Business data management method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a business data management method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining business data of a product, and performing feature extraction to obtain an extracted business feature vector; the business data comprises the order quantity, the serial number of the minimum stock unit of the product, the quantity of the minimum stock unit and the storage timestamp; clustering different products based on the service feature vectors to obtain a preliminary clustering result; using the product orders and the minimum inventory units as nodes to construct a graph neural network, and further clustering and combining the preliminary clustering results through the trained graph neural network to obtain a final clustering result; and according to the final clustering result, calculating a corresponding shortest sorting path. By adopting the method, the utilization rate of the business data can be improved, so that the cost is reduced, and the business data processing efficiency is improved.

Description

Technical Field

[0001] The present application relates to the technical field of e-commerce data management, and in particular to a business data management method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the rapid development of e-commerce, order management and inventory scheduling of e-commerce platforms have become key links in improving operational efficiency and customer experience. As technologies such as machine learning algorithms are gradually applied to e-commerce data management, intelligent order classification, merging and path optimization solutions are provided;

[0003] Although IoT technology has been widely used in inventory management and logistics tracking, there are still many challenges in data interaction and integration between different systems, especially in complex e-commerce networks with multiple warehouses and multiple nodes. The existing order data processing technology is not convenient for realizing refined real-time order analysis and picking path optimization in a dynamic environment, and cannot reflect the dynamic changes of order characteristics and SKU demand in real time, resulting in low rationality of order merging and low utilization efficiency of SKU resources. Summary of the invention

[0004] Based on this, it is necessary to provide a business data management method, apparatus, computer equipment, computer-readable storage medium and computer program product that can improve business data processing efficiency in response to the above technical problems.

[0005] In a first aspect, the present application provides a business data management method, comprising:

[0006] Acquire business data of the product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp;

[0007] Based on the business feature vector, different products are clustered to obtain preliminary clustering results;

[0008] The product orders and minimum inventory units are used as nodes to construct a graph neural network, and the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results;

[0009] According to the final clustering results, the corresponding shortest picking path is calculated.

[0010] In one embodiment, the method further comprises:

[0011] Obtain sample business data of the product and establish a prediction model based on a convolutional neural network;

[0012] The prediction model is trained based on the sample business data, and the parameters in the prediction model are continuously optimized using an optimization algorithm until the loss function of the prediction model no longer decreases, thereby obtaining a trained prediction model;

[0013] The business data generated by the product in the current period is input into the trained prediction model to obtain the future change trend of the business data of the product.

[0014] In one embodiment, clustering different products based on the business feature vector to obtain a preliminary clustering result includes:

[0015] The products corresponding to the business feature vectors are classified by the trained vector machine to obtain classification results; the training end condition of the trained vector machine is that the loss function value corresponding to the parameters in the vector machine is the minimum;

[0016] Based on the classification results obtained for each product, multiple products are clustered to obtain the cluster center;

[0017] Different products are clustered according to the cluster centers to which the products belong and the classification results of the products to obtain preliminary clustering results of the products.

[0018] In one embodiment, the trained graph neural network is used to further cluster and merge the preliminary clustering results to obtain the final clustering results, including:

[0019] Inputting the product order data corresponding to the preliminary clustering results into the trained graph neural network to obtain an order feature vector after integrating the product orders of all products under each cluster;

[0020] The cosine similarity between the order feature vectors corresponding to each two clusters in the preliminary clustering result is calculated, and two clusters whose cosine similarity is greater than a preset value are merged to obtain a final clustering result.

[0021] In one embodiment, the step of calculating the corresponding shortest picking path according to the final clustering result includes:

[0022] The number of the minimum inventory unit of each product under each cluster in the final clustering result is used as a path node, and a path formed by the path nodes is generated through a preset model;

[0023] Calculating the length of the path, and taking the reciprocal of the length as the fitness of the path;

[0024] The path is continuously iterated based on the genetic algorithm until the fitness stops increasing, and the path when the fitness stops increasing is used as the corresponding shortest picking path.

[0025] In one embodiment, the method further comprises:

[0026] In the case where the final clustering result changes, the corresponding shortest picking path is recalculated according to the changed final clustering result.

[0027] In a second aspect, the present application also provides a business data management device, including:

[0028] An acquisition module is used to acquire business data of a product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp;

[0029] A clustering module, used to cluster different products based on the business feature vector to obtain a preliminary clustering result;

[0030] The clustering module is further used to construct a graph neural network using product orders and minimum inventory units as nodes, and further cluster and merge the preliminary clustering results through the trained graph neural network to obtain a final clustering result;

[0031] The calculation module is used to calculate the corresponding shortest picking path according to the final clustering result.

[0032] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0033] Acquire business data of the product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp;

[0034] Based on the business feature vector, different products are clustered to obtain preliminary clustering results;

[0035] The product orders and minimum inventory units are used as nodes to construct a graph neural network, and the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results;

[0036] According to the final clustering results, the corresponding shortest picking path is calculated.

[0037] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0038] Acquire business data of the product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp;

[0039] Based on the business feature vector, different products are clustered to obtain preliminary clustering results;

[0040] The product orders and minimum inventory units are used as nodes to construct a graph neural network, and the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results;

[0041] According to the final clustering results, the corresponding shortest picking path is calculated.

[0042] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0043] Acquire business data of the product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp;

[0044] Based on the business feature vector, different products are clustered to obtain preliminary clustering results;

[0045] The product orders and minimum inventory units are used as nodes to construct a graph neural network, and the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results;

[0046] According to the final clustering results, the corresponding shortest picking path is calculated.

[0047] The above-mentioned business data management method, device, computer equipment, computer-readable storage medium and computer program product first obtain the business data of the product and perform feature extraction to obtain the extracted business feature vector; the business data includes the order quantity, the number of the minimum inventory unit of the product, the number of the minimum inventory unit and the storage timestamp; based on the business feature vector, different products are clustered to obtain preliminary clustering results; the product orders and minimum inventory units are used as nodes to construct a graph neural network, and the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results; according to the final clustering results, the corresponding shortest picking path is calculated. In this way, by making full use of the characteristics of the business data of the product for clustering, and calculating the shortest picking path according to the clustered results, the utilization rate of the business data is improved, the cost in the picking process is reduced, and the efficiency of order processing is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0049] FIG1 is an application environment diagram of a business data management method in one embodiment;

[0050] FIG2 is a schematic diagram of a flow chart of a business data management method in one embodiment;

[0051] FIG3 is a block diagram of a service data management device in one embodiment;

[0052] FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0054] The business data management method provided in the embodiment of the present application can be applied in the application environment shown in Figure 1. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the business data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car devices, projection devices, etc. Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The head-mounted device can be a virtual reality (VR) device, an augmented reality (AR) device, smart glasses, etc. The server 104 can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0055] In an exemplary embodiment, as shown in FIG2 , a service data management method is provided, which is described by taking the method applied to the terminal in FIG1 as an example, and includes the following steps 202 to 208. Among them:

[0056] Step 202, obtain the business data of the product and perform feature extraction to obtain the extracted business feature vector; the business data includes the order quantity, the number of the minimum inventory unit of the product, the quantity of the minimum inventory unit and the storage timestamp.

[0057] Among them, the storage timestamp is the time when the product is stored.

[0058] In some embodiments, the terminal collects the order quantity of the product, the number of the minimum inventory unit of the product, the number of the minimum inventory units and the storage timestamp, and the extracted order quantity of the product, the number of the minimum inventory unit of the product, the number of the minimum inventory units and the storage timestamp constitute a business feature vector.

[0059] Step 204: cluster different products based on the business feature vectors to obtain preliminary clustering results.

[0060] In some embodiments, the terminal classifies the service feature vectors through the trained vector machine to obtain classification results, clusters each product in the classification results to obtain a cluster center, and clusters the products according to the cluster center and the classification results to obtain preliminary clustering results of the products.

[0061] Step 206, constructing a graph neural network using product orders and minimum inventory units as nodes, and further clustering and merging the preliminary clustering results through the trained graph neural network to obtain the final clustering results.

[0062] In some embodiments, a graph neural network is constructed using product orders and minimum inventory units as nodes of the graph neural network, and the graph neural network is trained. The product order data corresponding to the preliminary clustering results are input into the trained graph neural network to obtain an order feature vector that integrates the product orders of all products under each cluster; the cosine similarity between the order feature vectors corresponding to each two clusters in the preliminary clustering results is calculated, and two clusters whose cosine similarity is greater than a preset value are merged to obtain the final clustering result.

[0063] Step 208: Calculate the corresponding shortest picking path according to the final clustering result.

[0064] In some embodiments, the number of the minimum inventory unit of each product under each cluster in the final clustering result is used as a path node, and a path formed by the path nodes is generated through a preset model; the length of the path is calculated, and the inverse of the length is used as the fitness of the path; the path is continuously iterated based on a genetic algorithm until the fitness no longer increases, and the path when the fitness no longer increases is used as the corresponding shortest picking path.

[0065] In the above business data management method, the business data of the product is obtained, and feature extraction is performed to obtain the extracted business feature vector; the business data includes the order quantity, the number of the minimum inventory unit of the product, the number of the minimum inventory unit and the storage timestamp; based on the business feature vector, different products are clustered to obtain preliminary clustering results; the product orders and minimum inventory units are used as nodes to construct a graph neural network, and the trained graph neural network is used to further cluster and merge the preliminary clustering results to obtain the final clustering results; according to the final clustering results, the corresponding shortest picking path is calculated. In this way, by making full use of the characteristics of the product's business data for clustering, and calculating the shortest picking path according to the clustering results, the utilization rate of business data is improved, the cost in the picking process is reduced, and the efficiency of order processing is improved.

[0066] In an exemplary embodiment, the business data management method also includes: obtaining sample business data of the product and establishing a prediction model based on a convolutional neural network; training the prediction model based on the sample business data, and using an optimization algorithm to continuously optimize the parameters in the prediction model until the loss function of the prediction model no longer decreases, thereby obtaining a trained prediction model; inputting the business data generated by the product in the current period into the trained prediction model to obtain the future change trend of the product's business data.

[0067] In actual implementation, sample business data of the product is obtained, a prediction model is established based on a convolutional neural network, the sample business data is input into the prediction model for continuous training, the cross entropy loss function is used to calculate the loss function predicted by the prediction model, and the gradient descent algorithm is used to continuously optimize the parameters in the prediction model until the loss function of the prediction model no longer decreases significantly, thus obtaining a trained prediction model.

[0068] Among them, the sample business data of the product is the business data of the historical product; the method for calculating the loss function includes but is not limited to the cross-entropy loss function, and can also be other methods that can calculate the loss function of the prediction model, which is not limited to the embodiments of the present application; the method for optimizing the parameters in the prediction model includes but is not limited to the gradient descent algorithm, and can also be other methods that can optimize the parameters in the prediction model, which is not limited to the embodiments of the present application.

[0069] Furthermore, the business data generated in the current time period is input into the trained prediction model to obtain the future change trend of the business data of the product. The time period for inputting the business data of the trained prediction model can be determined according to the needs.

[0070] In the above embodiment, the prediction model is used to predict the change trend of the business data of the product, so that preparations can be made in advance for possible situations.

[0071] In some exemplary embodiments, products corresponding to business feature vectors are classified by a trained vector machine to obtain classification results; the corresponding training end condition of the trained vector machine is that the loss function value corresponding to the parameters in the vector machine is minimized; for the classification result obtained for each product, multiple products are clustered to obtain cluster centers; different products are clustered according to the cluster centers to which the products belong and the classification results of the products to obtain preliminary clustering results of the products.

[0072] In actual implementation, the terminal uses sample business data to train the support vector machine SVM using the radial basis function RBF kernel (Gaussian kernel or square exponential kernel), and uses the gradient descent method to continuously optimize the parameters in the vector machine until the loss function value corresponding to the parameters in the vector machine is minimized.

[0073] The trained vector machine is used to classify the products corresponding to the business feature vectors. The specific expression is shown in formula (1).

[0074]

[0075] in, represents the classification result, B represents the total number of business feature vectors, represents the coefficient of the ith business characteristic vector, represents the classification label of the i-th business feature vector, represents the radial basis function RBF kernel function, and b represents the bias term.

[0076] Based on the classification results obtained for each product, multiple products are clustered. The specific formula is shown in (2), and the cluster center is obtained. The cluster center is shown in formula (3). The updated membership matrix is ​​specifically expressed as shown in formula (4).

[0077]

[0078]

[0079]

[0080] in, represents the clustering dispersion, n represents the total number of orders, and c represents the number of clusters. represents the membership of the ith business feature vector in the membership matrix to cluster j, m represents the fuzzy index, which is usually 2 based on historical data. represents the i-th order feature vector, represents the jth cluster center.

[0081] Use the algorithm to perform iterative update optimization. During the continuous iteration process, when the change of the cluster center no longer decreases significantly, stop the iteration to converge the final membership degree and cluster center. According to the cluster center to which the product belongs and the classification result of the product, cluster different products to obtain the preliminary clustering result of the product.

[0082] In the above embodiment, the order business feature vectors are preliminarily classified through the support vector machine model to determine whether the order is suitable for merging, and then the clustering algorithm is used to perform fuzzy clustering on the orders suitable for merging. Each order can reflect its similarity to different clusters according to the membership matrix. Fuzzy clustering can flexibly handle the differences between orders and realize dynamic grouping of orders while ensuring that the merger is reasonable.

[0083] In an exemplary embodiment, the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results, including: inputting the product order data corresponding to the preliminary clustering results into the trained graph neural network to obtain an order feature vector after integrating the product orders of all products under each cluster; calculating the cosine similarity between the order feature vectors corresponding to each two clusters in the preliminary clustering results, and merging two clusters with cosine similarity greater than a preset value to obtain the final clustering result.

[0084] In actual implementation, product orders and minimum inventory units are used as nodes to construct a graph neural network. It is the product order data, including the minimum stock unit quantity, storage timestamp and customer region information, the minimum stock unit node It includes the minimum stock unit number, demand frequency, and inventory information, and is connected to multiple product order nodes that include the minimum stock unit node.

[0085] The variable weight is defined to represent the degree of association between the product order node and the minimum inventory unit node. The specific formula is shown in (5).

[0086]

[0087] in, Indicates the number of minimum stock keeping unit nodes in the product order node. It represents the demand frequency of the minimum inventory unit node in the total orders, and m represents the total number of order nodes.

[0088] According to the number of structural layers of the neural network, the graph neural network is updated, which is expressed as formula (6).

[0089]

[0090] in, represents the association vector of product order node i in the g+1th layer, represents the ReLU activation function, N(i) represents the set of neighbor nodes of product order node i, and Represent the degree of the product order node and the neighbor node respectively (indicating the number of edges connecting the nodes), represents the weight of the g-th layer graph neural network, Represents the order feature vector of the g-th layer of product order node j.

[0091] Through weighted averaging, the features of neighboring product order nodes are passed to the target node to complete information integration. The graph neural network is trained using the training set, and the mean square error loss function is defined to minimize the association error between the product order node and the minimum inventory unit node. The model parameters are iteratively optimized through the gradient descent method. When the calculated error no longer decreases significantly during continuous iterations, the iteration is stopped to obtain the trained graph neural network.

[0092] The product order data corresponding to the preliminary clustering results are input into the trained graph neural network to obtain the order feature vector after integrating the product orders of all products under each cluster; the cosine similarity between the order feature vectors corresponding to each two clusters in the preliminary clustering results is calculated, and the two clusters with cosine similarity greater than the preset value are merged to obtain the final clustering result.

[0093] In the above embodiment, the order feature vectors of the product orders are calculated by the trained graph neural network, and then reasonably merged through cosine similarity to obtain the final clustering result, which makes the subsequent processing of the product orders more convenient and more efficient.

[0094] In an exemplary embodiment, the number of the minimum inventory unit of each product under each cluster in the final clustering result is used as a path node, and a path formed by the path nodes is generated through a preset model; the length of the path is calculated, and the inverse of the length is used as the fitness of the path; the path is continuously iterated based on a genetic algorithm until the fitness no longer increases, and the path when the fitness no longer increases is used as the corresponding shortest picking path.

[0095] In actual implementation, the traveling salesman problem TSP model is used to plan the shortest picking path, and the numbers of all minimum inventory units to be picked are defined as a path node set P, and the path nodes contain the location information corresponding to the numbers of the minimum inventory units.

[0096] The distance matrix D is defined according to the distance between the location information of different minimum inventory units. The matrix D represents the distance between each pair of minimum inventory unit storage locations. The optimization objective function is defined based on minimizing the total path length. The specific formula is shown in (7).

[0097]

[0098] Among them, MinZ represents the shortest picking path value, represents the distance between the i-th minimum inventory unit storage location and the j-th minimum inventory unit storage location, It represents the decision variable. If the picking path contains the path of the i-th minimum stock keeping unit storage location and the j-th minimum stock keeping unit storage location, it is represented as 1, otherwise it is represented as 0. h represents the total number of SKU storage locations.

[0099] The defined constraints include that each minimum inventory unit storage location can only enter and leave through one path, and sub-loop elimination is performed. This constraint ensures that the path does not form a sub-loop, thereby ensuring that all nodes are connected in series to form a complete closed loop;

[0100] A genetic algorithm is used to generate a path based on the minimum inventory unit number as the path node, and the total length of the path is calculated. The reciprocal of the total length is defined as the fitness of the corresponding path, and the path with high fitness is used as the parent path. Partial mapping cross PMX is performed based on the genetic algorithm, and the intervals formed by the path node pairs of the parent path are exchanged to generate the child path. The path is iteratively optimized until the fitness of the corresponding path is no longer significantly improved, then the iteration is stopped and the shortest picking path is output.

[0101] In the above embodiment, by ensuring that pickers pick products in the optimal order, the planning of the shortest picking path reduces the round-trip time for picking products in the warehouse, improves picking efficiency, and reduces the moving distance of manpower and equipment in the warehouse.

[0102] In an exemplary embodiment, the business data management method further includes: when the final clustering result changes, recalculating the corresponding shortest picking path according to the changed final clustering result.

[0103] In actual implementation, the terminal will monitor product changes, record the product path based on the inventory quantity of the minimum inventory unit, shelf number and change timestamp, update the product picking progress in real time, and recalculate the corresponding shortest picking path based on the final clustering result after the change.

[0104] In the above embodiment, by real-time monitoring of product changes, it is possible to ensure timely updating of inventory data, reduce the occurrence of repeated paths during the picking process, and improve picking efficiency.

[0105] In some embodiments, data interactions involved in the business data management process are all encrypted transmissions.

[0106] To explain the business data management method in the present application in detail, an embodiment is used for explanation below. By way of example, the present application is described with a business data management method in a specific e-commerce product picking scenario.

[0107] First, the terminal will collect the product's order quantity, the product's minimum inventory unit number, the number of minimum inventory units and the storage timestamp. The extracted product's order quantity, the product's minimum inventory unit number, the number of minimum inventory units and the storage timestamp constitute a business feature vector.

[0108] The terminal uses sample business data to train a support vector machine (SVM) using a radial basis function RBF kernel (Gaussian kernel or square exponential kernel), and uses the gradient descent method to continuously optimize the parameters in the vector machine until the loss function value corresponding to the parameters in the vector machine is minimized.

[0109] The trained vector machine is used to classify the products corresponding to the business feature vectors, as shown in formula (1). Based on the classification results obtained for each product, multiple products are clustered, as shown in formula (2), and the cluster center is obtained, as shown in formula (3). The updated membership matrix is ​​specifically expressed as shown in formula (4).

[0110] Use the algorithm to perform iterative update optimization. During the continuous iteration process, when the change of the cluster center no longer decreases significantly, stop the iteration to converge the final membership degree and cluster center. According to the cluster center to which the product belongs and the classification result of the product, cluster different products to obtain the preliminary clustering result of the product.

[0111] The product order and the minimum inventory unit are used as nodes to build a graph neural network. It is the product order data, including the minimum stock unit quantity, storage timestamp and customer region information, the minimum stock unit node It includes the minimum stock unit number, demand frequency, and inventory information, and is connected to multiple product order nodes that include the minimum stock unit node.

[0112] The variable weight is defined to represent the degree of association between the product order node and the minimum inventory unit node. The specific formula is shown in (5).

[0113] Through weighted averaging, the features of neighboring product order nodes are passed to the target node to complete information integration. The graph neural network is trained using the training set, and the mean square error loss function is defined to minimize the association error between the product order node and the minimum inventory unit node. The model parameters are iteratively optimized through the gradient descent method. When the calculated error no longer decreases significantly during continuous iterations, the iteration is stopped to obtain the trained graph neural network.

[0114] The product order data corresponding to the preliminary clustering results are input into the trained graph neural network to obtain the order feature vector after integrating the product orders of all products under each cluster; the cosine similarity between the order feature vectors corresponding to each two clusters in the preliminary clustering results is calculated, and the two clusters with cosine similarity greater than the preset value are merged to obtain the final clustering result.

[0115] In actual implementation, the traveling salesman problem TSP model is used to plan the shortest picking path, and the numbers of all minimum inventory units to be picked are defined as a path node set P, and the path nodes contain the location information corresponding to the numbers of the minimum inventory units.

[0116] The distance matrix D is defined according to the distance between the location information of different minimum inventory units. The matrix D represents the distance between each pair of minimum inventory unit storage locations. The optimization objective function is defined based on minimizing the total path length. The specific formula is shown in (7).

[0117] The defined constraints include that each minimum inventory unit storage location can only enter and leave through one path, and sub-loop elimination is performed. This constraint ensures that the path does not form a sub-loop, thereby ensuring that all nodes are connected in series to form a complete closed loop;

[0118] A genetic algorithm is used to generate a path based on the minimum inventory unit number as the path node, and the total length of the path is calculated. The reciprocal of the total length is defined as the fitness of the corresponding path, and the path with high fitness is used as the parent path. Partial mapping cross PMX is performed based on the genetic algorithm, and the intervals formed by the path node pairs of the parent path are exchanged to generate the child path. The path is iteratively optimized until the fitness of the corresponding path is no longer significantly improved, then the iteration is stopped and the shortest picking path is output.

[0119] The terminal will monitor product changes and record the product path based on the inventory quantity of the smallest inventory unit, shelf number and change timestamp, update the product picking progress in real time, and recalculate the corresponding shortest picking path based on the final clustering results after the change.

[0120] In this application, the combination of vector machines and clustering can achieve efficient classification and clustering of real-time products. The rapid classification processing and flexible clustering of the business characteristics of products improve the ability to handle product changes. After clustering is completed, the shortest picking path for the product is calculated and continuously updated according to the real-time changes of the product, making the picking efficiency of products higher during peak periods.

[0121] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0122] Based on the same inventive concept, the embodiment of the present application also provides a business data management device for implementing the business data management method involved above. The implementation solution provided by the device to solve the problem is similar to the implementation solution recorded in the above method, so the specific limitations in one or more business data management device embodiments provided below can refer to the limitations of the business data management method above, and will not be repeated here.

[0123] In an exemplary embodiment, as shown in FIG3 , a business data management device is provided, including: an acquisition module 301, a clustering module 302 and a calculation module 303, wherein:

[0124] An acquisition module is used to acquire business data of a product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp;

[0125] A clustering module, used to cluster different products based on the business feature vector to obtain a preliminary clustering result;

[0126] The clustering module is further used to construct a graph neural network using product orders and minimum inventory units as nodes, and further cluster and merge the preliminary clustering results through the trained graph neural network to obtain a final clustering result;

[0127] The calculation module is used to calculate the corresponding shortest picking path according to the final clustering result.

[0128] In some embodiments, the device also includes a prediction module for obtaining sample business data of a product and establishing a prediction model based on a convolutional neural network; training the prediction model based on the sample business data, and using an optimization algorithm to continuously optimize the parameters in the prediction model until the loss function of the prediction model no longer decreases, thereby obtaining a trained prediction model; inputting the business data generated by the product in the current period into the trained prediction model to obtain the future change trend of the product's business data.

[0129] In some embodiments, the above-mentioned clustering module is also used to classify the products corresponding to the business feature vector through the trained vector machine to obtain the classification results; the corresponding training end condition of the trained vector machine is that the loss function value corresponding to the parameters in the vector machine is the smallest; for the classification results obtained for each product, multiple products are clustered to obtain the cluster center; according to the cluster center to which the product belongs and the classification results of the product, different products are clustered to obtain preliminary clustering results of the product.

[0130] In some embodiments, the above-mentioned clustering module is also used to input the product order data corresponding to the preliminary clustering results into the trained graph neural network to obtain an order feature vector after integrating the product orders of all products under each cluster; calculate the cosine similarity between the order feature vectors corresponding to each two clusters in the preliminary clustering results, merge the two clusters whose cosine similarity is greater than a preset value, and obtain the final clustering result.

[0131] In some embodiments, the above-mentioned calculation module is also used to use the number of the minimum inventory unit of each product under each cluster in the final clustering result as a path node, and generate a path formed by the path nodes through a preset model; calculate the length of the path, and use the inverse of the length as the fitness of the path; based on the genetic algorithm, the path is continuously iterated until the fitness no longer increases, and the path when the fitness no longer increases is used as the corresponding shortest picking path.

[0132] In some embodiments, the above-mentioned calculation module is also used to recalculate the corresponding shortest picking path according to the changed final clustering result when the final clustering result changes.

[0133] Each module in the above-mentioned business data management device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module above.

[0134] In an exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be shown in FIG4. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store business data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a business data management method is implemented.

[0135] The display unit of the computer device is used to form a visually visible picture, and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball, or touchpad provided on the computer device housing, or an external keyboard, touchpad, or mouse, etc.

[0136] Those skilled in the art will understand that the structure shown in FIG. 4 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0137] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0138] Acquire business data of the product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp;

[0139] Based on the business feature vector, different products are clustered to obtain preliminary clustering results;

[0140] The product orders and minimum inventory units are used as nodes to construct a graph neural network, and the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results;

[0141] According to the final clustering results, the corresponding shortest picking path is calculated.

[0142] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0143] Acquire business data of the product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp;

[0144] Based on the business feature vector, different products are clustered to obtain preliminary clustering results;

[0145] The product orders and minimum inventory units are used as nodes to construct a graph neural network, and the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results;

[0146] According to the final clustering results, the corresponding shortest picking path is calculated.

[0147] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0148] Acquire business data of the product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp;

[0149] Based on the business feature vector, different products are clustered to obtain preliminary clustering results;

[0150] The product orders and minimum inventory units are used as nodes to construct a graph neural network, and the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results;

[0151] According to the final clustering results, the corresponding shortest picking path is calculated.

[0152] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0153] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0154] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0155] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A business data management method, characterized in that: The method comprises: Acquire business data of the product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp; Based on the business feature vector, different products are clustered to obtain preliminary clustering results; The product orders and minimum inventory units are used as nodes to construct a graph neural network, and the preliminary clustering results are further clustered and merged through the trained graph neural network to obtain the final clustering results; According to the final clustering results, the corresponding shortest picking path is calculated.

2. The method according to claim 1, characterized in that The method further comprises: Obtain sample business data of the product and establish a prediction model based on a convolutional neural network; The prediction model is trained based on the sample business data, and the parameters in the prediction model are continuously optimized using an optimization algorithm until the loss function of the prediction model no longer decreases, thereby obtaining a trained prediction model; The business data generated by the product in the current period is input into the trained prediction model to obtain the future change trend of the business data of the product.

3. The method according to claim 1, characterized in that: Based on the business feature vector, different products are clustered to obtain preliminary clustering results, including: The products corresponding to the business feature vectors are classified by the trained vector machine to obtain classification results; the training end condition of the trained vector machine is that the loss function value corresponding to the parameters in the vector machine is the minimum; Based on the classification results obtained for each product, multiple products are clustered to obtain the cluster center; Different products are clustered according to the cluster centers to which the products belong and the classification results of the products to obtain preliminary clustering results of the products.

4. The method according to claim 1, characterized in that: The trained graph neural network further clusters and merges the preliminary clustering results to obtain the final clustering results, including: Inputting the product order data corresponding to the preliminary clustering results into the trained graph neural network to obtain an order feature vector after integrating the product orders of all products under each cluster; The cosine similarity between the order feature vectors corresponding to each two clusters in the preliminary clustering result is calculated, and two clusters whose cosine similarity is greater than a preset value are merged to obtain a final clustering result.

5. The method according to claim 1, characterized in that The method of calculating the corresponding shortest picking path according to the final clustering result includes: The number of the minimum inventory unit of each product under each cluster in the final clustering result is used as a path node, and a path formed by the path nodes is generated through a preset model; Calculating the length of the path, and taking the reciprocal of the length as the fitness of the path; The path is continuously iterated based on the genetic algorithm until the fitness stops increasing, and the path when the fitness stops increasing is used as the corresponding shortest picking path.

6. The method according to claim 1, characterized in that The method further comprises: In the case where the final clustering result changes, the corresponding shortest picking path is recalculated according to the changed final clustering result.

7. A business data management device, characterized in that: The device comprises: An acquisition module is used to acquire business data of a product and perform feature extraction to obtain an extracted business feature vector; the business data includes order quantity, number of the minimum inventory unit of the product, quantity of the minimum inventory unit and storage timestamp; A clustering module, used to cluster different products based on the business feature vector to obtain a preliminary clustering result; The clustering module is further used to construct a graph neural network using product orders and minimum inventory units as nodes, and further cluster and merge the preliminary clustering results through the trained graph neural network to obtain a final clustering result; The calculation module is used to calculate the corresponding shortest picking path according to the final clustering result.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images