Cross-enterprise closed-loop order transfer and data cooperative processing method

Abstract

The invention relates to the technical field of data management, and provides a cross-enterprise closed-loop order transfer and data cooperative processing method, which comprises the steps of collecting multi-modal data, generating a unique distributed hash identifier for the multi-modal data based on a block chain technology, and performing intelligent association on the multi-modal data in combination with transfer node information; constructing a cross-enterprise order flow graph model by using the multi-modal data fusion; detecting an abnormal mode in real time by combining a graph embedding algorithm with a deep neural network; on the basis of the order flow graph model, a dynamic responsibility allocation algorithm is utilized, and in combination with real-time environmental factors when an anomaly occurs, a specific responsibility node where the anomaly occurs is positioned, and a corresponding responsibility weight is allocated; and generating an adaptive repair strategy based on a multi-objective optimization algorithm according to the responsibility weight. According to the invention, the transparency and cooperation efficiency of cross-enterprise order transfer are significantly improved, and the stability and efficiency of a supply chain system are ensured.

Description

Technical Field

[0001] The present application relates to the field of data management technology, and specifically, to a method for cross-enterprise closed-loop order circulation and data collaborative processing. Background Art

[0002] In recent years, with the diversification of inter-enterprise cooperation models and the globalization of supply chains, the demand for cross-enterprise order flow and data collaborative processing has increased. Traditional order management models mostly rely on centralized information systems, which have certain limitations in dealing with diversified data processing, order flow transparency, and cross-enterprise collaboration efficiency. In addition, with the gradual maturity of technologies such as the Internet of Things, blockchain, big data, and artificial intelligence, more and more companies are trying to use these emerging technologies to optimize the order management process.

[0003] However, the existing cross-enterprise order management system still has some obvious deficiencies in terms of transparency and collaborative efficiency. For example, data between enterprises is heterogeneous and diverse, and traditional methods are difficult to achieve dynamic association and sharing of multimodal data while ensuring data integrity and consistency; existing order flow monitoring methods are mostly based on static rules or simple models, which are difficult to deal with complex abnormal patterns that may occur during dynamic flow, and abnormal location and responsibility allocation lack scientificity and accuracy. In addition, existing repair strategies are usually fixed preset rules, which fail to fully combine actual environmental conditions to dynamically generate optimal strategies, thus affecting the timeliness and accuracy of problem solving. Summary of the invention

[0004] The embodiments of the present application provide a method for cross-enterprise closed-loop order circulation and data collaborative processing, thereby achieving efficient, secure and intelligent cross-enterprise order circulation and data collaborative processing at least to a certain extent.

[0005] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0006] According to one aspect of the present application, a cross-enterprise closed-loop order flow and data collaborative processing method is provided, including: collecting multimodal data, generating a unique distributed hash identifier for the multimodal data based on blockchain technology, and intelligently associating the multimodal data in combination with flow node information; using the multimodal data fusion to construct a cross-enterprise order flow graph model; detecting abnormal patterns in real time through a graph embedding algorithm combined with a deep neural network; based on the order flow graph model, using a dynamic responsibility allocation algorithm, combined with real-time environmental factors when the anomaly occurs, locating the specific responsibility node where the anomaly occurs, and assigning corresponding responsibility weights; based on the responsibility weights, generating an adaptive repair strategy based on a multi-objective optimization algorithm.

[0007] In the present application, based on the aforementioned scheme, a personalized dialysis scheme is generated according to the basic information, including: the intelligent association of multimodal data in combination with the flow node information, including: using the generated distributed hash identifier to intelligently associate multimodal data in combination with the upstream and downstream relationships of the flow nodes; by constructing an association matrix based on hash identifiers, different types of data are uniformly mapped to the time series chain of cross-enterprise order flow nodes.

[0008] In the present application, based on the aforementioned scheme, the order flow graph model includes: defining the key nodes of the order flow as nodes of the graph model; mapping the association relationship between the nodes as the edges of the graph model; and generating edge weights based on the multi-dimensional feature associations between the nodes.

[0009] In the present application, based on the aforementioned scheme, the abnormal pattern is detected in real time by combining a graph embedding algorithm with a deep neural network, including: using a graph embedding algorithm to extract features from the order flow graph model, and embedding node attributes and edge weights into a high-dimensional vector space; dynamically updating the order flow graph model in combination with the multimodal data and incremental data collected in real time; using historical order flow data to build an anomaly detection model based on a deep neural network; inputting a high-dimensional embedded feature vector to calculate the anomaly probability; and using the patterns of similar abnormal scenarios in historical data as a reference to determine the abnormal pattern.

[0010] In the present application, based on the aforementioned scheme, the embedding of node attributes and edge weights into a high-dimensional vector space includes: obtaining an order flow graph model, which includes node and edge information; for each node, using an embedding algorithm based on a graph convolutional network or a graph attention network to aggregate the information of the node itself and its neighbors layer by layer; embedding the features of the node into a high-dimensional vector space, and outputting a high-dimensional embedded feature vector for each node.

[0011] In the present application, based on the aforementioned scheme, the specific responsible nodes where the abnormality occurs are located, including: filtering potential abnormal nodes through the high abnormality probability nodes output by the anomaly detection model, combined with the priority in the dynamic responsibility allocation rule set; using the high-dimensional embedding features in the graph model to analyze the edge weights and feature changes of the target node's associated upstream and downstream nodes, confirming the abnormal nodes and propagation chains; determining the unique identifier of the abnormal node and its upstream and downstream node attribute information.

[0012] In the present application, based on the above-mentioned solution, the positioning of abnormal nodes in the flow path includes: the positioning result includes the unique identifier of the responsible node and the attribute information of the associated upstream and downstream nodes.

[0013] In the present application, based on the aforementioned scheme, the allocation of corresponding responsibility weights includes: defining the responsibility weight of each node as a weighted combination of multiple factors; for the associated upstream and downstream nodes of the responsible node, the weights are determined by the propagation effect and adjacency attributes: analyzing the intensity of abnormal propagation received by each associated node; for each associated node, checking its own attribute data to determine whether it has abnormal characteristics; combining the evaluation results of the propagation effect and attribute anomalies, and allocating responsibilities in a targeted manner by reasonably adjusting the proportions of the two evaluation results.

[0014] In the present application, based on the aforementioned solution, the adaptive repair strategy is generated based on a multi-objective optimization algorithm, including: the repair content includes flow path replanning and resource reallocation.

[0015] In the present application, based on the aforementioned scheme, the adaptive repair strategy generated based on the multi-objective optimization algorithm also includes: optimizing the path planning according to the responsibility node and upstream and downstream related node information, combining the responsibility weight, and re-planning the flow path using the shortest path algorithm or heuristic algorithm to give priority to repairing the paths associated with high-weight nodes; combining the responsibility allocation results and reallocating resources in combination with the real-time node load information, and the allocation rules include: giving priority to meeting the resource requirements of high-priority orders, balancing the use of resources across enterprises, and avoiding overloading of a single node or waste of resources.

[0016] According to one aspect of the present application, a computer-readable medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the cross-enterprise closed-loop order flow and data collaborative processing method as described in the above embodiments is implemented.

[0017] According to one aspect of the present application, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the cross-enterprise closed-loop order flow and data collaborative processing method as described in the above-mentioned embodiments.

[0018] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program including computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the cross-enterprise closed-loop order flow and data collaborative processing method provided in the above-mentioned various optional implementations.

[0019] In the technical solution of the present application, by collecting multimodal data and using blockchain to generate distributed hash identifiers, not only the integrity and immutability of the data are guaranteed, but also the dynamic association of cross-enterprise order flow data is realized; the constructed cross-enterprise order flow graph model, by modeling the key nodes of the order flow and their association relationships as a graph structure, provides a basis for the correlation analysis and anomaly detection of the entire process data.

[0020] By using graph embedding algorithms combined with deep neural networks, high-dimensional features of order flow graphs can be extracted in real time, and the accuracy and adaptability of anomaly detection can be enhanced through dynamic updates, thereby achieving timely identification of abnormal patterns in complex supply chains. Based on the dynamic responsibility allocation algorithm, combined with real-time environmental factors and anomaly detection results, it can accurately locate abnormal nodes and reasonably allocate responsibilities, providing a scientific basis for problem solving.

[0021] In addition, an adaptive repair strategy is generated through a multi-objective optimization algorithm to replan the flow path and resource allocation, prioritize the repair of key problem nodes and optimize resource utilization.

[0022] Overall, the present invention significantly improves the transparency and collaboration efficiency of cross-enterprise order flows, and ensures the stability and efficiency of the supply chain system.

[0023] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0025] Figure 1 A flowchart of a method for cross-enterprise closed-loop order flow and data collaborative processing in one embodiment of the present application is schematically shown.

[0026] Figure 2 A schematic diagram of an order flow graph model in an embodiment of the present application is schematically shown.

[0027] Figure 3 The following is a schematic diagram of detecting abnormal patterns in real time in one embodiment of the present application. DETAILED DESCRIPTION

[0028] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0029] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0030] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0031] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0032] The implementation details of the technical solution of this application are described in detail below:

[0033] Figure 1 A flowchart of a method for cross-enterprise closed-loop order transfer and data collaborative processing according to an embodiment of the present application is shown. Figure 1 As shown, the cross-enterprise closed-loop order circulation and data collaborative processing method includes at least steps S1 to S5, which are described in detail as follows:

[0034] S1: Collect multimodal data, generate a unique distributed hash identifier for the multimodal data based on blockchain technology, and intelligently associate the multimodal data with the flow node information.

[0035] Specifically, data is collected through multimodal data collection devices deployed at cross-enterprise order flow nodes, including text data (order status and operation log), image data (cargo status) and environmental data (temperature, humidity, vibration, etc.).

[0036] In one embodiment of the present application, the following steps may be specifically included:

[0037] S1.1: Obtain text data (order status, operation log), image data (cargo status) and environmental data (temperature, humidity, vibration, etc.) through multimodal data acquisition equipment.

[0038] The data is marked with timestamps and associated with node identifiers to ensure the continuity and real-time nature of data collection.

[0039] S1.2: Preprocess the collected multimodal data, including text data standardization (such as unified coding format), image data compression and noise filtering, and outlier removal of environmental data.

[0040] After data preprocessing, the flow node information is used to identify and remove duplicate data to generate a unique basic data set.

[0041] S1.3: Based on blockchain technology, a unique distributed hash identifier is generated for the preprocessed multimodal data.

[0042] Preferably, the hash tag uses a cross-enterprise encryption protocol, combined with the node's geographic location, data collection timestamp and device identification, to ensure the integrity and tamper-proof capabilities of data flowing between enterprises.

[0043] S1.4: Utilize the generated distributed hash identifiers and combine the upstream and downstream relationships of the flow nodes to intelligently associate multimodal data; by constructing an association matrix based on hash identifiers, map different types of data uniformly to the time series chain of cross-enterprise order flow nodes, providing consistent input for subsequent data analysis.

[0044] It should be noted that this technical solution simulates query and traceability operations in cross-enterprise scenarios by randomly selecting certain data nodes, and can use the distributed hash identifier to quickly locate associated multimodal data, and verify whether the data's integrity, accuracy, and associated logic meet cross-enterprise order flow requirements.

[0045] S2: Utilize the multimodal data fusion to build a cross-enterprise order flow graph model; detect abnormal patterns in real time by combining graph embedding algorithm with deep neural network.

[0046] In one embodiment of the present application, the following steps may be specifically included:

[0047] S2.1: Using the distributed hash identifier and multimodal data generated in step S1, define the key nodes of the order flow as nodes of the graph model, such as Figure 2As shown, the association relationship between nodes (such as upstream and downstream flow order) is mapped to the edge of the graph model, and the edge weight is generated according to the multi-dimensional feature association between the nodes.

[0048] Preferably, let the two nodes i and j in the order flow graph model represent two key nodes of the order flow, then the edge weight ω ij The calculation is as follows:

[0049]

[0050] Among them, ω ij is the association weight between nodes i and j, Δt ij is the time difference between nodes i and j, σ t is the standard deviation of the time difference, KL(p i ||p j ) is the multimodal distribution p of nodes i and j i and p j The Kullback-Leibler divergence between the two is used to measure the difference in data distribution, R ij is the historical transfer success rate between nodes, μ R It is the average of the historical flow success rate and is used to adjust the dynamics of the threshold.

[0051] S2.2: Use graph embedding algorithm to extract features from the order flow graph model and embed node attributes and edge weights into high-dimensional vector space.

[0052] The embedded features include circulation path efficiency, environmental parameter fluctuation range and order status stability.

[0053] Specifically, the order flow graph model is obtained, which includes node and edge information; node attributes include order creation time, processing time, completion status, current environment parameters, etc.

[0054] For each node, an embedding algorithm based on graph convolutional network or graph attention network is used to aggregate the information of the node itself and its neighbors layer by layer;

[0055] Embed the node features into a high-dimensional vector space and output the high-dimensional embedded feature vector of each node;

[0056] Among them, for node i, its high-dimensional embedding feature vector v i Expressed as:

[0057]

[0058] in, is the set of adjacent nodes of node i, h g is the hidden feature vector of the adjacent node g, a i and ag is the attribute vector of node i and node g, ||a i -a g || is the Euclidean distance between attribute vectors, τ is the normalization factor, ACT(.) is the nonlinear activation function (such as GELU), ω ig is the edge weight of the node.

[0059] S2.3: Combine the multimodal data in step S1 and the incremental data collected in real time to dynamically update the order flow graph model.

[0060] The dynamic update includes feature calculation of newly added nodes, edge weight adjustment and deletion of invalid nodes to ensure the real-time and integrity of the graph model.

[0061] Preferably, in an embodiment of the present application, the following operation steps may be specifically included:

[0062] Obtain incremental data from the order management system, including new order information, flow status updates, environmental parameter changes, etc.;

[0063] For newly added orders, a new node is created and an initial feature vector is generated based on its attributes, including the order's environmental parameters, creation time, etc.

[0064] Calculate the association strength between the node and the existing nodes, and generate the edges and edge weights related to them.

[0065] Furthermore, the attributes of existing nodes and edges are updated based on the incremental data, and the increment of edge weight is calculated by the newly added associated information, such as flow duration or success rate;

[0066] If the weight of an edge drops below a preset threshold, it means that the path is invalid and needs to be marked as invalid.

[0067] S2.4: Use historical order flow data to build an anomaly detection model based on deep neural network; use graph embedding features as training samples, and optimize model parameters through training to improve the accuracy of abnormal pattern detection.

[0068] Specifically, the high-dimensional embedding feature vector of the dynamically updated order flow graph model is input into the anomaly detection model, wherein the high-dimensional embedding feature vector v based on the dynamically updated i , abnormal probability P abn Calculated as:

[0069]

[0070] Among them, Z is the normalization factor, v i is the high-dimensional embedding feature of the target node, and M is the covariance matrix, which represents the correlation between the embedded features.

[0071] Preferably, nodes with abnormal probability higher than the abnormal threshold and their associated nodes are identified based on the calculated abnormal probability, and the changes in edge weights between nodes are analyzed to identify potential abnormal propagation paths.

[0072] Use patterns of similar abnormal scenarios in historical data as reference to identify abnormal patterns. For example:

[0073] Abnormal delay: The edge weights between associated nodes are significantly reduced;

[0074] Data consistency anomalies: There are obvious differences in the embedded feature vectors of adjacent nodes, etc.

[0075] S3: Based on the order flow graph model, a dynamic responsibility allocation algorithm is used to locate the specific responsibility node where the exception occurs, combined with real-time environmental factors when the exception occurs (such as logistics carrier load, node data consistency, etc.), and the corresponding responsibility weight is allocated.

[0076] In one embodiment of the present application, the following steps may be specifically included:

[0077] S3.1: Utilize the dynamically updated responsibility allocation rule set and anomaly detection results, combined with the order flow graph model, to accurately locate abnormal nodes and their upstream and downstream node attributes, and provide data support for abnormal propagation analysis and responsibility division.

[0078] The better one is, first, filtering potential abnormal nodes through the high abnormal probability nodes output by the anomaly detection model and combining the priorities in the dynamic responsibility allocation rule set;

[0079] By using the high-dimensional embedding features in the graph model, we analyze the edge weights and feature changes of the target node's associated upstream and downstream nodes. For example, if the delay degree and association strength are significantly reduced, we can further confirm the abnormal nodes and their propagation chains.

[0080] Finally, the unique identifier of the abnormal node and its upstream and downstream node attribute information (such as flow duration and correlation strength fluctuation) are determined.

[0081] It should be noted that the positioning result includes the unique identifier of the responsible node and the attribute information of the associated upstream and downstream nodes. For example, if node A is located as an abnormal node, the time delay and association strength of its upstream and downstream nodes B and C are recorded at the same time.

[0082] S3.2: Using the information of the located responsibility node and the upstream and downstream associated nodes, the responsibility weight of the responsibility node is calculated based on the dynamic responsibility allocation algorithm.

[0083] Specifically, the responsibility weight R(i) of each node i is defined as a weighted combination of multiple factors, and the calculation formula is as follows:

[0084] R(i)=β1W time (i)+β2W status (i)

[0085] Among them, β1~β2 are weight factors, W time (i) is the time delay weight of node i, which measures whether the processing time of the node significantly exceeds the normal range when an exception occurs:

[0086]

[0087] Where Δt(i) is the processing time deviation of node i, σ time is the standard deviation of the time deviation; W status (i) is the node status weight, which is used to measure the abnormal degree of abnormal nodes in the flow status. The formula is:

[0088] W status (i) = 1-S con (i)

[0089] Among them, S con (i) is the consistency score of the node status.

[0090] Furthermore, the weights of the associated upstream and downstream nodes of the responsible node are determined by the propagation effect and adjacency attributes, as follows:

[0091] For each associated node, analyze the strength of its reception anomaly propagation. The strength of the propagation effect can be evaluated based on characteristics such as the distance of the path, the stability of the connection, and the degree of attenuation of the abnormal signal. For example, if the node is closer to the responsible node or the interference on the path is small, the node is more likely to receive anomalies and its propagation effect is more significant.

[0092] In addition, for each associated node, check its own attribute data, such as operating status, performance indicators, or historical abnormal records, to determine whether it has abnormal characteristics. If the node's attributes deviate from the normal value range, its attribute abnormality is high and needs to be given a higher weight in the responsibility weight.

[0093] Finally, the evaluation results of the propagation effect and attribute anomaly are combined to reflect the comprehensive level of responsibility of the associated nodes in abnormal events. By reasonably adjusting the proportion of the two evaluation results (such as focusing on the propagation effect or the node's own anomaly), targeted allocation of responsibilities can be achieved.

[0094] Preferably, the responsibility nodes and corresponding responsibility weight results are associated with the order flow graph model and anomaly detection information and stored, and the output results include responsibility allocation logs and node attribute adjustment suggestions.

[0095] S4: According to the responsibility weights, an adaptive repair strategy is generated based on a multi-objective optimization algorithm.

[0096] Among them, the repair content includes re-planning of circulation routes and reallocation of resources.

[0097] In one embodiment of the present application, the following steps may be specifically included:

[0098] According to the calculated responsibility weight R(i), the abnormal responsibility is assigned to the relevant nodes, and the responsibility ratio of each node is recorded.

[0099] Preferably, according to the information of the responsible node and its upstream and downstream associated nodes, the path planning is optimized in combination with the responsibility weight, and the flow path is replanned using the shortest path algorithm or heuristic algorithm. For example, in the path planning process, the classic D ij The kstra algorithm or A* algorithm gives priority to repairing the paths associated with high-weight nodes.

[0100] Preferably, the planning result includes the adjustment order of nodes and the estimation of the total delay time of the path.

[0101] Furthermore, based on the responsibility allocation results and real-time node load information, resources such as logistics carriers and storage space are reallocated. The allocation rules include: giving priority to meeting the resource requirements of high-priority orders, balancing resource usage across enterprises, and avoiding overloading of a single node or waste of resources.

[0102] The re-planned flow path is associated with the responsible nodes and weights, and the priority repair order of nodes with high responsible weights is marked. Combined with the node load information, the actual feasibility of the resource allocation strategy is verified.

[0103] Furthermore, after executing the repair strategy, data is collected again to verify the repair effect; after the repair is completed, the abnormal data, repair strategy, etc. are written into the cross-enterprise dynamic rule base, and the rule base model is updated in real time; in the case of repair failure, the self-repair strategy iteration mechanism based on reinforcement learning is triggered to optimize the repair path and re-execute the repair.

[0104] In the technical solution of the present application, by collecting multimodal data and using blockchain to generate distributed hash identifiers, not only the integrity and immutability of the data are guaranteed, but also the dynamic association of cross-enterprise order flow data is realized; the constructed cross-enterprise order flow graph model, by modeling the key nodes of the order flow and their association relationships as a graph structure, provides a basis for the correlation analysis and anomaly detection of the entire process data.

[0105] By using graph embedding algorithms combined with deep neural networks, high-dimensional features of order flow graphs can be extracted in real time, and the accuracy and adaptability of anomaly detection can be enhanced through dynamic updates, thereby achieving timely identification of abnormal patterns in complex supply chains. Based on the dynamic responsibility allocation algorithm, combined with real-time environmental factors and anomaly detection results, it can accurately locate abnormal nodes and reasonably allocate responsibilities, providing a scientific basis for problem solving.

[0106] In addition, an adaptive repair strategy is generated through a multi-objective optimization algorithm to replan the flow path and resource allocation, prioritize the repair of key problem nodes and optimize resource utilization.

[0107] Overall, the present invention significantly improves the transparency and collaboration efficiency of cross-enterprise order flows, and ensures the stability and efficiency of the supply chain system.

[0108] The following introduces an embodiment of the device of the present application, which can be used to execute the cross-enterprise closed-loop order flow and data collaborative processing method in the above-mentioned embodiment of the present application. It can be understood that the device can be a computer program (including program code) running on a computer device, for example, the device is an application software; the device can be used to execute the corresponding steps in the method provided in the embodiment of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the cross-enterprise closed-loop order flow and data collaborative processing method mentioned above in the present application.

[0109] It should be noted that the computer-readable medium shown in the embodiment of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by an instruction execution system, device or device or used in combination with it. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a computer-readable computer program is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0110] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. Wherein, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0111] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. The names of these units do not, in some cases, constitute limitations on the units themselves.

[0112] According to one aspect of the present application, a computer program product or a computer program is provided, the computer program product or the computer program comprising computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in the above-mentioned various optional implementations.

[0113] As another aspect, the present application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device implements the cross-enterprise closed-loop order flow and data collaborative processing method described in the above embodiment.

[0114] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0115] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the implementation methods of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the implementation methods of the present application.

[0116] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not disclosed in the present application.

[0117] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for cross-enterprise closed-loop order circulation and data collaborative processing, characterized in that: include: Collect multimodal data, generate unique distributed hash identifiers for multimodal data based on blockchain technology, and intelligently associate multimodal data with flow node information; Constructing a cross-enterprise order flow graph model using the multimodal data fusion; Detect abnormal patterns in real time through graph embedding algorithms combined with deep neural networks; Based on the order flow graph model, a dynamic responsibility allocation algorithm is used to locate the specific responsibility node where the exception occurs, combined with the real-time environmental factors when the exception occurs, and the corresponding responsibility weight is allocated; According to the responsibility weights, an adaptive repair strategy is generated based on a multi-objective optimization algorithm.

2. The cross-enterprise closed-loop order flow and data collaborative processing method according to claim 1 is characterized in that: The intelligent association of multimodal data in combination with the flow node information includes: Using the generated distributed hash identifiers, the multimodal data is intelligently associated in combination with the upstream and downstream relationships of the flow nodes; By constructing a hash-based association matrix, different types of data are uniformly mapped to the time series chain of cross-enterprise order flow nodes.

3. The cross-enterprise closed-loop order flow and data collaborative processing method according to claim 1 is characterized in that: The order flow graph model includes: Define the key nodes of order flow as nodes of the graph model; Map the association relationship between nodes into edges of the graph model; Generate edge weights based on the multi-dimensional feature associations between nodes.

4. The cross-enterprise closed-loop order flow and data collaborative processing method according to claim 3 is characterized in that: The method combines a graph embedding algorithm with a deep neural network to detect abnormal patterns in real time, including: A graph embedding algorithm is used to extract features from the order flow graph model, embedding node attributes and edge weights into a high-dimensional vector space; Dynamically updating the order flow graph model by combining the multimodal data and the incremental data collected in real time; Use historical order flow data to build an anomaly detection model based on deep neural networks; input high-dimensional embedded feature vectors to calculate anomaly probabilities; Use the patterns of similar abnormal scenarios in historical data as reference to determine the abnormal patterns.

5. The cross-enterprise closed-loop order circulation and data collaborative processing method according to claim 4 is characterized in that: The embedding of node attributes and edge weights into a high-dimensional vector space includes: Obtaining an order flow graph model, wherein the order flow graph model includes information of nodes and edges; For each node, an embedding algorithm based on graph convolutional network or graph attention network is used to aggregate the information of the node itself and its neighbors layer by layer; The node features are embedded into a high-dimensional vector space, and the high-dimensional embedded feature vector of each node is output.

6. The cross-enterprise closed-loop order circulation and data collaborative processing method according to claim 1 is characterized in that: The specific responsible node where the positioning anomaly occurs includes: Filter potential abnormal nodes by using the high abnormal probability nodes output by the anomaly detection model and the priorities in the dynamic responsibility allocation rule set; Using the high-dimensional embedding features in the graph model, we analyze the edge weights and feature changes of the target node’s associated upstream and downstream nodes to identify abnormal nodes and propagation chains. Determine the unique identifier of the abnormal node and its upstream and downstream node attribute information.

7. The cross-enterprise closed-loop order flow and data collaborative processing method according to claim 6 is characterized in that: The locating of abnormal nodes in the flow path includes: The positioning result includes the unique identifier of the responsible node and the attribute information of the associated upstream and downstream nodes.

8. The cross-enterprise closed-loop order circulation and data collaborative processing method according to claim 1 is characterized in that: The allocation of corresponding responsibility weights includes: Define the responsibility weight of each node as a weighted combination of multiple factors; For the associated upstream and downstream nodes of the responsible node, the weights are determined by the propagation effect and adjacency attributes: Analyze the intensity of receiving abnormal propagation at each associated node; For each associated node, check its attribute data to determine whether it has abnormal characteristics; Combine the evaluation results of propagation effect and attribute anomaly, and allocate responsibilities in a targeted manner by reasonably adjusting the proportions of the two evaluation results.

9. The cross-enterprise closed-loop order circulation and data collaborative processing method according to claim 1 is characterized in that: The method of generating an adaptive repair strategy based on a multi-objective optimization algorithm includes: The repairs include re-planning of circulation routes and reallocation of resources.

10. The cross-enterprise closed-loop order flow and data collaborative processing method according to claim 9 is characterized in that: The generating of the adaptive repair strategy based on the multi-objective optimization algorithm also includes: According to the information of the responsible node and the upstream and downstream associated nodes, the path planning is optimized in combination with the responsibility weight, and the flow path is replanned using the shortest path algorithm or heuristic algorithm to give priority to repairing the paths associated with the high-weight nodes; Combining the responsibility allocation results and real-time node load information, resources are reallocated. The allocation rules include: giving priority to meeting the resource needs of high-priority orders, balancing resource usage across enterprises, and avoiding single node overload or resource waste.

Images