[0026] S45. During the local particle swarm optimization process, introduce the cooperation mechanism between variable subsets. By sharing part of the local best solution information, dynamically adjust the search directions of each subgroup, and the equipment scheduling, process flow and material flow co-evolve;

[0027] S46. After each round of particle swarm iteration is completed, evaluate the fitness of each particle in the particle swarm population according to the scheduling optimization objective function, screen out the local best scheduling sub-schemes of each subgroup, and integrate all local best scheduling sub-schemes to generate the optimal production scheduling sub-scheme.

[0028] Optionally, S5 specifically includes:

[0029] S51. Map the optimal production scheduling sub - scheme as an individual of the initial firefly population into the global search space, and each individual corresponds to a complete set of production scheduling scheme parameters;

[0030] S52. Assign a brightness value to each individual of the initial firefly population, and the brightness value is inversely proportional to the value of the scheduling optimization objective function corresponding to the individual;

[0031] S53. In the global search space, apply the improved firefly algorithm for iterative optimization. Specifically: in each round of iteration, for any two individuals and , if the brightness of individual is higher than that of individual , then individual adjusts its position according to the following improved position update formula: ;

[0032] where, is the position of the th individual in the th iteration, is the position of the th individual in the th iteration, is the position of the th individual in the th iteration, is the initial attractiveness, is the light intensity absorption coefficient, is the th Euclidean distance between the th and the th individuals, is the historical optimal position recorded by the individual during the global optimization process, and the individual is the complete production scheduling scheme individual after mapping the optimal production scheduling sub - scheme, is the incremental vector between individual and the current global optimal position, is the historical guidance weighting factor, and are the local perturbation and global perturbation coefficients respectively, and are the local perturbation term and global perturbation term respectively, following different - scale random distributions;

[0033] S54. After each update of the individual position, recalculate the brightness values of all population individuals, and continue to execute the attraction and movement process based on the brightness values until the set maximum number of iterations is reached or the global convergence condition is satisfied;

[0034] S55. After the optimization of the improved firefly algorithm is completed, screen the individual with the highest brightness value, and determine the corresponding production scheduling plan as the preliminary production scheduling plan.

[0035] Optionally, the preliminary production scheduling plan in step S55 is based on the dynamically updated digital model, with the optimal production scheduling sub-plan as the initial population, applying the improved firefly algorithm to optimize through multiple rounds of iteration. During the optimization process, the multi-objective optimization weight coefficients are adjusted in real time according to the changes in the production environment state, dynamically weighing the objectives of minimizing the delivery period, balancing the equipment load, and minimizing the energy consumption. Based on the brightness value, the comprehensive scheduling result is generated. The preliminary production scheduling plan includes the following three types of scheduling strategies: equipment processing capacity allocation strategy, reasonably allocating tasks and execution periods among equipment to improve resource utilization rate and balance the load; process processing sequence optimization strategy, dynamically determining the execution sequence and parallel relationship of processing nodes in the process flow path to shorten the overall production cycle; logistics transfer priority adjustment strategy, dynamically adjusting the transfer priority according to the current state of materials at the path nodes to reduce the residence time and energy consumption.

[0036] Optionally, step S6 specifically includes:

[0037] S61. Input the parameter set of the preliminary production scheduling plan into the dynamically updated digital model, call the simulation module built in the dynamically updated digital model, and perform simulation deduction on the execution effect of the preliminary production scheduling plan within the set simulation cycle. The simulation process is driven based on the dynamic attributes of equipment node processing capacity, process node processing sequence, logistics path transfer priority, and unit time energy consumption;

[0038] S62. After the simulation module completes the simulation, extract the simulation result data set, and the simulation result data set includes: total task completion time, average equipment load rate, total energy consumption, and average material residence time;

[0039] S63. Input the simulation result data set into the preset evaluation function, perform normalization processing according to the proportion of the total task completion time to the set maximum task time, the proportion of the total energy consumption to the set maximum energy consumption, the deviation proportion of the average equipment load rate from the ideal load rate, and the proportion of the average material residence time to the set maximum residence time. According to the preset weight coefficients, perform weighted summation on each normalized index to obtain a comprehensive evaluation value. The sum of the weight coefficients of each index is 1, and the comprehensive evaluation value measures the overall level of the preliminary production scheduling plan.

[0040] S64. Compare the comprehensive evaluation value with the set evaluation threshold If the comprehensive evaluation value is less than or equal to the evaluation threshold then confirm the preliminary production scheduling plan as the target production scheduling plan; if the comprehensive evaluation value is greater than the evaluation threshold then, based on the current preliminary production scheduling plan, return to the improved firefly algorithm optimization process, re-perform global scheduling optimization based on the simulation feedback information of the previous round, and generate a new preliminary production scheduling plan.

[0041] Optionally, the S8 specifically includes:

[0042] S81. Send the target production scheduling plan to the workshop production system, and control the equipment nodes, process nodes, and logistics transfer paths to perform actual production according to the equipment processing capacity allocation strategy, process processing sequence optimization strategy, and logistics transfer priority adjustment strategy set in the target production scheduling plan;

[0043] S82. During the production execution process, through the Internet of Things sensors and the production management system, collect the real-time data of the processing load of the equipment nodes , the real-time data of the processing progress of the process nodes , and the real-time data of the transfer status of the logistics path nodes , and synchronously update them to the digitally modeled after dynamic update;

[0044] S83. In the digitally modeled after dynamic update, periodically calculate the actual execution status deviation , and the deviation is defined as: ;

[0045] where represents the actual execution status deviation at time , is the deviation type index, represents the processing load deviation, represents the processing progress deviation, represents the logistics transfer deviation, is the actual collected th type of status data at time , corresponding to , , is the th type of expected status data set in the target production scheduling plan at time , corresponding to the expected data of the processing load of the equipment nodes , the expected data of the processing progress of the process nodes , the expected data of the transfer status of the logistics path nodes is the static weighting coefficient of the type of status deviation in the comprehensive deviation calculation, reflecting the influence weight of different deviations on the overall deviation, and satisfying , is the absolute value of the instantaneous static deviation of the type of status, is the static deviation amplification index, , is the weighted coefficient of the type of status change rate deviation, is the first derivative of the actual status of the type changing with time, representing the status change rate, is the absolute value of the type of status change rate, ;

[0046] S84. Compare the actually executed status deviation amount obtained by periodic calculation with the preset deviation tolerance threshold . If it satisfies , then continue to execute according to the target production scheduling plan . If there exists any moment such that , then based on the currently dynamically updated digital model and the real-time execution status, without changing the overall production scheduling framework, by locally adjusting the processing capacity parameters of equipment nodes, the processing order of process nodes, and the transfer priority of logistics path nodes, perform local fine-tuning and correction, and continue to execute the target production scheduling plan after dynamic correction.

[0047] A workshop production scheduling optimization system based on a digital model according to an embodiment of the present invention includes the following modules:

[0048] A data acquisition module for acquiring static structure data and real-time operation data of the workshop production system;

[0049] A modeling module for constructing an initial digital model based on the acquired data and dynamically updating it in combination with real-time data to form a digital model reflecting the current production status;

[0050] A scheduling optimization module for setting multi-objective scheduling optimization goals and generating a preliminary production scheduling plan using a cooperative particle swarm optimization algorithm and an improved firefly algorithm;

[0051] A simulation verification module for performing simulation and multi-dimensional evaluation on the preliminary production scheduling plan to determine whether it meets the preset criteria;

[0052] The execution control module is used to send the target production scheduling plan to the workshop system for execution and update the model status in real time;

[0053] The deviation monitoring module is used to monitor the actual execution deviation and trigger local scheduling correction when the tolerance is exceeded to ensure the dynamic and stable operation of the system.

[0054] The beneficial effects of the present invention are:

[0055] (1) The present invention constructs a high-precision digital workshop model that integrates static structure and dynamic attributes, which can reflect multi-dimensional information such as equipment status, process progress, and logistics path in real time, realizing the whole-process visualization and digital mapping of the production system. This model supports real-time update of multi-source heterogeneous data, overcomes the problem of slow response of traditional static models to on-site changes, and enhances the adaptability of the scheduling system to the actual production environment.

[0056] (2) The present invention introduces a two-stage scheduling optimization mechanism that combines cooperative particle swarm optimization and improved firefly algorithm. First, it improves the calculation efficiency through local grouped parallel optimization, and then achieves fine-grained search for the optimal scheduling plan by means of the global brightness attraction mechanism. This optimization mechanism can take into account multiple objectives such as the shortest delivery time, balanced equipment load, and minimum energy consumption at the same time, breaking through the limitations of traditional scheduling methods with single objective and weak adaptability.

[0057] (3) The present invention establishes a closed-loop scheduling control system from simulation verification to execution feedback, screens the optimal solution through simulation evaluation of the preliminary plan, and monitors the system operation status in real time based on the comprehensive deviation amount during the execution process. If the deviation exceeds the limit, the local scheduling parameters are corrected instead of recalculating the whole, ensuring the robustness and continuous optimization ability of the scheduling system, and effectively improving the scheduling stability and energy conservation in complex workshop environments. Description of the Drawings

[0058] The drawings are used to provide further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0059] Figure 1 is the flowchart of a workshop production scheduling optimization method based on a digital model proposed by the present invention;

[0060] Figure 2 is the cooperative particle swarm optimization structure diagram of a workshop production scheduling optimization method based on a digital model proposed by the present invention;

[0061] Figure 3 is the flowchart of the improved firefly algorithm of a workshop production scheduling optimization method based on a digital model proposed by the present invention. Detailed Embodiments

[0062] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, and therefore only showing the components related to the present invention.

[0063] Reference Figures 1 - 3 , a workshop production scheduling optimization method and system based on a digital model, comprising the following steps:

[0064] S1. Collect data of the workshop production system and establish an initial digital model of the workshop production system; S2. Deploy Internet of Things sensors to collect device data of the Internet of Things sensors in real time; S3. Input the device data into the initial digital model to generate a dynamically updated digital model, and set scheduling optimization objectives;

[0065] S4. According to the scheduling optimization objectives, use the cooperative particle swarm optimization algorithm to perform local variable grouping and parallel optimization on the dynamically updated digital model to generate an optimal production scheduling sub-scheme; S5. Map each optimal production scheduling sub-scheme to an initial population, and apply the firefly algorithm to further optimize in the global search space according to the brightness attraction mechanism to generate a preliminary production scheduling scheme; S6. Perform simulation verification on the preliminary production scheduling scheme. If the simulation verification result meets the preset evaluation criteria, it is determined as the target production scheduling scheme. If not, return to step S5 to regenerate the preliminary production scheduling scheme; S7. Send the target production scheduling scheme to the workshop production system for execution, and update the dynamically updated digital model in real time; S8. Based on the digitally modeled in real time, monitor the execution effect. If it is detected that the execution deviation exceeds the preset threshold, trigger a correction.

[0066] The present invention constructs a full-process scheduling optimization process covering data collection, dynamic modeling, intelligent optimization, simulation verification, and closed-loop execution monitoring. Compared with the existing static rules or single-stage optimization methods, the present invention realizes the real-time, intelligent, and flexible production scheduling in a multi-stage linkage manner, has strong robustness and adaptability, significantly improves the scheduling efficiency and resource utilization level, and provides a systematic solution for scheduling problems in complex manufacturing environments.

[0067] In this embodiment, S1 specifically includes: collecting data of the workshop production system, including equipment numbers, equipment types, processing capabilities, equipment load parameters, process step sequences, standard processing times, material source nodes, material destination nodes, material transfer paths, and unit time energy consumption data. Among them, the equipment numbers, equipment types, and processing capabilities describe the production resource status, the process step sequences and standard processing times describe the process flow constraints, the material source nodes, material destination nodes, and material transfer paths describe the material transfer relationships, and the unit time energy consumption data describes the equipment energy consumption status. Based on the collected data, an initial digital model of the workshop production system is established using the associated modeling method of nodes and edges. Here, nodes represent equipment or process steps, and the equipment numbers, equipment types, processing capabilities, equipment load parameters, process step sequences, standard processing times, and unit time energy consumption data are recorded in the node attributes. Edges represent material transfer paths, and the material source nodes, material destination nodes, and transfer path information are recorded in the edge attributes.

[0068] By defining refined parameters of the workshop production system and constructing an initial digital model in a structured manner of nodes and edges, the present invention realizes the unified modeling of multiple elements such as equipment, processes, and logistics. Compared with the limitations of traditional modeling methods that ignore material transfer paths or energy consumption factors, the present invention incorporates the unit time energy consumption into the edge attributes of the model, effectively supporting subsequent multi-objective optimization and enhancing the expression ability of the modeling level for actual scheduling logic and energy efficiency control.

[0069] In this embodiment, S2 specifically includes: in the workshop production system, the actual operating status, current processing capability, current load parameter, and current unit time energy consumption of the equipment corresponding to the equipment number are collected in real time, the process execution status and current processing time corresponding to the process step sequence are collected in real time, and the current position and transfer status of the material between the material source node and the material destination node are collected in real time. The status data collected in real time are respectively synchronously updated to the production resource status, material transfer relationship, and equipment energy consumption status in the initial digital model. The initial digital model has real-time dynamic attributes consistent with the actual workshop status before scheduling optimization.

[0070] By deploying Internet of Things sensing devices and management systems, the present invention collects the equipment status, process execution, and logistics transfer status at high frequency in real time and synchronously updates them to the digital model. Compared with the traditional periodic offline data update method, the present invention constructs a dynamic model with a real-time synchronization mechanism, enabling the scheduling algorithm to respond based on the latest status, and significantly improving the timeliness and accuracy of scheduling decisions.

[0071] In this embodiment, S3 specifically includes:

[0072] S31. Based on the device numbers collected in real time, locate each device node correspondingly in the initial digital model, update the attributes of the processing capacity, device load parameters, and energy consumption per unit time of the device node to the current real-time data. Based on the processing capacity and device load parameters, calculate the utilization rate of the device processing capacity, where the utilization rate of the processing capacity is the ratio of the device load parameter to the processing capacity. Based on the change of the device load parameter at two consecutive moments, calculate the load change rate. Based on the change of the energy consumption per unit time at two consecutive moments, calculate the energy consumption change rate. According to the preset weighting coefficients, perform a weighted sum of the utilization rate of the processing capacity, the load change rate, and the energy consumption change rate to generate a dynamic status label for the device node, which characterizes the changing trend of the comprehensive operating status of the device node;

[0073] S32. Based on the process step sequence collected in real time, locate each process node correspondingly in the initial digital model. Combine the standard processing time and the real-time completed processing time of the process node to calculate the remaining processing time of the process node. The remaining processing time is obtained by subtracting the completed processing time from the standard processing time, which reflects the current processing progress status of the process node in real time;

[0074] S33. Based on the material source node and the material destination node collected in real time, locate each material transfer path edge correspondingly in the initial digital model. Synchronously update the material transfer path and the current position data to the corresponding edge attributes. According to the path structure between the current position of the material and the material destination node, calculate the basic material transfer time, where the basic time is the distance from the current position to the destination node divided by the current transfer speed. At the same time, based on the material queuing quantity at each transfer node and the node's processing capacity per unit time, calculate the queuing residence time of the node. Add up the residence times of each node to predict the overall path residence time of the material, which is used as the dynamic attribute of the material transfer path edge;

[0075] S34. Integrate and write the dynamic status label of the device node, the remaining processing time of the process node, and the dynamic attributes of the material transfer path edge into the corresponding node and edge attributes respectively to form a dynamically updated digital model. The dynamically updated digital model combines the static basic attributes and the dynamically changing attributes of the workshop production system, and reflects the device status, process progress, and logistics transfer status in real time;

[0076] S35. Set a multi-objective scheduling optimization objective function, where the scheduling optimization objective function includes the objective of minimizing the delivery time, the objective of balancing the device load, and the objective of minimizing the energy consumption. Among them, the objective of minimizing the delivery time is optimized in the direction of minimizing the total task completion time of the process node, the objective of balancing the device load is optimized in the direction of minimizing the variance of the device node load parameters, and the objective of minimizing the energy consumption is optimized in the direction of minimizing the cumulative value of the energy consumption per unit time of the device node. Associate the set multi-objective scheduling optimization objective function with the dynamically updated digital model.

[0077] By extracting the key variables in the model after real-time update, calculating dynamic indicators such as processing capacity utilization rate, load change rate, and path residence time, and modeling based on a unified scheduling objective function, the present invention provides an accurate data basis for subsequent intelligent optimization. Compared with the traditional optimization method that relies on static parameters or a single objective function, the present invention constructs an optimization prestructure that integrates dynamic perception and multi-objective drive, improving the scientificity and applicability of the scheduling scheme generation.

[0078] In this embodiment, step S4 specifically includes:

[0079] S41. Based on the set multi-objective scheduling optimization objective function, extract the processing capacity of equipment nodes, equipment load parameters, energy consumption per unit time, standard processing time of process nodes, remaining processing time, and expected residence time of the material flow path edge in the digitally updated model after dynamic update as the decision variable set for scheduling optimization, denoted as ;

[0080] S42. Group the decision variable set locally according to the attribute category to form three variable subsets. The variables related to equipment nodes form the equipment group , the variables related to process nodes form the process group , and the variables related to the material flow path form the logistics group ;

[0081] S43. For each variable subset , and , initialize the corresponding particle swarm population respectively. Let the particle position represent a sub-scheduling scheme, and the particle velocity represent the change trend of the scheduling variable. Each particle swarm independently initializes the population size, position range, and velocity range;

[0082] S44. Inside each variable subset, use the cooperative particle swarm optimization algorithm for local optimization. Specifically: each particle adjusts its flight speed according to its own historical best position and the current global best position, updates its position, and the update rule is: ;

[0083] where is the position of particle at time , is the velocity of particle at time , is the position of particle at time , is the inertia weight, and are the learning factors, and are random numbers, is the historical optimal position of particle , and is the current global optimal position of the particle swarm;

[0084] S45. During the local particle swarm optimization process, introduce the cooperation mechanism between variable subsets. By sharing part of the local optimal solution information, dynamically adjust the search directions of each subgroup, and the equipment scheduling, process flow, and material flow evolve collaboratively;

[0085] S46. After each round of particle swarm iteration is completed, evaluate the fitness of each particle in the particle swarm population according to the scheduling optimization objective function, screen out the local optimal scheduling sub-schemes of each subgroup, and integrate all local optimal scheduling sub-schemes to generate the optimal production scheduling sub-scheme.

[0086] The present invention proposes a cooperative particle swarm optimization method, which divides the scheduling variables into equipment, process, and logistics subsets according to functional attributes, constructs sub-particle swarms respectively for parallel local optimization, and then generates sub-optimal solutions by sharing optimization information through the cooperation mechanism between subgroups. This method significantly enhances the search efficiency and local search depth, breaks through the bottleneck that the single-group particle search is easy to fall into local optimum, and realizes the distributed optimization strategy of efficient scheduling.

[0087] In this embodiment, the specific steps of S5 are as follows:

[0088] S51. Take the optimal production scheduling sub-scheme as the initial population individuals of fireflies, map them to the global search space, and each individual corresponds to a complete set of production scheduling scheme parameters;

[0089] S52. Assign a brightness value to each initial population individual of fireflies, and the brightness value is in an inverse relationship with the value of the scheduling optimization objective function corresponding to the individual;

[0090] S53. In the global search space, apply the improved firefly algorithm for iterative optimization. Specifically, in each round of iteration, for any two individuals and , if the brightness of individual is higher than that of individual , then individual adjusts its position according to the following improved position update formula: ;

[0091] where is the position of the th individual at the The position of the i-th iteration, is the position of the k-th individual at the i-th iteration, is the position of the m-th individual at the i-th iteration, is the initial attractiveness, is the light intensity absorption coefficient, is the Euclidean distance between the n-th and the p-th individuals, is the incremental vector between the individual and the current global optimal position, is the historical guiding weighting factor, 、 are the local perturbation and global perturbation coefficients respectively, 、 are the local perturbation term and global perturbation term respectively, following random distributions of different scales;

[0092] S54. After each update of the individual position, recalculate the brightness values of all population individuals, and continue to execute the attraction and movement process according to the brightness values until the set maximum number of iterations is reached or the global convergence condition is satisfied;

[0093] S55. After the optimization of the improved firefly algorithm is completed, select the individual with the highest brightness value, and the production scheduling plan corresponding to the individual is determined as the preliminary production scheduling plan.

[0094] Based on the sub-optimal solution, the present invention constructs an improved firefly algorithm with a brightness attraction mechanism as the core, and integrates local perturbation, global guidance and dynamic brightness function to achieve in-depth optimization of the global scheduling plan. This method introduces a richer dynamic adjustment mechanism and a fine search strategy compared with the traditional firefly algorithm, significantly improving the global quality and convergence stability of the scheduling solution, and is particularly suitable for global scheduling optimization in complex production constraint environments.

[0095] In this embodiment, the preliminary production scheduling plan in step S55 is based on the dynamically updated digital model. Using the optimal production scheduling sub-plan as the initial population, an improved firefly algorithm is applied to optimize through multiple rounds of iteration. During the optimization process, the multi-objective optimization weight coefficients are adaptively adjusted in real time according to the changes in the production environment state, dynamically weighing the objectives of minimizing the delivery period, balancing the equipment load, and minimizing the energy consumption. Based on the brightness value, the generated comprehensive scheduling results are screened. The preliminary production scheduling plan includes the following three types of scheduling strategies: the equipment processing capacity allocation strategy, which reasonably distributes tasks and execution periods among equipment to improve resource utilization and balance the load; the process processing sequence optimization strategy, which dynamically determines the execution sequence and parallel relationship of processing nodes in the process flow path to shorten the overall production cycle; and the logistics transfer priority adjustment strategy, which dynamically adjusts the transfer priority according to the current state of materials at the path nodes to reduce the residence time and energy consumption.

[0096] In the present invention, by introducing three types of scheduling strategies in the construction process of the preliminary scheduling plan, namely equipment processing capacity allocation, process sequence optimization, and logistics priority adjustment, the generated scheduling plan not only optimizes the task arrangement but also takes into account resource allocation and logistics rhythm, forming a multi-dimensional collaborative scheduling system. Compared with the traditional scheduling method based only on time or load, the present invention significantly improves the overall operation coordination and execution efficiency of the system.

[0097] In this embodiment, step S6 specifically includes:

[0098] S61. Input the parameter set of the preliminary production scheduling plan into the dynamically updated digital model, call the simulation module built in the dynamically updated digital model, and perform simulation deduction on the execution effect of the preliminary production scheduling plan within the set simulation period. The simulation process is driven based on the equipment node processing capacity, process node processing sequence, logistics path transfer priority, and dynamic attributes of energy consumption per unit time;

[0099] S62. After the simulation module completes the simulation, extract the simulation result data set, and the simulation result data set includes: the total task completion time, the average equipment load rate, the total energy consumption, and the average material residence time;

[0100] S63. Input the simulation result data set into a preset evaluation function, perform normalization processing according to the proportion of the total task completion time to the set maximum task time, the proportion of the total energy consumption to the set maximum energy consumption, the deviation ratio of the average equipment load rate to the ideal load rate, and the proportion of the average material residence time to the set maximum residence time. Perform weighted summation on each normalized index according to the preset weight coefficients to obtain a comprehensive evaluation value. The sum of the weight coefficients of each index is 1, and the comprehensive evaluation value measures the overall level of the preliminary production scheduling plan;

[0101] S64. Compare the comprehensive evaluation value with the set evaluation threshold If the comprehensive evaluation value is less than or equal to the evaluation threshold , then confirm the preliminary production scheduling plan as the target production scheduling plan; if the comprehensive evaluation value is greater than the evaluation threshold , then based on the current preliminary production scheduling plan, return to the improved firefly algorithm optimization process, re-perform global scheduling optimization based on the simulation feedback information of the previous round, and generate a new preliminary production scheduling plan.

[0102] In the present invention, by introducing a simulation module into the dynamic model and constructing an evaluation function using four-dimensional indicators of task completion time, load, energy consumption, and residence time, the preliminary plan is verified comprehensively. This simulation-evaluation-feedback mechanism enables the scheduling optimization process to have the ability of self-judgment and a quantitative basis for evaluating the quality of the plan. Compared with the traditional method that relies on expert experience to judge the feasibility of scheduling, the present invention realizes the automation, quantification, and iterative optimization of the scheduling evaluation process.

[0103] In this embodiment, the S8 specifically includes:

[0104] S81. Send the target production scheduling plan to the workshop production system, and control the equipment nodes, process nodes, and logistics transfer paths to perform actual production according to the equipment processing capacity allocation strategy, process processing sequence optimization strategy, and logistics transfer priority adjustment strategy set in the target production scheduling plan.

[0105] S82. During the production execution process, through the Internet of Things sensors and the production management system, real-time collect the real-time data of the processing load of the equipment nodes , the real-time data of the processing progress of the process nodes , and the real-time data of the transfer status of the logistics path nodes , and synchronously update them to the digitally updated digital model.

[0106] S83. In the digitally updated digital model, periodically calculate the actual execution status deviation , and the deviation is defined as:[[]] ;

[0107] Wherein,[[]] represents the actual execution status deviation at time , is the deviation type index,[[]] represents the processing load deviation,[[]] represents the processing progress deviation,[[]] represents the logistics transfer deviation,[[]] is the actual collected th type of status data at time , corresponding to respectively , , At time , the -th type of expected state data set in the target production scheduling plan respectively corresponds to the expected processing load data of the equipment node , the expected processing progress data of the process node , and the expected transfer status data of the logistics path node The -th type of static weighting coefficient in the comprehensive deviation calculation of the state deviation reflects the influence weight of different deviations on the overall deviation and satisfies . The -th type of absolute value of the instantaneous static deviation of the state , and is the static deviation amplification index . The -th type of state change rate deviation weighting coefficient . The -th type of first derivative of the actual state changing with time represents the state change rate . The -th type of absolute value of the state change rate , and is the dynamic change rate deviation amplification index

[0108] S84. Compare the actually executed state deviation amount calculated periodically with the preset deviation tolerance threshold . If is satisfied, continue to execute according to the target production scheduling plan . If there exists any time such that , then based on the currently dynamically updated digital model and the real-time execution state, without changing the overall production scheduling framework, perform local fine-tuning corrections by locally adjusting the processing capacity parameters of the equipment node, the processing order of the process node, and the transfer priority of the logistics path node, and continue to execute the target production scheduling plan after the dynamic correction.

[0109] The present invention proposes to introduce a real-time calculation mechanism for deviation amounts in the execution stage of the scheduling plan, construct a non-linear comprehensive deviation function based on multiple types of state data, and trigger a local fine-tuning mechanism for strategy correction if it is found that the deviation exceeds the threshold to avoid global recalculation. Compared with the problem that the traditional scheduling system lacks a feedback mechanism in the execution stage, the present invention constructs an adaptive operation correction ability, enables the scheduling system to have real-time response and continuous optimization characteristics, and enhances the stability and intelligence of the system operation.

[0110] A workshop production scheduling optimization system based on a digital model according to an embodiment of the present invention includes the following modules:

[0111] A data collection module for collecting static structure data and real-time operation data of the workshop production system;

[0112] A modeling module for constructing an initial digital model based on the collected data and dynamically updating it in combination with real-time data to form a digital model reflecting the current production status;

[0113] A scheduling optimization module for setting multi-objective scheduling optimization goals and generating a preliminary production scheduling plan using a cooperative particle swarm optimization algorithm and an improved firefly algorithm;

[0114] A simulation verification module for simulating and multi-dimensionally evaluating the preliminary production scheduling plan to determine whether it meets the preset criteria;

[0115] An execution control module for issuing the target production scheduling plan to the workshop system for execution and updating the model status in real time;

[0116] A deviation monitoring module for monitoring actual execution deviations and triggering local scheduling corrections when the tolerance is exceeded to ensure the dynamic stable operation of the system.

[0117] Example 1:

[0118] In the context of the continuous promotion of intelligent manufacturing, an automotive parts manufacturing enterprise A deployed multiple discrete production lines in the automated assembly workshop located in a certain province of our country, mainly producing core structural parts such as differential cases and brake brackets. This workshop has a large number of equipment, complex processing technologies, and frequent process switches, often resulting in problems such as long task waiting times, low equipment utilization rates, and redundant energy consumption due to unreasonable scheduling plans, seriously affecting the order delivery cycle and energy cost control.

[0119] To solve the above problems, enterprise A introduced the workshop production scheduling optimization method and system based on a digital model proposed by the present invention and deployed it on the joint architecture of the MES system and the industrial edge server for implementation and application. During the project implementation process, first, information such as the equipment numbers, types, processing capabilities, load parameters, historical energy consumption, and material flow paths of 10 core machining centers in the workshop was collected to construct an initial digital model. At the same time, high-precision power acquisition modules, real-time task tracking devices, and logistics label readers were installed on the equipment to achieve the input of all real-time data required for the dynamic model.

[0120] Through the deployed system, the enterprise can update the device status, process progress, and material logistics transfer in real time. The model automatically refreshes the status tags every 30 seconds and generates a scheduling optimization function based on the triple optimization goals of the shortest delivery time, equipment load balancing, and minimum energy consumption. Subsequently, the system uses the cooperative particle swarm algorithm to optimize the device variables, process variables, and logistics path variables in parallel, outputs three types of sub-optimal solutions, and then uses the improved firefly algorithm to perform global optimization to generate the final scheduling plan. The system simulates and validates each candidate solution and selects the optimal solution for execution based on the comprehensive score value.

[0121] Through a two-week on-site trial operation, the enterprise separately counted the key indicators of the workshop before and after optimization. From the results, without changing the equipment layout and product process, through the scheduling of the system of the present invention, the average load rate of 10 devices increased by about 12%, the average task completion time was shortened by about 3.7 minutes, the average energy consumption per unit time decreased by 0.42 kWh, and the average residence time of materials at the production line nodes decreased from the original range of 8 - 13 minutes to within 4 - 7 minutes, effectively alleviating the problems of production line congestion and material stacking during peak hours. The enterprise management feedback that the system has significantly improved the resource utilization rate and reduced the operating cost, and supports replication and deployment on more production lines in the future.

[0122] The following are the key data recorded during the application verification process:

[0123] Table 1: Comparison Table of the Implementation Effects of Digital Scheduling Optimization

[0124] It can be seen from the data in the table that after introducing the scheduling optimization system of the present invention, the task execution efficiency and energy consumption performance of 10 key devices have been significantly improved. The average task completion time of equipment numbers E001 to E010 was generally concentrated between 18.5 and 25.8 minutes before optimization, and generally decreased to about 18 minutes after optimization, with an average reduction of about 3.7 minutes. Especially for equipment E006, its average task completion time decreased from 24.17 minutes to 20.91 minutes, indicating that the scheduling scheme has good universality in improving the efficiency of task allocation.

[0125] In terms of load balancing, the average load rate of the devices after optimization has generally increased. Especially for equipment E003 and E009, they reached 94.57% and 93.79% respectively after optimization, compared with 79.28% and 85.71% before optimization, showing a significant increase. The improvement of the overall load level means a significant enhancement of resource utilization rate and a more sufficient release of production capacity, avoiding the waste of production capacity caused by the low-load operation of individual devices.

[0126] In terms of energy consumption indicators, the energy consumption per unit time after optimization has decreased. For example, the energy consumption per unit time of equipment E005 before optimization was 3.44 kWh, and it decreased to 2.93 kWh after optimization, with an average decrease of more than 14%. On the one hand, this benefits from the introduction of the energy consumption minimization objective function in the scheduling strategy of the present invention. On the other hand, it also stems from avoiding high-energy-consuming operations such as frequent task switching and idling waiting in the scheduling, which overall improves the energy utilization efficiency.

[0127] The improvement in logistics detention is also particularly obvious. Under the original system, the average residence time of materials at the production line nodes was concentrated between 8 and 13 minutes, which was prone to material stacking and queuing phenomena, affecting the start of subsequent processes. After optimization, the detention time generally decreased to the range of 4 to 7 minutes. Among them, the detention time of equipment E002 decreased from 12.89 minutes to 5.93 minutes, a decrease of more than 50%, reflecting the remarkable effectiveness of the logistics transfer priority strategy in improving the circulation speed.

[0128] Generally speaking, the present invention has achieved the comprehensive goals of "efficiency improvement, energy conservation, and detention reduction" in the manufacturing enterprise scenario, verifying its technical advantages of accurate scheduling modeling, efficient optimization mechanism, and timely feedback adjustment, and has extremely high promotion value and engineering adaptability. Especially in manufacturing scenarios with complex production rhythms and high resource coupling degrees, this system can be used as a standard scheduling intelligent upgrade solution to significantly improve the flexible manufacturing ability and operational efficiency of enterprises.

[0129] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. An optimization method for workshop production scheduling based on a digital model, characterized in that, Including: S1. Collect the data of the workshop production system and establish an initial digital model of the workshop production system; S2. Deploy Internet of Things sensors to collect the device data of the Internet of Things sensors in real time; S3. Input the device data into the initial digital model to generate a digitally modeled model after dynamic update, and set the scheduling optimization goal; S4. According to the scheduling optimization goal, use the cooperative particle swarm optimization algorithm to perform local variable grouping and parallel optimization on the digitally modeled model after dynamic update to generate an optimal production scheduling sub-scheme; S5. Map each optimal production scheduling sub-scheme to the initial population, and apply the firefly algorithm to further optimize in the global search space according to the brightness attraction mechanism to generate a preliminary production scheduling scheme; S6. Perform simulation verification on the preliminary production scheduling scheme. If the simulation verification result meets the preset evaluation criteria, it is determined as the target production scheduling scheme. If not, return to step S5 to regenerate the preliminary production scheduling scheme; S7. Send the target production scheduling scheme to the workshop production system for execution and update the digitally modeled model after dynamic update in real time; S8. Based on the digitally modeled model after real-time update, monitor the execution effect. If it is detected that the execution deviation exceeds the preset threshold, trigger correction.

2. The optimization method for workshop production scheduling based on a digital model according to claim 1, characterized in that The specific content of S3 includes: S31. Locate each device node in the initial digital model, update the attributes of the processing capacity, device load parameter and unit time energy consumption of the device node to the current real-time data. Based on the processing capacity and device load parameter, calculate the utilization rate of device processing capacity, load change rate and energy consumption change rate. According to the preset weighting coefficient, perform weighted summation on the utilization rate of device processing capacity, load change rate and energy consumption change rate to generate a dynamic status label for the device node; S32. Calculate the remaining processing time of the process node in the initial digital model; S33. Predict the overall path residence time of the material in the initial digital model as the dynamic attribute of the material flow path edge; S34. Integrate and write the dynamic status label of the device node, the remaining processing time of the process node and the dynamic attribute of the material flow path edge into the corresponding node and edge attributes respectively to form a digitally modeled model after dynamic update; S35. Set a multi-objective scheduling optimization objective function and associate the set multi-objective scheduling optimization objective function with the digitally modeled model after dynamic update.

3. The optimization method for workshop production scheduling based on a digital model according to claim 2, wherein The specific content of S4 includes: S41. Based on the set multi-objective scheduling optimization objective function, extract the equipment node processing capacity, equipment load parameters, unit time energy consumption, standard processing time of process nodes, remaining processing time, and the expected residence time of the material flow path edges in the dynamically updated digital model as the decision variable set for scheduling optimization, denoted as ; S42. Group the decision variable set locally according to the attribute category to form three variable subsets. The variables related to the equipment nodes form the equipment group , the variables related to the process nodes form the process group , and the variables related to the material flow path form the logistics group ; S43. For each variable subset , and , initialize the corresponding particle swarm population respectively. Assume that the particle position represents a scheduling sub-scheme, and the particle velocity represents the change trend of the scheduling variable. Each particle swarm independently initializes the population size, position range and velocity range; S44. Inside each variable subset, use the cooperative particle swarm optimization algorithm for local optimization. Specifically: each particle adjusts its flight speed according to its own historical optimal position and the current global optimal position, updates its position, and the update rule is: ； wherein, is the position of the particle at time ; is the velocity of the particle at time ; is the position of the particle at time ; is the inertia weight, and are the learning factors, and are random numbers, is the historical best position of the particle ; is the current global best position of the particle swarm; S45. During the local particle swarm optimization process, introduce the cooperation mechanism between variable subsets. By sharing some local optimal solution information, dynamically adjust the search direction of each subgroup, and the equipment scheduling, process flow and material flow co-evolve; S46. After each round of particle swarm iteration is completed, perform fitness evaluation on each particle in the particle swarm population according to the scheduling optimization objective function, screen out the local optimal scheduling sub-schemes of each subgroup, and integrate all local optimal scheduling sub-schemes to generate an optimal production scheduling sub-scheme.

4. The optimization method for workshop production scheduling based on a digital model according to claim 3, wherein, The specific content of S5 includes: S51. Map the optimal production scheduling sub-scheme as an individual of the initial population of fireflies into the global search space, and each individual corresponds to a complete set of production scheduling scheme parameters; S52. Assign a brightness value to each individual of the initial population of fireflies, and the brightness value is inversely proportional to the value of the scheduling optimization objective function corresponding to the individual; S53. In the global search space, the improved firefly algorithm is applied for iterative optimization, specifically: in each round of iteration, for any two individuals and , if the brightness of individual is higher than that of individual , then individual adjusts its position according to the following improved position update formula: ； Among them, is the position of the th individual at the th iteration, is the position of the th individual at the th iteration, is the position of the th individual at the th iteration, is the initial attractiveness, is the light intensity absorption coefficient, is the th Euclidean distance between the th individual and the th individual, is the historical optimal position recorded by the th individual during the global optimization process, and the individual is the complete production scheduling plan individual mapped by the optimal production scheduling sub - scheme, is the th individual's incremental vector with the current global optimal position, is the historical guiding weighting factor, is the global guiding weighting factor, and are the local perturbation and global perturbation coefficients respectively, and are the local perturbation term and global perturbation term respectively, and they follow random distributions with different scales; S54. After each update of the individual position, recalculate the brightness values of all population individuals, and continue to execute the attraction and movement process according to the brightness values until the set maximum number of iterations is reached or the global convergence condition is satisfied; S55. After the optimization of the improved firefly algorithm is completed, select the individual with the highest brightness value, and the production scheduling scheme corresponding to the individual is determined as the preliminary production scheduling scheme.

5. The optimization method for workshop production scheduling based on a digital model according to claim 4, wherein The preliminary production scheduling scheme in step S55 is based on the dynamically updated digital model, uses the optimal production scheduling sub-scheme as the initial population, applies the improved firefly algorithm to perform multi-round iterative optimization, and dynamically adjusts the multi-objective optimization weight coefficients in real time according to the changes in the production environment state during the optimization process, dynamically weighs the objectives of the shortest delivery time, equipment load balancing, and minimum energy consumption, and generates a comprehensive scheduling result based on the brightness value. The preliminary production scheduling scheme includes the following three types of scheduling strategies: equipment processing capacity allocation strategy, which allocates tasks and execution time periods among equipment; Process processing sequence optimization strategy, which dynamically determines the execution sequence and parallel relationship of processing nodes in the process flow path; Logistics transfer priority adjustment strategy, which dynamically adjusts the transfer priority according to the current state of materials at the path nodes.

6. The optimization method for workshop production scheduling based on a digital model according to claim 5, wherein The specific steps of S6 are as follows: S61. Input the parameter set of the preliminary production scheduling scheme into the dynamically updated digital model, and simulate and deduce the execution effect of the preliminary production scheduling scheme within the set simulation period; S62. After the simulation module completes the simulation, extract the simulation result data set, and the simulation result data set includes: total task completion time, average equipment load rate, total energy consumption, average material residence time; S63. Input the simulation result data set into the preset evaluation function, perform normalization processing according to the proportion of the total task completion time to the set maximum task time, the proportion of the total energy consumption to the set maximum energy consumption, the deviation ratio of the average equipment load rate to the ideal load rate, and the proportion of the average material residence time to the set maximum residence time, and perform weighted summation on each normalized index according to the preset weight coefficients to obtain a comprehensive evaluation value. The sum of the weight coefficients of each index is 1, and the comprehensive evaluation value measures the overall level of the preliminary production scheduling scheme; S64. Compare the comprehensive evaluation value with the set evaluation threshold If the comprehensive evaluation value is less than or equal to the evaluation threshold then confirm the preliminary production scheduling plan as the target production scheduling plan; if the comprehensive evaluation value is greater than the evaluation threshold then, based on the current preliminary production scheduling plan, return to the improved firefly algorithm optimization process, re-perform global scheduling optimization based on the feedback information of the previous round of simulation, and generate a new preliminary production scheduling plan.

7. The optimization method for workshop production scheduling based on a digital model according to claim 6, characterized in that The specific steps of S8 are as follows: S81. Send the target production scheduling scheme to the workshop production system; S82. During the production execution process, real-time data on the processing load of equipment nodes, real-time data on the processing progress of process nodes, and real-time data on the transfer status of logistics path nodes are collected in real time through Internet of Things sensors and the production management system, and synchronously updated to the digitized model after dynamic update; Real-time data on the processing progress of process nodes Real-time data on the transfer status of logistics path nodes , and synchronously updated to the digitized model after dynamic update; S83. In the digitally updated model after dynamic update, periodically calculate the deviation of the actual execution status. , where the deviation is defined as: ； Among them, represents the actual execution status deviation at time , is the deviation type index, represents the processing load deviation, represents the processing progress deviation, represents the logistics transfer deviation, is the -th type of actual collected status data at time , corresponding respectively to , , is the -th type of expected status data set in the target production scheduling plan at time , corresponding respectively to the expected processing load data of the equipment node , the expected processing progress data of the process node , and the expected transfer status data of the logistics path node is the static weighting coefficient of the -th type of status deviation in the comprehensive deviation calculation, reflecting the influence weight of different deviations on the overall deviation, satisfying , is the absolute value of the instantaneous static deviation of the -th type of status, is the static deviation amplification index, , is the weighted coefficient of the -th type of status change rate deviation, is the -th type of first-order derivative of the actual status changing with time, representing the status change rate, is the absolute value of the -th type of status change rate, is the dynamic change rate deviation amplification index, ; S84. Compare the actually executed state deviation calculated periodically with the preset deviation tolerance threshold . If it satisfies , continue to execute according to the target production scheduling plan . If there exists any moment such that , then based on the currently dynamically updated digital model and the real-time execution state, without changing the overall production scheduling framework, perform local fine-tuning corrections by locally adjusting the processing capacity parameters of equipment nodes, the processing order of process nodes, and the transfer priority of logistics path nodes, and continue to execute the target production scheduling plan after dynamic correction.

8. A workshop production scheduling optimization system based on a digital model, which is applied to a workshop production scheduling optimization method according to any one of claims 1 to 7, and is characterized in that, It includes the following modules: Data acquisition module, which is used to acquire the static structure data and real-time operation data of the workshop production system; Modeling module, which is used to build an initial digital model based on the acquired data and dynamically update it in combination with real-time data to form a digital model reflecting the current production state; Scheduling optimization module, which is used to set multi-objective scheduling optimization objectives and generate a preliminary production scheduling scheme using the cooperative particle swarm optimization algorithm and the improved firefly algorithm; The simulation verification module is used to perform simulation and multi-dimensional evaluation on the preliminary production scheduling plan to determine whether it meets the preset standards; The execution control module is used to issue the target production scheduling plan to the workshop system for execution and update the model status in real time; The deviation monitoring module is used to monitor the actual execution deviation and trigger local scheduling correction when the tolerance is exceeded to ensure the dynamic and stable operation of the system.

Images