A data generation method, device, apparatus, and storage medium

Abstract

The application discloses a data generation method and device, equipment and a storage medium. The method comprises the following steps: determining at least one product corresponding to a target material matching scene by using a data generation rule, and expanding production materials for each product in the at least one product step by step until the target level is reached, so as to obtain corresponding materials at each level; the data generation rule is generated based on real material matching data corresponding to the target material matching scene; the feature information of each material at each level is expanded by using the data generation rule, so as to obtain material feature information corresponding to each material; the material feature information at least comprises one or more of the following: inventory, use priority, and material type; and the materials at each level and the material feature information corresponding to each material at each level are determined as material information corresponding to the target material matching scene.

Description

Technical Field

[0001] This application relates to the field of data simulation technology, and in particular to a data generation method, apparatus, device and storage medium. Background Technology

[0002] In industrial production, material kitting is a crucial step to ensure the availability of required materials. Traditional methods for solving material kitting problems rely on heuristic rules, which require extensive data for simulation experiments and parameter tuning during the design process. However, the amount of real-world data in industrial production is limited, and acquiring, compiling, and organizing this data is challenging. Furthermore, data confidentiality requirements can hinder its use. Besides traditional heuristic methods, machine learning methods can also be used to solve material kitting problems, improving efficiency and accuracy. However, machine learning methods are also data-driven, placing higher demands on the quantity and quality of data, potentially leading to data insufficiency and compromising the security of real-world data. Summary of the Invention

[0003] This application aims to provide a data generation method, apparatus, device, and storage medium.

[0004] The technical solution of this application embodiment is implemented as follows:

[0005] This application provides a data generation method that uses data generation rules to determine at least one product corresponding to a target material completeness scenario, and expands the production materials for each product in the at least one product level by level until the target level is reached, thereby obtaining the corresponding materials at each level; the data generation rules are generated based on the actual material completeness data corresponding to the target material completeness scenario.

[0006] Using the data generation rules, feature information is expanded for each material in each level of materials to obtain the material feature information corresponding to each material; the material feature information includes at least one or more of the following: inventory quantity, usage priority, and material type;

[0007] The materials at each level, and the material feature information corresponding to each material at each level, are determined as the material information corresponding to the target material completeness scenario.

[0008] This application provides a data generation apparatus, including:

[0009] The determination module is used to use data generation rules to determine at least one product corresponding to the target material completeness scenario, and to expand the production materials for each product in the at least one product step by step until the target level is reached to obtain the corresponding materials at each level; the data generation rules are generated based on the real material completeness data corresponding to the target material completeness scenario.

[0010] An expansion module is used to expand the feature information of each material in each level of materials using the data generation rules, so as to obtain the material feature information corresponding to each material; the material feature information includes at least one or more of the following: inventory quantity, usage priority, and material type;

[0011] The determining module is further configured to determine the materials at each level, and the material feature information corresponding to each material at each level, as the material information corresponding to the target material matching scenario.

[0012] This application provides a data generation device, which includes: a processor, a memory, and a communication bus;

[0013] The communication bus is used to realize the communication connection between the processor and the memory;

[0014] The processor is used to execute the computer program stored in the memory to implement the above-described data generation method.

[0015] This application provides a computer-readable storage medium storing one or more computer programs that can be executed by one or more processors to implement the above-described data generation method. Attached Figure Description

[0016] Figure 1 A flowchart illustrating a data generation method provided in an embodiment of this application;

[0017] Figure 2 This is a schematic diagram of an exemplary material expansion process provided for an embodiment of this application;

[0018] Figure 3 A schematic diagram illustrating an exemplary process for generating scene data, provided as an embodiment of this application;

[0019] Figure 4 A schematic diagram of an exemplary order group structure provided for an embodiment of this application;

[0020] Figure 5 This application provides an exemplary flowchart for generating data generation rules. Figure 1 ;

[0021] Figure 6 This application provides an exemplary flowchart for generating data generation rules. Figure 2 ;

[0022] Figure 7 A schematic diagram illustrating an exemplary model training process provided in this application embodiment;

[0023] Figure 8 A schematic diagram illustrating an exemplary data filtering process provided in this application embodiment;

[0024] Figure 9 This is a schematic diagram illustrating an exemplary process for generating target scene data, provided as an embodiment of this application.

[0025] Figure 10 This is a schematic flowchart illustrating an exemplary process for determining a target solution model, provided as an embodiment of this application.

[0026] Figure 11 This is a schematic diagram of the structure of a data generation device provided in an embodiment of this application;

[0027] Figure 12 This is a schematic diagram of the structure of a data generation device provided in an embodiment of this application. Detailed Implementation

[0028] The technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. It should be understood that the specific embodiments described herein are merely for explaining the relevant application and not for limiting the application. Furthermore, it should be noted that, for ease of description, only the parts relevant to the application are shown in the accompanying drawings.

[0029] This application provides a data generation method, implemented by a data generation device, such as... Figure 1 As shown, the process includes the following steps S101 to S103:

[0030] Step S101: Using data generation rules, determine at least one product corresponding to the target material completeness scenario, and expand the production materials for each product in the at least one product step by step until the target level is reached to obtain the corresponding materials at each level; the data generation rules are generated based on the real material completeness data corresponding to the target material completeness scenario.

[0031] In the embodiments of this application, the data generation device is an electronic device with data generation function. For example, the data generation device may be an electronic device such as a computer device or a smart terminal device.

[0032] In the embodiments of this application, the target material completeness scenario can be the material completeness scenario of various production industries and enterprises. For example, the target material completeness scenario can be scenarios such as food, electronics, automobiles, and manufacturing. The data generation device can use data generation rules to determine at least one product corresponding to the target material completeness scenario. The data generation rules are generated based on the real material completeness data corresponding to the target material completeness scenario. For example, if the target material completeness scenario is an electronic material completeness scenario, then the at least one product determined based on the data generation rules can be: mouse, keyboard, host, computer, headphones, etc.

[0033] In the embodiments of this application, after identifying at least one product for a target material completeness scenario, data generation rules are used to progressively expand the production materials for each product within the at least one product until the target level is reached, thus obtaining the corresponding materials at each level. An exemplary expansion method is described in [reference needed]. Figure 2 As shown, for the expansion of final product A (product), the first level is expanded: material 1-1 and material 1-2. Then, for the expanded material 1-1 and material 1-2, a second level expansion is performed. Material 1-1 expands to material 2-1, and material 1-2 expands to material 2-2 and material 2-3, and so on, expanding to the target level to obtain the materials at each level corresponding to product A. There are production dependencies between the materials at each level; see [link to documentation]. Figure 2 The lines connecting the materials (solid and dashed lines); dashed lines indicate that the level is a newly added material.

[0034] Step S102: Using data generation rules, expand the feature information of each material in each level of materials to obtain the material feature information corresponding to each material; the material feature information includes at least one or more of the following: inventory quantity, usage priority, and material type.

[0035] In the embodiments of this application, the data generation device will continue to use data generation rules to expand the feature information of each material in each level of materials corresponding to each product; wherein, the material feature information includes at least one or more of the following: inventory quantity, usage priority, and material type.

[0036] Step S103: Determine the material information corresponding to the target material matching scenario as the material information of each level of material and the material characteristic information of each material in each level of material.

[0037] In the embodiments of this application, the data generation device determines the material information corresponding to the target material completeness scenario as the material information of each level of material and the material feature information of each material in each level of material.

[0038] For example, in the material information generation stage, this stage uses finished goods (products) as units, constructing a sub-bill of materials (each material in each level of material) of a specific size for each finished product. Starting from each finished product, the material information is gradually (level by level) expanded using a queue method. Specifically, materials whose information needs to be improved are maintained using a queue (level by level). The material at the head of the queue is dequeued, and its inventory, type, and other characteristics (material characteristic information) are determined according to preset data generation rules (data generation rules). Sub-materials required for its production are generated (continuous expansion). The newly generated sub-materials are added to the queue for information improvement, and iterative material information expansion continues until data of a specified size is obtained (until expansion reaches the target level). See the flowchart of the exemplary material expansion process. Figure 2 As shown.

[0039] Compared to the difficulties in data acquisition and data confidentiality issues encountered in related technologies regarding target material completeness scenarios, this application can generate data generation rules that characterize scenario features based on real material completeness data of the target material completeness scenario, and then generate simulation data that conforms to the target material completeness scenario based on the data generation rules, thereby reducing the difficulty of data acquisition and improving the security of real material completeness data.

[0040] In some embodiments, the data generating device may also perform the following steps S301 to S303:

[0041] Step S301: Using data generation rules, generate at least one order for the target material completeness scenario, and generate corresponding order feature information for each order in the at least one order; the order can be: formal order, stock preparation order, replenishment order or forecast order; the order feature information includes at least the materials required for the order.

[0042] In the embodiments of this application, the data generation device generates at least one order based on data generation rules for the scenario of complete set of target materials; wherein, at least one order may include one or more of the following: formal order, stock preparation order, replenishment order, and forecast order. Of course, at least one order may also be one of the above order types, such as at least one order being a formal order, a stock preparation order, a replenishment order, or a forecast order.

[0043] In the embodiments of this application, after the data generation device generates at least one order for the target material completeness scenario, it generates corresponding order feature information for each order in the at least one order; wherein, the order feature information includes at least the materials required for the order. Of course, the order feature information may also include information such as demand quantity, priority, and completion deadline. The specific order feature information can be set according to the application scenario and actual needs, and this application does not limit it.

[0044] Step S302: Classify at least one order according to the required materials to obtain at least one set of orders.

[0045] In embodiments of this application, after receiving at least one order, the data generation device classifies the at least one order according to the required materials to obtain at least one group of orders, wherein each group of orders in the at least one group of orders requires the same materials.

[0046] In the embodiments of this application, after generating material information, the data generation device generates orders according to data generation rules, and then manages the orders in groups. To facilitate the classification of orders, each order requires only one material. If an order requires multiple materials, the data generation device will split the order into multiple orders, each requiring only one material. For example, if order A requires material B and material C, then order A is split into order A1 and order A2. Order A1 requires material B, and order A2 requires material C. The data generation device groups orders requiring the same material together, ensuring that at least one group of orders requires the same material. See [link to relevant documentation]. Figure 4 As shown, orders 1, 2, ..., and 13 in order group 1 require the same materials and include order characteristic information for each order, such as demand quantity and priority.

[0047] Step S303: Based on the required materials of each order in at least one set of orders, perform production relationship matching between each order and the corresponding materials included in the material information in each set of orders to generate scenario data corresponding to the target material completeness scenario.

[0048] In the embodiments of this application, the data generation device matches the production relationships between each order and the corresponding materials included in the material information of each order in at least one group of orders, and generates scene data corresponding to the target material completeness scenario, based on the required materials of each group of orders. For example, assuming that all orders in order group 1 require material A, the data generation device will establish production relationships between orders 1, orders 2, ..., orders 13 included in order group 1 and material A to generate scene data corresponding to the target material completeness scenario.

[0049] For example, the data generation device matches the generated materials and orders. The matching method is to use a random strategy to specify a material as the target material for each order. Depending on the scenario requirements, a heuristic matching strategy based on material and order characteristics can also be used to replace the random strategy.

[0050] In some embodiments, the data generating device may also perform the following steps S501 to S503:

[0051] Step S501: Obtain actual material completeness data.

[0052] In the embodiments of this application, the data generation device acquires real material kitting data for the target material kitting scenario.

[0053] Step S502: Analyze the standard data that characterizes the target material availability scenario in the real material availability data; the standard data shall include at least one or more of the following: the inventory distribution of different materials, the demand distribution of different orders, the probability distribution of different materials, and the distribution of materials required for order production.

[0054] In the embodiments of this application, after obtaining real material completeness data, the data generation device analyzes the standard data in the real material completeness data that characterizes the characteristics of the target material completeness scenario. The standard data includes at least one or more of the following: the inventory distribution of different materials, the demand distribution of different orders, the probability distribution of different materials, and the distribution of materials required for order production.

[0055] For example, real data (real material kitting data) is obtained from the material kitting production scenario to be simulated (target material kitting scenario), and statistical analysis is performed on this data to obtain specific statistics (standard data). These statistics include: the inventory distribution of different categories of materials (inventory distribution of different materials), the demand distribution of different categories of orders (demand distribution of different orders), and the distribution of the number of material types required for material production. Different statistics can be flexibly designed for different scenarios to better fit the scenario characteristics. If the amount of accessible real data for the scenario is limited, or if real data has not yet been generated for the scenario to be simulated, various statistical indicators can be approximated through manual estimation.

[0056] Step S503: Based on standard data, generate data generation rules for the target material completeness scenario.

[0057] In the embodiments of this application, the data generation device generates data generation rules for the target material completeness scenario based on standard data.

[0058] In some embodiments, the data generation rules include at least one or more of the following: data distribution fitting rules, data association fitting rules, and special data generation rules. The data generation device performing step S503 may include at least one or more of the following steps S601 to S603:

[0059] Step S601: Using the distribution generation function, generate data distribution fitting rules based on different distribution data in the standard data.

[0060] In embodiments of this application, the data generation device can utilize a distributional generation function to generate data distribution fitting rules based on different distributions in standard data. For example, the data distribution fitting (data distribution fitting rule) design and generation process uses a distributional generation function, such as the probability distribution of material types or the distribution of material inventory, to align the generated simulated distribution with the actual distribution.

[0061] Step S602: Generate data association fitting rules based on the correlation between data with different distributions in the standard data.

[0062] In embodiments of this application, the data generation device generates data association fitting rules based on the correlations between different distributional data in standard data. For example, data association fitting (data association fitting rules) studies the correlations between different distributional indicators, such as the correlation between material types and the number of material categories required for production, to model the correlation information of real-world scenarios.

[0063] Step S603: Based on the distribution of abnormal data in the standard data, set special data generation rules.

[0064] In embodiments of this application, the data generation device sets special data generation rules based on the distribution of abnormal data in standard data. For example, the special data generation (special data generation rules) studies marginal distributions in scenarios, such as orders with abnormally high or low demand, to include special cases in scenarios where target materials are readily available.

[0065] In the embodiments of this application, the rules included in the above data generation rules can be flexibly adjusted according to scene characteristics and simulation requirements, and this application does not limit them.

[0066] In some embodiments, the data generation device may include the following steps when performing step S102: determining the initial material characteristic information of each material according to the data distribution fitting rule; adjusting the initial material characteristic information of each material according to the data association fitting rule to obtain the adjusted material characteristic information of each material in each level of material; and generating special data for the adjusted material characteristic information of each material according to the special data generation rule to obtain the material characteristic information corresponding to each material.

[0067] In the embodiments of this application, the data generation device first determines the initial material characteristic information of each material according to the data distribution fitting rule, then adjusts the generated initial material characteristic information according to the data association fitting rule to obtain the adjusted material characteristic information, and finally generates special data according to the special data generation rule to obtain the corresponding material characteristic information.

[0068] In some embodiments, the data generation device may include the following steps when performing step S301: determining the initial order feature information of each order according to the data distribution fitting rule; adjusting the initial order feature information of each order according to the data association fitting rule to obtain the adjusted order feature information of each order; and generating special data on the adjusted order feature information of each order according to the special data generation rule to obtain the order feature information of each order.

[0069] In the embodiments of this application, the data generation device first determines the initial order feature information of each order according to the data distribution fitting rules, then adjusts the generated initial order feature information according to the data association fitting rules to obtain the adjusted order feature information, and finally generates special data according to the special data generation rules to obtain the corresponding order feature information.

[0070] In the embodiments of this application, to ensure that the generated simulation examples can be used to optimize the solution effects of heuristic and machine learning methods and to guarantee the quality of data generation, this application proposes a three-stage simulation example construction method: First, generate simulation material information (see steps S101 to S103 above); then, generate simulation order information (see steps S301 to 302 above); finally, match the two (see step S303 above). Thus, this application uses three aspects of scenario alignment methods: data distribution fitting, data association fitting, and special data generation. These methods, while ensuring the confidentiality of real data, enable the generated data to possess the characteristics of specific real production scenarios.

[0071] In some embodiments, after performing step S303, the data generating device may further perform steps S701 to S703:

[0072] Step S701: For the scenario data, establish a mixed integer programming problem model, and call the solver to solve the mixed integer programming problem model to obtain the optimal solution.

[0073] In the embodiments of this application, the data generation device establishes a mixed integer programming (MIP) problem model for the scene data, and calls the solver (Solving Constraint Integer Programs, SCIP) to solve the mixed integer programming problem model to obtain the optimal solution.

[0074] Step S702: Select the scenario data whose solution time for the mixed integer programming problem model is within the target solution time range as the target scenario data.

[0075] In embodiments of this application, the data generation device filters scenario data whose solution time for the mixed-integer programming problem model falls within the target solution time range as target scenario data. To address the degradation problem in the generated simulation examples (scenario data), data filtering is necessary. For example, see [link to relevant documentation]. Figure 8 As shown, the data filtering method is as follows: the simulation example (scene data) 81 is modeled into a mixed integer programming problem model (MIP example) 83 using the MIP modeling module 82; the mathematical programming solver (MIP solver) 84 is used to solve the problem of the MIP example, and the solution time is recorded 85; according to the upper and lower bounds of the preset expected solution time (target solution time range), data with a solution time exceeding the limit is removed (data filtering) 86, and the required simulation data (target scene data) 87 is obtained.

[0076] Step S703: Using the target scenario data and the optimal solution, train the model to be trained for the target material completeness scenario to obtain the target solution model, and use the target solution model to solve the real data under the target material completeness scenario.

[0077] In the embodiments of this application, after obtaining the target scene data and the optimal solution for the target scene data, the data generation device uses the target scene data and the optimal solution to train the model to be trained for the target material completeness scenario, thereby obtaining the target solution model, and then using the target solution model to solve the real data under the target material completeness scenario.

[0078] For example, such as Figure 9 As shown, this application also provides a specific flowchart for generating target scene data. Figure 9 As shown, generating target scene data includes the following steps S901 to S904:

[0079] Step S901: Setting rules to align with real data features.

[0080] like Figure 9 As shown, the data generation device extracts features representing the target material completeness scenario from the acquired real data (real material completeness data) and sets corresponding rules (data generation rules) 90; among which, the data generation rules include: data distribution fitting 91, data association fitting 92, and feature data generation rules 93.

[0081] Step S902: Generation of material matching simulation sample.

[0082] like Figure 9As shown, the data generation device generates a material kitting simulation example (scenario data) 94 based on the data generation rule 90 set in step S901 above. For example, material information 96 is generated using the material information generation module 95 and order information 98 is generated using the order information generation module 97. Then, the material-order matching module 99 is used to match the production relationship between materials and orders to generate the material kitting simulation example 94.

[0083] Step S903: Material completeness data screening.

[0084] like Figure 9 As shown, the data generation device uses the MIP modeling module 910 to establish a mixed integer programming problem model for the generated material kit simulation example 94, and calls the solver to solve it. Then, the data filtering module 911 removes data whose solution time exceeds the upper and lower bounds of the preset expected solution time, and filters out data whose solution time is within the upper and lower bounds of the preset expected solution time.

[0085] Step S904: Obtain target scene data.

[0086] like Figure 9 As shown, the target scenario data obtained after screening is confidential and can be desensitized from real data; it is simulated and can reflect the characteristics of material matching problems; it is of high quality and is the data obtained after screening out degraded data.

[0087] For example, such as Figure 10 As shown, the process of determining the target solution model can be as follows: embedding the material kitting problem simulation sample (target scenario data) 11 into a graph convolutional neural network 12, identifying the neural network prediction result 13, performing MIP modeling on the material kitting problem simulation sample 11, and calling the MIP solver 14 to solve for the optimal solution label (optimal solution) 15, then calculating the loss information 16 between the neural network prediction result 13 and the optimal solution label 15, and then training the model to be trained based on the loss information 16 to obtain the target solution model, so as to use the target solution model to solve the real data under the target material kitting scenario. Thus, the simulation sample generation method (data generation method) provided in this application fills the shortcomings of existing data generation methods and can construct high-quality material kitting problem simulation data. The model enhancement method proposed in this patent can use the generated simulation data to optimize model parameters, improve problem-solving efficiency, and is of great significance for decision-makers to optimize problem-solving algorithms and improve production planning, thereby improving production resource utilization and satisfaction.

[0088] In the embodiments of this application, the simulation data generated by the data generation device can be used to train machine learning models and perform model enhancement. The model enhancement process (determining the target and solving the model) is as follows: Figure 10 As shown, firstly, mixed-integer programming is performed on the simulation data (target scenario data), and the training labels (optimal solutions) are obtained using the mathematical programming solver SCIP. Then, the simulation data is input into the machine learning model (the model to be trained), and the model's output (neural network prediction results) is obtained. The loss is calculated using the training labels and model output, the model parameters are updated, and this process is repeated until model training is complete. The resulting machine learning model (target solution model) can solve material kitting problems, improving solution efficiency.

[0089] This application provides a data generation method. Using data generation rules, at least one product corresponding to a target material completeness scenario is determined. For each of the at least one product, production materials are progressively expanded until the target level is reached, resulting in corresponding material levels. The data generation rules are generated based on real material completeness data corresponding to the target material completeness scenario. Using the data generation rules, feature information is expanded for each material at each level, resulting in material feature information for each material. The material feature information includes at least one or more of the following: inventory level, usage priority, and material type. Each level of material, and the corresponding material feature information for each material at each level, are determined as the material information corresponding to the target material completeness scenario. The data generation method provided in this application can generate data generation rules representing scenario characteristics based on real material completeness data of the target material completeness scenario, and then generate simulation data that conforms to the target material completeness scenario based on the data generation rules, reducing the difficulty of data acquisition and improving the security of real material completeness data.

[0090] This application provides a data generation apparatus, such as... Figure 11 As shown, it includes:

[0091] The determination module 1101 is used to determine at least one product corresponding to the target material completeness scenario using data generation rules, and to expand the production materials of each product in the at least one product step by step until the target level is reached to obtain the corresponding materials at each level; the data generation rules are generated based on the real material completeness data corresponding to the target material completeness scenario.

[0092] The expansion module 1102 is used to expand the feature information of each material in each level of materials using data generation rules to obtain the material feature information corresponding to each material; the material feature information includes at least one or more of the following: inventory quantity, usage priority, and material type.

[0093] The determination module 1101 is also used to determine the material information corresponding to the target material matching scenario as the material information of each level of material and the material characteristic information of each material in each level of material.

[0094] In one embodiment of this application, the determining module 1101 is further configured to generate at least one order for the target material completeness scenario using data generation rules, and generate corresponding order feature information for each order in the at least one order; the order can be: a formal order, a backup order, a replenishment order, or a forecast order; the order feature information includes at least the materials required for the order; the at least one order is classified according to the required materials to obtain at least one group of orders; based on the required materials of each group of orders in the at least one group of orders, the production relationship of each order in each group of orders is matched with the corresponding materials included in the material information to generate scenario data corresponding to the target material completeness scenario.

[0095] In one embodiment of this application, the data generation device further includes a generation module (not shown in the figure). The generation module is used to acquire real material completeness data; analyze standard data in the real material completeness data that characterizes the features of the target material completeness scenario; the standard data includes at least one or more of the following: the inventory distribution of different materials, the demand distribution of different orders, the probability distribution of different materials, and the distribution of materials required for order production; and generate data generation rules for the target material completeness scenario based on the standard data.

[0096] In one embodiment of this application, the data generation rules include at least one or more of the following: data distribution fitting rules, data association fitting rules, and special data generation rules; the generation module is further configured to use a distribution generation function to generate data distribution fitting rules based on different distribution data in the standard data; generate data association fitting rules based on the association relationship between different distribution data in the standard data; and set special data generation rules based on the distribution of abnormal data in the standard data.

[0097] In one embodiment of this application, the determining module 1101 is further configured to determine the initial material characteristic information of each material according to the data distribution fitting rule; adjust the initial material characteristic information of each material according to the data association fitting rule to obtain the adjusted material characteristic information of each material in each level of material; and generate special data for the adjusted material characteristic information of each material according to the special data generation rule to obtain the material characteristic information corresponding to each material.

[0098] In one embodiment of this application, the determining module 1101 is further configured to determine the initial order feature information of each order according to the data distribution fitting rule; adjust the initial order feature information of each order according to the data association fitting rule to obtain the adjusted order feature information of each order; and generate special data for the adjusted order feature information of each order according to the special data generation rule to obtain the order feature information of each order.

[0099] In one embodiment of this application, the data generation device further includes a training module (not shown in the figure). The training module is used to establish a mixed integer programming problem model for the scenario data, and call a solver to solve the mixed integer programming problem model to obtain the optimal solution; to filter the scenario data whose solution time for the mixed integer programming problem model is within the target solution time range as target scenario data; and to use the target scenario data and the optimal solution to train the solution model to be trained for the target material completeness scenario to obtain the target solution model, so as to use the target solution model to solve the real data under the target material completeness scenario.

[0100] This application provides a data generation device, such as... Figure 12 As shown, the data generation device includes: a processor 1201, a memory 1202, and a communication bus 1203;

[0101] The communication bus 1203 is used to realize the communication connection between the processor 1201 and the memory 1202;

[0102] The processor 1201 is used to execute the computer program stored in the memory 1202 to implement the above-described data generation method.

[0103] This application provides a data generation device that uses data generation rules to determine at least one product corresponding to a target material completeness scenario. For each of the at least one product, production materials are progressively expanded until the target level is reached, resulting in corresponding material levels. The data generation rules are generated based on real material completeness data corresponding to the target material completeness scenario. Using the data generation rules, feature information is expanded for each material at each level, resulting in material feature information for each material. The material feature information includes at least one or more of the following: inventory level, usage priority, and material type. Each level of material, and the corresponding material feature information for each material at each level, are determined as the material information corresponding to the target material completeness scenario. The data generation device provided in this application can generate data generation rules representing scenario characteristics based on real material completeness data of the target material completeness scenario, and then generate simulation data that conforms to the target material completeness scenario based on the data generation rules, reducing the difficulty of data acquisition and improving the security of real material completeness data.

[0104] This application provides a computer-readable storage medium storing one or more computer programs, which can be executed by one or more processors to implement the data generation method described above. The computer-readable storage medium can be volatile memory, such as random-access memory (RAM); or non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD); or it can be a device including one or any combination of the above-mentioned memories, such as a mobile phone, computer, tablet device, personal digital assistant, etc.

[0105] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0106] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0107] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0108] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0109] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A data generation method, the method comprising: Using data generation rules, at least one product corresponding to the target material completeness scenario is identified, and production materials are expanded level by level for each product in the at least one product until the target level is reached to obtain the corresponding materials at each level; the data generation rules are generated based on the real material completeness data corresponding to the target material completeness scenario. Using the data generation rules, feature information is expanded for each material in each level of materials to obtain material feature information corresponding to each material; the material feature information includes at least one or more of the following: inventory quantity, usage priority, and material type; The materials at each level, and the material feature information corresponding to each material at each level, are determined as the material information corresponding to the target material completeness scenario; Based on the data generation rules, order information corresponding to the target material completeness scenario is generated, and the material information and the order information are matched to obtain scenario data; A mixed integer programming problem model is established for the scenario data, and a solver is called to solve it. Scenario data whose solution time is within the target solution time range are selected as target scenario data. The target scenario data is used to train the solution model for the target material set scenario and determine the target solution model.

2. The method according to claim 1, wherein generating order information corresponding to the target material completeness scenario according to the data generation rules, and matching the material information and the order information to obtain scenario data, includes: Using the data generation rules, at least one order is generated for the target material completeness scenario, and corresponding order feature information is generated for each order in the at least one order; The orders are: formal orders, stock preparation orders, replenishment orders, or forecast orders; the order characteristic information includes at least the materials required for the order; The at least one order is classified according to the required materials to obtain at least one set of orders; Based on the required materials of each order in the at least one set of orders, the production relationship of each order in each set of orders is matched with the corresponding materials included in the material information to generate scene data corresponding to the target material completeness scenario.

3. The method according to claim 1 or 2, further comprising: Obtain the actual material kitting data; Analyze the standard data characterizing the target material completeness scenario from the real material completeness data; The standard data includes at least one or more of the following: the inventory distribution of different materials, the demand distribution of different orders, the probability distribution of different materials, and the distribution of materials required for order production. Based on the standard data, the data generation rules are generated for the scenario where the target material is fully equipped.

4. The method according to claim 3, wherein the data generation rules include at least one or more of data distribution fitting rules, data association fitting rules, and special data generation rules; the generation of data generation rules based on the standard data for the target material completeness scenario includes at least one or more of the following: Using a distribution generation function, the data distribution fitting rule is generated based on different distribution data in the standard data; Based on the correlation between different distributions of data in the standard data, the data correlation fitting rules are generated; Based on the distribution of abnormal data in the standard data, the special data generation rules are set.

5. The method according to claim 4, wherein the step of using the data generation rule to expand the feature information of each material in each level of materials to obtain the material feature information corresponding to each material includes one or more of the following: Based on the data distribution fitting rules, the initial material characteristic information of each material is determined; According to the data association fitting rules, the initial material feature information of each material is adjusted to obtain the adjusted material feature information of each material in each level of material. According to the special data generation rules, special data is generated for the adjusted material characteristic information of each material to obtain the material characteristic information corresponding to each material.

6. The method according to claim 2, wherein generating corresponding order feature information for each order in the at least one order using the data generation rule includes one or more of the following: Based on the data distribution fitting rules, the initial order feature information of each order is determined; According to the data association fitting rules, the initial order feature information of each order is adjusted to obtain the adjusted order feature information of each order; According to special data generation rules, special data is generated for the adjusted order feature information of each order to obtain the order feature information of each order.

7. The method according to claim 1, wherein establishing a mixed-integer programming problem model for the scene data, calling a solver to solve it, and selecting scene data whose solution time is within the target solution time range as target scene data includes: For the aforementioned scenario data, a mixed integer programming problem model is established, and the solver is invoked to solve the mixed integer programming problem model to obtain the optimal solution; The scenario data whose solution time for the mixed integer programming problem model is within the target solution time range are selected as the target scenario data; Accordingly, after selecting the scenario data whose solution time falls within the target solution time range as the target scenario data, the method further includes: Using the target scenario data and the optimal solution, the model to be trained for the target material completeness scenario is trained to obtain the target solution model, which is then used to solve the real data under the target material completeness scenario.

8. A data generation apparatus, comprising: The determination module is used to use data generation rules to determine at least one product corresponding to the target material completeness scenario, and to expand the production materials for each product in the at least one product step by step until the target level is reached to obtain the corresponding materials at each level; the data generation rules are generated based on the real material completeness data corresponding to the target material completeness scenario. An expansion module is used to expand the feature information of each material in each level of the material using the data generation rules, so as to obtain the material feature information corresponding to each material; the material feature information includes at least one or more of the following: inventory quantity, usage priority, and material type; The determining module is further configured to determine the materials at each level, and the material feature information corresponding to each material at each level, as the material information corresponding to the target material matching scenario; The determining module is further configured to generate order information corresponding to the target material completeness scenario according to the data generation rules, and match the material information and the order information to obtain scenario data; The training module is used to establish a mixed integer programming problem model for the scene data, call the solver to solve it, and select the scene data whose solution time is within the target solution time range as the target scene data; The target scenario data is used to train the solution model for the target material set scenario and determine the target solution model.

9. A data generation device, comprising: Processor, memory, and communication bus; The communication bus is used to realize the communication connection between the processor and the memory; The processor is configured to execute a computer program stored in the memory to implement the data generation method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more computer programs, which can be executed by one or more processors to implement the data generation method as described in any one of claims 1 to 7.

Images