Injection Molding Scheduling Optimization Method, Device, Electronic Device and Computer Storage Medium

Abstract

This application relates to the field of advanced scheduling technology, and particularly to an injection molding scheduling optimization method, device, electronic device and computer storage medium. By obtaining injection molding scheduling data and constructing a set of rule strategies, an initial solution is generated as a preliminary scheduling plan based on the rule strategies and the corresponding first strategy parameter weights in the set. The task processing order of the preliminary scheduling plan is optimized to determine the optimal processing order. The second strategy parameter weights corresponding to the fine-tuning algorithm strategy are adjusted to generate a neighborhood solution, and its first index value is calculated to approximate the local optimal solution. The third strategy parameter weights corresponding to the equipment strategy are fine-tuned to generate a global optimal solution based on the local optimal solution, and its second index value is calculated. When the second index value meets the preset termination condition, the optimal scheduling plan is output according to the optimal processing order, realizing the efficient optimization of injection molding scheduling.

Description

Technical Field

[0001] This application relates to the field of advanced scheduling technology, and particularly to an injection molding scheduling optimization method, device, electronic device, and computer storage medium. Background Art

[0002] As the core process of plastic product manufacturing, the scheduling problem in injection molding production has always been a key factor restricting production efficiency. Most traditional injection molding scheduling methods rely on fixed rules or experience, and it is difficult to adapt to complex and changeable production environments. Especially during the frequent order changes, fixed rule or experience methods often fail to effectively adapt to new situations, making it difficult to improve scheduling efficiency.

[0003] In the actual production process, the scheduling of injection molding production lines needs to consider various factors, such as machine selection, mold configuration, diverse orders, material supply, production time, changeover time, process constraints, etc. The mutual influence among these factors increases the complexity of the scheduling problem, making it difficult to handle solely by rule or experience methods. At the same time, these factors also have dynamic and diverse data aspects, making it difficult for traditional scheduling methods to obtain satisfactory scheduling results. Summary of the Invention

[0004] In order to overcome the deficiencies of the prior art, the present invention provides an injection molding scheduling optimization method, device, electronic device, and computer storage medium, which adaptively adjusts the scheduling strategy according to the changes in the actual production environment, and improves the flexibility and efficiency of scheduling.

[0005] The first aspect of this application provides an injection molding scheduling optimization method, and the method includes:

[0006] Obtain injection molding scheduling data, and construct a rule strategy set according to the injection molding scheduling data;

[0007] Generate an initial solution according to each rule strategy and the corresponding first strategy parameter weight order in the rule strategy set, and determine the initial solution as a preliminary scheduling plan;

[0008] Optimize the task processing order of the preliminary scheduling plan to determine the optimal processing order;

[0009] Fine-tune the second strategy parameter weight corresponding to the fine-tuning algorithm strategy to generate a neighborhood solution based on the initial solution, and calculate the first index value of the neighborhood solution, so that the neighborhood solution gradually approaches the local optimal solution;

[0010] Fine-tune the third strategy parameter weight corresponding to the equipment strategy to generate a global optimal solution based on the local optimal solution, and calculate the second index value of the global optimal solution;

[0011] When it is determined that the second index value meets the preset termination condition, an optimal scheduling plan is output according to the optimal processing sequence for the global optimal solution.

[0012] In an alternative embodiment, the rule strategy set includes a task sorting rule, a task clustering rule, a task classification rule, and a task order optimization rule. The generation of the initial solution according to each rule strategy and the corresponding first strategy parameter weight order in the rule strategy set includes:

[0013] Obtain the locked tasks within the locked time period, and prioritize the rule strategies in the above rule strategy set according to the first strategy parameter weights;

[0014] Sort the locked tasks according to the task sorting rule corresponding to the sorted first weight to obtain an ordered task list;

[0015] Group the ordered task list according to the task clustering rule corresponding to the sorted second weight to obtain multiple clustered task groups;

[0016] Decompose each clustered task group according to the task classification rule corresponding to the sorted third weight to obtain multiple independent task sets;

[0017] Allocate a first task set and a second task set to each injection molding machine according to the multiple independent task sets,

[0018] Adjust the order of the task sets for the first task set and the second task set according to the task order optimization rule to generate the preliminary scheduling plan.

[0019] In an alternative embodiment, the optimization of the task processing sequence of the preliminary scheduling plan to determine the optimal processing sequence includes:

[0020] Use the particle swarm algorithm to optimize the task processing sequence on each injection molding machine. The preliminary scheduling plan includes the task processing sequence corresponding to each injection molding machine among multiple injection molding machines, and each particle represents a processing sequence plan;

[0021] Optimize the task processing sequence by iteratively updating the speed and position of the particles to output the optimal processing sequence, and the optimal processing sequence is the task sequence plan corresponding to the global optimum.

[0022] In an alternative embodiment, the method further includes:

[0023] Determine the fitness function through the following formula:

[0024] ;

[0025] Where, is the fitness function, is the completion time of a single injection molding machine, is the overdue penalty coefficient;

[0026] The completion time of the single injection molding machine is determined by the following formula:

[0027] ;

[0028] where is the completion time of the injection molding machine j and is the processing time of task k on the injection molding machine j and is the switching time between task k and the next task.

[0029] In an alternative embodiment, updating the velocity of the particle includes:

[0030] The velocity of the particle is updated by the following formula:

[0031] ;

[0032] where w is the inertia weight, , are the learning factors, , are random numbers in the range of 0 to 1, represents the position difference.

[0033] In an alternative embodiment, calculating the first index value of the neighborhood solution includes:

[0034] The first index value includes multiple objective values, and the first index value is determined by the following formula:

[0035] ;

[0036] where represents different objective values among the multiple objective values, represents the weight coefficient corresponding to each objective value among the different objective values, represents the sum of the different objective values.

[0037] In an alternative embodiment, the method further includes:

[0038] determining an adjustment value of the second policy weight parameter according to the second policy parameter weight and the first index value;

[0039] Update the second policy parameter weight according to the adjustment value.

[0040] The second aspect of this application provides an injection molding scheduling optimization device, which includes:

[0041] An acquisition module, configured to acquire injection molding scheduling data and construct a rule policy set according to the injection molding scheduling data;

[0042] A preliminary generation module, configured to generate an initial solution according to each rule policy and the corresponding first policy parameter weight in the rule policy set in sequence, and determine the initial solution as a preliminary scheduling plan;

[0043] An optimization module, configured to optimize the task processing order of the preliminary scheduling plan to determine the optimal processing order;

[0044] A first adjustment module, configured to fine-tune the second policy parameter weight corresponding to the algorithm policy to generate a neighborhood solution based on the initial solution, and calculate the first index value of the neighborhood solution, so that the neighborhood solution gradually approaches the local optimal solution;

[0045] A second adjustment module, configured to fine-tune the third policy parameter weight corresponding to the equipment policy to generate a global optimal solution based on the local optimal solution, and calculate the second index value of the global optimal solution;

[0046] A final generation module, configured to, when it is determined that the second index value meets the preset termination condition, output an optimal scheduling plan according to the optimal processing order for the global optimal solution.

[0047] The third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the steps of the injection molding scheduling optimization method are implemented.

[0048] The fourth aspect of this application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above injection molding scheduling optimization method are implemented.

[0049] In summary, the injection molding scheduling optimization method, device, electronic device, and computer storage medium provided by this application have at least one of the following beneficial effects:

[0050] 1. After obtaining the injection molding scheduling data, construct a set of rule strategies. Generate an initial solution according to each rule strategy and the corresponding first strategy parameter weight order, and determine it as the preliminary scheduling plan. Instead of simply relying on fixed rules or experience, multiple rule strategy combinations are constructed based on the scheduling data, and initial solutions are generated through different strategy parameter weight orders, increasing the diversity and flexibility of the scheduling plan, providing a wider selection space for subsequent optimization, and being able to better adapt to complex and changeable production environments, especially in the case of frequent order changes;

[0051] 2. After obtaining the preliminary scheduling plan, optimize the task processing order. By considering the impact of the task processing order on the scheduling efficiency in actual production, a more reasonable task execution order can be found through optimization, avoiding unreasonable order problems caused by fixed rules or experience, and further improving the adaptability of the scheduling plan to the production environment;

[0052] 3. Fine-tune the second strategy parameter weight corresponding to the algorithm strategy, generate a neighborhood solution based on the initial solution, and calculate the first index value of the neighborhood solution to make the neighborhood solution gradually approach the local optimal solution. In injection molding scheduling, the algorithm strategy can comprehensively consider various factors. By fine-tuning the parameter weight, it can be adjusted and optimized at the algorithm level according to the characteristics and mutual influence relationships of different factors, making the generated neighborhood solution more in line with the actual production requirements, gradually approaching the local optimal solution, and effectively coping with the complexity problems brought by the mutual influence of multiple factors;

[0053] 4. After obtaining the local optimal solution, further fine-tune the third strategy parameter weight corresponding to the equipment strategy, generate a global optimal solution based on the local optimal solution, and calculate the second index value of the global optimal solution. And the equipment strategy involves key factors such as machine selection and mold configuration. By fine-tuning the parameter weight, the equipment strategy can be optimized according to the operating state, performance characteristics of the equipment in actual production and the impact of different factors on equipment use, so as to generate a global optimal solution. By considering the dynamics of multiple factors and the diversity at the data level, a more satisfactory scheduling result can be obtained;

[0054] 5. When it is determined that the second index value meets the preset termination condition, output the optimal scheduling plan according to the optimal processing order for the global optimal solution. The preset termination condition can ensure obtaining the optimal scheduling plan within reasonable computing resources and time, avoiding over-computation or getting stuck in the local optimal solution and being unable to obtain the global optimal solution. By combining the optimal processing order and the global optimal solution, an optimal scheduling plan that comprehensively considers multiple factors and adapts to complex production environments can be output, effectively solving the problem that traditional methods are difficult to cope with the influence of multiple factors and obtain satisfactory scheduling results. Description of the Drawings

[0055] Figure 1It is a schematic flowchart of an injection molding scheduling optimization method shown in an embodiment of the present application;

[0056] Figure 2 It is a schematic flowchart of a method for generating a preliminary scheduling plan shown in an embodiment of the present application;

[0057] Figure 3 It is a schematic architecture diagram of an injection molding scheduling optimization system shown in an embodiment of the present application;

[0058] Figure 4 It is a functional module diagram of an injection molding scheduling optimization device shown in an embodiment of the present application;

[0059] Figure 5 It is a schematic structural diagram of an electronic device shown in an embodiment of the present application. Detailed implementation manners

[0060] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0061] The concept, specific structure and technical effects of the present invention will be clearly and completely described below in conjunction with the embodiments and the accompanying drawings, so as to fully understand the purpose, features and effects of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, other embodiments obtained by those skilled in the art without creative efforts shall fall within the scope of protection of the present invention. In addition, all the connection / connection relationships involved in the patent do not refer to the direct connection of components alone, but refer to the more optimal connection structure that can be formed by adding or reducing connection accessories according to the specific implementation situation. Each technical feature in the present invention can be combined interactively without conflicting with each other.

[0062] Refer to Figure 1 As shown, it is a schematic flowchart of an injection molding scheduling optimization method shown in an embodiment of the present application, and the injection molding scheduling optimization method includes the following steps.

[0063] S11, obtain injection molding scheduling data, and construct a rule strategy set according to the injection molding scheduling data.

[0064] In some embodiments, the electronic device first obtains injection molding scheduling data during the injection molding production process, including information such as production batches, downtime plans, mold repair orders, offline lead times, machine-mold matching, and equipment type priorities, and integrates the injection molding scheduling data to construct a data model. Among them, the data model covers basic data such as task sets, injection molding machine sets, and mold sets, as well as key parameters such as injection molding cycles, task switching times, and mold switching times. In addition, it also considers constraints in actual production such as task due dates, injection molding machine priorities, and mold repair plans, not only describing the basic situation of production tasks and resources, but also reflecting time, space, and process constraints in the production process, providing strong support for subsequent optimization calculations. Specifically, the electronic device first constructs a data model for the injection molding production process to objectively describe various types of data in the injection molding production process. The data model may include, but is not limited to: task sets, injection molding machine sets, mold sets, injection molding cycles, task switching times, mold switching times, model change times, machine-mold matching, task-mold matching relationships, injection molding machine priorities, task due dates, task lead times, injection molding machine downtime plans, mold repair plans, task product types, task type statuses, task-specified injection molding machines, task-specified molds, task auxiliary processes, and task bills of materials. Among them, the task set refers to the aggregate of production scheduling task batches, which is a one-dimensional data model and can be represented by ; the injection molding machine set refers to the aggregate of all available injection molding machines, which is a one-dimensional data model and can be represented by ; the mold set refers to the aggregate of all available molds, which is a one-dimensional data model and can be represented by ; the injection molding cycle refers to the injection molding cycles of all tasks on all molds, which is a two-dimensional data model; the task switching time refers to the time from task P to task The changeover time is a two-dimensional data model; the mold changeover time refers to the changeover time of all molds and is a one-dimensional data model; the style change time refers to the style change time of all tasks on all molds and is a two-dimensional data model; the machine-mold matching refers to the matching relationship between all molds and all injection molding machines and is a two-dimensional data model; the task-mold matching relationship refers to the matching relationship between all tasks and all molds and is a two-dimensional data model; the injection molding machine priority means that the smaller the data, the higher the priority and is a one-dimensional data model; the task due date refers to the start date of task delivery and is a one-dimensional data model; the task lead time refers to the time after subtracting the injection molding offline lead time from the task due date and is a one-dimensional data model; the injection molding machine shutdown plan refers to the unavailable time period of the injection molding machine and is a two-dimensional data model; the mold repair plan refers to the unavailable time period of the mold and is a two-dimensional data model; the task product type refers to the product types of all tasks and is a one-dimensional data model; the task type status refers to the types of all tasks, including statuses such as normal, trial mold, and normal lock, and is a one-dimensional data model; the task-specified injection molding machine refers to the injection molding machines locked for all tasks, and those not specified are set to null, and is a one-dimensional data model; the task-specified mold refers to the molds locked for all tasks, and those not specified are set to null, and is a one-dimensional data model; the task auxiliary process refers to whether all tasks require auxiliary equipment (nut machine / pad printer) and is a one-dimensional data model; the task bill of materials refers to the bill of materials release status of all tasks and is a one-dimensional data model. The production model is described from aspects such as tasks, resources, time, and process constraints through the data model. Based on the constructed data model, various time costs and various constraints in the production process can be accurately calculated.

[0065] While constructing the data model, before constructing the data model, or after completing the construction of the data model, the electronic device can also construct a set of rule strategies based on the injection molding scheduling data for the injection molding production process, which is used to refine the rules for generating the initial solution in the injection molding scheduling process. The set of rule strategies includes multiple rule strategies, which can include, but are not limited to: task sequencing rules, task clustering rules, task decomposition rules, mold selection strategies, injection molding machine selection strategies, and task order optimization strategies. In the rule parsing stage, the electronic device can perform refined processing on various rule strategies in the injection molding production process. For example, the task sequencing rule can be set according to factors such as the delivery start date and the machine cleaning date to ensure that tasks with higher priorities can be processed first. The task clustering rule can be divided according to factors such as the mold status and the material color, so as to classify similar tasks and improve the scheduling efficiency. Specifically, the task sequencing rule can refer to determining the task priority according to the delivery start date, the machine cleaning date, etc.; the task clustering rule can refer to classifying tasks according to the mold, insert status, material color, material, whether HF, etc.; the task decomposition rule can refer to decomposing tasks according to the delivery time gap, production cycle, machine closing strategy, etc.; the mold selection strategy can refer to selecting a mold using the load balancing or earliest completion strategy; the injection molding machine selection strategy can refer to selecting an injection molding machine using the load balancing or earliest completion strategy; the task order optimization strategy can refer to optimizing the task order by comprehensively considering factors such as the delivery date, order fulfillment rate, and the number of mold changeovers.

[0066] S12. Generate an initial solution according to each rule strategy and the corresponding first strategy parameter weight order in the set of rule strategies, and determine the initial solution as the preliminary scheduling plan.

[0067] In some embodiments, after each rule policy in the constructed rule policy set is completed, a corresponding rule ID and weight coefficient can be generated during initialization. Specifically, the electronic device can set the weights of different rule policies in the generation of the preliminary scheduling plan according to the production reality and rule requirements during the initialization of the rule policy, that is, the policy parameter weights. Among them, the rule ID refers to the unique identifier assigned by the electronic device to each rule policy, which is used to accurately call and manage the rules in the algorithm. For example, the task sorting rule (Rule_001), the mold selection strategy (Rule_002), and each rule ID corresponds to a weight parameter, which is used to determine the importance of each rule policy when generating the preliminary scheduling plan. It should be noted that the weight coefficient of the rule is determined based on domain knowledge, application frequency, and multiple debuggings. The sum of the weight coefficients is always 1. If a certain rule policy is more critical in actual production, its corresponding weight coefficient may be higher, so it will be given priority when generating the preliminary scheduling plan. In addition, after determining the policy weight coefficient, the policy weight coefficient can be adjusted according to actual needs. For example, in actual use, if the delivery date is urgent, the policy weight coefficient of the task sorting rule (by delivery date) is set higher.

[0068] After determining the policy parameter weights corresponding to each rule policy in the rule policy set, the electronic device can generate an initial solution, that is, a preliminary scheduling plan, according to the policy parameter weights and the order of the rule policy set.

[0069] In an alternative embodiment, the rule policy set includes a task sorting rule, a task clustering rule, a task classification rule, and a task order optimization rule. Generating an initial solution according to the order of each rule policy and the corresponding first policy parameter weight in the rule policy set includes:

[0070] Obtain the locked tasks within the locked time period, and prioritize the rule policies in the above rule policy set according to the first policy parameter weight;

[0071] Sort the locked tasks according to the task sorting rule corresponding to the sorted first weight to obtain an ordered task list;

[0072] Group the ordered task list according to the task clustering rule corresponding to the sorted second weight to obtain multiple clustered task groups;

[0073] Decompose each clustered task group according to the task classification rule corresponding to the sorted third weight to obtain multiple independent task sets;

[0074] Allocate a first task set and a second task set to each injection molding machine according to the multiple independent task sets

[0075] Adjust the order of the first task set and the second task set according to the task order optimization rule to generate the preliminary scheduling plan.

[0076] Refer to Figure 2 , the electronic device can first select according to the weight priority of the first policy parameter weight, that is, the rule policy with a higher policy parameter weight is preferentially selected. In the embodiment of the present application, the rule policy set may include a task sorting rule, a task clustering rule, a task classification rule, and a task order optimization rule. The electronic device first determines the input data, that is, determines the locked task set , including task attributes. Among them, the task attributes may include delivery date, mold type, material, priority, etc., and determine the rule policy set , and the first policy parameter weight , for example , , , . Then, sort the rule policies in the rule policy set in descending order of weight to generate an execution order queue . For the first weight after sorting, that is, the task sorting rule corresponding to the highest policy parameter weight, the tasks in the generated order can be sorted according to factors such as task priority, urgency, processing time, etc. Through sorting, it can be ensured that important tasks are processed first, and an ordered task list can be obtained . For the second weight, that is, the task clustering rule corresponding to the second highest policy parameter weight, on the basis of the task sorting rule, similar tasks are grouped together to reduce auxiliary times such as mold change and material change, and improve the utilization rate of the equipment. For example, tasks with the same mold or material are merged, and a clustered task group list can be output , where the second weight is less than the first weight and is adjacent to the first weight. Next, resource allocation is performed, that is, considering factors such as the applicability of the mold, the performance of the injection molding machine, and the availability of the equipment according to the mold selection strategy and the injection molding machine selection strategy, a suitable mold and injection molding machine are allocated to each task. Specifically, according to the injection molding machine / mold transient strategy, a task set with only one set of molds is allocated, which is called the first task set, and the task set that does not require mold change is allocated to compatible equipment, and the task set with multiple sets of molds is allocated, which is called the second task set, to process the task set that requires mold change. Considering the mold changeover cost, the task group can be split into smaller units, that is, based on the task decomposition rule corresponding to the third weight and on the basis of the task clustering rule, the clustered task set is split into the first task set and the second task set, where the third weight is less than the second weight and is adjacent to the second weight. Finally, according to the task order optimization rule corresponding to the fourth weight, the order of the first task set and the second task set is adjusted, that is, the task order is optimized for each piece of equipment in each task set, for example, minimizing the total mold change time, and the task execution plan corresponding to the optimized order sorting is output, that is, the preliminary scheduling plan.

[0077] S13, optimizing the task processing order of the preliminary scheduling plan to determine the optimal processing order.

[0078] In some embodiments, during the injection molding production process, the electronic device can optimize in multiple dimensions such as task processing order, equipment usage, and resource allocation. Specifically, in order to obtain a more efficient processing flow to optimize the task processing order on the injection molding set in the preliminary scheduling plan, first, the particle swarm algorithm is used to optimize each task on each injection molding machine one by one. In the injection molding scheduling, each particle represents a possible processing order. By calculating the fitness value of the particle (that is, the optimization objective function value, and the target value can be set according to the shortest completion time), the speed and position of the particle are continuously adjusted to output the optimal processing order. Specifically, in order to obtain a more efficient processing flow, the electronic device can optimize the task processing order on the injection molding machine in the initial solution. First, the particle swarm algorithm is used to optimize each task on each injection molding machine one by one. In the injection molding scheduling, each particle represents a possible processing order. By calculating the fitness value of the particle (and the optimization objective function value, and the target value can be set according to the shortest completion time), by continuously adjusting the speed and position of the particle, the optimal processing order is finally found.

[0079] Specifically, the fitness value is determined by calculating the makespan of the scheduling plan (that is, the completion time of the last task among all injection molding machines), including:

[0080] Step 1: Analyze the processing order represented by the particle. Each particle corresponds to a task allocation and processing order scheme. The particle encoding can be expressed as: 1) Which injection molding machine the task is allocated to; 2) The processing order of the tasks on each injection molding machine.

[0081] Step 2: Calculate the completion time of a single injection molding machine. For each injection molding machine , its task sequence is , and in addition, the injection cycle and switching time (mold change, model change, and material change time between tasks) of the tasks need to be calculated in sequence.

[0082] Calculate the completion time of a single injection molding machine through the following formula:

[0083] ;

[0084] Calculate the particle fitness (global maximum completion time) through the following formula:

[0085] ;

[0086] Among them, refers to the overdue penalty coefficient, and the overdue time refers to the difference between the task's time completion time and the delivery date.

[0087] After determining the fitness value, continuously adjust the velocity and position of the particles until the optimal processing sequence is found. Specifically, it includes the following steps:

[0088] 1. Initialization: Perform position encoding and velocity encoding, generate the initial particle swarm, and calculate the fitness value;

[0089] 2. Update the individual best (pbest) and the global best (gbest): Record the optimal position of each particle in the historical iteration (i.e., the processing sequence corresponding to the minimum fitness value), and record the optimal processing sequence among all particles (the position corresponding to the global minimum fitness value).

[0090] 3. Update the velocity: Update the velocity based on the individual and group experience.

[0091] Update the particle velocity through the following formula by means of a probabilistic exchange operation:

[0092] ;

[0093] Among them, w is the inertia weight, which is used to control the influence of the historical velocity; , is the learning factor, which is used to adjust the weights of the historical velocity and the group experience; , is a random number with a range of 0 to 1; represents the position difference, such as the set of exchange operations.

[0094] 4. Update the position: Perform exchange or insertion operations according to the velocity.

[0095] (1) Probabilistic exchange, which decides whether to perform an exchange operation according to the probability value in the velocity vector.

[0096] (2) Insertion operation, which sorts the velocities and selects a high-probability position for task insertion.

[0097] 5. Mutation operation: Perturb the position with a low probability;

[0098] 6. Termination condition: Reaching the maximum number of iterations or the fitness value converges.

[0099] By iteratively updating the particle velocities and positions, the task processing order can be continuously optimized. When the preset iteration termination condition is met, such as reaching the maximum number of iterations, or the global optimal fitness value has not been improved in consecutive N times (for example, 20 times), the optimal processing order can be output, where the optimal processing order is the task order scheme corresponding to the global optimum.

[0100] S14, fine-tuning the weight of the second strategy parameter corresponding to the algorithm strategy to generate a neighborhood solution based on the initial solution, and calculating the first index value of the neighborhood solution, so that the neighborhood solution gradually approaches the local optimal solution.

[0101] Among them, the algorithm strategy. After obtaining the preliminary scheduling plan, the electronic device can further optimize the preliminary production scheduling plan. In the embodiments of the present application, the electronic device utilizes a variety of algorithms and technologies in the optimization engine, such as the particle swarm algorithm, the local search algorithm, and the genetic algorithm, etc., and through mutual cooperation, jointly searches for the optimal production scheduling plan. For example, the particle swarm algorithm can find the optimal solution by simulating the foraging behavior of a bird flock, while the genetic algorithm can perform global search by simulating the biological evolution process. In addition, during the optimization process, an operator adaptive selection mechanism is introduced. By recording the changes in the strategy parameter weights and the corresponding index values of the better solutions, the strategy parameter weights are dynamically adjusted to achieve the adaptive selection of the operator, so as to automatically adjust the optimization strategy according to the actual situation and improve the efficiency and accuracy of the optimization.

[0102] When the task processing order is initially optimized, the electronic device can further generate a neighborhood solution by fine-tuning the weight of the strategy parameter of the algorithm strategy, that is, the second strategy parameter weight, and calculate the index value corresponding to each neighborhood solution (referred to as the first index value or the domain index value) to evaluate the quality of each neighborhood solution. Among them, the neighborhood solution is a new solution set obtained by fine-tuning the second strategy parameter weight (for example, the inertia weight in the particle swarm algorithm w , the learning factor , , or the crossover probability, mutation probability, etc. in the genetic algorithm) based on the current solution. By adjusting the second strategy parameter, the behavior of the algorithm in local search can be changed. For example, increasing the inertia weight wMore historical speeds will be retained to enhance the global search ability. The neighborhood search algorithm is a strategy for searching near the solution to find the optimal solution within a local range. To perform local search, the electronic device can utilize the neighborhood search algorithm to gradually approximate the local optimal solution by continuously fine-tuning the policy parameter weights and calculating the metric values (referred to as local optimal solution metric values).

[0103] Specifically, the electronic device selects an initial solution as the starting point of the search based on the neighborhood search algorithm and defines the search space of the problem, that is, the set of all possible solutions. Among them, the search space is usually high-dimensional, and each solution corresponds to a parameter vector. Then, based on the second policy weight, certain behaviors in the search process are adjusted, such as step size, direction, or the balance of the search strategy, to generate neighborhood solutions. Then, the first metric value corresponding to each neighborhood solution is calculated, where the first metric value is a function for evaluating the quality of the neighborhood solution.

[0104] Specifically, the electronic device can determine the first metric value through the following formula:

[0105] ;

[0106] Among them, the first metric value includes multiple objective values, such as device utilization rate, production capacity utilization rate, order fulfillment rate, total number of production line changeovers, total time consumed for production line changeovers, etc., represents different objective values among the multiple objective values, represents the weight coefficient corresponding to each objective value among the different objective values, represents the sum of the different objective values.

[0107] By continuously iteratively updating the second policy parameter weight within the preset conditions, after determining the first metric value, the electronic device can determine whether the first metric value meets the preset conditions. When it is determined that the preset conditions are met, the neighborhood solution corresponding to the best metric value is obtained as the output, that is, the optimal solution found in the search process, that is, the solution that meets the local optimal conditions. In some embodiments, the preset conditions may be: (1) the preset number of iterations to terminate the loop, such as 1000 times; (2) comparing the metric value calculated each time with the previous metric value, if the best metric value has not changed for consecutive N times (for example, 20 times); (3) the preset iteration loop termination time, such as 5 minutes.

[0108] In some embodiments, when the policy parameter weight and the first metric value are obtained, the electronic device can store them using an efficient data structure for quick access and query during the subsequent policy parameter weight update process to improve the calculation efficiency.

[0109] Through the above optional implementation manners, by combining multiple algorithms in the optimization engine and introducing an operator adaptive selection mechanism, generating a neighborhood solution using the second strategy parameter weight of the fine-tuning algorithm strategy and calculating its first metric value, gradually approaching the local optimal solution, the optimization efficiency and accuracy are improved, and the optimal solution is ensured to be found within a reasonable range through preset conditions, effectively enhancing the optimization effect of the production scheduling plan.

[0110] S15, fine-tune the third strategy parameter weight corresponding to the equipment strategy to generate a global optimal solution based on the local optimal solution, and calculate the second metric value of the global optimal solution.

[0111] S16, when it is determined that the second metric value meets the preset termination condition, output the optimal scheduling plan according to the optimal processing sequence for the global optimal solution.

[0112] Among them, the equipment strategy refers to optimization rules related to equipment such as machine selection, mold configuration, and load balancing. By adjusting the third strategy parameter weight of the equipment strategy (such as machine priority weight, mold change cost weight), the global scheduling plan is optimized to ensure the maximization of equipment utilization rate and meet production constraints. Based on the local optimal solution, the electronic device fine-tunes the strategy parameter weight of the equipment strategy, that is, the third strategy parameter weight (for example, the load balancing weight and mold change time weight in the injection molding machine selection strategy), to generate the final solution, and uses the genetic algorithm for global search. In injection molding scheduling optimization, the module regards each solution in the solution set as an individual, evaluates its pros and cons by calculating the metric value of the individual, and uses the crossover, mutation, and selection operations of the genetic algorithm for global search to find the global optimal solution, that is, the optimal scheduling plan. Exemplarily, if a certain injection molding machine needs to reduce its priority due to a maintenance plan, the task assignment can be reduced by adjusting its weight to avoid downtime conflicts.

[0113] According to the above embodiments for determining the same first metric value, the second metric value can be determined. By continuously iterating and updating the third strategy parameter weight within the preset conditions, when the second metric value is determined, the electronic device can determine whether the second metric value meets the preset conditions. When it is determined that the conditions are met, the global optimal solution corresponding to the best metric value is obtained as the output.

[0114] In some embodiments, the electronic device can refine and optimize the data model, rule strategy parsing, initial solution generation, optimization engine, operator adaptive selection, etc. according to different production environments and production requirements to better adapt to actual production needs.

[0115] In some embodiments, after the optimal scheduling plan is output, the electronic device can analyze the optimal scheduling plan, allocate specific execution times and execution devices for each task, including detailed information such as task allocation (injection molding machine / mold), processing sequence, processing time, etc., and output the optimal production scheduling plan in a visual or report form for production management personnel to reference and execute. In addition, the electronic device can also analyze detailed information such as task allocation, processing sequence, equipment usage, etc., and present it to the production scheduler in a visual or report form, so that the production scheduler can clearly understand the specific content of the production scheduling plan and thus perform production scheduling operations according to the plan.

[0116] Through the above optional implementation manners, by fine-tuning the weights of the device policy parameters to generate the global optimal solution and using the genetic algorithm for global search, the optimal scheduling plan can be accurately output, and it can also be refined and optimized according to production requirements, and the plan can be analyzed in detail and presented in a visual or report form, providing a clear basis for production scheduling and improving production efficiency and accuracy.

[0117] In an optional implementation manner, the method further includes:

[0118] Determine an adjustment value of the second policy weight parameter according to the second policy parameter weight and the first index value;

[0119] Update the second policy parameter weight according to the adjustment value.

[0120] In some embodiments, the electronic device can update the policy based on a learning mechanism, for example, algorithms such as linear interpolation, exponential smoothing, Bayesian optimization, etc. Through the update policy, calculate the adjustment value of the policy parameter weight according to historical data and new feedback information, and dynamically adjust the weight value according to the adjustment value to achieve real-time control of the optimization process, making the weight update more accurate and reasonable. Specifically, the electronic device can record in real time the set of policy parameter weights and the corresponding index values of the better solutions during the fine-tuning process, including the index values of the neighborhood solutions and the corresponding policy parameter weights, the index values of the local optimal solutions and the corresponding policy parameter weights. In addition, the electronic device can use an efficient data structure to record and store the policy parameter weights and index values, and can quickly access and query these records during subsequent weight update processes to improve calculation efficiency. At the same time, the electronic device can also preprocess and screen the data to be stored to ensure that the recorded data is representative and effective. For example, for duplicate or redundant data, the module will perform deduplication or merging operations; for abnormal data that significantly deviates from the average value or standard deviation, the module will perform deletion or correction operations.

[0121] In some embodiments, during the process of updating the weights of policy parameters, the electronic device can also set limiting conditions and thresholds to prevent over-adjustment or oscillation during the weight update process. It should be noted that the limiting conditions and thresholds can be set and adjusted according to the actual problem scenario. Similarly, the weights of the third policy parameters are updated according to the same embodiment method of adjusting the weights of the second policy parameters.

[0122] In the embodiments of the present application, with reference to Figure 3 , the electronic device can provide a front-end interaction layer, which can include a front-end interface for providing user operation entrances and functions such as visual scheduling dashboards, parameter configuration, and result display. In addition, the basic data (including static data such as management molds, equipment, and materials), scheduling tasks (including the current scheduling task list and status (such as locked, to be optimized, completed)), KPI indicators (including equipment utilization rate, order on-time delivery rate), and order pool are visually displayed through the front-end interface. Among them, the front-end interface communicates with the back-end algorithm service through the HTTPS protocol, sends requests (such as scheduling trigger instructions) and receives responses (such as optimization solutions). When the back-end algorithm service layer receives a user operation instruction from the front-end interface, it can standardize the front-end and back-end data interaction formats (such as JSON) using a data interface to ensure data transmission consistency and build a data model. Among them, the algorithm service layer can be divided into multiple data processing modules, which can include a scheduling optimization module and a parsing module, etc. The scheduling optimization module integrates optimization algorithms (such as particle swarm algorithm, genetic algorithm) to generate an optimal scheduling plan according to the weights of policy parameters and support adaptive weight adjustment (such as dynamic update based on Bayesian optimization); the parsing module parses the optimization results into a structure that can be displayed on the front end (such as Gantt chart data). In addition, the back end also provides a data support layer to provide a data center, where the data middle platform is used as a unified data management platform to record and store master data, such as core static data (such as equipment models, mold specifications). At the same time, it is also used to record and store offline data, such as historical orders and scheduling records, for training prediction models or analyzing trends, and generating model data, such as process parameters (such as injection molding cycle time, mold change rules), and real-time data, such as the current production status (such as equipment fault alarms, task progress). The algorithm service layer can obtain the required data (such as real-time equipment status affecting scheduling feasibility) from the data center through read operations. With reference to Figure 4 shown, it is a functional module diagram of the injection molding scheduling optimization device shown in the embodiments of the present application.

[0123] In some embodiments, the injection molding scheduling optimization device 40 can include multiple functional modules composed of computer program segments. The computer programs of each program segment of the injection molding scheduling optimization device 40 can be stored in the memory of the electronic device and executed by at least one processor to execute (see details in Figure 1Function for optimizing injection molding scheduling. According to the functions it performs, it can be divided into multiple functional modules. The functional modules may include: an acquisition module 401, a preliminary generation module 402, an optimization module 403, a first adjustment module 404, a second adjustment module 405, and a final generation module 406. The module referred to in this application means a series of computer program segments that can be executed by at least one processor and can complete fixed functions, and are stored in the memory. In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

[0124] The acquisition module 401 is used to acquire injection molding scheduling data and construct a rule strategy set according to the injection molding scheduling data.

[0125] The preliminary generation module 402 is used to generate an initial solution according to each rule strategy and the corresponding first strategy parameter weight order in the rule strategy set, and determine the initial solution as a preliminary scheduling plan.

[0126] The optimization module 403 is used to optimize the task processing order of the preliminary scheduling plan to determine the optimal processing order.

[0127] The first adjustment module 404 is used to fine-tune the second strategy parameter weight corresponding to the algorithm strategy to generate a neighborhood solution based on the initial solution, and calculate the first index value of the neighborhood solution to make the neighborhood solution gradually approach the local optimal solution.

[0128] The second adjustment module 405 is used to fine-tune the third strategy parameter weight corresponding to the equipment strategy to generate a global optimal solution based on the local optimal solution, and calculate the second index value of the global optimal solution.

[0129] The final generation module 406 is used to output an optimal scheduling plan according to the optimal processing order for the global optimal solution when it is determined that the second index value meets the preset termination condition.

[0130] The preliminary generation module 402 is further specifically configured to: obtain the locked tasks within the locked time period, and prioritize the rule policies in the above-mentioned rule policy set according to the first policy parameter weight; sort the locked tasks according to the task sorting rule corresponding to the sorted first weight to obtain an ordered task list; group the ordered task list according to the task clustering rule corresponding to the sorted second weight to obtain multiple clustered task groups; decompose each of the clustered task groups according to the task classification rule corresponding to the sorted third weight to obtain multiple independent task sets; allocate a first task set and a second task set to each injection molding machine according to the multiple independent task sets, and adjust the order of the task sets for the first task set and the second task set according to the task order optimization rule to generate the preliminary scheduling plan.

[0131] The optimization module 403 is further specifically configured to: optimize the task processing order on each injection molding machine by using the particle swarm algorithm, the preliminary scheduling plan includes the task processing order corresponding to each injection molding machine among multiple injection molding machines, and each particle represents a processing order plan; optimize the task processing order by iteratively updating the speed and position of the particles to output the optimal processing order, and the optimal processing order is the task order plan corresponding to the global optimum.

[0132] The optimization module 403 is further specifically configured to: determine the fitness function through the following formula:

[0133] ;

[0134] where is the fitness function, is the completion time of a single injection molding machine, is the overdue penalty coefficient;

[0135] Determine the completion time of the single injection molding machine through the following formula:

[0136] ;

[0137] where is the injection molding machine j 's completion time, is the task k 's processing time on the injection molding machine j , is the task k 's switching time with the next task.

[0138] The optimization module 403 is further specifically configured to: update the particle velocity through the following formula:

[0139] ;

[0140] Among them, w is the inertia weight, , are the learning factors, , are random numbers, ranging from 0 to 1, represents the position difference.

[0141] The first adjustment module 404 is further specifically configured to: The first index value includes multiple target values, and the first index value is determined through the following formula:

[0142] ;

[0143] Among them, represents different target values among the multiple target values, represents the weight coefficient corresponding to each target value among the different target values, represents the sum of the different target values.

[0144] The first adjustment module 404 is further configured to: determine an adjustment value of the second policy weight parameter according to the second policy parameter weight and the first index value; update the second policy parameter weight according to the adjustment value.

[0145] It should be understood that the various change modes and specific embodiments in the injection molding scheduling optimization method provided in the above embodiments are equally applicable to the injection molding scheduling optimization device in this embodiment. Through the foregoing detailed description of the injection molding scheduling optimization method, those skilled in the art can clearly know the implementation method of the injection molding scheduling optimization device in this embodiment. For the sake of simplicity of the specification, it will not be elaborated herein.

[0146] Refer to Figure 5 shown, which is a schematic structural diagram of an electronic device shown in an embodiment of the present application. In a preferred embodiment of the present application, the electronic device 5 includes a memory 51, at least one processor 52, and at least one communication bus 53.

[0147] Those skilled in the art should understand that Figure 5 the structure of the electronic device shown does not constitute a limitation on the embodiments of the present application. It can be a bus structure or a star structure. The electronic device 5 may further include more or fewer other hardware or software than shown, or different component arrangements.

[0148] In some embodiments, the electronic device 5 is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions. Its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits, programmable gate arrays, digital signal processors, and embedded devices, etc. The electronic device 5 may also include a user device, which includes, but is not limited to, any electronic product that can interact with a user through means such as a keyboard, mouse, remote control, touchpad, or voice control device. For example, personal computers, tablet computers, smartphones, digital cameras, etc.

[0149] In the above embodiments provided in the present application, it should be understood that the disclosed methods, devices, computer-readable storage media, and electronic devices can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods. For example, multiple components or modules can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, indirect couplings or communication connections of devices or components or modules, which can be electrical, mechanical, or other forms.

[0150] The components described as separate components may or may not be physically separated. The components displayed as components may or may not be physical modules, that is, they may be located in one place, or they may be distributed to multiple network modules. Some or all of the components can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0151] In addition, in each embodiment of the present invention, the functional modules can be integrated in one processing module, or each component can exist physically alone, or two or more modules can be integrated in one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0152] When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0153] It should be noted that for the foregoing method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be carried out in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily all essential to the present invention.

[0154] In the above embodiments, the descriptions of the various embodiments have their own focuses. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0155] The above is a specific description of the preferred embodiments of the present invention, but the present invention is not limited to the described embodiments. Those skilled in the art can also make various equivalent deformations or substitutions without departing from the spirit of the present invention, and these equivalent deformations or substitutions are all included within the scope defined by the claims of this application.

Claims

1. An injection molding scheduling optimization method, characterized in that, The method includes: Obtaining injection molding scheduling data and constructing a set of rule strategies according to the injection molding scheduling data, where the set of rule strategies includes a task sorting rule, a task clustering rule, a task classification rule, and a task order optimization rule; Generating an initial solution according to each rule strategy and the corresponding first strategy parameter weight order in the set of rule strategies, and determining the initial solution as a preliminary scheduling plan, including: obtaining locked tasks within a locked time period, and performing priority sorting on the rule strategies in the above set of rule strategies according to the first strategy parameter weight; sorting the locked tasks according to the task sorting rule corresponding to the sorted first weight to obtain an ordered task list; grouping the ordered task list according to the task clustering rule corresponding to the sorted second weight to obtain multiple clustered task groups; decomposing each clustered task group according to the task classification rule corresponding to the sorted third weight to obtain multiple independent task sets; allocating a first task set and a second task set to each injection molding machine according to the multiple independent task sets, and adjusting the task set order of the first task set and the second task set according to the task order optimization rule to generate the preliminary scheduling plan; Optimizing the task processing order of the preliminary scheduling plan to determine the optimal processing order; Fine-tune the weight of the second policy parameter corresponding to the algorithm policy to generate a neighborhood solution based on the initial solution, and calculate the first index value of the neighborhood solution so that the neighborhood solution gradually approaches the local optimal solution, where calculating the first index value of the neighborhood solution includes: the first index value includes multiple target values, and the first index value is determined by the following formula: ; where represents different target values among the multiple target values, represents the weight coefficient corresponding to each target value among the different target values, represents the sum of the different target values; Fine-tuning the third strategy parameter weight corresponding to the equipment strategy to generate a global optimal solution based on the local optimal solution, and calculating the second index value of the global optimal solution; When it is determined that the second index value meets the preset termination condition, outputting an optimal scheduling plan according to the optimal processing order for the global optimal solution.

2. The injection molding scheduling optimization method according to claim 1, wherein The optimizing the task processing order of the preliminary scheduling plan to determine the optimal processing order includes: Using a particle swarm algorithm to optimize the task processing order on each injection molding machine, where the preliminary scheduling plan includes the task processing order corresponding to each injection molding machine among multiple injection molding machines, and each particle represents a processing order plan; Optimizing the task processing order by iteratively updating the speed and position of the particles to output the optimal processing order, where the optimal processing order is the task order plan corresponding to the swarm optimum.

3. The injection molding scheduling optimization method according to claim 2, wherein The method further includes: Determining a fitness function through the following formula: ； Among them, is the fitness function, is the completion time of a single injection molding machine, is the overdue penalty coefficient; Determining the completion time of the single injection molding machine through the following formula: ； Among them, is the processing time of the task k on the injection molding machine j , and is the switching time of the task k with the next task.

4. The injection molding scheduling optimization method according to claim 2, characterized in that, The updating the speed of the particles includes: Updating the particle speed through the following formula: ； Among them, w is the inertia weight, , are the learning factors, , are random numbers, ranging from 0 to 1, represents the position difference.

5. The injection molding scheduling optimization method according to claim 1, characterized in that The method further includes: Determining an adjustment value of the second strategy weight parameter according to the second strategy parameter weight and the first index value; Updating the second strategy parameter weight according to the adjustment value.

6. An injection molding scheduling optimization device, characterized in that, The device includes: An obtaining module, configured to obtain injection molding scheduling data and construct a set of rule strategies according to the injection molding scheduling data; A preliminary generation module, configured to generate an initial solution according to each rule policy and the corresponding first policy parameter weight order in the rule policy set, and determine the initial solution as a preliminary scheduling plan, including: obtaining locked tasks within a locked time period, and performing priority sorting on the rule policies in the rule policy set according to the first policy parameter weight; sorting the locked tasks according to the task sorting rule corresponding to the sorted first weight to obtain an ordered task list; grouping the ordered task list according to the task clustering rule corresponding to the sorted second weight to obtain multiple clustered task groups; decomposing each clustered task group according to the task classification rule corresponding to the sorted third weight to obtain multiple independent task sets; allocating a first task set and a second task set to each injection molding machine according to the multiple independent task sets, and adjusting the order of the task sets of the first task set and the second task set according to the task order optimization rule to generate the preliminary scheduling plan; An optimization module, configured to optimize the task processing order of the preliminary scheduling plan to determine an optimal processing order; The first adjustment module is used to finely adjust the weight of the second policy parameter corresponding to the algorithm policy to generate a neighborhood solution based on the initial solution, and calculate the first index value of the neighborhood solution, so that the neighborhood solution gradually approaches the local optimal solution, where calculating the first index value of the neighborhood solution includes: the first index value includes multiple target values, and the first index value is determined by the following formula: ; where represents different target values among the multiple target values, represents the weight coefficient corresponding to each target value among the different target values, represents the sum of the different target values; A second adjustment module, configured to finely adjust the third policy parameter weight corresponding to the device policy to generate a global optimal solution based on the local optimal solution, and calculate a second index value of the global optimal solution; A final generation module, configured to, when it is determined that the second index value meets a preset termination condition, output an optimal scheduling plan according to the optimal processing order for the global optimal solution.

7. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the steps of the injection molding scheduling optimization method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the injection molding scheduling optimization method according to any one of claims 1 to 5 are implemented.

Images