A production scheduling method for resource-constrained parallel machines based on intensive strategy coordination

Abstract

The purpose of this invention is to provide a model scheduling method for resource-constrained parallel machines based on intensive strategy coordination. This method includes: acquiring order data; calculating a model scheduling scheme based on the order data using a domain-based heuristic algorithm; and optimizing the model scheduling scheme using an intelligent algorithm based on model-address intensive optimization. This invention addresses the challenge of efficiently generating scientifically effective model scheduling schemes in resource-constrained parallel machine scenarios, solving the computational complexity issue arising from the complex mapping between orders, addresses, and models; and considering the problem of solution space fragmentation caused by multiple constraints. Simultaneously, it proactively considers delivery, reducing the optimization difficulty of delivery scheduling, thereby improving production efficiency and reducing delivery costs.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a method for scheduling the production of resource-constrained parallel machines based on intensive strategy coordination. Background Technology

[0002] Overall, scale and cost advantages are crucial factors for the development of the home appliance industry. However, changing environments are placing increasingly stringent demands on the industry. Specifically, firstly, the market demand for diverse, small-batch production of various products is gradually replacing the previous demand for fewer varieties and larger batches. This has led to high production costs for traditional, inventory-oriented mass production models due to product diversification and rapid iteration. Secondly, to reduce finished goods warehouse space, companies have adopted a direct-shipment model where completed orders are shipped directly to the desired address. However, facing fragmented order demands and highly dispersed delivery locations, this mass production model results in orders at the same address being completed in a scattered manner, increasing distribution costs. Therefore, it is necessary to improve the current mass production model in light of product diversification and the need for integrated production and distribution. However, most manufacturers rely on third-party logistics for distribution, leading to information asymmetry between manufacturers and distributors. This makes it difficult to jointly consider integrated production and distribution, further increasing revenue pressure on companies.

[0003] While balancing product diversity and intensive production, considering intensive distribution to multiple locations and developing production scheduling plans undoubtedly greatly increases the difficulty of scheduling for planners. The decision-making dimensions, factors to be considered, and the amount of data required in the scheduling process for planners increase exponentially, increasing the scheduling time for planners and the risk of oversights. Summary of the Invention

[0004] The purpose of this invention is to provide a resource-constrained parallel machine model scheduling method based on intensive strategy coordination. This method efficiently generates scientific and effective model scheduling schemes for resource-constrained parallel machine scenarios, solving the problem of complex computational complexity caused by the complex mapping between orders, addresses, and models; and addressing the issue of solution space fragmentation caused by multiple constraints. Simultaneously, it proactively considers delivery, reducing the optimization difficulty of delivery scheduling, thereby improving production efficiency and reducing delivery costs.

[0005] A production scheduling method for resource-constrained parallel machines based on intensive strategy coordination includes:

[0006] Obtain order data;

[0007] A domain-based heuristic algorithm is used to calculate the model production scheduling scheme based on the order data;

[0008] The production scheduling scheme for the model is optimized using an intelligent algorithm based on model address aggregation optimization.

[0009] Preferably, obtaining order data includes:

[0010] Obtain information on the number of machines in the factory, the models that the machines can produce, the machine start-up time, the production time required for each model, the resources required for each model, and customer order information.

[0011] Preferably, the step of using a domain-based heuristic algorithm to calculate the model production scheduling scheme based on the order data includes:

[0012] Retrieve order information, model information, resource information, work calendar information, and configuration information;

[0013] Order information, model information, resource information, work calendar information, and configuration information are converted into integer data by establishing key-value pair matching.

[0014] The order information is summarized by month and model to obtain the set of models that need to be produced;

[0015] Based on the data preprocessing workflow, redundant data in the model information and resource information of the current version are removed;

[0016] Initialize the list of idle machines;

[0017] Determine if the current shift is the first shift;

[0018] Choose between a continuous down-model strategy or an up-model strategy based on domain knowledge and constraints, depending on the current shift information.

[0019] Output the production scheduling plan for each model.

[0020] Preferably, the step of selecting a continuous takeover strategy for the machine or a takeover strategy based on domain knowledge and constraints based on the current shift information includes:

[0021] If the current shift is the first shift, select the continuous takeover strategy for the machine, including:

[0022] Obtain the production schedule of the previous shift and pre-allocate it as the mold layout plan for the current shift.

[0023] Based on the previous shift's production schedule, determine whether the molds that have already been installed have reached the maximum number of continuous production days;

[0024] If the maximum continuous production days are reached, verify the relationship between the remaining demand for the mold model and the current capacity. The current available capacity = number of shifts for mold installation * maximum number of molds for the current model * UPH for the current model. If the number of machines with molds installed for the current model cannot meet the minimum continuous production days limit when production continues, and the remaining demand can be completed within the monthly range, then remove the mold from one of the machines and re-verify the current step.

[0025] If the maximum continuous production days have not been reached, determine whether the molds that have been installed have reached the maximum continuous production days. If the maximum continuous production days have been reached, check whether all orders within the monthly range for the corresponding mold model have been completed. If all orders have been completed, proceed with the mold removal process. If all orders have not been completed, do not proceed with the mold removal process.

[0026] The door body resources of the mold layout plan for the current shift are checked. If the door body resources are insufficient to support the production of the corresponding model of the mold, the mold is removed. If the door body resources are sufficient to support the production of the model, the mold is not removed.

[0027] The remaining model information is checked for production capacity. The check is made to see if the demand can be met within the available production capacity of the model. If it cannot be met, the model with the constraint is selected first, and then the mold with more upper molds is selected for the lower mold.

[0028] Save the set of models for which the mold has been completed.

[0029] Preferably, the step of selecting a continuous takeover strategy for the machine or a takeover strategy based on domain knowledge and constraints based on the current shift information includes:

[0030] If the current shift is not the first shift, select a modeling strategy based on domain knowledge and constraints, including:

[0031] Summarize the list of all available machines in the current shift, and prioritize them according to the number of orders in the current month for the same model, from highest to lowest.

[0032] Prioritize models that are subject to model exclusivity and machine production constraints, then select models with higher model priority, and sort the set of models to be produced according to model priority to obtain the list of candidate models for the current shift's upper mold.

[0033] During the pre-allocation of production plans for the previous shift, the set of models that were assigned to the lower mold are removed from the list of candidate models for the upper mold in the current shift.

[0034] Summarize the list of molds that have been installed on non-idle machines in the current shift, and remove mutually exclusive models from the list of alternative mold models for the current shift.

[0035] The current shift's list of available mold models is arranged from highest to lowest. The corresponding mold model is selected for mold processing based on the remaining order demand. The maximum number of products that can be processed = min(box body mold production capacity, door body resource supply capacity) / daily production capacity of a single mold.

[0036] Determine if the number of mold types exceeds the maximum number of mold types. If it does, select a new model for mold installation and remove the model corresponding to the current mold from the list of available mold models for the current shift.

[0037] If the maximum number of mold types is not exceeded, determine whether there is a mutual exclusion relationship between the molds corresponding to the model. If the mold corresponding to the model to be newly processed is mutually exclusive with other molds already processed in the current shift, then remove the model corresponding to the mold from the current shift's mold candidate model list P, and at the same time reselect the model for mold processing.

[0038] If there is no mutual exclusion between the molds corresponding to the model, determine whether there is a mold application area restriction on the machine and mold currently being molded. If there is a mold application area restriction, select a model from the mold application candidate list for the current shift and process the mold application.

[0039] If there is no restriction on the area where the mold can be installed, determine whether there are any idle machines in the current shift. If there are idle machines in the current shift, and the list of alternative mold models for the current shift has not been completely traversed, then continue to select a model for mold installation.

[0040] If there are no idle machines in the current shift or the list of candidate models for the upper mold in the current shift has been completely traversed, then jump to the next shift and re-perform the production scheduling plan pre-allocation and upper mold strategy based on domain knowledge and constraints from the previous shift.

[0041] Preferably, the optimization of the model production scheduling scheme using an intelligent algorithm based on model address aggregation includes:

[0042] Obtain order information, model information, resource information, work calendar information, and production scheduling plan for the next higher-level model;

[0043] Initialize the number of iterations;

[0044] The order information is summarized by month, model, and address to obtain a set of models and addresses that need to be produced;

[0045] Based on the address aggregation algorithm, the address allocation sequence is obtained;

[0046] Based on the address allocation sequence, the production scheduling plan of the previous level model, and the set of models and addresses to be produced, determine the number of units of each model produced by each machine in each shift and the addresses to be shipped to, summarize the information, and update the set of models and addresses to be produced.

[0047] Based on the model address joint optimization algorithm, combined with the model production scheduling plan of the previous level and the model address shipment summary information, a new model production scheduling plan is obtained, and the resource information is used to verify whether the resources can meet the requirements.

[0048] If resources are sufficient, output an optimized production scheduling plan for the model.

[0049] Preferably, obtaining the address allocation sequence based on the address aggregation algorithm includes:

[0050] Obtain configuration information such as iteration count, population size, and mutation probability;

[0051] Based on the configured population size, an encoding matrix is ​​created according to the address and month, and A different numbers of addresses are randomly generated according to the month to fill the matrix;

[0052] Calculate the fitness value and sort each chromosome S in the population according to the fitness value;

[0053] Following the method of roulette, select two different chromosomes S1 and S2;

[0054] Randomly select a gene from S1 as the crossover point C1, record the row information of C1, randomly select a gene from S2 as the crossover point C2, and perform position crossover on C1 and C2;

[0055] Detect whether there are duplicate or missing digits in the encoded sequence of S1 and S2 after crossover;

[0056] If there are duplicate or missing numbers in the encoded sequence, randomly select a duplicate number and replace the missing number.

[0057] Check if the current situation meets the mutation conditions. If it does, regenerate a new chromosome and insert it into the population, and remove those chromosomes with low fitness evaluation indicators.

[0058] Check if the current situation has reached the upper limit of the number of iterations. If it has, output the optimal result obtained in the current iteration.

[0059] If the maximum number of iterations has not been reached, continue iterating to find the best option.

[0060] Preferably, decoding the chromosome using an address reduction optimization algorithm includes:

[0061] Extract month information and the order information of each address in each month from the chromosome, and index each month one by one;

[0062] Obtain the production schedule for the current month, index the addresses sequentially based on the address order information, and obtain the current address information;

[0063] Order information belonging to the current month and current address is summarized by month, quantity required, model, and address, and sorted in descending order of the quantity of the model to be produced within the same address;

[0064] By combining the shift and shift duration information provided by the work calendar of the foaming line for the cabinet with the hourly capacity information in the cabinet mold information table, the number of units that can be produced is calculated as: Model UPH * Address aggregation time * Number of molds on the machine corresponding to the model. This gives us how many units of each model can be produced within the time frame of the consolidated shipment, and the number of models that can be produced is sorted from largest to smallest.

[0065] The system iterates through the collected information and selects machines within the same address and time range in sequence according to the obtained information for allocation until all orders under the current address in the current month are allocated or there is no capacity available for allocation within the same address and time range. At the same time, the information is updated so that machines that have already been occupied can no longer be allocated.

[0066] After allocating the collected information, find the next address and model to allocate;

[0067] After completing the address allocation for the current month, we will continue to search for addresses for the next month.

[0068] Preferably, the iterative strategy of the joint optimization algorithm for model address includes:

[0069] Obtain information on iteration count, model concentration emphasis, and decision step size configuration;

[0070] Calculate whether each machine has undergone mold changes compared to the previous shift under each shift, and count the number of machines undergoing mold changes in each shift;

[0071] Count the number of addresses delivered under each shift, and calculate the degree of optimization of the coded addresses under each shift dimension based on the importance of model clustering;

[0072] Based on the framework solution of production scheduling by model and the mapping relationship of address intensive allocation algorithm, the number of units produced in each address of each mold corresponding to each model is summarized, and the addresses are sorted according to the number of units produced.

[0073] Select the model mold and address in sequence. Reselect and assign the molds corresponding to the model within the shift and their priority order. For each assigned mold, check whether the model can be produced simultaneously, the mold quantity limit, and the machine production constraints. If the constraint check passes, use the corresponding mold. If the constraint check fails, select the next mold in the sequence for mold installation.

[0074] If it is possible to reduce mold switching and the number of units required for the address can support single-shift allocation, then allocate the currently selected mold and address;

[0075] If it is possible to reduce mold switching, but the number of units required at the address is less than the number of units that the machine can produce in a shift, then allocate the remaining number of units to be produced at the current address and the number of units required at the next sequential address in the current mold.

[0076] If it is possible to reduce mold switching but there is no corresponding mold address for the required number of units that need to be allocated, select the next sequence of molds for allocation.

[0077] Preferably, after optimizing the production scheduling scheme for the model using an intelligent algorithm based on model address aggregation, the process further includes evaluating the optimized production scheduling scheme, specifically:

[0078] Based on the average number of days each address is allocated Number of model changes between adjacent shifts A comprehensive evaluation was conducted, resulting in the following comprehensive evaluation:

[0079]

[0080] The optimal production scheduling scheme for each model across multiple iterations is used as the final production scheduling scheme.

[0081] Where m represents month information, a represents address information, i represents shift information, j represents machine information, p represents model information, inf2 represents the number of model changes when inputting the model and address joint optimization algorithm, inf3 represents the average number of days allocated when inputting the model and address joint optimization algorithm, nf2 represents the number of model changes in the solution result under the current optimization, and nf3 represents the average number of days allocated in the solution result under the current optimization.

[0082] The beneficial effects of this invention are as follows: The heuristic algorithm based on domain rules is fundamentally based on configuring relevant business parameters according to actual production environment conditions. These parameters are then used to apply molds to the required models shift by shift according to the work calendar, ensuring model continuity while maintaining product diversity. The intelligent algorithm based on model address optimization is also fundamental. Based on the first part of the model production scheduling plan, addresses are allocated, and the model and address are optimized using intelligent algorithms, thereby improving address continuity. Attached Figure Description

[0083] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0084] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0085] Figure 1 This is a flowchart of a resource-constrained parallel machine model scheduling method based on intensive strategy coordination according to the present invention;

[0086] Figure 2 This is a schematic diagram of the domain rule-based heuristic algorithm flow of the present invention;

[0087] Figure 3 This is a schematic diagram of the model address aggregation optimization algorithm based on intelligent algorithms of the present invention;

[0088] Figure 4 This is a schematic diagram of the address aggregation algorithm decoding allocation strategy of the present invention;

[0089] Figure 5 This is a schematic diagram illustrating the iterative process of the model address joint optimization algorithm of the present invention. Detailed Implementation

[0090] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0091] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0092] Furthermore, the use of terms such as "first" and "second" in this invention is for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" and "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but only on the basis of being achievable by those skilled in the art. When the combination of technical solutions is contradictory or impossible to implement, such a combination of technical solutions should be considered non-existent and not within the scope of protection claimed by this invention.

[0093] While balancing product diversity and intensive production, considering intensive distribution to multiple locations and developing production scheduling plans undoubtedly greatly increases the difficulty of scheduling for planners. The decision-making dimensions, factors to be considered, and the amount of data required in the scheduling process for planners increase exponentially, increasing the scheduling time for planners and the risk of oversights.

[0094] This invention employs a domain-rule-based heuristic algorithm. Its core is to configure relevant business parameters based on actual production environment conditions. These parameters are then used to apply molds to the required models shift by shift according to the work calendar, ensuring model continuity while maintaining product diversity. The intelligent algorithm based on model address optimization, on the other hand, allocates addresses based on the first model production scheduling plan and optimizes both models and addresses using intelligent algorithms, thereby improving address continuity.

[0095] A production scheduling method for resource-constrained parallel machines based on intensive strategy coordination includes:

[0096] S100, retrieve order data;

[0097] The S200 uses a domain-based heuristic algorithm to calculate the model production scheduling scheme based on order data;

[0098] The S300 uses an intelligent algorithm based on model address aggregation optimization to optimize the model production scheduling scheme.

[0099] Preferably, in step S100, obtaining order data includes:

[0100] Obtain information on the number of machines in the factory, the models that the machines can produce, the machine start-up time, the production time required for each model, the resources required for each model, and customer order information.

[0101] Because actual production data is configured by planners, misconfigurations and omissions can occur. For example, to meet market demand, a company might launch a new model for sale, and customers might place orders for it, but the corresponding supply route for this new model cannot be found in the system. To ensure the efficiency and stability of the algorithm and to shield against interference from abnormal data, this module performs a data preprocessing process before the production data enters the algorithm.

[0102] By configuring the user's quick code, information such as the maximum and minimum continuous production days and the machine's production matrix can be obtained. The data is then processed by establishing key-value pair relationships to achieve standardized processing that converts character information into integer information.

[0103] The products in the order information table are summarized by month to obtain the set N of models that need to be produced.

[0104] Based on the required production model set N, check if a corresponding relationship exists in the information table with the model field. If a single model exists, and the data is not found in the model-related information table but is present in the production model set N, then delete the corresponding data to improve the efficiency of the algorithm's data retrieval. If a single model exists, and the data is not found in the model-related information table but is present in the production model set N, then transfer the order corresponding to that model to a new order set D, and provide the relevant information to the user after completing the entire production scheduling process.

[0105] Preferably, in step S200, the model scheduling scheme is calculated based on order data using a domain-based heuristic algorithm, including:

[0106] Retrieve order information, model information, resource information, work calendar information, and configuration information;

[0107] Order information, model information, resource information, work calendar information, and configuration information are converted into integer data by establishing key-value pair matching.

[0108] The order information is summarized by month and model to obtain the set of models that need to be produced;

[0109] Based on the data preprocessing workflow, redundant data in the model information and resource information of the current version are removed;

[0110] Initialize the list of idle machines;

[0111] Determine if the current shift is the first shift;

[0112] Choose between a continuous down-model strategy or an up-model strategy based on domain knowledge and constraints, depending on the current shift information.

[0113] Output the production scheduling plan for each model.

[0114] This module primarily utilizes the experience of production schedulers to develop a heuristic algorithm based on domain knowledge. Following this algorithm's flow, a production scheduling plan for each model is derived. This ensures product diversity and production continuity, avoiding infeasible solutions arising from constraints. The module mainly comprises shift scheduling decisions, continuous mold-taking strategies, and mold-up strategies based on domain knowledge and constraints. The relevant processes are as follows: Figure 2 As shown.

[0115] Pseudocode for a domain-knowledge-based production-intensive model scheduling algorithm:

[0116] Input: Order information (Order_Info), Model information (Fac_Info), Resource information (Res_Info), Work calendar information (Data_Info), Configuration information (Config_Info);

[0117] The order information (Order_Info), model information (Fac_Info), resource information (Res_Info), work calendar information (Data_Info), and configuration information (Config_Info) are converted into integer data by establishing key-value pair matching.

[0118] Summarize the order information Order_Info by month and model to obtain the set N of models that need to be produced;

[0119] for i = 1: length(N)

[0120] According to the data preprocessing process, irrelevant data in the model information Fac_Info and resource information Res_Info of the current version are removed;

[0121] end for

[0122] for i = 1:length(Data_Info)

[0123] Initialize the idle machine list (idle_List);

[0124] If the current bus is not the first bus.

[0125] Inherit the mold information from the previous shift to the current shift, and perform mold removal on the machine according to the continuous mold removal strategy, while updating the idle machine list idle_List;

[0126] end if

[0127] Based on the domain knowledge and constraint-based modeling strategy, the idle machine list idle_List is modeled.

[0128] end for

[0129] Output the production scheduling scheme Mould_scheme1;

[0130] Preferably, selecting a continuous takeover strategy for the machine or a takeover strategy based on domain knowledge and constraints based on the current shift information includes:

[0131] If the current shift is the first shift, select the continuous takeover strategy for the machine, including:

[0132] Obtain the production schedule of the previous shift and pre-allocate it as the mold layout plan for the current shift.

[0133] Based on the previous shift's production schedule, determine whether the molds that have already been installed have reached the maximum number of continuous production days;

[0134] If the maximum continuous production days are reached, verify the relationship between the remaining demand for the mold model and the current capacity. The current available capacity = number of shifts for mold installation * maximum number of molds for the current model * UPH for the current model. If the number of machines with molds installed for the current model cannot meet the minimum continuous production days limit when production continues, and the remaining demand can be completed within the monthly range, then remove the mold from one of the machines and re-verify the current step.

[0135] If the maximum continuous production days have not been reached, determine whether the molds that have been installed have reached the maximum continuous production days. If the maximum continuous production days have been reached, check whether all orders within the monthly range for the corresponding mold model have been completed. If all orders have been completed, proceed with the mold removal process. If all orders have not been completed, do not proceed with the mold removal process.

[0136] The door body resources of the mold layout plan for the current shift are checked. If the door body resources are insufficient to support the production of the corresponding model of the mold, the mold is removed. If the door body resources are sufficient to support the production of the model, the mold is not removed.

[0137] The remaining model information is checked for production capacity. The check is made to see if the demand can be met within the available production capacity of the model. If it cannot be met, the model with the constraint is selected first, and then the mold with more upper molds is selected for the lower mold.

[0138] Save the set of models for which the mold has been completed.

[0139] Shift scheduling decision: This decision primarily determines whether the current shift can inherit shift information. If it can, it is the first shift; otherwise, it is not. If the current shift is the first shift, a model-based strategy based on domain knowledge and constraints is executed. If the current shift is not the first shift, the scheduling information of the previous shift is retrieved, and the current shift is pre-assigned according to the scheduling plan of the previous shift.

[0140] Preferably, selecting a continuous takeover strategy for the machine or a takeover strategy based on domain knowledge and constraints based on the current shift information includes:

[0141] If the current shift is not the first shift, select a modeling strategy based on domain knowledge and constraints, including:

[0142] Summarize the list of all available machines in the current shift, and prioritize them according to the number of orders in the current month for the same model, from highest to lowest.

[0143] Prioritize models that are subject to model exclusivity and machine production constraints, then select models with higher model priority, and sort the set of models to be produced according to model priority to obtain the list of candidate models for the current shift's upper mold.

[0144] During the pre-allocation of production plans for the previous shift, the set of models that were assigned to the lower mold are removed from the list of candidate models for the upper mold in the current shift.

[0145] Summarize the list of molds that have been installed on non-idle machines in the current shift, and remove mutually exclusive models from the list of alternative mold models for the current shift.

[0146] The current shift's list of available mold models is arranged from highest to lowest. The corresponding mold model is selected for mold processing based on the remaining order demand. The maximum number of products that can be processed = min(box body mold production capacity, door body resource supply capacity) / daily production capacity of a single mold.

[0147] Determine if the number of mold types exceeds the maximum number of mold types. If it does, select a new model for mold installation and remove the model corresponding to the current mold from the list of available mold models for the current shift.

[0148] If the maximum number of mold types is not exceeded, determine whether there is a mutual exclusion relationship between the molds corresponding to the model. If the mold corresponding to the model to be newly processed is mutually exclusive with other molds already processed in the current shift, then remove the model corresponding to the mold from the current shift's mold candidate model list P, and at the same time reselect the model for mold processing.

[0149] If there is no mutual exclusion between the molds corresponding to the model, determine whether there is a mold application area restriction on the machine and mold currently being molded. If there is a mold application area restriction, select a model from the mold application candidate list for the current shift and process the mold application.

[0150] If there is no restriction on the area where the mold can be installed, determine whether there are any idle machines in the current shift. If there are idle machines in the current shift, and the list of alternative mold models for the current shift has not been completely traversed, then continue to select a model for mold installation.

[0151] If there are no idle machines in the current shift or the list of candidate models for the upper mold in the current shift has been completely traversed, then jump to the next shift and re-perform the production scheduling plan pre-allocation and upper mold strategy based on domain knowledge and constraints from the previous shift.

[0152] Preferably, in step S300, the intelligent algorithm based on model address aggregation optimization is used to optimize the model production scheduling scheme, including:

[0153] Obtain order information, model information, resource information, work calendar information, and production scheduling plan for the next higher-level model;

[0154] Initialize the number of iterations;

[0155] The order information is summarized by month, model, and address to obtain a set of models and addresses that need to be produced;

[0156] Based on the address aggregation algorithm, the address allocation sequence is obtained;

[0157] Based on the address allocation sequence, the production scheduling plan of the previous level model, and the set of models and addresses to be produced, determine the number of units of each model produced by each machine in each shift and the addresses to be shipped to, summarize the information, and update the set of models and addresses to be produced.

[0158] Based on the model address joint optimization algorithm, combined with the model production scheduling plan of the previous level and the model address shipment summary information, a new model production scheduling plan is obtained, and the resource information is used to verify whether the resources can meet the requirements.

[0159] If resources are sufficient, output an optimized production scheduling plan for the model.

[0160] Since the higher-level production scheduling scheme considers the continuity of model production while ensuring product diversity, local optimization based on the higher-level production scheduling scheme to ensure the intensive completion of orders at the same address can effectively improve algorithm efficiency and solution quality. The model address intensive optimization algorithm based on intelligent algorithms specifically establishes a mapping relationship between addresses and models based on the higher-level production scheduling scheme, and then uses this mapping relationship to perform intensive coordination and optimization of the model and address production scheduling scheme. Its core mainly consists of an address intensive allocation algorithm and a model address joint scheduling algorithm, and the relevant process is as follows: Figure 3 As shown.

[0161] Model scheduling search algorithm for intensive production and distribution coordination:

[0162] Input: Order information (Order_Info), model information (Fac_Info), resource information (Res_Info), work calendar information (Data_Info), and the parent model production scheduling scheme (Mould_scheme1);

[0163] Initialize the number of iterations (count)

[0164] The order information Order_Info is summarized by month, model and address to obtain the set N_add of the models and addresses that need to be produced;

[0165] Based on the address set allocation algorithm, the address allocation sequence M is obtained;

[0166] for i = 1: length(M)

[0167] Based on the address allocation sequence M, the upper-level model production scheduling scheme Mould_scheme1, and the set of models and addresses to be produced N_add, determine the number of units of each model produced by each machine in each shift and the addresses to be shipped to, summarize the information Add_scheme, and update the set of models and addresses to be produced N_add.

[0168] end for

[0169] for i = 1: length(count)

[0170] Based on the model address joint optimization algorithm, combined with the previous level model production scheduling scheme Mould_scheme1 and the model address shipment summary information Add_scheme, a new model production scheduling scheme Mould_scheme2 is obtained, and the resource information Res_Info is used to verify whether the resources can meet the requirements.

[0171] end for

[0172] Output model production scheduling scheme Mould_scheme2;

[0173] Preferably, the address allocation sequence obtained according to the address aggregation algorithm includes:

[0174] Obtain configuration information such as iteration count, population size, and mutation probability;

[0175] Based on the configured population size, an encoding matrix is ​​created according to the address and month, and A different numbers of addresses are randomly generated according to the month to fill the matrix;

[0176] Calculate the fitness value and sort each chromosome S in the population according to the fitness value;

[0177] Following the method of roulette, select two different chromosomes S1 and S2;

[0178] Randomly select a gene from S1 as the crossover point C1, record the row information of C1, randomly select a gene from S2 as the crossover point C2, and perform position crossover on C1 and C2;

[0179] Detect whether there are duplicate or missing digits in the encoded sequence of S1 and S2 after crossover;

[0180] If there are duplicate or missing numbers in the encoded sequence, randomly select a duplicate number and replace the missing number.

[0181] Check if the current situation meets the mutation conditions. If it does, regenerate a new chromosome and insert it into the population, and remove those chromosomes with low fitness evaluation indicators.

[0182] Check if the current situation has reached the upper limit of the number of iterations. If it has, output the optimal result obtained in the current iteration.

[0183] If the maximum number of iterations has not been reached, continue iterating to find the best option.

[0184] Preferably, decoding the chromosome using an address reduction optimization algorithm includes:

[0185] Extract month information and the order information of each address in each month from the chromosome, and index each month one by one;

[0186] Obtain the production schedule for the current month, index the addresses sequentially based on the address order information, and obtain the current address information;

[0187] Order information belonging to the current month and current address is summarized by month, quantity required, model, and address, and sorted in descending order of the quantity of the model to be produced within the same address;

[0188] By combining the shift and shift duration information provided by the work calendar of the foaming line for the cabinet with the hourly capacity information in the cabinet mold information table, the number of units that can be produced is calculated as: Model UPH * Address aggregation time * Number of molds on the machine corresponding to the model. This gives us how many units of each model can be produced within the time frame of the consolidated shipment, and the number of models that can be produced is sorted from largest to smallest.

[0189] The system iterates through the collected information and selects machines within the same address and time range in sequence according to the obtained information for allocation until all orders under the current address in the current month are allocated or there is no capacity available for allocation within the same address and time range. At the same time, the information is updated so that machines that have already been occupied can no longer be allocated.

[0190] After allocating the collected information, find the next address and model to allocate;

[0191] After completing the address allocation for the current month, we will continue to search for addresses for the next month.

[0192] For example, suppose address 1 requires 1.5X units of model 1, 0.5X units of model 2, and 2X units of model 3; address 2 requires 1.5X units of model 1 and 0.5X units of model 2. The allocation action is as follows: Figure 4 As shown.

[0193] Preferably, refer to Figure 5 The iterative strategy for the joint optimization algorithm of model address includes:

[0194] Obtain information on iteration count, model concentration emphasis, and decision step size configuration;

[0195] Calculate whether each machine has undergone mold changes compared to the previous shift under each shift, and count the number of machines undergoing mold changes in each shift;

[0196] Count the number of addresses delivered under each shift, and calculate the degree of optimization of the coded addresses under each shift dimension based on the importance of model clustering;

[0197] Based on the framework solution of production scheduling by model and the mapping relationship of address intensive allocation algorithm, the number of units produced in each address of each mold corresponding to each model is summarized, and the addresses are sorted according to the number of units produced.

[0198] Select the model mold and address in sequence. Reselect and assign the molds corresponding to the model within the shift and their priority order. For each assigned mold, check whether the model can be produced simultaneously, the mold quantity limit, and the machine production constraints. If the constraint check passes, use the corresponding mold. If the constraint check fails, select the next mold in the sequence for mold installation.

[0199] If it is possible to reduce mold switching and the number of units required for the address can support single-shift allocation, then allocate the currently selected mold and address;

[0200] If it is possible to reduce mold switching, but the number of units required at the address is less than the number of units that the machine can produce in a shift, then allocate the remaining number of units to be produced at the current address and the number of units required at the next sequential address in the current mold.

[0201] If it is possible to reduce mold switching but there is no corresponding mold address for the required number of units that need to be allocated, select the next sequence of molds for allocation.

[0202] Preferably, after optimizing the model scheduling scheme using an intelligent algorithm based on model address aggregation, the process further includes evaluating the optimized scheduling scheme, specifically:

[0203] Based on the average number of days each address is allocated Number of model changes between adjacent shifts A comprehensive evaluation was conducted, resulting in the following comprehensive evaluation:

[0204]

[0205] The optimal production scheduling scheme for each model across multiple iterations is used as the final production scheduling scheme.

[0206] Where m represents month information, a represents address information, i represents shift information, j represents machine information, p represents model information, inf2 represents the number of model changes when inputting the model and address joint optimization algorithm, inf3 represents the average number of days allocated when inputting the model and address joint optimization algorithm, nf2 represents the number of model changes in the solution result under the current optimization, and nf3 represents the average number of days allocated in the solution result under the current optimization.

[0207] This module aims to evaluate the merits of the algorithm based on the iteration results, retain individuals with better fitness to improve the optimization effect, and at the same time perform relevant robustness analysis to demonstrate the robustness of the algorithm.

[0208] Table 1: What model is produced in machine j during shift i?

[0209]

[0210]

[0211] The proposed address allocation scheme is shown in Table 2. The number before the comma in the table represents the shipping address, and the number after the comma represents the number of units shipped. For example, the data [2,77] means that 77 refrigerators will be shipped to address 2.

[0212] Table 2: Proposed Address Allocation Scheme

[0213]

[0214]

[0215] Robustness analysis of production scheduling scheme. (Using formulas...) A robustness analysis was conducted on the production scheduling plan for the model. Among them, f m f represents the maximum value. avg f represents the average value. dev The deviation of the objective function is represented. The data robustness analysis is shown in Table 3.

[0216] Table 3: Data Robustness Analysis

[0217]

[0218] This invention employs a domain-rule-based heuristic algorithm. Its core is to configure relevant business parameters based on actual production environment conditions. These parameters are then used to apply molds to the required models shift by shift according to the work calendar, ensuring model continuity while maintaining product diversity. The intelligent algorithm based on model address optimization, on the other hand, allocates addresses based on the first model production scheduling plan and optimizes both models and addresses using intelligent algorithms, thereby improving address continuity.

[0219] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A production scheduling method for resource-constrained parallel machines based on intensive strategy coordination, characterized in that, include: Obtain order data; A domain-based heuristic algorithm is used to calculate the model production scheduling scheme based on the order data; The production scheduling scheme for the aforementioned model is optimized using an intelligent algorithm based on model address aggregation optimization. Based on the model address joint optimization algorithm, combined with the model production scheduling plan of the previous level and the model address shipment summary information, a new model production scheduling plan is obtained, and the resource information is used to verify whether the resources can meet the requirements. If resources are sufficient, output an optimized production scheduling plan for the model; The iterative strategy of the joint optimization algorithm for model address includes: Obtain information on iteration count, model concentration emphasis, and decision step size configuration; Calculate whether each machine has changed molds compared to the previous shift under each shift, and count the number of machines that changed molds in each shift; Count the number of addresses delivered under each shift, and calculate the degree of optimization of the coded addresses under each shift dimension based on the importance of model clustering; Based on the framework solution of production scheduling for each model and the mapping relationship of the address intensive allocation algorithm, the number of units produced in each address of each mold corresponding to each model is summarized, and the addresses are sorted according to the number of units produced. Select the model mold and address in sequence. Reselect and assign the molds corresponding to the model within the shift and their priority order. For each assigned mold, check whether the model can be produced simultaneously, the mold quantity limit, and the machine production constraints. If the constraint check passes, use the corresponding mold. If the constraint check fails, select the next mold in the sequence for mold installation. If it is possible to reduce mold switching and the number of units required for the address can support single-shift allocation, then allocate the currently selected mold and address; If it is possible to reduce mold switching, but the number of units required at the address is less than the number of units that the machine can produce in a shift, then allocate the remaining number of units to be produced at the current address and the number of units required at the next sequential address in the current mold. If it is possible to reduce mold switching but there is no corresponding mold address for the required number of units that need to be allocated, select the next sequence of molds for allocation.

2. The resource-constrained parallel machine model scheduling method based on intensive strategy coordination according to claim 1, characterized in that, The acquisition of order data includes: Obtain information on the number of machines in the factory, the models that the machines can produce, the machine start-up time, the production time required for each model, the resources required for each model, and customer order information.

3. The resource-constrained parallel machine model scheduling method based on intensive strategy coordination according to claim 1, characterized in that, The method of using a domain-based heuristic algorithm to calculate the model production scheduling scheme based on the order data includes: Retrieve order information, model information, resource information, work calendar information, and configuration information; Order information, model information, resource information, work calendar information, and configuration information are converted into integer data by establishing key-value pair matching. The order information is summarized by month and model to obtain the set of models that need to be produced; According to the data preprocessing process, redundant data in the model information and resource information of the current version are removed; Initialize the list of idle machines; Determine if the current shift is the first shift; Choose between a continuous down-model strategy or an up-model strategy based on domain knowledge and constraints, depending on the current shift information. Output the production scheduling plan for each model.

4. The resource-constrained parallel machine model scheduling method based on intensive strategy coordination according to claim 3, characterized in that, The strategy of selecting a continuous down-shift strategy or an up-shift strategy based on domain knowledge and constraints based on the current shift information includes: If the current shift is the first shift, select the continuous takeover strategy for the machine, including: Obtain the production schedule of the previous shift and pre-allocate it as the mold layout plan for the current shift. Based on the previous shift's production schedule, determine whether the molds that have already been installed have reached the maximum number of continuous production days; If the maximum continuous production days are reached, verify the correspondence between the remaining demand of the mold model and the current capacity. The current available capacity = number of shifts on which the mold can be installed * maximum number of molds for the current model * UPH for the current model. If the number of machines on which the mold is installed for the current model cannot meet the minimum continuous production days limit when production continues, and the remaining demand can be completed within the monthly range, then remove the mold from one of the machines and re-verify the current step. If the maximum number of continuous production days has not been reached, check whether all orders for the corresponding model of the mold within the monthly range have been completed. If all orders have been completed, proceed with the mold removal process; otherwise, do not proceed with the mold removal process. The door body resources of the mold layout plan for the current shift are checked. If the door body resources are insufficient to support the production of the corresponding model of the mold, the mold is removed. If the door body resources are sufficient to support the production of the model, the mold is not removed. The remaining model information is checked for production capacity. The check is made to see if the demand can be met within the available production capacity of the model. If it cannot be met, the model with the constraint is selected first, and then the mold with more upper molds is selected for the lower mold. Save the set of models for which the mold has been completed.

5. The resource-constrained parallel machine model scheduling method based on intensive strategy coordination according to claim 3, characterized in that, The strategy of selecting a continuous down-shift strategy or an up-shift strategy based on domain knowledge and constraints based on the current shift information includes: If the current shift is not the first shift, select a modeling strategy based on domain knowledge and constraints, including: Summarize the list of all available machines in the current shift, and prioritize them according to the number of orders in the current month for the same model, from highest to lowest. Prioritize models that are subject to model exclusivity and machine production constraints, then select models with higher model priority according to model priority. After sorting the set of models to be produced according to model priority, the list of candidate models for the upper mold in the current shift is obtained. During the pre-allocation of production plans for the previous shift, the set of models that were assigned to the lower mold are removed from the list of candidate models for the upper mold in the current shift. Summarize the list of molds that have been installed on non-idle machines in the current shift, and remove mutually exclusive models from the list of alternative mold models for the current shift. The current shift's list of available mold models is arranged from highest to lowest. The corresponding mold model is selected for mold processing based on the remaining order demand. The maximum number of products that can be processed is min(box body mold production capacity, door body resource supply capacity) / daily production capacity of a single mold. Determine if the number of mold types exceeds the maximum number of mold types. If it does, select a new model for mold installation and remove the model corresponding to the current mold from the mold installation candidate list for the current shift. If the maximum number of mold types is not exceeded, determine whether there is a mutual exclusion relationship between the molds corresponding to the model. If the mold corresponding to the model to be newly processed is mutually exclusive with other molds already processed in the current shift, then remove the model corresponding to the mold from the current shift's mold candidate model list P, and at the same time reselect the model for mold processing. If there is no mutual exclusion between the molds corresponding to the model, determine whether there is a mold application area restriction on the machine and mold currently being molded. If there is a mold application area restriction, select a model from the mold application candidate list for the current shift and process the mold application. If there is no restriction on the area where the mold can be installed, determine whether there are any idle machines in the current shift. If there are idle machines in the current shift, and the list of alternative mold models for the current shift has not been completely traversed, then continue to select a model for mold installation. If there are no idle machines in the current shift or the list of candidate models for the upper mold in the current shift has been completely traversed, then jump to the next shift and re-perform the production scheduling plan pre-allocation and upper mold strategy based on domain knowledge and constraints from the previous shift.

6. The resource-constrained parallel machine model scheduling method based on intensive strategy coordination according to claim 1, characterized in that, The optimization of the model scheduling scheme using an intelligent algorithm based on model address aggregation includes: Obtain order information, model information, resource information, work calendar information, and production scheduling plan for the previous model; Initialize the number of iterations; The order information is summarized by month, model, and address to obtain a set of models and addresses that need to be produced; Based on the address aggregation algorithm, the address allocation sequence is obtained; Based on the address allocation sequence, the production scheduling plan of the previous level model, and the set of models and addresses to be produced, determine the number of units of each model produced by each machine in each shift and the addresses to be shipped to, summarize the information, and update the set of models and addresses to be produced.

7. The resource-constrained parallel machine model scheduling method based on intensive strategy coordination according to claim 6, characterized in that, The address allocation sequence obtained based on the address aggregation algorithm includes: Obtain configuration information such as iteration count, population size, and mutation probability; Based on the configured population size, an encoding matrix is ​​created according to the address and month, and A different numbers of addresses are randomly generated according to the month to fill the matrix; Calculate the fitness value and sort each chromosome S in the population according to the fitness value; Following the method of roulette, select two different chromosomes S1 and S2; Randomly select a gene from S1 as the crossover point C1, record the row information of C1, randomly select a gene from S2 as the crossover point C2, and perform position crossover on C1 and C2; Detect whether there are duplicate or missing digits in the encoded sequence of S1 and S2 after crossover; If there are duplicate or missing numbers in the encoded sequence, randomly select a duplicate number and replace the missing number. Check if the current situation meets the mutation conditions. If it does, regenerate a new chromosome and insert it into the population, and remove those chromosomes with low fitness evaluation indicators. Check if the current situation has reached the upper limit of the number of iterations. If it has, output the optimal result obtained in the current iteration. If the maximum number of iterations has not been reached, continue iterating to find the best option.

8. The resource-constrained parallel machine model scheduling method based on intensive strategy coordination according to claim 7, characterized in that, Decoding chromosomes using address reduction optimization algorithms includes: Extract month information and the order information of each address in each month from the chromosome, and index each month one by one; Obtain the production schedule for the current month, index the addresses sequentially based on the address order information, and obtain the current address information; Order information belonging to the current month and current address is summarized by month, quantity required, model, and address, and sorted in descending order of the quantity of the model to be produced within the same address; By combining the shift and shift duration information provided by the work calendar of the foaming line for the box body with the hourly capacity information in the box body mold information table, the number of units that can be produced is calculated as: Model UPH * Address aggregation time * Number of molds on the machine corresponding to the model. This gives us how many units of each model can be produced within the time frame of the consolidated shipment, and the number of models that can be produced is sorted from largest to smallest. The system iterates through the collected information and selects machines within the same address and time range in sequence according to the obtained information for allocation until all orders under the current address in the current month are allocated or there is no capacity available for allocation within the same address and time range. At the same time, the information is updated so that machines that have already been occupied can no longer be allocated. After allocating the collected information, find the next address and model to allocate; After traversing all addresses for the current month, continue searching for addresses for the next month.

9. The resource-constrained parallel machine model scheduling method based on intensive strategy coordination according to claim 1, characterized in that, After optimizing the production scheduling scheme for the model using an intelligent algorithm based on model address aggregation optimization, the process also includes evaluating the optimized production scheduling scheme.

Images