A supply chain risk prediction method and system

Abstract

The application belongs to the technical field of supply chain data analysis, and discloses a supply chain risk prediction method and system.The method comprises the following steps: based on a cloud data center, collecting historical supply chain big data and performing preprocessing; performing clustering processing and data dimension reduction; using an artificial intelligence algorithm, constructing a corresponding supply chain risk prediction model and a supply chain optimization model; collecting real-time supply chain data of a current supply chain node, obtaining a matching supply chain link, and performing data dimension reduction; using the matching supply chain risk prediction model, performing supply chain risk prediction to obtain a real-time supply chain risk prediction result; using the matching supply chain optimization model, performing matching supply chain optimization to obtain a real-time supply chain optimization strategy, and returning to the corresponding supply chain node to execute the real-time supply chain optimization strategy.The application solves the problems of difficult data collection and integration, poor risk prediction effect and inflexible response mechanism in the prior art.

Description

Technical Field

[0001] This invention belongs to the field of supply chain data analysis technology, specifically relating to a supply chain risk prediction method and system. Background Technology

[0002] In today's globalized business environment, supply chain management has become a key factor in corporate competitiveness. Supply chain stability directly impacts a company's production efficiency, cost control, and customer satisfaction. However, supply chains face various internal and external risks, which may stem from market changes, political instability, natural disasters, supplier issues, and other factors. Existing supply chain risk prediction technologies have the following limitations:

[0003] 1) Data collection and integration is difficult. Supply chain data is usually scattered across different systems and departments, and the data formats and standards are not uniform, which makes data collection and integration difficult.

[0004] 2) Poor risk prediction performance. Existing technologies rely heavily on qualitative risk assessment methods, such as expert scoring and the Delphi method. These methods are highly subjective, lack quantitative analysis, and are difficult to fully cover the diversified risks of the supply chain.

[0005] 3) Inflexible response mechanism: When risks occur, traditional supply chain risk management lacks a rapid response mechanism, resulting in delayed risk response measures. The supply chain lacks flexibility and adaptability, making it difficult to adjust according to real-time conditions. Summary of the Invention

[0006] To address the problems of data collection and integration difficulties, poor risk prediction performance, and inflexible response mechanisms in existing technologies, the present invention aims to provide a supply chain risk prediction method and system.

[0007] The technical solution adopted in this invention is as follows:

[0008] A supply chain risk prediction method includes the following steps:

[0009] Based on cloud data centers, heterogeneous historical supply chain big data is collected from different data sources and includes different supply chain links. The historical supply chain big data is preprocessed to obtain several homogeneous preprocessed historical supply chain data.

[0010] Clustering is performed on several preprocessed historical supply chain data to obtain several cluster centers and corresponding data clusters. Then, data dimensionality reduction is performed on several data clusters to obtain the dimensionality-reduced historical supply chain dataset and corresponding key indicator set for each supply chain link.

[0011] Based on the dimensionality-reduced historical supply chain dataset, artificial intelligence algorithms are used to construct corresponding supply chain risk prediction models and supply chain optimization models. After traversing the dimensionality-reduced historical supply chain datasets of all supply chain links, the supply chain risk prediction model and supply chain optimization model for each supply chain link are obtained.

[0012] Collect real-time supply chain data of the current supply chain node, obtain the corresponding matching supply chain links based on the Euclidean distance between the real-time supply chain data and several cluster centers, and perform data dimensionality reduction on the real-time supply chain data based on the matching key indicator set corresponding to the matching supply chain links to obtain dimensionality-reduced real-time supply chain data.

[0013] Extract the matching supply chain risk prediction model and matching supply chain optimization model corresponding to the matching supply chain links, and use the matching supply chain risk prediction model to predict supply chain risks based on the dimensionality-reduced real-time supply chain data to obtain real-time supply chain risk prediction results.

[0014] Based on the real-time supply chain risk prediction results, the matching supply chain optimization model is used to perform matching supply chain optimization, obtain the real-time supply chain optimization strategy, return the real-time supply chain optimization strategy to the corresponding supply chain node, execute the real-time supply chain optimization strategy, and return to the real-time supply chain data collection step.

[0015] Furthermore, based on a cloud data center, heterogeneous historical supply chain big data is collected from different data sources and includes historical supply chain links. This historical supply chain big data is then preprocessed to obtain several homogeneous preprocessed historical supply chain data sets, including the following steps:

[0016] Based on cloud data centers, heterogeneous historical supply chain big data is collected from different data sources and includes historical supply chain data from different supply chain links.

[0017] Data filtering was performed on historical supply chain big data to obtain a number of cleaned historical supply chain data.

[0018] Based on the data format of the cloud data center, a unified data model is constructed, and based on the unified data model, corresponding historical data mapping and transformation strategies are set for each data source;

[0019] Based on historical data mapping and transformation strategies, several heterogeneous cleaned historical supply chain data are mapped and transformed to obtain several homogeneous transformed historical supply chain data.

[0020] Furthermore, clustering is performed on several preprocessed historical supply chain data to obtain several cluster centers and corresponding data clusters. Dimensionality reduction is then performed on these data clusters to obtain the dimensionality-reduced historical supply chain dataset and corresponding key indicator set for each supply chain link. This process includes the following steps:

[0021] Using supply chain links as clustering targets and differences in data indicators as clustering criteria, the FCM clustering algorithm was used to cluster several pre-processed historical supply chain data to obtain several cluster centers.

[0022] Based on the Euclidean distance between several preprocessed historical supply chain data and each cluster center, the preprocessed historical supply chain data is divided into data clusters corresponding to each cluster center.

[0023] Each data cluster is converted into a data matrix, and the PCA method is used to reduce the dimensionality of the data matrix to obtain the dimensionality-reduced historical supply chain dataset and corresponding key indicator set for each supply chain link.

[0024] Furthermore, based on the dimensionality-reduced historical supply chain dataset, artificial intelligence algorithms are used to construct corresponding supply chain risk prediction models and supply chain optimization models. After traversing the dimensionality-reduced historical supply chain datasets of all supply chain links, the supply chain risk prediction model and supply chain optimization model for each supply chain link are obtained, including the following steps:

[0025] Based on the dimensionality-reduced historical supply chain dataset of the same supply chain link, a corresponding supply chain risk prediction model is constructed using deep learning algorithms, and several historical supply chain risk prediction results are generated.

[0026] Based on several historical supply chain risk prediction results of the same supply chain link, a corresponding supply chain optimization model is constructed using reinforcement learning algorithms.

[0027] By traversing the dimensionality-reduced historical supply chain datasets of all supply chain links, a supply chain risk prediction model and a supply chain optimization model are constructed for each supply chain link.

[0028] The supply chain links are used as search tags for supply chain risk prediction models and supply chain optimization models, and several supply chain risk prediction models and several supply chain optimization models with search tags are stored.

[0029] Furthermore, the supply chain risk prediction model is built based on the IFWA-DBN algorithm.

[0030] Furthermore, the supply chain optimization model is constructed based on the DQN algorithm.

[0031] Furthermore, real-time supply chain data of the current supply chain nodes is collected. Based on the Euclidean distance between the real-time supply chain data and several cluster centers, the corresponding matching supply chain links are obtained. Then, based on the matching key indicator set corresponding to the matching supply chain links, the real-time supply chain data is dimensionality reduced to obtain dimensionality-reduced real-time supply chain data. This includes the following steps:

[0032] Collect real-time supply chain data stored on the data server of the current supply chain node, and set corresponding real-time data mapping and transformation strategies according to the data format of the current supply chain node and the unified data model;

[0033] Based on the real-time data mapping and transformation strategy, the real-time supply chain data is mapped and transformed to obtain the transformed real-time supply chain data;

[0034] Obtain the Euclidean distance between the transformed real-time supply chain data and several cluster centers, and use the supply chain link of the cluster center with the closest Euclidean distance as the matching supply chain link;

[0035] Extract the set of key matching indicators corresponding to the matching supply chain links, and perform data dimensionality reduction on the transformed real-time supply chain data based on the set of key matching indicators to obtain dimensionality-reduced real-time supply chain data.

[0036] Furthermore, the matching supply chain risk prediction model and matching supply chain optimization model corresponding to the matching supply chain links are extracted. Based on the dimensionality-reduced real-time supply chain data, the matching supply chain risk prediction model is used to predict supply chain risks and obtain real-time supply chain risk prediction results. This includes the following steps:

[0037] Based on the matching supply chain links, search among several search tags to obtain the corresponding target search tags;

[0038] The supply chain risk prediction model and supply chain optimization model corresponding to the target search tags are used as the matching supply chain risk prediction model and matching supply chain optimization model;

[0039] Input the dimensionality-reduced real-time supply chain data into the matching supply chain risk prediction model, and use the matching supply chain risk prediction model to extract the real-time data features of the dimensionality-reduced real-time supply chain data.

[0040] Based on real-time data characteristics, supply chain risk prediction is performed to obtain real-time supply chain risk prediction results.

[0041] Furthermore, based on the real-time supply chain risk prediction results, a matching supply chain optimization model is used to perform matching supply chain optimization, resulting in a real-time supply chain optimization strategy. This strategy is then returned to the corresponding supply chain node for execution, and the real-time supply chain data collection step is completed, including the following steps:

[0042] Based on the real-time supply chain risk prediction results, update the model parameters of the matching supply chain optimization model to obtain the updated model parameters;

[0043] Based on the updated model parameters, the matching supply chain optimization model is used to perform matching supply chain optimization and obtain a real-time supply chain optimization strategy.

[0044] The real-time supply chain optimization strategy is returned to the corresponding supply chain node, and the decision support system of the supply chain node is used to execute the real-time supply chain optimization strategy and return the real-time supply chain data collection step.

[0045] A supply chain risk prediction system is provided for implementing a supply chain risk prediction method. The system includes a cloud data center and several supply chain nodes. The supply chain nodes are all connected to the cloud data center and include a data server and a decision support system.

[0046] The cloud data center includes a big data acquisition unit, a data processing unit, a model building unit, a supply chain segment classification unit, a supply chain risk prediction unit, and a supply chain optimization unit, which are connected in sequence. The big data acquisition unit and the supply chain segment classification unit are respectively connected to the data servers of several supply chain nodes, and the supply chain optimization unit is respectively connected to the decision support system of several supply chain nodes.

[0047] The beneficial effects of this invention are as follows:

[0048] This invention discloses a supply chain risk prediction method and system. Utilizing a cloud data center, it efficiently collects and processes heterogeneous historical supply chain big data from different data sources, achieving data unification and standardization, and reducing the difficulty of data collection and integration. Through clustering and dimensionality reduction techniques, this invention can extract key indicators from complex supply chain data, simplifying model construction complexity and improving prediction accuracy and effectiveness. The use of artificial intelligence algorithms to construct supply chain risk prediction and optimization models significantly improves prediction accuracy and the effectiveness of optimization strategies. By feeding back supply chain optimization strategies to supply chain nodes for execution, a closed-loop management mechanism is formed, ensuring the continuity and effectiveness of supply chain risk prediction and management. The construction of corresponding supply chain risk prediction and optimization models based on different supply chain links ensures the relevance and practicality of risk prediction and optimization strategies. Furthermore, a rapid response mechanism prevents delayed risk response measures when risks occur, improving the flexibility and adaptability of the supply chain and enabling optimization based on real-time conditions. Through cloud data centers and artificial intelligence algorithms, a high degree of integration of supply chain risk management tools and systems is achieved, breaking down information silos and improving overall management efficiency.

[0049] Other beneficial effects of the present invention will be further explained in the specific embodiments. Attached Figure Description

[0050] Figure 1 This is a flowchart of the supply chain risk prediction method in this invention.

[0051] Figure 2 This is a structural block diagram of the supply chain risk prediction system in this invention. Detailed Implementation

[0052] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0053] Example 1:

[0054] like Figure 1 As shown, this embodiment provides a supply chain risk prediction method, including the following steps:

[0055] S1: Based on a cloud data center, collect heterogeneous historical supply chain big data from different data sources and including different supply chain links, and preprocess the historical supply chain big data to obtain several homogeneous preprocessed historical supply chain data, including the following steps:

[0056] S1-1: Based on cloud data centers, collect heterogeneous historical supply chain big data from different data sources and including different supply chain links;

[0057] The supply chain includes planning, procurement, production, delivery, and transportation. Each stage involves different data indicators. The planning stage is the initial planning process for supply chain activities, involving data indicators related to materials, procurement, production, warehousing, distribution, sales, fulfillment, and inventory. Good planning can guide and monitor the supply chain execution process, ensuring the achievement of goals. The procurement stage is based on the procurement plan and includes data indicators such as price inquiry, supplier management, negotiation, contract signing, purchase orders, delivery, receipt, acceptance, and settlement. The production stage includes commodity production and order production. Commodity production is usually completed in factories or processing centers, manufacturing raw materials into finished products, and includes various commodity production data indicators. Order production involves the packaging and dispatch of customer orders from warehouses, distribution centers, or stores, and includes various order production data indicators. The delivery stage is the process of delivering products from upstream suppliers to downstream customers, including data indicators such as logistics and distribution process management and timeliness control. The distribution stage includes data indicators related to the process of goods moving from suppliers through manufacturers, wholesalers and logistics companies, retailers, etc., to finally reaching consumers. Therefore, the data structures of the supply chain vary greatly across different stages, making unified data analysis impossible.

[0058] S1-2: Data filtering is performed on historical supply chain big data to obtain a number of cleaned historical supply chain data;

[0059] Data filtering includes deduplication, error removal, and noise reduction to improve data quality and support subsequent data analysis and processing, thereby improving the speed and effectiveness of model training.

[0060] S1-3: Based on the data format of the cloud data center, construct a unified data model, and based on the unified data model, set corresponding historical data mapping and transformation strategies for each data source, including the following steps:

[0061] S1-3-1: Analyze historical supply chain data from different data sources to obtain several core data elements for each data source;

[0062] S1-3-2: Based on several core data elements from different data sources, define the data structure of a unified data model, such as tables, fields, and relationships;

[0063] S1-3-3: Based on the data structure of the unified data model, define the data relationships between core data elements, including one-to-one, one-to-many, and many-to-many relationships;

[0064] S1-3-4: Based on the core data elements of different data sources, the data structure of the unified data model, and the data relationships between the core data elements, set the data constraints for each core data element, such as data type, value range, default value, etc.

[0065] S1-3-5: Based on the core data elements of different data sources, the data structure of the unified data model, the data relationships between the core data elements, and the data constraints of each core data element, construct the corresponding unified data model;

[0066] S1-3-6: Based on the unified data model, set corresponding historical data mapping and transformation strategies for each data source;

[0067] Because different hardware systems and data formats are used at different nodes in the supply chain, the collected supply chain big data includes several heterogeneous data sets. These heterogeneous data sets pose obstacles to subsequent data analysis and processing. In this embodiment, the data mapping and transformation strategy includes a mapping strategy and a data transformation strategy. The mapping strategy is used to determine how to map heterogeneous data to a unified data model. The data transformation strategy is used to convert heterogeneous data into data in a unified data model based on a preset transformation algorithm or script. The data mapping and transformation strategy is used to handle data heterogeneity and solve inconsistencies in data format, data type, data semantics, etc., in different data sources.

[0068] S1-4: Based on the historical data mapping and transformation strategy, perform data format mapping and transformation on several heterogeneous cleaned historical supply chain data to obtain several homogeneous transformed historical supply chain data.

[0069] S2: Cluster the preprocessed historical supply chain data to obtain several cluster centers and corresponding data clusters. Then, perform dimensionality reduction on these data clusters to obtain the dimensionality-reduced historical supply chain dataset and corresponding key indicator set for each supply chain link. This includes the following steps:

[0070] S2-1: Using supply chain links as the clustering target and data indicator differences as the clustering standard, the Fuzzy C-means (FCM) clustering algorithm is used to cluster several preprocessed historical supply chain data to obtain several cluster centers. This includes the following steps:

[0071] S2-1-1: Select clustering parameters, including fuzzy factor The total number of cluster centers is calculated, and based on the clustering parameters, the Fuzzy C-means (FCM) clustering algorithm is used to cluster several preprocessed historical supply chain data to obtain several initial cluster centers.

[0072] S2-1-2: Based on the Euclidean distance between each preprocessed historical supply chain data and several initial cluster centers, set a corresponding fuzzy membership degree for each initial cluster center;

[0073] The formula is:

[0074]

[0075] In the formula, For the first i Preprocessed historical supply chain data and the first j Euclidean distance between cluster centers; For the first Preprocessed historical supply chain data; For the first j Cluster center; For data indication quantity; This is an indicator of the cluster center.

[0076] S2-1-3: Update the cluster centers based on the fuzzy membership degree to obtain several updated cluster centers;

[0077] The formula for updating fuzzy membership is:

[0078]

[0079] In the formula, For the first Preprocessed historical supply chain data; For data indication quantity; All are cluster center indicators; The total number of cluster centers; For the first i Preprocessed historical supply chain data to the next level j , k Distance to cluster centers; For the first i Preprocessed historical supply chain data belongs to the first The updated fuzzy membership of cluster centers;

[0080] The formula for updating cluster centers is:

[0081]

[0082] In the formula, For the first j Updated cluster centers; For fuzzy factors; For data indication quantity; The total number of data; This is an indicator of the cluster center. For the first Preprocessed historical supply chain data; For the first i Preprocessed historical supply chain data belongs to the first Fuzzy membership degree of cluster centers;

[0083] S2-1-4: Using the Lagrange multiplier method, the combined function is calculated to obtain the combined function value and the change value of the combined function. The formula is as follows:

[0084]

[0085] In the formula, For the first t , t The combined function value of the Lagrange multiplier method after -1 iterations; The corresponding change value; For the first i Feature parameters; t This is an indicator of the number of iterations. For the first i Preprocessed historical supply chain data belongs to the first Membership degree of cluster centers; For fuzzy factors; For data indication quantity; The total number of data; This is an indicator of the cluster center. The total number of cluster centers; For the first i Preprocessed historical supply chain data to the next level j Euclidean distance between cluster centers;

[0086] S2-1-5: If the value of the merge function is greater than the function threshold, or the change value of the merge function is greater than the change value threshold, then continue to update the cluster center; otherwise, take the current cluster center as the final cluster center.

[0087] S2-2: Based on the Euclidean distance between several preprocessed historical supply chain data and each cluster center, the preprocessed historical supply chain data is divided into data clusters corresponding to each cluster center;

[0088] S2-3: Convert each data cluster into a data matrix, and use Principal Component Analysis (PCA) to reduce the dimensionality of the data matrix, obtaining the dimensionality-reduced historical supply chain dataset and corresponding key indicator set for each supply chain link. This includes the following steps:

[0089] S2-3-1: Perform matrix transformation on the data cluster to obtain the corresponding data matrix. ,in, For the first p A preprocessed historical supply chain data row vector p For row vector indicators, n The total number of historical supply chain data after preprocessing; the initial row vector of the data matrix is ​​the historical supply chain data after preprocessing, and the initial column vector of the data matrix is ​​the historical supply chain indicator data;

[0090] S2-3-2: Standardize the data matrix to obtain the corresponding standardized data matrix;

[0091] The formula is:

[0092]

[0093] In the formula, This is a standardized data matrix; The mean of the data matrix; Let be the variance of the data matrix;

[0094] S2-3-3: Obtain the covariance matrix of the standardized data matrix, and obtain the corresponding candidate principal component matrix based on the standardized data matrix and the covariance matrix; the candidate row vector of the candidate principal component matrix is ​​the preprocessed historical supply chain data, and the candidate column vector of the candidate principal component matrix is ​​the historical supply chain candidate indicator data.

[0095] The formula is:

[0096]

[0097] In the formula, The covariance matrix of the standardized data matrix; For candidate principal component matrices; This is the transformation matrix; The unit eigenvector matrix; n This represents the total number of historical supply chain data after preprocessing.

[0098]

[0099] In the formula, For candidate principal component matrices; This is the transformation matrix; This is a standardized data matrix;

[0100] S2-3-4: Select the principal component matrix Several candidate column vectors with the top 90% cumulative contribution rate of Chinese difference As corresponding principal component column vectors, the dimensionality-reduced data matrix is ​​obtained by using these principal component column vectors. The key row vectors of the dimension-reduced data matrix are the historical supply chain data after dimension reduction, and the key column vectors of the dimension-reduced data matrix are the key indicator data of the historical supply chain.

[0101] The formula is:

[0102]

[0103] In the formula, Cumulative contribution rate of variance; No. One candidate principal component The variance; Indicator values ​​for alternative principal components; The total number of candidate principal components; The total number of principal components;

[0104] S2-3-5: Convert the dimensionality-reduced data matrix into several corresponding dimensionality-reduced historical supply chain datasets, and integrate the key supply chain indicators corresponding to the key column vectors to obtain the corresponding key indicator set;

[0105] S3: Based on the dimensionality-reduced historical supply chain dataset, use artificial intelligence algorithms to construct corresponding supply chain risk prediction models and supply chain optimization models. After traversing the dimensionality-reduced historical supply chain datasets of all supply chain links, obtain the supply chain risk prediction model and supply chain optimization model for each supply chain link, including the following steps:

[0106] S3-1: Based on the dimensionality-reduced historical supply chain dataset of the same supply chain link, use the Improved Fireworks Algorithm (IFWA) - Deep Belief Network (DBN) algorithm to construct the corresponding supply chain risk prediction model and generate several historical supply chain risk prediction results, including the following steps:

[0107] S3-1-1: Add labels to the cluster centers of the dimensionality-reduced historical supply chain dataset for the same supply chain link, and spread the added labels to the corresponding data clusters to obtain a supply chain risk prediction sample set with real supply chain risk labels.

[0108] S3-1-2: Divide the supply chain risk prediction sample set of the same supply chain link into a model training sample set and a model testing sample set in a 7:3 ratio;

[0109] S3-1-3: Using several unlabeled, dimensionality-reduced historical supply chain data, the DBN network is pre-trained to build an initial supply chain risk prediction model.

[0110] S3-1-4: The initial network parameters of the DBN network are used as the optimization target of the IFWA optimization algorithm. Based on the optimization target, the IFWA optimization algorithm is used to obtain the optimal initial network parameters of the DBN network. This includes the following steps:

[0111] S3-1-4-1: Use the initial network parameters of the DBN network as the optimization target of the IFWA optimization algorithm;

[0112] S3-1-4-3: Set the IFWA population parameters, maximum number of iterations, and fitness function of the IFWA optimization algorithm, and use the optimization target as the position of the IFWA individual in the IFWA population;

[0113] S3-1-4-3: Based on the IFWA population parameters, the IFWA population is initialized using the Circle chaotic mapping sequence to obtain the initialized IFWA population, as shown in the formula:

[0114]

[0115] In the formula, The initial IFWA individuals for the Circle chaotic mapping; The initial IFWA individuals are randomly generated; For indication;

[0116] S3-1-4-4: Calculate the fitness value of IFWA individuals in the initialized IFWA population based on the fitness function. The formula is:

[0117]

[0118] In the formula, For the initial IFWA individuals fitness value; The mean square error function for prediction; For predicted values ​​and actual values;

[0119] S3-1-4-5: Obtain the explosion radius and spark count of each initial firework in the initial firework set, using the following formula:

[0120]

[0121] In the formula, For the initial IFWA individuals The number of sparks; It is a constant; The maximum fitness value in the initialized IFWA population; For the initial IFWA individuals fitness value; It is an infinitesimal constant;

[0122]

[0123] In the formula, For the initial IFWA individuals Explosion radius; This is a constant used to adjust the explosion radius. The minimum fitness value in the initialized IFWA population;

[0124] S3-1-4-6: Based on the explosion radius and spark count of each initial IFWA individual in the initialized IFWA population, firework explosions are performed to obtain an updated IFWA population. The formula is:

[0125]

[0126] In the formula, For updated IFWA individuals; A random number between -1 and 1; For the initial IFWA individual;

[0127] S3-1-4-7: Use the Gaussian mutation algorithm to perform Gaussian mutation on the initialized IFWA population to generate a Gaussian-mutated IFWA population. The formula is as follows:

[0128]

[0129] In the formula, An IFWA individual with Gaussian mutation; These are random numbers distributed according to a Gaussian distribution with a mean and variance of 1.

[0130] S3-1-4-8: Using a dynamic back-learning algorithm, the initialized IFWA population is dynamically back-learned to generate a dynamically back-learned IFWA population. The formula is as follows:

[0131]

[0132] In the formula, For dynamically reversed IFWA individuals; This represents a positive candidate optimal solution; γ is the decreasing inertia coefficient. These are the maximum and minimum values ​​in the vector space, respectively.

[0133] S3-1-4-9: Calculate the fitness value of all IFWA individuals in the updated IFWA population, the Gaussian-mutated IFWA population, and the dynamically reversed IFWA population, and take the IFWA individual with the lowest fitness value as the best individual.

[0134] S3-1-4-10: If the number of iterations reaches the threshold or the fitness value of the best individual meets the requirements, then the optimal solution corresponding to the current best individual is output to obtain the optimal initial network parameters of the DBN network.

[0135] S3-1-5: Based on the optimal initial network parameters of the DBN network, optimize the network structure of the initial supply chain risk prediction model, input the model training sample set, perform optimization training, and obtain the optimized supply chain risk prediction model.

[0136] S3-1-6: Input the model test sample set, test the optimized supply chain risk prediction model, and obtain the model test accuracy.

[0137] S3-1-7: If the model test accuracy is greater than the model test accuracy threshold, output the optimal supply chain risk prediction model; otherwise, continue optimization training.

[0138] S3-2: Based on several historical supply chain risk prediction results for the same supply chain link, use the Deep Q Network (DQN) algorithm to construct the corresponding supply chain optimization model, including the following steps:

[0139] S3-2-1: Using the generation of supply chain optimization solutions as the simulation environment for the DQN algorithm, construct intelligent agents and experience replay pools;

[0140] S3-2-2: Define the state space of the DQN algorithm based on each type of supply chain status corresponding to the historical supply chain risk prediction results. The parameters of the state space correspond to each supply chain status. For example, the supply chain status in the planning stage includes procurement planning status, fulfillment planning status, inventory planning status, etc. The above planning status is used to characterize the status involved in the planning stage in the supply chain data analysis process.

[0141] S3-2-3: Define the action space of the DQN algorithm according to the actions required by the supply chain optimization strategy; for example, the actions in the planning stage include adjusting the start time of the procurement plan in the procurement plan state and adjusting the completion time of the fulfillment plan in the fulfillment plan state. Through the above actions, the various states of the supply chain can be adjusted, thereby achieving supply chain optimization.

[0142] S3-2-4: Based on the potential impact of each action in the action space, define the reward function of the DQN algorithm to evaluate the quality or impact of the action;

[0143] S3-2-5: Construct the input layer, several hidden layers, and output layer of the deep Q-network. Connect the input layer to the state space and the output layer to the action space.

[0144] S3-2-6: Based on the state space, action space, and reward function, and according to several historical supply chain risk prediction results, the deep Q network and agent are optimized and trained to build a supply chain optimization model, and the generated supply chain optimization experience is stored in the experience replay pool.

[0145] S3-3: Traverse the dimensionality-reduced historical supply chain dataset of all supply chain links, and construct a supply chain risk prediction model and a supply chain optimization model for each supply chain link.

[0146] As described in step S1-1, the supply chain includes planning, procurement, production, delivery, and transportation. Each link involves different data indicators and has a large difference in data structure, making it impossible to perform unified data analysis. Existing artificial intelligence models require strict definition of input features, which means that the same supply chain risk prediction model and supply chain optimization model cannot analyze supply chain data from different supply chain links. In this embodiment, corresponding supply chain risk prediction models and supply chain optimization models are trained for different supply chain links, which improves the adaptability and matching degree of data, and improves the efficiency and accuracy of data analysis.

[0147] S3-4: Use supply chain links as search tags for supply chain risk prediction models and supply chain optimization models, and store several supply chain risk prediction models and several supply chain optimization models with search tags set.

[0148] S4: Collect real-time supply chain data of the current supply chain node, obtain the corresponding matching supply chain links based on the Euclidean distance between the real-time supply chain data and several cluster centers, and perform data dimensionality reduction on the real-time supply chain data based on the matching key indicator set corresponding to the matching supply chain links to obtain dimensionality-reduced real-time supply chain data. This includes the following steps:

[0149] S4-1: Collect real-time supply chain data stored on the data server of the current supply chain node, and set corresponding real-time data mapping and transformation strategies according to the data format of the current supply chain node and the unified data model;

[0150] S4-2: Based on the real-time data mapping and transformation strategy, perform data mapping and transformation on the real-time supply chain data to obtain the transformed real-time supply chain data;

[0151] S4-3: Obtain the Euclidean distance between the transformed real-time supply chain data and several cluster centers, and use the supply chain link of the cluster center with the closest Euclidean distance as the matching supply chain link;

[0152] Pre-analyze supply chain data to obtain information on its links, providing support for subsequent model retrieval;

[0153] S4-4: Extract the set of key matching indicators corresponding to the matching supply chain links, and perform data dimensionality reduction on the transformed real-time supply chain data based on the set of key matching indicators to obtain dimensionality-reduced real-time supply chain data.

[0154] S5: Extract the matching supply chain risk prediction model and matching supply chain optimization model corresponding to the matching supply chain links, and use the matching supply chain risk prediction model to perform supply chain risk prediction based on the dimensionality-reduced real-time supply chain data to obtain the real-time supply chain risk prediction results. This includes the following steps:

[0155] S5-1: Based on the matching supply chain links, search among several search tags to obtain the corresponding target search tag;

[0156] S5-2: Use the supply chain risk prediction model and supply chain optimization model corresponding to the target search tag as the matching supply chain risk prediction model and matching supply chain optimization model;

[0157] S5-3: Input the dimensionality-reduced real-time supply chain data into the matching supply chain risk prediction model, and use the matching supply chain risk prediction model to extract the real-time data features of the dimensionality-reduced real-time supply chain data.

[0158] S5-4: Based on real-time data characteristics, perform supply chain risk prediction and obtain real-time supply chain risk prediction results;

[0159] S6: Based on the real-time supply chain risk prediction results, use the matching supply chain optimization model to perform matching supply chain optimization, obtain the real-time supply chain optimization strategy, return the real-time supply chain optimization strategy to the corresponding supply chain node, execute the real-time supply chain optimization strategy, and return the real-time supply chain data collection steps, including the following steps:

[0160] S6-1: Based on the real-time supply chain risk prediction results, update the model parameters of the matching supply chain optimization model to obtain the updated model parameters;

[0161] The updated model parameters include the updated state space. and updated action space ,in, For the updated number State value, For status indication, The total number of dimensions in the state space. For the updated number Action value, For action indication quantity; This represents the total number of dimensions in the action space.

[0162] S6-2: Based on the updated model parameters, use the matching supply chain optimization model to perform matching supply chain optimization and obtain a real-time supply chain optimization strategy, including the following steps:

[0163] S6-2-1: Update the state space As input to the automatic control model, a deep Q-network is used to generate an updated action space. The Q value for each possible action;

[0164] S6-2-2: Use the reward function to obtain the reward value of each possible action in the updated action space, and update the Q value of the possible action according to the reward value to obtain the updated Q value of the possible action.

[0165] The formula is:

[0166]

[0167] In the formula, For the updated state value and updated action values The corresponding updated Q value; State value and action value The corresponding predicted Q value; The learning rate; The highest predicted Q value;

[0168] S6-2-3: Repeat the above steps until the iteration threshold is reached. Use a greedy strategy to select the possible action corresponding to the highest updated Q value as the action to be executed, and output the action as the real-time supply chain optimization strategy. The real-time supply chain optimization strategy includes adjusting the start time of the procurement plan under the procurement plan state and adjusting the completion time of the fulfillment plan under the fulfillment plan state.

[0169] S6-3: Return the real-time supply chain optimization strategy to the corresponding supply chain node, and use the decision support system of the supply chain node to execute the real-time supply chain optimization strategy and return to the real-time supply chain data collection step; the decision support system adjusts the start time of the procurement plan in the procurement plan state and the completion time of the fulfillment plan in the fulfillment plan state according to the detailed strategies involved in the real-time supply chain optimization strategy, and generates corresponding real-time supply chain optimization decisions, such as increasing the procurement plan start time parameter and shortening the fulfillment plan completion time parameter. Through the above optimization, the cost of the procurement plan in the planning stage can be improved, and the node time of the fulfillment plan in the planning stage can be advanced, thereby solving the risk of high cost and low efficiency in the real-time supply chain risk prediction results.

[0170] Example 2:

[0171] like Figure 2 As shown, this embodiment provides a supply chain risk prediction system for implementing a supply chain risk prediction method. The system includes a cloud data center and several supply chain nodes. The supply chain nodes are all communicatively connected to the cloud data center, and each supply chain node includes a data server and a decision support system.

[0172] Data servers are used to store historical / real-time supply chain data, providing data support for supply chain risk prediction and supply chain optimization.

[0173] The decision support system is used to execute real-time supply chain optimization strategies, make decisions and adjustments to the supply chain, and affect the collection and storage of real-time supply chain data.

[0174] The cloud data center includes a big data acquisition unit, a data processing unit, a model building unit, a supply chain segment classification unit, a supply chain risk prediction unit, and a supply chain optimization unit connected in sequence. The big data acquisition unit and the supply chain segment classification unit are respectively connected to the data servers of several supply chain nodes, and the supply chain optimization unit is respectively connected to the decision support system of several supply chain nodes.

[0175] The big data acquisition unit is used to collect heterogeneous historical supply chain big data from different data sources and including different supply chain links, and to preprocess the historical supply chain big data to obtain several homogeneous preprocessed historical supply chain data.

[0176] The data processing unit is used to perform clustering on several pre-processed historical supply chain data to obtain several cluster centers and corresponding data clusters, and to perform data dimensionality reduction on several data clusters to obtain the dimensionality-reduced historical supply chain dataset and corresponding key indicator set for each supply chain link.

[0177] The model building unit is used to construct corresponding supply chain risk prediction models and supply chain optimization models based on the dimensionality-reduced historical supply chain dataset using artificial intelligence algorithms. After traversing the dimensionality-reduced historical supply chain dataset of all supply chain links, the supply chain risk prediction model and supply chain optimization model for each supply chain link are obtained.

[0178] The supply chain segment classification unit is used to collect real-time supply chain data of the current supply chain node. Based on the Euclidean distance between the real-time supply chain data and several cluster centers, the corresponding matching supply chain segments are obtained. Based on the matching key indicator set corresponding to the matching supply chain segments, the real-time supply chain data is dimensionality reduced to obtain dimensionality-reduced real-time supply chain data.

[0179] The supply chain risk prediction unit is used to extract the matching supply chain risk prediction model and matching supply chain optimization model corresponding to the matching supply chain links, and to use the matching supply chain risk prediction model to perform supply chain risk prediction based on the dimensionality-reduced real-time supply chain data, so as to obtain the real-time supply chain risk prediction results.

[0180] The supply chain optimization unit is used to perform matching supply chain optimization based on the real-time supply chain risk prediction results and the matching supply chain optimization model to obtain the real-time supply chain optimization strategy, and then return the real-time supply chain optimization strategy to the corresponding supply chain node.

[0181] This invention discloses a supply chain risk prediction method and system. Utilizing a cloud data center, it efficiently collects and processes heterogeneous historical supply chain big data from different data sources, achieving data unification and standardization, and reducing the difficulty of data collection and integration. Through clustering and dimensionality reduction techniques, this invention can extract key indicators from complex supply chain data, simplifying model construction complexity and improving prediction accuracy and effectiveness. The use of artificial intelligence algorithms to construct supply chain risk prediction and optimization models significantly improves prediction accuracy and the effectiveness of optimization strategies. By feeding back supply chain optimization strategies to supply chain nodes for execution, a closed-loop management mechanism is formed, ensuring the continuity and effectiveness of supply chain risk prediction and management. The construction of corresponding supply chain risk prediction and optimization models based on different supply chain links ensures the relevance and practicality of risk prediction and optimization strategies. Furthermore, a rapid response mechanism prevents delayed risk response measures when risks occur, improving the flexibility and adaptability of the supply chain and enabling optimization based on real-time conditions. Through cloud data centers and artificial intelligence algorithms, a high degree of integration of supply chain risk management tools and systems is achieved, breaking down information silos and improving overall management efficiency.

[0182] This invention is not limited to the optional embodiments described above, and anyone can derive other various forms of products based on the inspiration of this invention. The specific embodiments described above should not be construed as limiting the scope of protection of this invention; the scope of protection of this invention should be determined by the claims, and the specification can be used to interpret the claims.

Claims

1. A supply chain risk prediction method, characterized in that: Includes the following steps: Based on cloud data centers, heterogeneous historical supply chain big data is collected from different data sources and includes historical supply chain links. The historical supply chain big data is then preprocessed to obtain several homogeneous preprocessed historical supply chain data sets, including the following steps: Based on cloud data centers, heterogeneous historical supply chain big data is collected from different data sources and includes historical supply chain data from different supply chain links. Data filtering was performed on historical supply chain big data to obtain a number of cleaned historical supply chain data. Based on the data format of the cloud data center, a unified data model is constructed, and based on the unified data model, corresponding historical data mapping and transformation strategies are set for each data source, including the following steps: By analyzing historical supply chain data from different data sources, several core data elements of each data source are obtained. Based on several core data elements from different data sources, a unified data model data structure is defined, including the data structures for reports, fields, and relationships. Based on the data structure of the unified data model, define the data relationships between core data elements, including one-to-one, one-to-many, and many-to-many data relationships; Based on the core data elements of different data sources, the data structure of the unified data model, and the data relationships between the core data elements, set data constraints for each core data element, including data constraints on data type, value range, and default value. Based on the core data elements of different data sources, the data structure of the unified data model, the data relationships between the core data elements, and the data constraints of each core data element, construct the corresponding unified data model. Based on a unified data model, set corresponding historical data mapping and transformation strategies for each data source; Based on the historical data mapping and transformation strategy, several heterogeneous cleaned historical supply chain data are mapped and transformed to obtain several homogeneous transformed historical supply chain data. Clustering is performed on several preprocessed historical supply chain data to obtain several cluster centers and corresponding data clusters. Then, data dimensionality reduction is performed on several data clusters to obtain the dimensionality-reduced historical supply chain dataset and corresponding key indicator set for each supply chain link. Based on the dimensionality-reduced historical supply chain dataset, artificial intelligence algorithms are used to construct corresponding supply chain risk prediction models and supply chain optimization models. After traversing the dimensionality-reduced historical supply chain datasets of all supply chain links, the supply chain risk prediction model and supply chain optimization model for each supply chain link are obtained. Collect real-time supply chain data of the current supply chain node, obtain the corresponding matching supply chain links based on the Euclidean distance between the real-time supply chain data and several cluster centers, and perform data dimensionality reduction on the real-time supply chain data based on the matching key indicator set corresponding to the matching supply chain links to obtain dimensionality-reduced real-time supply chain data. Extract the matching supply chain risk prediction model and matching supply chain optimization model corresponding to the matching supply chain links, and use the matching supply chain risk prediction model to predict supply chain risks based on the dimensionality-reduced real-time supply chain data to obtain real-time supply chain risk prediction results. Based on the real-time supply chain risk prediction results, the matching supply chain optimization model is used to perform matching supply chain optimization, obtain the real-time supply chain optimization strategy, return the real-time supply chain optimization strategy to the corresponding supply chain node, execute the real-time supply chain optimization strategy, and return to the real-time supply chain data collection step.

2. The supply chain risk prediction method according to claim 1, characterized in that: Clustering is performed on several preprocessed historical supply chain data sets to obtain several cluster centers and corresponding data clusters. Then, dimensionality reduction is performed on several data clusters to obtain the dimensionality-reduced historical supply chain dataset and corresponding key indicator set for each supply chain link. The process includes the following steps: The FCM clustering algorithm was used to cluster several pre-processed historical supply chain data to obtain several cluster centers. Based on the Euclidean distance between several preprocessed historical supply chain data and each cluster center, the preprocessed historical supply chain data is divided into data clusters corresponding to each cluster center. Each data cluster is converted into a data matrix, and the PCA method is used to reduce the dimensionality of the data matrix to obtain the dimensionality-reduced historical supply chain dataset and corresponding key indicator set for each supply chain link.

3. The supply chain risk prediction method according to claim 1, characterized in that: Based on the dimensionality-reduced historical supply chain dataset, artificial intelligence algorithms are used to construct corresponding supply chain risk prediction models and supply chain optimization models. After traversing the dimensionality-reduced historical supply chain datasets of all supply chain links, the supply chain risk prediction model and supply chain optimization model for each supply chain link are obtained, including the following steps: Based on the dimensionality-reduced historical supply chain dataset of the same supply chain link, a corresponding supply chain risk prediction model is constructed using deep learning algorithms, and several historical supply chain risk prediction results are generated. Based on several historical supply chain risk prediction results of the same supply chain link, a corresponding supply chain optimization model is constructed using reinforcement learning algorithms. By traversing the dimensionality-reduced historical supply chain datasets of all supply chain links, a supply chain risk prediction model and a supply chain optimization model are constructed for each supply chain link. The supply chain links are used as search tags for supply chain risk prediction models and supply chain optimization models, and several supply chain risk prediction models and several supply chain optimization models with search tags are stored.

4. The supply chain risk prediction method according to claim 3, characterized in that: The supply chain risk prediction model described above is constructed based on the IFWA-DBN algorithm.

5. The supply chain risk prediction method according to claim 4, characterized in that: The supply chain optimization model described above is constructed based on the DQN algorithm.

6. The supply chain risk prediction method according to claim 1, characterized in that: The process involves collecting real-time supply chain data from current supply chain nodes, identifying corresponding matching supply chain links based on the Euclidean distance between the real-time supply chain data and several cluster centers, and then performing dimensionality reduction on the real-time supply chain data based on the set of matching key indicators corresponding to the matching supply chain links. This yields the dimensionality-reduced real-time supply chain data. The steps include: Collect real-time supply chain data stored on the data server of the current supply chain node, and set corresponding real-time data mapping and transformation strategies according to the data format of the current supply chain node and the unified data model; Based on the real-time data mapping and transformation strategy, the real-time supply chain data is mapped and transformed to obtain the transformed real-time supply chain data; Obtain the Euclidean distance between the transformed real-time supply chain data and several cluster centers, and use the supply chain link of the cluster center with the closest Euclidean distance as the matching supply chain link; Extract the set of key matching indicators corresponding to the matching supply chain links, and perform data dimensionality reduction on the transformed real-time supply chain data based on the set of key matching indicators to obtain dimensionality-reduced real-time supply chain data.

7. The supply chain risk prediction method according to claim 4, characterized in that: Extract the matching supply chain risk prediction model and matching supply chain optimization model corresponding to the matching supply chain links, and use the matching supply chain risk prediction model to predict supply chain risks based on the dimensionality-reduced real-time supply chain data to obtain real-time supply chain risk prediction results. The process includes the following steps: Based on the matching supply chain links, search among several search tags to obtain the corresponding target search tags; The supply chain risk prediction model and supply chain optimization model corresponding to the target search tags are used as the matching supply chain risk prediction model and matching supply chain optimization model; Input the dimensionality-reduced real-time supply chain data into the matching supply chain risk prediction model, and use the matching supply chain risk prediction model to extract the real-time data features of the dimensionality-reduced real-time supply chain data. Based on real-time data characteristics, supply chain risk prediction is performed to obtain real-time supply chain risk prediction results.

8. The supply chain risk prediction method according to claim 5, characterized in that: Based on the real-time supply chain risk prediction results, a matching supply chain optimization model is used to perform matching supply chain optimization, resulting in a real-time supply chain optimization strategy. This strategy is then returned to the corresponding supply chain nodes for execution, and the real-time supply chain data collection steps are completed, including the following steps: Based on the real-time supply chain risk prediction results, update the model parameters of the matching supply chain optimization model to obtain the updated model parameters; Based on the updated model parameters, the matching supply chain optimization model is used to perform matching supply chain optimization and obtain a real-time supply chain optimization strategy. The real-time supply chain optimization strategy is returned to the corresponding supply chain node, and the decision support system of the supply chain node is used to execute the real-time supply chain optimization strategy and return the real-time supply chain data collection step.

9. A supply chain risk prediction system for implementing the supply chain risk prediction method as described in any one of claims 1-8, characterized in that: The system includes a cloud data center and several supply chain nodes, all of which are communicatively connected to the cloud data center. Each supply chain node includes a data server and a decision support system. The cloud data center includes a big data acquisition unit, a data processing unit, a model building unit, a supply chain segment classification unit, a supply chain risk prediction unit, and a supply chain optimization unit connected in sequence. The big data acquisition unit and the supply chain segment classification unit are each connected to the data servers of several supply chain nodes, and the supply chain optimization unit is connected to the decision support systems of several supply chain nodes.

Images