Dynamic inventory scheduling method and system based on ERP (Enterprise Resource Planning) system

Abstract

The invention provides a dynamic inventory scheduling method and system based on an ERP system, and relates to the technical field of inventory scheduling. The inventory and order information is obtained from the ERP system in real time, it is ensured that the data basis of scheduling decision is newest and accurate, scheduling errors caused by information lag are effectively reduced, meanwhile, the cargo correlation degree is comprehensively evaluated by applying a multi-dimensional correlation analysis and weighted decision model, and the scheduling efficiency is improved. Optimization of inventory layout and reduction of transportation time and cost are realized, intelligence of inventory scheduling is realized through a graph neural network and an adaptive clustering algorithm, high-correlation cargo combinations can be automatically identified, differentiated scheduling schemes can be generated, scheduling efficiency and accuracy are improved, and the method has flexibility and expandability and is suitable for popularization and application. According to the technical scheme, adjustment and optimization can be carried out according to actual conditions to adapt to different service requirements and scene changes, and then inventory resource configuration is effectively optimized through implementation of collaborative scheduling and differential scheduling schemes.

Description

Technical Field

[0001] The present invention relates to the technical field of inventory scheduling, and in particular to a dynamic inventory scheduling method and system based on an ERP system. Background Art

[0002] In the operation and management of modern enterprises, inventory scheduling is a key link between production and sales. Its efficiency and accuracy are directly related to the enterprise's operating costs, customer satisfaction and market competitiveness.

[0003] Traditional inventory scheduling methods often rely on manual recording and updating of inventory information, which is not only inefficient but also prone to errors. Due to the untimely update of information, schedulers are often unable to obtain the latest inventory and order information, resulting in outdated and inaccurate data basis for scheduling decisions. This problem of information lag and inaccurate data seriously affects the timeliness and effectiveness of scheduling decisions, which in turn leads to frequent inventory backlogs, stockouts and other problems.

[0004] Traditional inventory scheduling methods usually adopt fixed scheduling strategies, such as simple classification and scheduling according to cargo type, storage location, etc. However, this single scheduling strategy often fails to fully consider complex factors such as the correlation between cargoes, transportation needs, and customer needs, resulting in low scheduling efficiency and difficulty in meeting diverse business needs. In addition, traditional methods lack flexibility and cannot be adjusted and optimized in real time according to market changes and customer needs, further limiting the improvement of scheduling effects.

[0005] Traditional inventory scheduling methods lack intelligent decision support, and the scheduling process mainly relies on manual experience and judgment. This manual scheduling method is not only inefficient, but also easily affected by human factors, resulting in instability and unpredictability of scheduling results. In addition, traditional methods lack automated technical support, such as automated data collection, processing and analysis, making the scheduling process cumbersome and error-prone.

[0006] Since traditional inventory scheduling methods cannot fully evaluate the correlation between goods, inventory layout is often unreasonable. The correlation between goods is ignored. Secondly, due to the lack of flexibility in scheduling strategies, companies often find it difficult to make real-time inventory adjustments and optimizations based on market and customer needs, further exacerbating the problems of unreasonable inventory layout and high transportation costs.

[0007] Therefore, it is necessary to provide a dynamic inventory scheduling method and system based on an ERP system to solve the above technical problems. Summary of the invention

[0008] In order to solve the above technical problems, the present invention provides a dynamic inventory scheduling method and system based on an ERP system to solve the problems of delayed inventory scheduling information and inaccurate data, single scheduling strategy and lack of flexibility, low intelligence and lack of automation support.

[0009] The present invention provides a dynamic inventory scheduling method based on an ERP system, comprising the following steps: The inventory scheduling method comprises the following steps: Obtain the current inventory static data and dynamic order information from the ERP system in real time to build an information matrix of goods to be dispatched; Based on the information matrix of goods to be dispatched, multi-dimensional correlation analysis is used to output the correlation coefficient matrix through the weighted decision model; Use graph neural networks to calculate node centrality and quantify the coordination coefficient between goods; An adaptive clustering algorithm is used to map the coordination coefficient into different levels of scheduling labels, where the different levels of scheduling labels specifically include first-level coordination, second-level coordination, and third-level coordination; Based on different levels of scheduling labels, corresponding differentiated scheduling plans are generated.

[0010] Preferably, the steps of acquiring the static data of the current inventory and the dynamic order information from the ERP system in real time and constructing the information matrix of the goods to be dispatched are as follows: Obtain the current inventory static data in real time through the API interface or database connection of the ERP system, where the static data includes the inventory quantity, storage location or type of goods; Obtain dynamic order information from the ERP system, including order number, customer name, and order time, and analyze order demand from the dynamic order information; The acquired static data and dynamic order information are integrated to construct an information matrix of goods to be dispatched, where the goods information matrix has goods types as rows and inventory quantities, storage locations, and order requirements as columns.

[0011] Preferably, the method is based on the information matrix of the goods to be dispatched, uses multi-dimensional correlation analysis, and outputs a correlation coefficient matrix through a weighted decision model, and the specific steps are: Conduct multi-dimensional correlation analysis on the information matrix of the goods to be dispatched, and use a weighted decision model to calculate the correlation coefficient between each pair of goods, where the multiple dimensions include physical dimension, transportation dimension, and timeliness dimension; The calculated correlation coefficients between each pair of goods are organized into a correlation coefficient matrix.

[0012] Preferably, the graph neural network is used to calculate the node centrality and quantify the collaborative scheduling coefficient between goods, and the specific steps are: The information matrix of goods to be dispatched and the correlation coefficient matrix are converted into a graph structure, where the goods are used as nodes and the correlation coefficients are used as edge weights between nodes; Use graph neural network to process the graph structure and calculate the centrality of each node; Based on the calculated centrality of each node, the centrality of the node is converted into a collaborative scheduling coefficient between goods.

[0013] Preferably, the adaptive clustering algorithm is used to map the coordination coefficient to scheduling labels of different levels, wherein the scheduling labels of different levels specifically include primary coordination, secondary coordination and tertiary coordination, and the specific steps are: Use the K-means algorithm or DBSCAN algorithm in the adaptive clustering algorithm to cluster the goods according to the collaborative scheduling coefficient and obtain the clustering results; According to a plurality of preset clustering threshold segments, the clustering results are mapped to scheduling labels of different levels, wherein each cluster corresponds to a scheduling label.

[0014] Preferably, the specific steps of generating corresponding differentiated scheduling schemes based on scheduling labels of different levels are as follows: For first-level collaboration, a joint picking task list is generated to group together goods with a high correlation exceeding a preset threshold for picking; For secondary collaboration, adjust the lighting frequency and path indication of the electronic tag picking system, optimize according to the relevance and picking sequence of the goods, and reduce the moving distance and time of the pickers; For level 3 collaboration, activate the redundant packaging line standby stations and adjust the packaging line workload according to the picking progress and packaging requirements of the goods.

[0015] A dynamic inventory scheduling system based on an ERP system, the scheduling system comprising: The information acquisition module is used to obtain the static data of the current inventory and the dynamic order information from the ERP system in real time and build the information matrix of the goods to be dispatched; The correlation analysis module is used to apply multi-dimensional correlation analysis based on the information matrix of the goods to be dispatched, and output the correlation coefficient matrix through the weighted decision model; The collaborative quantification module is used to calculate the node centrality using graph neural networks and quantify the collaborative scheduling coefficient between goods; A label partitioning module is used to map the coordination coefficient into different levels of scheduling labels using an adaptive clustering algorithm, where the different levels of scheduling labels specifically include first-level coordination, second-level coordination, and third-level coordination; The scheduling generation module is used to generate corresponding differentiated scheduling plans based on scheduling labels of different levels.

[0016] Compared with the related art, the dynamic inventory scheduling method and system based on the ERP system provided by the present invention has the following beneficial effects: The present invention obtains inventory and order information from the ERP system in real time, ensuring that the data basis for scheduling decisions is up-to-date and accurate, effectively reducing scheduling errors caused by information lags. At the same time, it uses multi-dimensional correlation analysis and weighted decision models to comprehensively evaluate the correlation between goods, optimize inventory layout and reduce transportation time and costs. Through graph neural networks and adaptive clustering algorithms, intelligent inventory scheduling is achieved, which can automatically identify high-correlation cargo combinations and generate differentiated scheduling plans, thereby improving scheduling efficiency and accuracy. The present invention has flexibility and scalability, and can be adjusted and optimized according to actual conditions to adapt to different business needs and scenario changes. Secondly, through the implementation of collaborative scheduling and differentiated scheduling plans, inventory resource allocation is effectively optimized, operating costs are reduced, and customer order requirements are met more quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flow chart of a dynamic inventory scheduling method based on an ERP system of the present invention; Figure 2 The system block diagram of a dynamic inventory scheduling system based on an ERP system of the present invention. DETAILED DESCRIPTION

[0018] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0019] Embodiment 1 like Figure 1 As shown, a dynamic inventory scheduling method based on an ERP system includes the following steps: S1. Obtain the current inventory static data and dynamic order information from the ERP system in real time to build an information matrix of goods to be dispatched; S2. Based on the information matrix of the goods to be dispatched, multi-dimensional correlation analysis is used to output the correlation coefficient matrix through the weighted decision model; S3. Use graph neural network to calculate node centrality and quantify the coordination coefficient between goods; S4. Adopting an adaptive clustering algorithm to map the coordination coefficient into different levels of scheduling labels, wherein the different levels of scheduling labels specifically include first-level coordination, second-level coordination, and third-level coordination; S5. Generate corresponding differentiated scheduling plans based on scheduling labels of different levels.

[0020] In the specific implementation process, the specific steps of step S1 are: S101. Obtaining static data of current inventory in real time through an API interface or database connection of an ERP system, wherein the static data includes inventory quantity, storage location or type of goods.

[0021] Specifically, through the API interface or database connection of the ERP system, the API interface method: the system calls the API interface provided by the ERP system, sends requests according to the interface specifications, and receives the inventory static data returned by the ERP system; the database connection method: the system directly connects to the database of the ERP system and obtains the inventory static data through SQL query statements.

[0022] S102. Obtain dynamic order information from the ERP system, including order number, customer name, and order time, and analyze order demand from the dynamic order information.

[0023] Specifically, the latest dynamic order information is extracted from the ERP system, including order number, customer name, order time, etc. The order information is parsed through text recognition to analyze and identify the specific requirements of each order, that is, what goods the customer needs and how many goods, etc. The analyzed order requirements are associated with the static inventory data to provide a basis for subsequent goods scheduling.

[0024] S103, integrating the acquired static data and dynamic order information to construct an information matrix of goods to be dispatched, wherein the goods information matrix has goods types as rows and inventory quantities, storage locations, and order requirements as columns.

[0025] Specifically, the static data obtained in step S101 and the order demand analyzed in step S102 are integrated to construct an information matrix of goods to be dispatched. Each element in the matrix represents specific information of a certain type of goods in a certain aspect, providing a data basis for subsequent multi-dimensional correlation analysis and dispatching decisions.

[0026] Exemplarily, an information matrix of goods to be scheduled is constructed, in which the first row represents goods A, and the columns represent the inventory (100 pieces), storage location (area A of warehouse 1), and order demand (30 pieces); the second row represents goods B, and the columns represent the inventory (50 pieces), storage location (area B of warehouse 2), and order demand (20 pieces); the third row represents goods C, and the columns represent the inventory (200 pieces), storage location (area C of warehouse 1), and order demand (100 pieces).

[0027] In the specific implementation process, the specific steps of step S2 are: S201. Perform multi-dimensional correlation analysis on the information matrix of the goods to be dispatched, and use a weighted decision model to calculate the correlation coefficient between each pair of goods, where the multiple dimensions include physical dimension, transportation dimension, and timeliness dimension.

[0028] Specifically, multi-dimensional correlation analysis technology is used to evaluate the correlation between goods from multiple angles such as physical dimension, transportation dimension and time dimension; physical dimension: consider the physical properties of goods such as size, weight, shape, and the spatial relationship between storage locations, and evaluate the mutual influence of goods during the picking and handling process; transportation dimension: analyze factors such as the transportation route, transportation tools, and transportation time of the goods, and evaluate the coordination and efficiency of the goods during the transportation process; time dimension: consider time factors such as the order demand time and inventory shelf life of the goods, and evaluate the coordination of goods in meeting customer needs and maintaining inventory turnover.

[0029] In this embodiment, the steps of applying the weighted decision model are: 1. Score each dimension: Specific physical dimension: for example, whether both batches of goods need to be refrigerated (yes = 1 point, no = 0 points), or whether the size can be consolidated (yes = 1 point, no = 0 points); transportation dimension: calculate the distance between the delivery addresses of the two orders, for example, 5 kilometers gets 0.8 points, 10 kilometers gets 0.5 points (the closer the distance, the higher the score); timeliness dimension: if both orders require delivery within 2 hours, the urgency is high, and each gets 0.9 points; if one is urgent and the other is not, the score difference is large.

[0030] 2. Preset weights for each dimension, where physical compatibility accounts for 30% (weight 0.3), transportation distance accounts for 50% (weight 0.5), and time urgency accounts for 20% (weight 0.2).

[0031] 3. For example, the physical score of goods A and B is 0.8, the transportation score is 0.6, and the timeliness score is 0.7. The correlation coefficient between them calculated by the weighted decision model is = 0.8×0.3+0.6×0.5+0.7×0.2=0.24+0.3+0.14=0.68.

[0032] S202. Arrange the calculated correlation coefficients between each pair of goods into a correlation coefficient matrix.

[0033] In the specific implementation process, the specific steps of step S3 are: S301, converting the to-be-scheduled cargo information matrix and the correlation coefficient matrix into a graph structure, wherein the cargo is used as a node and the correlation coefficient is used as an edge weight between nodes.

[0034] Specifically, the information matrix of goods to be dispatched and the correlation coefficient matrix are read, wherein the information matrix of goods to be dispatched contains key information such as the type, inventory, storage location, and order requirements of the goods, while the correlation coefficient matrix reflects the strength of correlation between the goods.

[0035] In this embodiment, the information matrix of goods to be dispatched contains three kinds of goods A, B, and C, whose inventory, storage location, and order requirements are different. At the same time, the correlation coefficient matrix shows that the correlation coefficient between goods A and B is 0.8, the correlation coefficient between goods A and C is 0.6, and the correlation coefficient between goods B and C is 0.7; after being converted into a graph structure, goods A, B, and C become three nodes in the graph, and the edge weights between them are set to 0.8, 0.6, and 0.7, respectively.

[0036] S302. Use graph neural network to process the graph structure and calculate the centrality of each node.

[0037] Specifically, after the graph structure is constructed, it is processed using a graph neural network, which is a neural network model specifically used to process graph structure data. It can capture the complex relationships between nodes and calculate the feature representations of the nodes. Finally, the centrality of each node is calculated. The centrality is an indicator to measure the importance of a node in the graph structure. It reflects the connection strength and influence between a node and other nodes.

[0038] S303. Based on the calculated centrality of each node, the centrality of the node is converted into a coordination scheduling coefficient between goods.

[0039] Specifically, after obtaining the centrality of each node, it is converted into a collaborative scheduling coefficient between goods. The collaborative scheduling coefficient is calculated by normalizing or weighting the node centrality, and finally a collaborative scheduling coefficient matrix is ​​obtained.

[0040] For example, assuming that the normalized collaborative scheduling coefficient ranges from 0 to 1, the collaborative scheduling coefficient between goods A and B is 0.85 (calculated based on the association coefficient and centrality between them), the collaborative scheduling coefficient between goods A and C is 0.7 (also calculated based on the association coefficient and centrality), and the collaborative scheduling coefficient between goods B and C may be 0.75.

[0041] In the specific implementation process, the specific steps of step S4 are: S401. Use the K-means algorithm or DBSCAN algorithm in the adaptive clustering algorithm to cluster the goods according to the collaborative scheduling coefficient to obtain a clustering result.

[0042] Specifically, in this embodiment, the K-means algorithm in the adaptive clustering algorithm is adopted. First, a cluster number K is preset, and then K initial cluster centers are randomly selected. Next, the distance from each cargo to each cluster center is calculated, and the cargo is assigned to the nearest cluster center. After that, the center position of each cluster is recalculated, and this process is iterated until the cluster center no longer changes significantly or the preset number of iterations is reached.

[0043] For example, assuming that there are 10 different kinds of goods, and their coordinated scheduling coefficients have been calculated through the previous steps, the K-means algorithm is used, and the number of clusters K=3 is preset. After iterative calculation, these 10 kinds of goods are divided into three groups. The goods in each group are relatively close in coordinated scheduling coefficient, while the goods between different groups are quite different.

[0044] S402 . Mapping clustering results to scheduling labels of different levels according to a plurality of pre-set clustering threshold segments, wherein each cluster corresponds to a scheduling label.

[0045] Specifically, a series of clustering threshold segments are set, wherein the series of clustering threshold segments should be able to cover all possible clustering result ranges, and then each cluster is mapped to a specific dispatch label according to the range of its collaborative dispatch coefficient, which is specifically achieved by comparing the average collaborative dispatch coefficient of the cluster center or the goods in the cluster with the threshold segment. In this embodiment, three threshold segments are set, specifically: high (collaborative dispatch coefficient greater than 0.8), medium (between 0.5 and 0.8) and low (less than 0.5).

[0046] For example, after clustering in step S401, among the three clusters obtained, the average coordination coefficient of the first cluster is 0.9, the second is 0.6, and the third is 0.4. Therefore, the first cluster can be mapped to a first-level coordination label, the second to a second-level coordination label, and the third to a third-level coordination label.

[0047] In the specific implementation process, the specific steps of step S5 are: For the first-level collaboration, a joint picking task list is generated, and the goods with a high correlation exceeding the preset threshold are grouped together for picking.

[0048] For example, if there are three types of goods A, B, and C in the warehouse, and their correlation coefficient is extremely high, exceeding the preset threshold, the system will combine these three types of goods into one picking batch and generate a joint picking task list. The pickers can complete the picking of these three types of goods at one time, greatly improving the picking efficiency.

[0049] For secondary collaboration, the lighting frequency and path indication of the electronic tag picking system are adjusted, and optimization is performed based on the relevance of the goods and the picking order to reduce the moving distance and time of the pickers.

[0050] Specifically, the lighting frequency of the electronic tag picking system is dynamically adjusted according to the relevance of the goods and the picking order. Goods with high relevance will be picked first, so the corresponding electronic tags will light up in advance; then, according to the storage location and picking order of the goods, the optimal path instructions are provided to the pickers.

[0051] For example, there are two types of goods D and E in the warehouse, which have a high correlation but do not meet the first-level coordination standard. During the picking process, the lighting frequency and path indication of the electronic tag picking system are dynamically adjusted according to the correlation and storage location of the two types of goods. The picker can first pick the highly correlated goods D, and then pick the goods E to reduce the moving distance and time.

[0052] For level 3 collaboration, redundant packaging line standby stations are activated, and the workload of the packaging line is adjusted according to the picking progress and packaging requirements of the goods. By optimizing the picking and packaging process, the operational efficiency and accuracy of the warehouse can be improved.

[0053] Specifically, when the picking progress and packaging requirements of the goods exceed the processing capacity of the current packaging line, the system will automatically activate the redundant packaging line standby station. For example, there are two kinds of goods F and G in the warehouse, and their picking progress is fast and the packaging requirements are large. When the current packaging line cannot meet these requirements, the redundant packaging line standby station is automatically activated and the workload of the packaging line is adjusted, so that the two kinds of goods F and G can be packaged in a timely and accurate manner and sent to the next logistics link.

[0054] Embodiment 2 like Figure 2 As shown, a dynamic inventory scheduling system based on an ERP system is applied to a dynamic inventory scheduling method based on an ERP system, comprising: The information acquisition module is used to obtain the static data of the current inventory and the dynamic order information from the ERP system in real time and build the information matrix of the goods to be dispatched; The correlation analysis module is used to apply multi-dimensional correlation analysis based on the information matrix of the goods to be dispatched, and output the correlation coefficient matrix through the weighted decision model; The collaborative quantification module is used to calculate the node centrality using graph neural networks and quantify the collaborative scheduling coefficient between goods; A label partitioning module is used to map the coordination coefficient into different levels of scheduling labels using an adaptive clustering algorithm, where the different levels of scheduling labels specifically include first-level coordination, second-level coordination, and third-level coordination; The scheduling generation module is used to generate corresponding differentiated scheduling plans based on scheduling labels of different levels.

[0055] The present invention obtains inventory and order information from the ERP system in real time, ensuring that the data basis for scheduling decisions is up-to-date and accurate, effectively reducing scheduling errors caused by information lags. At the same time, it uses multi-dimensional correlation analysis and weighted decision models to comprehensively evaluate the correlation between goods, optimize inventory layout and reduce transportation time and costs. Through graph neural networks and adaptive clustering algorithms, intelligent inventory scheduling is achieved, which can automatically identify high-correlation cargo combinations and generate differentiated scheduling plans, thereby improving scheduling efficiency and accuracy. The present invention has flexibility and scalability, and can be adjusted and optimized according to actual conditions to adapt to different business needs and scenario changes. Secondly, through the implementation of collaborative scheduling and differentiated scheduling plans, inventory resource allocation is effectively optimized, operating costs are reduced, and customer order requirements are met more quickly.

[0056] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0057] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0058] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A dynamic inventory scheduling method based on ERP system, characterized in that: The inventory scheduling method comprises the following steps: Obtain the current inventory static data and dynamic order information from the ERP system in real time to build an information matrix of goods to be dispatched; Based on the information matrix of goods to be dispatched, multi-dimensional correlation analysis is used to output the correlation coefficient matrix through the weighted decision model; Use graph neural networks to calculate node centrality and quantify the coordination coefficient between goods; An adaptive clustering algorithm is used to map the coordination coefficient into different levels of scheduling labels, where the different levels of scheduling labels specifically include first-level coordination, second-level coordination, and third-level coordination; Based on different levels of scheduling labels, corresponding differentiated scheduling plans are generated.

2. The dynamic inventory scheduling method based on ERP system according to claim 1 is characterized in that: The steps of obtaining the static data of current inventory and dynamic order information from the ERP system in real time and building the information matrix of goods to be dispatched are as follows: Obtain the current inventory static data in real time through the API interface or database connection of the ERP system, where the static data includes the inventory quantity, storage location or type of goods; Obtain dynamic order information from the ERP system, including order number, customer name, and order time, and analyze order demand from the dynamic order information; The acquired static data and dynamic order information are integrated to construct an information matrix of goods to be dispatched, where the goods information matrix has goods types as rows and inventory quantities, storage locations, and order requirements as columns.

3. The dynamic inventory scheduling method based on ERP system according to claim 1 is characterized in that: Based on the information matrix of the goods to be dispatched, multi-dimensional correlation analysis is used to output the correlation coefficient matrix through the weighted decision model. The specific steps are: Conduct multi-dimensional correlation analysis on the information matrix of the goods to be dispatched, and use a weighted decision model to calculate the correlation coefficient between each pair of goods, where the multiple dimensions include physical dimension, transportation dimension, and timeliness dimension; The calculated correlation coefficients between each pair of goods are organized into a correlation coefficient matrix.

4. The dynamic inventory scheduling method based on ERP system according to claim 1 is characterized in that: The specific steps of using graph neural network to calculate node centrality and quantify the collaborative scheduling coefficient between goods are as follows: The information matrix of goods to be dispatched and the correlation coefficient matrix are converted into a graph structure, where the goods are used as nodes and the correlation coefficients are used as edge weights between nodes; Use graph neural network to process the graph structure and calculate the centrality of each node; Based on the calculated centrality of each node, the centrality of the node is converted into a collaborative scheduling coefficient between goods.

5. The dynamic inventory scheduling method based on ERP system according to claim 1 is characterized in that: The adaptive clustering algorithm is used to map the coordination coefficient into different levels of scheduling labels, wherein the different levels of scheduling labels specifically include primary coordination, secondary coordination and tertiary coordination, and the specific steps are: Use the K-means algorithm or DBSCAN algorithm in the adaptive clustering algorithm to cluster the goods according to the collaborative scheduling coefficient and obtain the clustering results; According to a plurality of pre-set clustering threshold segments, the clustering results are mapped to scheduling labels of different levels, wherein each cluster corresponds to a scheduling label.

6. The dynamic inventory scheduling method based on ERP system according to claim 1 is characterized in that: The specific steps of generating corresponding differentiated scheduling schemes based on scheduling labels of different levels are as follows: For first-level collaboration, a joint picking task list is generated to group together goods with a high correlation exceeding a preset threshold for picking; For secondary collaboration, adjust the lighting frequency and path indication of the electronic tag picking system, optimize according to the relevance and picking sequence of the goods, and reduce the moving distance and time of the pickers; For level 3 collaboration, activate the redundant packaging line standby stations and adjust the packaging line workload according to the picking progress and packaging requirements of the goods.

7. A dynamic inventory scheduling system based on an ERP system, applied to a dynamic inventory scheduling method based on an ERP system according to any one of claims 1 to 6, characterized in that: The dispatching system comprises: The information acquisition module is used to obtain the static data of the current inventory and the dynamic order information from the ERP system in real time and build the information matrix of the goods to be dispatched; The correlation analysis module is used to apply multi-dimensional correlation analysis based on the information matrix of the goods to be dispatched, and output the correlation coefficient matrix through the weighted decision model; The collaborative quantification module is used to calculate the node centrality using graph neural networks and quantify the collaborative scheduling coefficient between goods; A label partitioning module is used to map the coordination coefficient into different levels of scheduling labels using an adaptive clustering algorithm, where the different levels of scheduling labels specifically include first-level coordination, second-level coordination, and third-level coordination; The scheduling generation module is used to generate corresponding differentiated scheduling plans based on scheduling labels of different levels.

Images