A hybrid intelligent decision-based PCB production real-time scheduling method and system

Abstract

The application provides a PCB production real-time scheduling method and system based on hybrid intelligent decision-making, the method comprising: acquiring production information of PCB production, and constructing a corresponding digital twin for the production information; identifying explicit process tasks and implicit process tasks in a production work order based on the state of the digital twin; acquiring a current business policy and loading a hard constraint rule set and a soft constraint rule set based on the current business policy; comprehensively integrating the explicit process tasks, the implicit process tasks, the hard constraint rule set and the soft constraint rule set to obtain a scheduling problem instance; converting the scheduling problem instance into an objective function, and performing multi-objective optimization solving on the objective function to output an initial scheduling plan; and performing real-time monitoring and dynamic incremental rescheduling circulation on the initial scheduling plan to complete the PCB production real-time scheduling. The application deeply integrates domain expert rules, operations research optimization algorithms and real-time data feedback in a closed-loop system to achieve dynamic self-adaptive optimization of production scheduling.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of industrial automation and information technology, and specifically relates to a PCB production real-time scheduling method and system based on hybrid intelligent decision-making. BACKGROUND

[0002] In the manufacturing field, production scheduling is a core link, and its purpose is to allocate production tasks (work orders) to production resources (machines) under the premise of meeting a series of constraints (such as delivery time, resources, process), so as to achieve one or more optimization goals (such as shortest production cycle, highest equipment utilization, lowest production cost). The technical means for realizing scheduling usually includes:

[0003] Manufacturing execution system (MES): mainly responsible for the execution, monitoring and management of the production process. Some MES systems contain simple scheduling modules, but their core is oriented to "execution" rather than "planning". Its scheduling function is usually based on local, preset dispatching rules (such as first-in-first-out FIFO, emergency work order priority), and lacks overall view and prediction ability of the global state of the factory, so its decision is often local and suboptimal.

[0004] Enterprise resource planning (ERP): focuses on enterprise-level resource macro planning, such as order management, material requirement planning (MRP), etc. Its scheduling function is usually a rough plan with "infinite capacity", or a bottleneck process evaluation based on "rough capacity planning (RCCP)", which cannot handle real-time constraints and dynamic changes of hundreds of specific resources such as machines, tooling, personnel, etc. in the workshop.

[0005] Advanced planning and scheduling (APS) system: a professional planning and scheduling software that bridges the gap between ERP macro planning and MES workshop execution. This kind of system usually uses operations research algorithms such as mathematical programming (such as mixed integer programming MIP) or constraint programming (CP) to model and solve scheduling problems containing complex constraints, in order to find the mathematically optimal solution.

[0006] For example, the patent with publication number CN113610233A discloses a "flexible job shop scheduling method based on improved genetic algorithm". The technical solution of this patent focuses on the specific design of the coding method, crossover and mutation operators of the genetic algorithm to improve the performance and convergence speed of the algorithm in solving the static flexible job shop scheduling problem (FJSP). However, this patent technical solution is essentially an offline optimization algorithm for static problem instances, and it does not disclose a system architecture that can respond to real-time disturbances in the production field, nor does it involve the deep integration of configurable business rules and optimization solving process. SUMMARY

[0007] In order to solve the above technical problems, the application provides a PCB production real-time scheduling method and system based on hybrid intelligent decision-making, which is used to solve the technical problems in the prior art.

[0008] In a first aspect, the application provides the following technical solution: a PCB production real-time scheduling method based on hybrid intelligent decision-making, comprising:

[0009] obtaining production information of PCB production, and constructing a corresponding digital twin for the production information;

[0010] identifying explicit process tasks and implicit process tasks in a production work order based on the state of the digital twin;

[0011] obtaining a current business policy and loading a hard constraint rule set and a soft constraint rule set based on the current business policy;

[0012] integrating the explicit process tasks, the implicit process tasks, the hard constraint rule set and the soft constraint rule set to obtain a scheduling problem instance;

[0013] converting the scheduling problem instance into an objective function, and performing multi-objective optimization solving on the objective function to output an initial scheduling plan;

[0014] performing real-time monitoring and dynamic incremental rescheduling on the initial scheduling plan to complete PCB production real-time scheduling.

[0015] Compared with the prior art, the application has the following beneficial effects: the application realizes a paradigm shift from 'offline batch processing' to 'online real-time decision-making', revolutionarily improves production agility, and the prior art is generally offline and static. Through real-time data sensing and incremental planning technology based on event triggering, the application shortens the decision-making cycle from 'hour level' to'minute level'. When machine failure or urgent order insertion occurs, the system can automatically generate a globally coordinated response plan, so that the production plan is changed from rigid instructions to dynamically adaptive 'living' entities, and the flexibility and market response ability of the production system are fundamentally improved.

[0016] The application significantly improves the quality and on-site feasibility of scheduling results: the application uses a hybrid rule engine to explicitly convert implicit expert knowledge into a series of quantifiable soft constraints, and deeply embeds them into the scoring function of the optimization algorithm. This enables the algorithm to 'understand' the subtle preferences of the business (such as reducing line changes as much as possible and balancing equipment load) at every step of optimization. The generated plan is not only mathematically optimal, but also more reasonable in production practice, effectively improving the comprehensive OEE (overall equipment effectiveness) and reducing production costs.

[0017] The present application greatly enhances the flexibility and maintainability of the system: the existing APS system rules are solidified, and the change of business process requires expensive and time-consuming secondary development, the present application adopts the design of decoupling business rules and core algorithm, and parameterizes and configures the rules and weights, so that business personnel (such as production managers) can adjust the weight of different optimization targets (such as delivery time and cost) through the interface, and quickly respond to changes in enterprise strategy.

[0018] The present application comprehensively improves the efficiency of scheduling work and the intelligent level of enterprise operation: the present application adopts advanced meta-heuristic optimization algorithm, which can find high-quality solutions in acceptable time when dealing with large-scale and high-complexity PCB scheduling problems. More importantly, the whole closed-loop process of "perception-decision-execution-feedback" is highly automated, which completely liberates the scheduling personnel from tedious, repetitive and error-prone manual planning and adjustment work, so that they can change into the "commander" of the production site, and significantly improve the intelligent level of the whole production operation.

[0019] Preferably, the step of constructing a corresponding digital twin for the production information of the PCB production includes:

[0020] Obtaining work order data, resource data, material and tool data, and real-time events of the PCB production to obtain production information;

[0021] Constructing a data structure for consumable tools in the production information, the data structure including a static ID and dynamically updated consumption data to obtain a micro digital twin model;

[0022] Constructing a macro digital twin model for other data in the production information, and integrating the micro digital twin model and the macro digital twin model to obtain a corresponding digital twin.

[0023] Preferably, the step of identifying explicit process tasks and implicit process tasks in the production work order based on the state of the digital twin includes:

[0024] Splitting the work order in the digital twin into a plurality of process tasks arranged according to the process route;

[0025] Using available resource information, available tool information and order information in the scheduling process as decision information, and associating the decision information with the process tasks to obtain explicit process tasks;

[0026] According to the process requirement analysis of the process task, if there is a process requirement analysis unqualified condition, an implicit task meeting the process requirement analysis is created to obtain an implicit process task.

[0027] Preferably, in the step of obtaining the current business strategy and loading the hard constraint rule set and the soft constraint rule set based on the current business strategy, the hard constraint rule set includes machine capacity constraints, process sequence constraints, resource exclusivity constraints, and tool availability constraints, and the soft constraint rule set includes delivery time optimization rules, line changeover minimization rules, load balancing rules, and machine optimization rules.

[0028] Preferably, the step of converting the scheduling problem instance into an objective function and performing multi-objective optimization on the objective function to output an initial scheduling plan includes:

[0029] Transform the scheduling problem instance into an objective function. :

[0030] ;

[0031] ; ;

[0032] ;

[0033] ;

[0034] ;

[0035] In the formula, These are respectively the order set, the machine set, and the process task set. , , , These represent the penalty items corresponding to delivery delay penalties, changeover time penalties, tool-task mismatch penalties, and other soft constraint rules, respectively. , Orders Start time, required delivery date, For the machine On the task Execute the task immediately upon completion. Required changeover time These are the first, second, and third binary variables, respectively. Representing tasks Start time, task End time, task End time, For orders Priority weights, For a set of priority relations, These are the first soft constraint weight, the second soft constraint weight, and the third soft constraint weight, respectively.

[0036] adding operators and solving the objective function using the operators with an adaptive learning mechanism to obtain an initial scheduling plan.

[0037] Preferably, in the step of adding operators and solving the objective function using the operators with an adaptive learning mechanism to obtain an initial scheduling plan, the operators include a batch merging moving operator, a tool-task collaborative moving operator, and a high-low precision task pair switching moving operator.

[0038] The batch merging moving operator is used to process tasks of the same type in batches and reduce tool change tasks caused by tool replacement, so as to reduce the value of the tool change time penalty in the objective function.

[0039] The tool-task collaborative moving operator is used to control the use of existing tools on the machine and avoid the generation of tool change tasks, so as to reduce the value of the tool change time penalty in the objective function.

[0040] The high-low precision task pair switching moving operator is used to identify task resource mismatch conditions in the current solution of the objective function, and exchange the resource allocation of tasks according to the resource mismatch conditions, so as to reduce the value of the tool-task mismatch penalty in the objective function.

[0041] Preferably, the step of monitoring the initial scheduling plan in real time and dynamically incrementally rescheduling to complete real-time scheduling of PCB production includes:

[0042] monitoring the state changes and external event streams of the production site in real time, and determining whether a rescheduling condition is triggered based on the state changes and the external event streams;

[0043] If the rescheduling condition is not triggered, production processing is performed according to the initial scheduling plan.

[0044] If the rescheduling condition is triggered, the digital twin is updated, all completed process tasks within a rolling frozen time window are locked, and the state changes and the external event streams are updated to the current scheduling problem instance.

[0045] The updated scheduling problem instance is converted into an updated objective function, and the first soft constraint weight, the second soft constraint weight, and the third soft constraint weight in the updated objective function are updated.

[0046] ;

[0047] ;

[0048] wherein, is the first soft constraint weight or the second soft constraint weight or the third soft constraint weight after the i-th rescheduling iteration, ​the first soft constraint weight, the second soft constraint weight, or the third soft constraint weight is updated, respectively represent the first , the second , the current error of the first respectively represent the proportional action weight, the integral action weight, and the differential action weight.

[0049] The weight of the operator is updated as follows:

[0050]

[0051] wherein, is a learning rate, is the weight of the updated operator, is the weight of the operator before updating, is the target function value of the operator in the current week;

[0052] An adaptive learning mechanism is used to solve the updated target function based on the updated first soft constraint weight, the second soft constraint weight, the third soft constraint weight, and the weight of the updated operator, to obtain an updated scheduling plan and publish the updated scheduling plan.

[0053] In a second aspect, the present application provides the following technical solution, a PCB production real-time scheduling system based on hybrid intelligent decision-making, the system comprises:

[0054] An acquisition module is configured to acquire production information of PCB production, and construct a corresponding digital twin for the production information.

[0055] A task module is configured to identify explicit process tasks and implicit process tasks in a production work order based on the state of the digital twin.

[0056] A constraint module is configured to acquire a current business policy and load a hard constraint rule set and a soft constraint rule set based on the current business policy.

[0057] An instance module is configured to synthesize the explicit process tasks, the implicit process tasks, the hard constraint rule set, and the soft constraint rule set to obtain a scheduling problem instance.

[0058] A solving module is configured to convert the scheduling problem instance into a target function, and perform multi-objective optimization solving on the target function to output an initial scheduling plan.

[0059] A loop module is configured to perform real-time monitoring and dynamic incremental rescheduling on the initial scheduling plan to complete the real-time scheduling of PCB production.

[0060] ​In a third aspect, the present application provides a computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the hybrid intelligent decision-based real-time scheduling method for PCB production when executing the computer program.

[0061] In a fourth aspect, the present application provides a storage medium, and the storage medium stores a computer program, and the computer program is executed by a processor to implement the hybrid intelligent decision-based real-time scheduling method for PCB production. BRIEF DESCRIPTION OF DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiments or prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0063] Figure 1 A flow chart of the hybrid intelligent decision-based real-time scheduling method for PCB production provided by the first embodiment of the present application;

[0064] Figure 2 A structural block diagram of the hybrid intelligent decision-based real-time scheduling system for PCB production provided by the second embodiment of the present application;

[0065] Figure 3 A hardware structure schematic diagram of the computer provided by another embodiment of the present application.

[0066] The embodiments of the present application will be further described below with reference to the drawings. DETAILED DESCRIPTION

[0067] The embodiments of the present application will be described in detail below, and examples of the embodiments are shown in the drawings, wherein the same or similar reference signs represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are intended to explain the embodiments of the present application, and cannot be understood as a limitation of the present application.

[0068] Embodiment one

[0069] In the first embodiment of the present application, as shown in Figure 1 A hybrid intelligent decision-based real-time scheduling method for PCB production comprises the following steps:

[0070] S1, obtaining production information of PCB production, and constructing a corresponding digital twin for the production information;

[0071] Specifically, the main sources of production information are:

[0072] MES / ERP / WMS system: provides work order data (process route, delivery date, priority), material inventory, etc.

[0073] Equipment IoT platform: provides real-time status of machines (running, idle, failure), capability parameters, historical processing yield, etc.

[0074] Tool / quality management system: provides available state, remaining useful life (RUL) and quality abnormal events of drills and other tools.

[0075] Meanwhile, the various constraints in the following text also come from the above platforms and systems.

[0076] The step S1 includes:

[0077] S11, obtaining work order data, resource data, material and tool data, and real-time events of PCB production to obtain production information;

[0078] The work order data includes work order number, product material number, quantity, process route (including alternative process), customer priority, promised delivery date, technical requirements (such as board thickness, copper thickness), etc. The resource data includes machine ID, machine group, capability parameters (such as drilling speed, accuracy level, pressing temperature range), current state (running, idle, failure, planned maintenance, line change), current processing task, assigned task queue, estimated available time point, historical processing yield, etc. The material and tool data includes inventory, location, and estimated arrival time of PCB substrates, chemical solutions, drills, and other materials; and available state, life, and location of drills, film, and other tools. Real-time events include machine failure alarms, quality abnormal stop (QC Hold), urgent order insertion requests, order cancellation or modification, material delivery in place signals, operator login / logout, etc. The implementation method is to use a service-oriented architecture (SOA) combined with RESTful API interface, database change data capture (CDC), and distributed message queue (such as Kafka, RabbitMQ) to ensure that any state change in the production site can be perceived within seconds or sub-seconds, providing a solid real-time data foundation for system decision-making.

[0079] S12, constructing a data structure for consumable tools in the production information, the data structure including a static ID and dynamically updated consumption data to obtain a micro digital twin model.

[0080] Specifically, the consumable tool here takes a drill bit as an example, a digital twin object is created for each independent drill bit, and the data structure thereof contains a static ID and dynamically updated consumption attributes such as a hit count and a remaining useful life (RUL). The RUL is calculated in real time based on the number of holes processed by the drill bit, the cumulative depth, and the hardness coefficient of the processed plate material through a preset wear model. Here, the remaining useful life is not static, but has the ability of online self-learning and correction. The system will continuously collect event data related to the actual scrap or quality degradation of the drill bit (for example, a “burr exceeds the standard” event fed back by a quality management system, or a “drill bit broken” event entered by an operator). When the state of a drill bit twin is updated to “end of life”, the system will compare the difference between its “predicted life” and “actual service life”. The difference will be used as an error signal to fine-tune the core parameters of the drill bit type in the wear model (such as the hardness wear coefficient of different plates) through an adaptive algorithm (for example, a simple least squares method or a more complex Bayesian update). In this way, the digital twin model can continuously learn from historical data, making its predictions more and more close to the physical reality, thereby realizing the transition from “passive mapping” to “active evolution”.

[0081] S13, constructing a macro digital twin model for other data in the production information, and integrating the micro digital twin model and the macro digital twin model to obtain a corresponding digital twin;

[0082] Specifically, through the above processing, the relational data and event data are instantiated into a set of interrelated, high-fidelity standardized object models including macro devices, micro tools, and invisible preparation activities, thereby providing a solid data foundation for achieving accurate and practical scheduling optimization. In actual situations, the state of the twin is updated in real time. Once a state event such as a drilling task completion event is received, the twin automatically updates its hitCount and RUL values according to the processing hole number and other information carried in the event. When the RUL is lower than the warning threshold, the twin will actively issue a “tool life warning event” to provide a basis for subsequent scheduling decisions.

[0083] S2, identifying explicit process tasks and implicit process tasks in the production work order based on the state of the digital twin;

[0084] The step S2 includes:

[0085] S21, splitting the work order in the digital twin into a plurality of process tasks arranged according to a process route;

[0086] Specifically, the process task here is a planning entity, specifically refers to the object to be scheduled, and each work order is divided into a series of process tasks sorted by process route. The task object not only contains its basic processing time, but also contains attributes such as setupDuration (changeover time, which may depend on the previous task), requiredToolType (required tool type), and the core planning variable is the assigned machine (Machine) and the start execution time (StartTime).

[0087] S22, use the available resource information, available tool information, and order information in the scheduling process as decision information, associate the decision information with the process task to obtain an explicit process task;

[0088] Specifically, the decision information here refers to information that is fixed or used as a basis for decision-making in scheduling. For example, the machine (Machine) object has attributes including a skillSet (ability tag set) and a maintenanceSchedule (maintenance plan time window). The Order object has attributes including priorityLevel (priority) and dueDate (delivery date).

[0089] S23, perform process requirement analysis according to the process task, and if there is a process requirement analysis unqualified condition, create an implicit task that meets the process requirement analysis to obtain an implicit process task;

[0090] In this application, when the work order is parsed into a process task (Task), the system analyzes the process requirements of the task (such as the required drill type). If the required drill does not match the currently installed machine, or the RUL of the current drill cannot meet the precision requirements of the task, the system will automatically inject an independent "tool change task (ToolChangeTask)" before the process task. This tool change task has independent duration, resource requirements, and other attributes, and participates in the subsequent optimization solution as an independent planning entity;

[0091] Specifically, in addition to the "tool change task", the implicit process task can also include but is not limited to: necessary machine preheating tasks, specific chemical solution replacement and debugging tasks, or automatic cleaning tasks required between processing two products with large differences. The system can automatically identify and inject these necessary preparation activities to ensure production quality and continuity according to the preset process knowledge base combined with the attribute differences of the previous and subsequent process tasks, so as to ensure the completeness and actual executability of the scheduling plan.

[0092] S3, obtain the current business policy and load the hard constraint rule set and the soft constraint rule set based on the current business policy;

[0093] Specifically, the hard constraint rule set includes machine capacity constraint, process sequence constraint, resource exclusive constraint, tool availability constraint, the soft constraint rule set includes delivery optimization rule, line change minimization rule, load balancing rule, preferred machine rule;

[0094] Among them, the hard constraint rule is a rule that must be strictly followed, and violation of any rule will result in the scheduling scheme being judged as invalid. For example: machine capacity constraint, process sequence constraint, resource exclusive constraint, tool availability constraint, etc.

[0095] The soft constraint rule is a rule that is desired to be satisfied as much as possible, but is not mandatory. Each violation of the soft constraint will result in a corresponding penalty score according to its importance. The optimization goal is to minimize the total penalty score. For example: delivery optimization rule, line change minimization rule, load balancing rule, preferred machine rule, etc.

[0096] At the same time, in this application, the above rules are realized as dynamic soft constraints for production scenarios, and their weights can be self-adaptively adjusted according to the real-time situation of the workshop, which are respectively rule logic and scenario adaptive weight:

[0097] Rule logic: when the remaining useful life (RUL) of a drill bit enters the end of life, if the system assigns it to a low-precision or non-critical product drilling task, the scheduling scheme obtains a reward score; otherwise, if it is assigned to a high-precision task, a penalty score is generated;

[0098] Scenario adaptive weight: the penalty weight of this rule is not fixed. When the number of high-precision orders accumulated in the system exceeds the threshold, in order to prioritize quality, the system will automatically increase the penalty weight of this constraint to force high-precision tasks to use new tools. Conversely, when the production pressure is low, the weight is appropriately reduced to encourage the consumption of old tools and save costs.

[0099] S4, combine the explicit process task, the implicit process task, the hard constraint rule set and the soft constraint rule set to obtain a scheduling problem instance;

[0100] Specifically, in this step, based on the process requirements and the state of the tool twin, the necessary implicit process tasks are automatically identified and injected as independent planning entities. Finally, all planning entities including explicit process tasks and implicit process tasks, as well as all resources and constraints, are converted into a complete scheduling problem instance.

[0101] S5, convert the scheduling problem instance into a target function, and perform multi-objective optimization solution on the target function to output an initial scheduling plan;

[0102] Wherein, the step S5 includes:

[0103] S51. Convert the scheduling problem instance into an objective function. :

[0104] ;

[0105] ; ;

[0106] ;

[0107] ;

[0108] ;

[0109] In the formula, These are respectively the order set, the machine set, and the process task set. , , , These represent the penalty items corresponding to delivery delay penalties, changeover time penalties, tool-task mismatch penalties, and other soft constraint rules, respectively. , Orders Start time, required delivery date, For the machine On the task Execute the task immediately upon completion. Required changeover time These are the first, second, and third binary variables, respectively. Representing tasks Start time, task End time, task End time, For orders Priority weights, For a set of priority relations, These are the first soft constraint weight, the second soft constraint weight, and the third soft constraint weight, respectively.

[0110] Among them, for the first, second, and third binary variables, when the process task Assigned to machine During execution, The value of is 1 if it is 1, otherwise it is 0; the same applies to the second binary variable, when the process task Assigned to machine During execution, The value is 1 if the task is not specified, and 0 otherwise. and process tasks At the machine 1 if the next job is arranged to be processed in the immediate sequence, otherwise 0, used to determine whether to generate a changeover time;

[0111] Specifically, the penalty term corresponding to the other soft constraint rule here can be a penalty term corresponding to load balancing, preferred machine constraint, etc. Meanwhile, in actual situations, the above objective function also needs tool availability constraint, i.e. during the execution of a task, the tool used by the task must be available.

[0112] S52, adding an operator and using an adaptive learning mechanism to solve the objective function using the operator to obtain an initial scheduling plan;

[0113] The operator includes a batch merging moving operator, a tool-task collaborative moving operator, and a high-low precision task pair moving operator.

[0114] Specifically, in the actual solving process, two steps are performed. The first step is high-quality initial solution construction. The optimization is not started from a random solution, but a domain-customized constructive heuristic algorithm is first run. The algorithm follows the "bottleneck first" and "critical chain" principles, and preferentially arranges process tasks that are expected to be the most time-consuming or use the most stressed resources (such as high-precision machines required for special plates) to their optimal positions, thereby quickly generating an initial solution with a good skeleton. The second step is adaptive large neighborhood search iterative optimization. It is the core iterative optimization phase, and its framework includes two types of core operators: a destruction operator responsible for removing a part of the tasks from the current solution to create a "repair" space. A repair operator responsible for reinserting the removed tasks into the solution in a more optimal way.

[0115] The batch merging moving operator, the tool-task collaborative moving operator, and the high-low precision task pair moving operator can all be used as destruction operators and repair operators. In the solving process, the frequency and amplitude of successful improvement of the solution in the recent iteration of each "destruction-repair" operator combination are counted in real time. According to the statistical results, the probability of selecting each operator is dynamically adjusted, so that the algorithm can learn by itself and automatically tend to use more effective operators in the current solving stage, thereby greatly improving the search efficiency.

[0116] The batch merging moving operator is used to process tasks of the same type in batches and reduce the tool change tasks caused by tool replacement, so as to reduce the value of the changeover time penalty in the objective function.

[0117] Among them, the batch merging moving operator is a high-level operator, which can identify tasks using the same tool (such as the same type of drill bit) but arranged at different times or on different machines, and will "destroy" these tasks, and then "repair" them in the "repair" stage to arrange them continuously on the same machine, and the optimization goal is to maximize the reduction of "tool changing tasks" caused by tool replacement by "batch processing" similar tasks, thereby reducing the value of the tool change time penalty in the objective function.

[0118] The tool-task cooperative moving operator is used to reduce the value of the tool change time penalty in the objective function by controlling the use of existing tools on the machine and avoiding the generation of tool changing tasks.

[0119] Among them, the tool-task cooperative moving operator is a heuristic logic operator, which is used as a repair operator in this application, and when it needs to find a new insertion position for a task, it will give priority to machines that have installed the required tool, and the optimization goal is similar to "batch merging", which aims to avoid tool changing by "compromise" with the existing tools on the machine, thereby reducing the value of the tool change time penalty in the objective function.

[0120] The high-low precision task pair switching moving operator is used to identify task resource mismatch conditions in the current solution of the objective function, and exchange the resource allocation of the tasks according to the resource mismatch conditions, thereby reducing the value of the tool-task mismatch penalty in the objective function.

[0121] Among them, the high-low precision task pair switching moving operator is a highly targeted composite operator that actively seeks a specific "resource mismatch" condition in the current solution: a high-precision task uses an old tool at the end of its life, while a low-precision task occupies a brand new tool. After finding such a "mismatch pair", it will exchange the machine or tool allocation of the two tasks, and the optimization goal is to directly improve the score of the "tool life end value maximization" soft constraint rule by ensuring that "good tools are used on sharp edges", thereby reducing the value of the tool-task mismatch penalty in the objective function, achieving dual optimization of quality and cost.

[0122] Specifically, to implement the high-low precision task pair switching moving operator, the system will automatically mark each process task with "high precision" or "regular precision" properties according to the technical requirements (such as specific drilling tolerance, line precision level, or customer-specified key process) in the obtained work order data in advance. This property serves as an objective basis for accurately identifying resource mismatch conditions during the solution process, thereby ensuring the effectiveness of the operator's execution.

[0123] In summary, at each step of the iterative optimization, the algorithm tries to apply these operators to produce new solutions, if a new solution makes the total score (total penalty) of the objective function Z reduced, then this "move" is a successful optimization, through tens of thousands of such attempts, the entire scheduling scheme will gradually approach the optimal.

[0124] S6, real-time monitoring and dynamic incremental rescheduling cycle of the initial scheduling plan to complete the real-time scheduling of PCB production.

[0125] Among them, step S6 includes:

[0126] S61, real-time monitoring of the state changes and external event streams of the production site, and determining whether to trigger a rescheduling condition based on the state changes and the external event streams;

[0127] Specifically, the rescheduling condition here can be periodic triggering, such as performing a rolling optimization every 15 minutes, or event triggering, such as receiving a machine failure alarm or inserting a high-priority work order.

[0128] S62, if the rescheduling condition is not triggered, the initial scheduling plan is followed for production processing;

[0129] S63, if the rescheduling condition is triggered, the digital twin is updated and all completed process tasks within a rolling frozen time window are locked, and the state changes and the external event streams are updated to the current scheduling problem instance;

[0130] Specifically, if triggered, the system immediately updates the digital twin model, automatically marks all planned tasks within a preset, rolling frozen time window as 'locked' to ensure production stability, and updates new changes (such as a failure machine becoming unavailable within a specific time period, a newly added work order task) to the current scheduling problem instance.

[0131] S64, converting the updated scheduling problem instance into an updated objective function, and updating the first soft constraint weight, the second soft constraint weight, and the third soft constraint weight in the updated objective function:

[0132] ;

[0133] ;

[0134] wherein, is the first soft constraint weight or the second soft constraint weight or the third soft constraint weight after the nth rescheduling iteration, is the first soft constraint weight or the second soft constraint weight or the third soft constraint weight after the nth rescheduling iteration, is the first soft constraint weight or the second soft constraint weight or the third soft constraint weight after the nth rescheduling iteration, respectively represent the first the current error of the n th re-scheduling iteration, respectively represent the proportional action weight, the integral action weight, and the differential action weight;

[0135] Specifically, in the updating process here, the process of iterative updating, that is, updating the current result based on the result of the last reordering after each reordering, in this step, the PID algorithm is used to update the soft constraint weight accordingly, and in the actual process, the updating method based on the key performance indicator is introduced for updating. First, set a target value (for example, 98%) for the core business target (such as "high-priority order on-time delivery rate"), and after each scheduling, the achievement of the KPI will be predicted based on the scheduling result. The strategy controller will continuously compare the "predicted KPI value" with the "target KPI value". If the predicted value is lower than the target value, the controller will automatically and smoothly increase the weight of the soft constraint (such as "delivery date optimization rule") directly related to the KPI in the input of the next re-scheduling. On the contrary, if the predicted value is much higher than the target, the weight can be appropriately reduced to release optimization space for other optimization targets (such as cost). This mechanism changes the adjustment of business strategy from discrete, threshold-based judgment to continuous, feedback-based, and automated closed-loop control;

[0136] Here, the current error of the n th re-scheduling iteration is precisely defined as the normalized deviation between the KPI predicted value and the target KPI value of the scheduling plan, which can be specifically represented as = (KPItarget-KPIpredicted,n) / KPItarget. Wherein, KPIpredicted,n is the predicted delivery on-time rate calculated based on the current scheduling result, that is, the predicted KPI value, and KPItarget is the target KPI value.

[0137] And the calculation process of the predicted value of the KPI here is:

[0138] Traverse all orders marked as "high priority" in the scheduling scheme, for each order, find the completion time of the last process task, compare the completion time with the required delivery date of the order, if the completion time ≤ the required delivery date, the order is judged as on time, finally, the predicted KPI value is: (the number of high-priority orders completed on time / the total number of high-priority orders) x 100%.

[0139] S65, update the weight of the operator:

[0140] ;

[0141] wherein, is the learning rate, is the weight of the updated operator, is the weight of the operator before update, is the objective function value of the operator in the current week;

[0142] Specifically, in the initial stage, an initial weight is assigned to each operator (or each "break-fix" operator combination), at the beginning, the probability of being selected by all operators can be equal, in each iteration of optimization, a pair of "break-fix" operators needs to be selected to modify the current solution, this selection process is based on probability, usually using "roulette" and other methods, the higher the score of the operator, the greater the probability of being selected;

[0143] After the operator combination is applied, a new scheduling solution is generated, the quality (i.e., the total penalty score) of the new solution is evaluated, and the score of the selected operator is updated in real time according to the evaluation result, if the new solution is much better than the current optimal solution, the operator combination obtains a high score reward; if the new solution is slightly improved, a lower reward is obtained; in some algorithm mechanisms (such as simulated annealing), a poor solution may also be accepted in order to jump out of the local optimum. In this case, the score of the operator may not change or increase slightly; if the new solution is not improved and is rejected, the operator does not score.

[0144] The learning process is carried out in "segments" (for example, every 100 iterations is a segment), after the end of a segment, the system updates the selection weights of the operators in the next segment according to the scores accumulated by each operator during this period, and the update formula is as follows.

[0145] S66, based on the updated first soft constraint weight, the second soft constraint weight, the third soft constraint weight and the updated weight of the operator, an adaptive learning mechanism is used to solve the updated objective function to obtain an updated scheduling plan and publish the updated scheduling plan.

[0146] The PCB production real-time scheduling method based on hybrid intelligent decision provided by the embodiment one of the application has the beneficial effects that, compared with the prior art, the application realizes the paradigm shift from 'offline batch processing' to 'online real-time decision', revolutionarily improves production agility, and the prior art is generally offline and static, while the application shortens the decision period from 'hour level' to'minute level' through real-time data sensing and incremental planning technology based on event triggering, when machine failure or urgent order insertion occurs, the system can automatically generate a globally coordinated response scheme, so that the production plan is changed from rigid instructions to dynamically adaptive 'living' entities, and the flexibility and market response ability of the production system are fundamentally improved;

[0147] The application significantly improves the quality and field feasibility of scheduling results: the application uses a hybrid rule engine to make implicit expert knowledge explicit as a series of quantifiable soft constraints, and deeply embeds them into the scoring function of the optimization algorithm, which enables the algorithm to "understand" the subtle preferences of the business (such as minimizing line changes and balancing equipment load) at every step of the optimization, generating solutions that are not only mathematically optimal but also more reasonable in production practice, effectively improving overall OEE (overall equipment effectiveness) and reducing production costs.

[0148] The application greatly enhances the flexibility and maintainability of the system: existing APS systems have fixed rules, and changes in business processes require expensive and time-consuming secondary development. The application decouples business rules from core algorithms, parameterizes and configures rules and weights, allowing business personnel (such as production managers) to adjust the weights of different optimization objectives (such as delivery time and cost) through the interface to quickly respond to changes in enterprise strategy.

[0149] The application improves the efficiency of scheduling work and the intelligent level of enterprise operation: the application uses advanced meta-heuristic optimization algorithms to find high-quality solutions in acceptable time when dealing with large-scale and high-complexity PCB scheduling problems. More importantly, the entire "perception-decision-execution-feedback" closed-loop process is highly automated, freeing scheduling personnel from tedious, repetitive, and error-prone manual planning and adjustment work, allowing them to become "commanders" in the production field, significantly improving the intelligent level of the entire production operation.

[0150] Embodiment two

[0151] As shown in Figure 2 Embodiment two of the application provides a PCB production real-time scheduling system based on hybrid intelligent decision-making, which includes:

[0152] The acquisition module 1 is used to acquire production information of PCB production, and to construct a corresponding digital twin for the production information;

[0153] The task module 2 is used to identify explicit process tasks and implicit process tasks in the production work order based on the state of the digital twin;

[0154] The constraint module 3 is used to acquire the current business strategy and load the hard constraint rule set and the soft constraint rule set based on the current business strategy;

[0155] The instance module 4 is used to integrate the explicit process tasks, the implicit process tasks, the hard constraint rule set, and the soft constraint rule set to obtain a scheduling problem instance;

[0156] A solving module 5 is configured to convert the scheduling problem instance into a target function, and perform multi-objective optimization solving on the target function to output an initial scheduling plan;

[0157] A circulation module 6 is configured to perform real-time monitoring and dynamic incremental rescheduling circulation on the initial scheduling plan to complete real-time scheduling of PCB production;

[0158] The obtaining module 1 comprises:

[0159] An information sub-module is configured to obtain work order data, resource data, material and tool data, and real-time events of PCB production to obtain production information;

[0160] A structure sub-module is configured to construct a data structure for consumable tools in the production information, the data structure comprising a static ID and dynamically updated consumption data to obtain a micro digital twin model;

[0161] A twin sub-module is configured to construct a macro digital twin model for other data in the production information, and to integrate the micro digital twin model and the macro digital twin model to obtain a corresponding digital twin.

[0162] The task module 2 comprises:

[0163] A splitting sub-module is configured to split the work order in the digital twin into a plurality of process tasks arranged according to a process route;

[0164] An association sub-module is configured to use available resource information, available tool information, and order information as decision information in a scheduling process, and to associate the decision information with the process tasks to obtain explicit process tasks;

[0165] An analysis sub-module is configured to perform process requirement analysis according to the process tasks, and to create an implicit task that meets the process requirement analysis if there is a process requirement analysis unqualified condition to obtain implicit process tasks.

[0166] The solving module 5 comprises:

[0167] A function sub-module is configured to convert the scheduling problem instance into a target function :

[0168] ;

[0169] ; ;

[0170] ;

[0171] ;

[0172] ;

[0173] wherein, respectively represent a set of orders, a set of machines, a set of process tasks, 、 、 、 respectively represent a penalty term corresponding to a delivery delay penalty, a changeover time penalty, a tool-task mismatch penalty, and other soft constraint rules, 、 respectively represent a start time of an order , a required delivery date, represent a changeover time required for executing a task immediately after a task is completed on a machine , respectively represent a first, a second, and a third binary variable, respectively represent a start time of a task , an end time of a task , and an end time of a task , is a priority weight of an order , is a set of priority relations, respectively represent a first, a second, and a third soft constraint weight;

[0174] a solving submodule, configured to add an operator, and use the operator to solve the objective function by using an adaptive learning mechanism to obtain an initial scheduling plan.

[0175] The loop module 6 comprises:

[0176] a monitoring submodule, configured to monitor a state change and an external event flow of a production site in real time, and determine whether a rescheduling condition is triggered based on the state change and the external event flow;

[0177] a condition submodule, configured to perform production and processing according to the initial scheduling plan if the rescheduling condition is not triggered;

[0178] a locking submodule, configured to update a digital twin and lock all completed process tasks within a rolling frozen time window if the rescheduling condition is triggered, and update the state change and the external event flow to a current scheduling problem instance;

[0179] a first updating submodule, configured to convert the updated scheduling problem instance into an updated objective function, and update a first soft constraint weight, a second soft constraint weight, and a third soft constraint weight in the updated objective function:

[0180] ;

[0181] ;

[0182] In the formula, For the first The first, second, or third soft constraint weights after the second rescheduling iteration For the updated first soft constraint weight, second soft constraint weight, or third soft constraint weight, They represent the first sequence The current error of the second rescheduling iteration. These are the proportional action weight, integral action weight, and differential action weight, respectively.

[0183] The second update submodule is used to update the weights of the operators:

[0184] ;

[0185] In the formula, For learning rate, For the weights of the updated operator, The weights of the operators before the update. This represents the objective function value of the operator this week.

[0186] The third update submodule is used to solve the update objective function using an adaptive learning mechanism based on the updated first soft constraint weights, second soft constraint weights, third soft constraint weights, and updated operator weights, so as to obtain the updated scheduling plan and publish the updated scheduling plan.

[0187] In other embodiments of the present invention, the present invention provides the following technical solution: a computer, including a memory 102, a processor 101, and a computer program stored on the memory 102 and executable on the processor 101, wherein the processor 101 executes the computer program to implement the PCB production real-time scheduling method based on hybrid intelligent decision-making as described above.

[0188] Specifically, the processor 101 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of the present invention.

[0189] The memory 102 can include mass storage for data or instructions. By way of example, and not limitation, the memory 102 can include a Hard Disk Drive (HDD), a floppy disk drive, a solid-state drive (SSD), a flash drive, a Compact Disk Drive (CD Drive), a DVD Drive, a tape drive, a USB drive, or any combination of two or more of these. The memory 102 can be removable and / or non-removable (or fixed) as appropriate. The memory 102 can be internal or external as appropriate. In certain embodiments, the memory 102 is a non-volatile memory. In certain embodiments, the memory 102 includes a Read-Only Memory (ROM) and a Random Access Memory (RAM). The ROM can be a mask-programmed ROM, a Programmable ROM (PROM), an Erasable PROM (EPROM), an Electrically Erasable PROM (EEPROM), an Electrically Alterable ROM (EAROM), or a FLASH memory, or any combination of two or more of these, as appropriate. The RAM can be a Static Random-Access Memory (SRAM) or a Dynamic Random-Access Memory (DRAM), which can be a Fast Page Mode Dynamic Random-Access Memory (FPMDRAM), an Extended Data Output Dynamic Random-Access Memory (EDODRAM), a Synchronous Dynamic Random-Access Memory (SDRAM), or the like, as appropriate.

[0190] The memory 102 can be used to store or buffer various data files needed for processing and / or communication, and possible computer program instructions executed by the processor 101.

[0191] The processor 101 realizes the above-mentioned real-time scheduling method for PCB production based on hybrid intelligent decision by reading and executing the computer program instructions stored in the memory 102.

[0192] In some embodiments, the computer can further include a communication interface 103 and a bus 100. As shown, the processor 101, the memory 102, and the communication interface 103 are connected through the bus 100 and complete communication with each other. Figure 3

[0193] The communication interface 103 is used to realize the communication between the modules, devices, units and / or equipment in the embodiments of the present application. The communication interface 103 can also realize data communication with other components, such as external devices, image / data acquisition devices, databases, external storage, image / data processing workstations, etc.

[0194] ​Bus 100 includes hardware, software, or both, to couple components of computer device to each other and to couple components of computer device to other devices. Bus 100 includes, but is not limited to, at least one of the following: a data bus, an address bus, a control bus, an expansion bus, a local bus, and the like. By way of example and not limitation, bus 100 can include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand (IB) interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or another suitable bus or a combination of two or more of these. Where appropriate, bus 100 can include one or more buses. Although the present embodiments describe and show a particular bus, the present embodiments contemplate any suitable bus or interconnect.

[0195] The computer can execute the hybrid intelligent decision-making based PCB production real-time scheduling method based on the hybrid intelligent decision-making based PCB production real-time scheduling system, so as to realize the hybrid intelligent decision-making based PCB production real-time scheduling.

[0196] In some embodiments of the present application, in combination with the hybrid intelligent decision-making based PCB production real-time scheduling method described above, the present embodiments provide the following technical solutions: a storage medium, wherein the storage medium stores a computer program, and the computer program is executed by a processor to realize the hybrid intelligent decision-making based PCB production real-time scheduling method.

[0197] Those skilled in the art will appreciate that the logic and / or steps represented in the flow diagrams, or otherwise described herein, for example, can be thought of as a list of executable instructions for implementing logic functions, and can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions, or a combination thereof. For the purposes of this specification, a "computer-readable medium" can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.

[0198] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example, via an optical scanner, then compiled, interpreted, or otherwise processed, and stored in a computer memory in a form that is then employable by a computer. In this context, the above-mentioned media are considered machine-readable media.

[0199] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the above embodiments, various steps or methods can be implemented in software or firmware that is stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any of the following technologies, known in the art, or a combination thereof, can be used: discrete logic circuitry having logic gates for implementing logic functions upon data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and so forth.

[0200] The technical features of the above-described embodiments can be combined in any manner, and for the sake of brevity, not all possible combinations are described, however, any combination of the technical features is considered to be within the scope of the present specification.

[0201] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent. It should be noted that for ordinary skilled persons in the art, without departing from the concept of the present application, several modifications and improvements can be made, which are within the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

Claims

1. A hybrid intelligent decision based real-time scheduling method for PCB production, characterized in that, The method comprises the following steps: obtaining production information of PCB production, and constructing a corresponding digital twin for the production information; identifying explicit process tasks and implicit process tasks in a production work order based on a state of the digital twin; obtaining a current business policy and loading a hard constraint rule set and a soft constraint rule set based on the current business policy; integrating the explicit process tasks, the implicit process tasks, the hard constraint rule set and the soft constraint rule set to obtain a scheduling problem instance; converting the scheduling problem instance into an objective function, and performing multi-objective optimization solving on the objective function to output an initial scheduling plan; real-time monitoring and dynamic incremental rescheduling cycle are performed on the initial scheduling plan to complete real-time scheduling of PCB production; the step of identifying explicit process tasks and implicit process tasks in a production work order based on a state of the digital twin comprises: splitting the work order in the digital twin into a plurality of process tasks arranged according to a process route; using available resource information, available tool information and order information as decision information, and associating the decision information with the process tasks to obtain explicit process tasks; performing process requirement analysis according to the process tasks, and if there is a process requirement analysis unqualified condition, creating an implicit task meeting the process requirement analysis to obtain implicit process tasks; in the step of obtaining a current business policy and loading a hard constraint rule set and a soft constraint rule set based on the current business policy, the hard constraint rule set comprises machine capacity constraints, process sequence constraints, resource exclusive constraints and tool availability constraints, and the soft constraint rule set comprises due date optimization rules, line change minimization rules, load balancing rules and preferred machine rules; the step of real-time monitoring and dynamic incremental rescheduling cycle are performed on the initial scheduling plan to complete real-time scheduling of PCB production comprises: real-time monitoring of state changes and external event streams in the production site, and determining whether a rescheduling condition is triggered based on the state changes and the external event streams; if the rescheduling condition is not triggered, production processing is performed according to the initial scheduling plan; if the rescheduling condition is triggered, the digital twin is updated, all completed process tasks within a rolling frozen time window are locked, and the state changes and the external event streams are updated to the current scheduling problem instance; the updated scheduling problem instance is converted into an updated objective function, and the first soft constraint weight, the second soft constraint weight and the third soft constraint weight in the updated objective function are updated: ； ； In the formula, is the first soft constraint weight or the second soft constraint weight or the third soft constraint weight after the first rescheduling iteration, is the first soft constraint weight or the second soft constraint weight or the third soft constraint weight after the first rescheduling iteration, is the updated first soft constraint weight or the updated second soft constraint weight or the updated third soft constraint weight, respectively represent the current error of the first rescheduling iteration, respectively represent the current error of the first rescheduling iteration, respectively represent the current error of the first rescheduling iteration, respectively represent the proportional action weight, the integral action weight, and the differential action weight. the weight of the operator is updated: ； wherein is the learning rate, is the weight of the updated operator, is the weight of the operator before the update, is the objective function value of the operator in the current week; an adaptive learning mechanism is used to solve the updated objective function based on the updated first soft constraint weight, the second soft constraint weight and the third soft constraint weight and the updated weight of the operator to obtain an updated scheduling plan and publish the updated scheduling plan.

2. The hybrid intelligence decision based real-time scheduling method for PCB production according to claim 1, wherein, The step of obtaining production information of PCB production and constructing a corresponding digital twin for the production information comprises: obtaining work order data, resource data, material and tool data and real-time events of PCB production to obtain production information; constructing a data structure for consumable tools in the production information, the data structure including a static ID and dynamically updated consumption data to obtain a micro digital twin model; constructing a macro digital twin model for other data in the production information, and synthesizing the micro digital twin model and the macro digital twin model to obtain a corresponding digital twin.

3. The hybrid intelligence decision based real-time scheduling method for PCB production according to claim 1, wherein, The step of converting the scheduling problem instance into a target function, and performing multi-objective optimization solving on the target function to output an initial scheduling plan includes: converting the scheduling problem instance into an objective function : ； ； ； ； ； ； wherein, respectively represent a set of orders, a set of machines, a set of process tasks, , , , respectively represent a penalty item corresponding to a delivery delay penalty, a changeover time penalty, a tool-task mismatch penalty, and other soft constraint rules, , respectively represent a start time of an order , a required delivery date, respectively represent a changeover time required for executing a task immediately after a task is completed on a machine , respectively represent a first, a second, and a third binary variable, respectively represent a start time of a task , an end time of a task , and an end time of a task , respectively represent a priority weight of an order , respectively represent a set of priority relations, respectively represent a first, a second, and a third soft constraint weight; adding an operator and using the operator to solve the target function by using an adaptive learning mechanism to obtain an initial scheduling plan.

4. The hybrid intelligence decision based real-time scheduling method for PCB production according to claim 3, wherein, In the step of adding an operator and using the operator to solve the target function by using an adaptive learning mechanism to obtain an initial scheduling plan, the operator includes a batch merging moving operator, a tool-task collaborative moving operator, and a high-low precision task pair switching moving operator. The batch merging moving operator is used to process tasks of the same type in batches and reduce tool change tasks caused by tool replacement, so as to reduce the value of the tool change time penalty in the target function. The tool-task collaborative moving operator is used to control the use of existing tools by the machine and avoid the generation of tool change tasks, so as to reduce the value of the tool change time penalty in the target function. The high-low precision task pair switching moving operator is used to identify task resource mismatching in the current solution of the target function, and exchange the resource allocation of tasks according to the resource mismatching, so as to reduce the value of the tool-task mismatching penalty in the target function.

5. A hybrid intelligent decision based real-time scheduling system for PCB production, which adopts the hybrid intelligent decision based real-time scheduling method for PCB production according to claim 1, characterized in that, The system includes: an acquisition module configured to acquire production information of PCB production, and construct a corresponding digital twin for the production information; a task module configured to identify explicit process tasks and implicit process tasks in a production work order based on a state of the digital twin; a constraint module configured to acquire a current business policy and load a hard constraint rule set and a soft constraint rule set based on the current business policy; an instance module configured to synthesize the explicit process tasks, the implicit process tasks, the hard constraint rule set, and the soft constraint rule set to obtain a scheduling problem instance; a solving module configured to convert the scheduling problem instance into a target function, and perform multi-objective optimization solving on the target function to output an initial scheduling plan; a loop module configured to perform real-time monitoring and dynamic incremental rescheduling on the initial scheduling plan to complete real-time scheduling of PCB production.

6. A computer comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the hybrid intelligent decision-based real-time scheduling method for PCB production according to any one of claims 1 to 4 when executing the computer program.

7. A storage medium, characterized by The storage medium has the computer program stored thereon, and the computer program implements the hybrid intelligent decision-based real-time scheduling method for PCB production according to any one of claims 1 to 4 when executed by the processor.

Images