Supply chain multi-dimensional fluctuation prediction and inventory optimization method based on artificial intelligence

Abstract

The invention provides a supply chain multi-dimensional fluctuation prediction and inventory optimization method based on artificial intelligence, and relates to the technical field of supply chain management, and the method comprises the steps: obtaining multi-dimensional data, carrying out the preprocessing of the multi-dimensional data, generating a standardized time series data set and a knowledge graph, carrying out the multi-scale decomposition through wavelet transformation, and extracting fluctuation features; and constructing a state matrix by combining a graph attention network and a multi-head self-attention mechanism, predicting a risk conduction path based on historical fluctuation pattern matching and a graph convolutional network, and generating a risk prevention and control scheme. According to the invention, the fluctuation characteristics and the risk conduction path in the supply chain system can be accurately identified, and the early warning accuracy and the response timeliness are improved.

Description

Technical Field

[0001] The present invention relates to supply chain management technology, and in particular to a supply chain multi-dimensional fluctuation prediction and inventory optimization method based on artificial intelligence. Background Art

[0002] As the global supply chain network becomes increasingly complex and multi-dimensional data surges, the volatility risk faced by the supply chain system continues to increase. Existing supply chain forecasting methods mainly use time series analysis and statistical modeling, which cannot effectively deal with data missing and abnormal problems, and have limitations in multi-dimensional data fusion, resulting in insufficient extraction of volatility features, and the forecast results are difficult to reflect the actual volatility of the supply chain system.

[0003] Existing supply chain fluctuation prediction methods often only focus on single-dimensional data analysis, ignoring the mutual influence and correlation between multiple elements in the supply chain system, resulting in insufficient accuracy and reliability of the prediction results. This single-dimensional analysis method is difficult to fully reflect the overall operating status of the supply chain system and cannot effectively identify potential risk factors. Most of the current prediction models use static data processing methods, fail to fully consider the dynamic evolution characteristics of the supply chain system, and have weak early warning capabilities for emergencies and abnormal fluctuations. This static analysis method is difficult to capture changes in the market environment in a timely manner, resulting in the lag of early warning signals. Traditional inventory optimization methods lack in-depth analysis of the supply chain network structure and fail to effectively use advanced technologies such as knowledge graphs to mine implicit knowledge and empirical rules in the supply chain. This makes the formulation of risk prevention and control plans lack of systematicity and foresight, and it is difficult to achieve accurate inventory control and risk management.

[0004] Therefore, there is an urgent need for a supply chain fluctuation prediction and inventory optimization method that can integrate multi-dimensional data and make full use of artificial intelligence technology and domain knowledge, so as to accurately predict the fluctuation trend of the supply chain system and accurately identify the risk transmission path, thereby improving the risk prevention and control capabilities of the supply chain system. Summary of the invention

[0005] The embodiment of the present invention provides an artificial intelligence supply chain multi-dimensional fluctuation prediction and inventory optimization method, which can solve the problems in the prior art.

[0006] According to a first aspect of the embodiments of the present invention, Provides an artificial intelligence supply chain multi-dimensional fluctuation prediction and inventory optimization method, including: Acquire and preprocess multi-dimensional data in the supply chain system, use piecewise linear interpolation to supplement missing data, generate standardized time series data sets, and build a supply chain knowledge graph based on the standardized time series data sets; Based on the standardized time series data set, the wavelet transform method is used to decompose the data at multiple scales, extract the trend component, periodic component and random component in the time series data, and calculate the node fluctuation characteristic index value; The node fluctuation characteristic index value is used as input, and the feature aggregation is performed through the graph attention network enhanced by the knowledge graph to obtain the fluctuation characteristic vector. The correlation strength between different nodes is calculated based on the multi-head self-attention mechanism, and a state matrix reflecting the overall fluctuation trend of the supply chain is constructed based on the fluctuation characteristic vector and the correlation strength. Match the historical fluctuation pattern with the current state matrix in time series and output the risk probability distribution; The risk probability distribution is input into the graph convolutional network, and feature propagation is performed based on the supply chain knowledge graph. The local structural feature matrix in the supply chain network is extracted through multi-layer spatial convolution operations. Combined with the expert rules in the supply chain knowledge graph, the key path of risk transmission is identified, and a risk warning signal containing the transmission probability is generated. According to the risk warning signal, the preset emergency response strategy library is called to generate the corresponding risk prevention and control plan.

[0007] In an optional embodiment, Acquire and preprocess multi-dimensional data in the supply chain system, use piecewise linear interpolation to supplement missing data, generate standardized time series data sets, and build a supply chain knowledge graph based on the standardized time series data sets, including: Acquire multi-dimensional data in the supply chain system, including supplier production data, logistics and transportation data, warehouse inventory data, and market demand data; Establish a feature distribution matrix according to the multi-dimensional data, obtain a density feature vector using a local density calculation method, calculate a local deviation value of a data point based on the density feature vector, and compare it with an adaptive threshold to identify the first round of abnormal data, and record the location information of the first round of abnormal data in an abnormal index list; Extract the time series data segments around the first round of abnormal data according to the abnormal index list, calculate the change trend coefficients between the data points in the time series data segments, construct a data change feature matrix, use the data change feature matrix to perform secondary verification on the first round of abnormal data, and generate a final abnormal data set; Based on the location information of the final abnormal data set, a data interval to be supplemented is divided in the multi-dimensional data, the numerical difference and time span of the data points at both ends of the data interval to be supplemented are calculated, and the supplementary data weight is adaptively generated according to the numerical difference and time span, and the supplementary data weight is used to perform piecewise linear interpolation calculation to obtain a supplementary data set; Filling the supplementary data set into the multi-dimensional data, normalizing the filled data to generate a standardized data matrix; Based on the standardized data matrix, supply chain entity features are extracted, the multidimensional correlation between entities is calculated to obtain a correlation strength matrix, and a supply chain knowledge graph is constructed according to the correlation strength matrix, wherein nodes represent supply chain entities and the weights of connecting edges are determined by the correlation strength matrix.

[0008] In an optional embodiment, Based on the standardized time series data set, the wavelet transform method is used to decompose the data at multiple scales, extract the trend component, periodic component and random component in the time series data, and calculate the node fluctuation characteristic index values ​​including: Calculate the energy concentration and information entropy of different wavelet basis functions based on the standardized supply chain time series data set, construct a multi-objective optimization function according to the energy concentration and information entropy, and obtain the optimal wavelet basis function by solving the multi-objective optimization function; The standardized supply chain time series data set is subjected to an adaptive multi-level decomposition, and the scale coefficient and wavelet coefficient are calculated at each decomposition level. The trend component is reconstructed according to the scale coefficient, and the periodic component and random component are reconstructed according to the wavelet coefficient. The energy distribution characteristics of the trend component, periodic component and random component at different scales are used as fluctuation characteristic indicator values.

[0009] In an optional embodiment, Based on the standardized supply chain time series data set, the energy concentration and information entropy of different wavelet basis functions are calculated, and a multi-objective optimization function is constructed according to the energy concentration and information entropy. The optimal wavelet basis function is obtained by solving the multi-objective optimization function, including: Constructing a wavelet basis function candidate set, and using each wavelet basis function in the wavelet basis function candidate set to perform wavelet transform on the supply chain time series data in the standardized supply chain time series data set to obtain corresponding wavelet coefficients; Calculating the energy value of each wavelet basis function based on the wavelet coefficient and performing normalization processing to obtain an energy distribution probability, and calculating the energy concentration of each wavelet basis function according to the energy distribution probability; Calculating the information entropy of each wavelet basis function based on the energy distribution probability, and obtaining the relative entropy characteristics of each wavelet basis function using the ratio of the information entropy to the theoretical maximum entropy; Reconstructing a signal using the wavelet coefficients and corresponding wavelet basis functions to obtain a reconstructed signal, and calculating a reconstruction error between the reconstructed signal and the supply chain time series data; The energy concentration, relative entropy characteristics and reconstruction error are weightedly combined to construct a multi-objective optimization function, and the multi-objective optimization function is iteratively optimized using the gradient descent method. When the adjacent iteration difference of the multi-objective optimization function is less than a preset convergence threshold and the function variance meets the stability requirement, the optimal wavelet basis function is selected from the wavelet basis function candidate set.

[0010] In an optional embodiment, The node fluctuation characteristic index value is used as input, and the fluctuation characteristic vector is obtained by feature aggregation through the graph attention network enhanced by the knowledge graph. The correlation strength between different nodes is calculated based on the multi-head self-attention mechanism. The state matrix reflecting the overall fluctuation trend of the supply chain is constructed based on the fluctuation characteristic vector and the correlation strength, including: Obtaining a fluctuation characteristic index value of a supply chain node, embedding the fluctuation characteristic index value and position coding information to obtain a node representation vector; Extracting node association information from the supply chain knowledge graph, building a knowledge-enhanced graph attention network based on the node association information, the knowledge-enhanced graph attention network fuses the node association information with the node representation vector to calculate the attention weights between nodes, and performing feature aggregation on the node representation vector according to the attention weights to generate a node-level fluctuation feature vector; The fluctuation feature vector is input into a multi-head self-attention network, and the multi-head self-attention network performs feature transformation on the fluctuation feature vector through multiple sets of feature mapping matrices to obtain a query vector, a key vector and a value vector, calculates the association strength between different nodes based on the query vector and the key vector, and performs weighted fusion on the value vector using the association strength to generate an association feature vector of the node; A node state vector is constructed based on the fluctuation eigenvector and the associated eigenvector, and the node state vectors are combined to form a state matrix. Multi-scale consistency constraints and temporal consistency constraints are imposed on the state matrix for optimization to obtain a state matrix that reflects the overall fluctuation trend of the supply chain.

[0011] In an optional embodiment, Match the historical fluctuation pattern with the current state matrix in time series, and output the risk probability distribution including: The historical fluctuation data is split into a trend component, a period component and a random component by data decomposition, and the trend component, the period component and the random component are respectively calculated to obtain trend features, period features and random features, so as to construct a historical fluctuation pattern feature set; The feature weight is determined by calculating the information entropy of each feature in the historical fluctuation pattern feature set, and the historical fluctuation pattern feature set is weightedly fused using the feature weight to obtain a historical fluctuation pattern matrix; Construct a multi-layer progressive similarity calculation network, calculate the similarity between the current state matrix and the historical fluctuation pattern matrix in each layer of the network, and use the similarity result of each layer as the calculation weight of the next layer to obtain the progressive time series matching weight between layers; Adaptively adjust the time window size according to the distribution characteristics of the timing matching weight, resample the current state matrix based on the time window size, and perform multi-layer progressive similarity calculation at different scales, and stop the calculation when the difference between the timing matching weights obtained by two adjacent calculations is less than a preset weight threshold; The time series matching weights at different scales are nonlinearly combined to obtain a comprehensive matching score, which is used to perform probability mapping on the risk indicators in the historical volatility pattern matrix to output the risk probability distribution.

[0012] In an optional embodiment, The risk probability distribution is input into the graph convolutional network, and feature propagation is performed based on the supply chain knowledge graph. The local structural feature matrix in the supply chain network is extracted through multi-layer spatial convolution operations. Combined with the expert rules in the supply chain knowledge graph, the key path of risk transmission is identified, and risk warning signals containing transmission probability are generated, including: The risk probability distribution is input into the graph convolutional network, and an adjacency matrix is ​​constructed based on the entity relationship in the supply chain knowledge graph. The adjacency matrix uses bidirectional adaptive weights to dynamically adjust the node relationship. Multi-layer spatial convolution operations are performed on the node features through the adjacency matrix. Each layer of convolution introduces a residual structure and a skip-layer connection with a gating mechanism. The local structure feature matrix is ​​obtained through weight matrix and nonlinear activation function processing; The structural similarity between nodes is calculated using the local structural feature matrix, the importance of historical conduction rules is decayed based on the time series decay function, and the structural similarity and the decayed expert rule score are combined through dynamic weights to obtain the rule confidence; Based on the local structural feature matrix, a multi-hop transmission link between nodes is established, a transmission path set is constructed from the risk source node to the target node, and the risk diffusion coefficient of each transmission path is calculated. The risk diffusion coefficient is calculated by the risk probability size of the nodes on the path, the structural similarity between the nodes, and the rule confidence. The transmission path with the largest risk diffusion coefficient is selected as the key risk transmission path; For adjacent nodes on the critical path of risk conduction, the conduction strength is calculated by combining the temporal conduction mode and the spatial conduction mode between the nodes, and the multiplication value of the conduction strength and the risk probability distribution of the node is used as the path conduction probability to generate a risk warning signal including the critical path of risk conduction and the path conduction probability.

[0013] In an optional embodiment, For the adjacent nodes on the critical risk transmission path, the transmission intensity is calculated by combining the temporal transmission mode and the spatial transmission mode between the nodes, including: For adjacent nodes on the critical path of risk transmission, the historical fluctuation pattern of the adjacent nodes is extracted based on the sliding time window, and the historical fluctuation pattern is matched with the current state matrix in a multi-dimensional time series through fluctuation frequency matching, fluctuation amplitude matching and fluctuation trend matching. The fluctuation mutation detection is introduced to obtain the fluctuation mutation coefficient, and the multi-dimensional time series matching result and the fluctuation mutation coefficient are weightedly combined to obtain the time series transmission characteristics; Based on the current state matrix, the business association strength, capital flow strength and information transmission strength of the adjacent nodes are extracted, and the conduction impedance coefficient is calculated in combination with the time series conduction characteristics. The conduction impedance coefficient represents the risk resistance of the node. The business association strength, capital flow strength and information transmission strength are corrected by the conduction impedance coefficient to obtain the spatial conduction characteristics; Adaptively weighting and combining the temporal conduction feature and the spatial conduction feature to obtain a basic conduction strength, calculating a conduction threshold value based on the basic conduction strength, and calculating a cascade influence coefficient when the basic conduction strength exceeds the conduction threshold value; The basic conduction strength and the cascade influence coefficient are combined to obtain the conduction strength between adjacent nodes.

[0014] According to a second aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0015] According to a third aspect of the embodiments of the present invention, A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0016] In this embodiment, by constructing a supply chain knowledge graph and combining the graph attention network and multi-head self-attention mechanism, the multi-dimensional fluctuation characteristics and complex correlations between nodes in the supply chain system can be effectively captured, improving the accuracy and interpretability of fluctuation prediction. Wavelet transform is used for multi-scale decomposition, and local structural features are extracted through graph convolutional networks. In combination with expert rules, risk transmission paths can be identified, which can timely detect potential risks and track their propagation paths, effectively improving the risk prevention and control capabilities of the supply chain system. A preset emergency response strategy library has been established, which can quickly generate corresponding prevention and control plans based on risk warning signals, improve the emergency response efficiency of the supply chain system for emergencies, reduce inventory management costs, and enhance the overall resilience of the supply chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of a method for multi-dimensional fluctuation prediction and inventory optimization of a supply chain based on artificial intelligence according to an embodiment of the present invention; Figure 2 A schematic diagram of constructing a supply chain knowledge graph according to an embodiment of the present invention; Figure 3 This is a distribution diagram of the fluctuation trend prediction accuracy of an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0020] Figure 1 Schematic diagram of the process of the artificial intelligence supply chain multi-dimensional fluctuation prediction and inventory optimization method according to the embodiment of the present invention. Figure 1 As shown, the method includes: S101. Obtain and preprocess multi-dimensional data in the supply chain system, use piecewise linear interpolation to supplement missing data, generate a standardized time series data set, and build a supply chain knowledge graph based on the standardized time series data set; Based on the standardized time series data set, the wavelet transform method is used to decompose the data at multiple scales, extract the trend component, periodic component and random component in the time series data, and calculate the node fluctuation characteristic index value; The node fluctuation characteristic index value is used as input, and the feature aggregation is performed through the graph attention network enhanced by the knowledge graph to obtain the fluctuation characteristic vector. The correlation strength between different nodes is calculated based on the multi-head self-attention mechanism, and a state matrix reflecting the overall fluctuation trend of the supply chain is constructed based on the fluctuation characteristic vector and the correlation strength. Match the historical fluctuation pattern with the current state matrix in time series and output the risk probability distribution; The risk probability distribution is input into the graph convolutional network, and feature propagation is performed based on the supply chain knowledge graph. The local structural feature matrix in the supply chain network is extracted through multi-layer spatial convolution operations. Combined with the expert rules in the supply chain knowledge graph, the key path of risk transmission is identified, and a risk warning signal containing the transmission probability is generated. According to the risk warning signal, the preset emergency response strategy library is called to generate the corresponding risk prevention and control plan.

[0021] Among them, the emergency response strategy library is a set of preset risk response plans, which contains response measures for different supply chain risks, such as inventory adjustment, supplier replacement, logistics optimization, etc. When the system detects a risk, it can automatically call the most appropriate strategy from the emergency response strategy library and provide it to decision makers for reference so that they can take corresponding risk control measures. For example, in the case of a supplier's delivery delay, the system can select "purchase from alternative suppliers" or "adjust production plans" from the strategy library as a response strategy.

[0022] In an optional implementation, multi-dimensional data in the supply chain system is obtained and preprocessed, missing data is supplemented by a piecewise linear interpolation method, a standardized time series data set is generated, and a supply chain knowledge graph is constructed based on the standardized time series data set, including: Acquire multi-dimensional data in the supply chain system, including supplier production data, logistics and transportation data, warehouse inventory data, and market demand data; Establish a feature distribution matrix according to the multi-dimensional data, obtain a density feature vector using a local density calculation method, calculate a local deviation value of a data point based on the density feature vector, and compare it with an adaptive threshold to identify the first round of abnormal data, and record the location information of the first round of abnormal data in an abnormal index list; Extract the time series data segments around the first round of abnormal data according to the abnormal index list, calculate the change trend coefficients between the data points in the time series data segments, construct a data change feature matrix, use the data change feature matrix to perform secondary verification on the first round of abnormal data, and generate a final abnormal data set; Based on the location information of the final abnormal data set, a data interval to be supplemented is divided in the multi-dimensional data, the numerical difference and time span of the data points at both ends of the data interval to be supplemented are calculated, and the supplementary data weight is adaptively generated according to the numerical difference and time span, and the supplementary data weight is used to perform piecewise linear interpolation calculation to obtain a supplementary data set; Filling the supplementary data set into the multi-dimensional data, normalizing the filled data to generate a standardized data matrix; Based on the standardized data matrix, supply chain entity features are extracted, the multidimensional correlation between entities is calculated to obtain a correlation strength matrix, and a supply chain knowledge graph is constructed according to the correlation strength matrix, wherein nodes represent supply chain entities and the weights of connecting edges are determined by the correlation strength matrix.

[0023] Exemplarily, to obtain multi-dimensional data in a supply chain system, it is necessary to collect supply chain related information from different data sources. The multi-dimensional data includes supplier production data, logistics and transportation data, warehousing and inventory data, and market demand data. Supplier production data involves information such as production progress, capacity utilization, and raw material inventory, which is used to evaluate the supplier's production capacity and supply stability. Logistics and transportation data includes information such as transportation time, transportation cost, route planning, and vehicle scheduling, which reflects the efficiency and cost of material flow in the supply chain. Warehousing and inventory data covers inventory levels, inventory turnover, safety stock, etc., which helps to judge the rationality of supply chain inventory management. Market demand data involves information such as sales forecasts, customer order volumes, and market trends, providing market guidance for supply chain decisions.

[0024] A feature distribution matrix is ​​established based on the multi-dimensional data. The feature distribution matrix is ​​used to describe the statistical characteristics of the data, including eigenvalues ​​such as mean, variance, skewness, and kurtosis, for subsequent analysis. The density feature vector is obtained by using a local density calculation method. This method measures the relative density of data points in the overall data by calculating the distribution density of data points in a local area. The local deviation value of the data point is calculated based on the density feature vector. The local deviation value reflects the degree of deviation of the data point relative to the adjacent data and is used to identify abnormal data. When the local deviation value exceeds the set adaptive threshold, the data point is determined to be abnormal data. The adaptive threshold is dynamically adjusted according to the overall distribution characteristics of the data to ensure the accuracy of anomaly detection. Data points determined to be abnormal are recorded in the abnormal index list.

[0025] Extract the time series data segments around the first round of abnormal data according to the abnormal index list. The time series data segment refers to a set of data points containing multiple time points before and after the abnormal data. The purpose is to further verify the rationality of the abnormal data by observing the change trend of the data. Calculate the change trend coefficient between the data points in the time series data segment. This coefficient is used to measure the change amplitude and direction of adjacent data points, so as to identify whether the abnormal data conforms to the overall trend. Construct a data change feature matrix, which is used to comprehensively describe the change pattern of the data, including data change rate, change amplitude, trend direction and other characteristics. Use the data change feature matrix to perform a second verification of the first round of abnormal data. By comparing the historical data pattern and the current data trend, further screen out the real abnormal data and generate the final abnormal data set.

[0026] Based on the location information of the final abnormal data set, the data interval to be supplemented is divided in the multi-dimensional data. The data interval to be supplemented refers to a continuous time period containing abnormal data, and its two ends are bounded by normal data points. The numerical difference and time span of the data points at both ends of the data interval to be supplemented are calculated. The numerical difference represents the amplitude of change of the data in the time dimension, and the time span represents the interval of the abnormal data on the time axis. The supplementary data weight is adaptively generated according to the numerical difference and time span. The supplementary data weight is used to adjust the calculation result of the supplementary data to make it conform to the overall change trend of the data. The supplementary data weight is used to perform piecewise linear interpolation calculation. Piecewise linear interpolation is a data filling method that obtains a supplementary data set by connecting adjacent normal data points and performing interpolation estimation based on weights.

[0027] The supplementary data set is filled into the multi-dimensional data, and the filled data is normalized to generate a standardized data matrix. The purpose of normalization is to eliminate the numerical differences between different data dimensions so that the data can be calculated and compared on the same scale. Common normalization methods include minimum-maximum normalization, mean normalization, etc. The standardized data matrix is ​​a structured data format that has been normalized to facilitate subsequent feature extraction and analysis.

[0028] Supply chain entity characteristics are extracted based on the standardized data matrix. Supply chain entity characteristics include information such as supplier supply capacity, logistics distribution efficiency, inventory management level, market demand fluctuations, etc. The multi-dimensional correlation between entities is calculated to obtain an association strength matrix, which is used to quantify the degree of mutual influence between different supply chain entities. When calculating the correlation strength, multiple factors such as geographical location, transaction frequency, and supply chain partnership can be considered. A supply chain knowledge graph is constructed based on the association strength matrix, in which nodes represent supply chain entities, such as suppliers, warehouses, logistics companies, retailers, etc., and the weights of connecting edges are determined by the association strength matrix, indicating the degree of dependence between entities. The supply chain knowledge graph can intuitively describe the supply chain structure, which is convenient for subsequent risk assessment and optimization decisions.

[0029] like Figure 2As shown in the figure, the process of building a supply chain knowledge graph based on multi-dimensional data is demonstrated. Among them, supplier A is the core node, connected to logistics provider B through a transportation relationship, and connected to warehouse C through a storage relationship, forming the basic business network of the supply chain. Nodes D and E are connected to logistics provider B through supply and collaboration relationships, and nodes F and G are connected to warehouse C through scheduling and distribution relationships, forming a complete supply chain network structure. The weights of the connection edges between nodes are determined by the association strength matrix, which calculates the multi-dimensional association between entities based on the standardized data matrix. In this way, the knowledge graph not only describes the topological structure between supply chain entities, but also quantifies the degree of business association between entities, providing a basis for subsequent risk transmission analysis.

[0030] In this embodiment, the data quality and integrity are improved through anomaly detection and supplementation of multi-dimensional data, providing a reliable data basis for subsequent analysis. The accuracy of anomaly detection is improved by adopting the dual verification method of local density and change trend. The supplementary method of piecewise linear interpolation combined with adaptive weights ensures the continuity and rationality of the data. Standardization eliminates the dimensional differences of data of different dimensions and makes the data comparable. The multi-dimensional correlation calculation comprehensively depicts the complex relationship between supply chain entities, and the constructed knowledge graph intuitively shows the supply chain network structure. The design of adaptive thresholds and weights improves the robustness of the method, enabling it to adapt to the dynamic changes of data distribution. The construction of the knowledge graph provides strong support for supply chain optimization and decision-making. Compared with the traditional simple mean filling or interpolation method, it can better restore the time series characteristics of the data. By constructing a supply chain knowledge graph, the correlation between each link of the supply chain is quantified, so that supply chain risk assessment and optimization can be based on more comprehensive data support, improving the scientificity and feasibility of supply chain management decisions.

[0031] In an optional implementation, based on a standardized time series data set, a wavelet transform method is used to perform multi-scale decomposition on the data, and trend components, periodic components, and random components in the time series data are extracted. The node fluctuation characteristic index value is calculated, including: Calculate the energy concentration and information entropy of different wavelet basis functions based on the standardized supply chain time series data set, construct a multi-objective optimization function according to the energy concentration and information entropy, and obtain the optimal wavelet basis function by solving the multi-objective optimization function; The standardized supply chain time series data set is subjected to an adaptive multi-level decomposition, and the scale coefficient and wavelet coefficient are calculated at each decomposition level. The trend component is reconstructed according to the scale coefficient, and the periodic component and random component are reconstructed according to the wavelet coefficient. The energy distribution characteristics of the trend component, periodic component and random component at different scales are used as fluctuation characteristic indicator values.

[0032] For example, the supply chain time series data set is standardized by subtracting the mean and dividing by the standard deviation to make the data distribution more standardized for subsequent processing. For example, if the original data is [100, 120, 95, 130], after standardization, standardized values ​​such as [-0.5, 0.2, -0.8, 1.1] can be obtained.

[0033] Then, different types of wavelet basis functions, including Haar, Daubechies, Symlets, etc., are selected to decompose the standardized data. The energy concentration of each wavelet basis function is calculated by calculating the ratio of the sum of squares of coefficients to the total energy at different scales. At the same time, the information entropy is calculated to reflect the uniformity of the coefficient distribution. For example, the energy distribution of a wavelet basis function at three scales is [0.6, 0.3, 0.1], and the corresponding information entropy can be calculated to be 1.45.

[0034] A multi-objective optimization function is constructed, and the maximization of energy concentration and the minimization of information entropy are taken as optimization goals. The optimal wavelet basis function is solved by optimization methods such as genetic algorithms. For example, the optimization results of 10 different wavelet basis functions show that the db4 wavelet has the best comprehensive performance.

[0035] The optimal wavelet basis function is used to perform multi-level decomposition of the data. The scale coefficient and wavelet coefficient are obtained at each decomposition level. The scale coefficient reflects the low-frequency trend characteristics of the data, and the wavelet coefficient reflects the high-frequency detail characteristics. Taking the three-level decomposition as an example, three groups of scale coefficients [a3, a2, a1] and wavelet coefficients [d3, d2, d1] are obtained.

[0036] The trend component is reconstructed based on the scale coefficient to reflect the overall change trend of the data. The periodic component and random component are reconstructed based on wavelet coefficients at different levels to reflect the periodic fluctuation and random disturbance characteristics. For example, the trend component of a certain time series data after decomposition shows an upward trend, the periodic component shows a periodicity of about 30 days, and the random component fluctuates within a range of ±0.2.

[0037] Calculate the energy distribution characteristics of the trend component, period component and random component at different scales as fluctuation characteristic indicators, including the energy proportion and energy concentration of each component. For example, the energy proportion of the trend component is 0.7, the period component is 0.2, and the random component is 0.1, which reflects that the data is mainly affected by the trend.

[0038] In this embodiment, the multi-scale decomposition of time series data is achieved through wavelet transform, and the trend, cycle and random characteristics in the data are accurately extracted, providing a reliable basis for supply chain fluctuation analysis. The multi-objective optimization function is constructed by using energy concentration and information entropy to select the optimal wavelet basis function, which improves the decomposition effect and makes the decomposed components more physically meaningful. A fluctuation index system is constructed based on the energy distribution characteristics of components at different scales to comprehensively characterize the fluctuation characteristics of the supply chain and provide a quantitative basis for supply chain risk prevention and control.

[0039] In an optional implementation, energy concentration and information entropy of different wavelet basis functions are calculated based on a standardized supply chain time series data set, a multi-objective optimization function is constructed according to the energy concentration and information entropy, and the optimal wavelet basis function is obtained by solving the multi-objective optimization function, including: Constructing a wavelet basis function candidate set, and using each wavelet basis function in the wavelet basis function candidate set to perform wavelet transform on the supply chain time series data in the standardized supply chain time series data set to obtain corresponding wavelet coefficients; Calculating the energy value of each wavelet basis function based on the wavelet coefficient and performing normalization processing to obtain an energy distribution probability, and calculating the energy concentration of each wavelet basis function according to the energy distribution probability; Calculating the information entropy of each wavelet basis function based on the energy distribution probability, and obtaining the relative entropy characteristics of each wavelet basis function using the ratio of the information entropy to the theoretical maximum entropy; Reconstructing a signal using the wavelet coefficients and corresponding wavelet basis functions to obtain a reconstructed signal, and calculating a reconstruction error between the reconstructed signal and the supply chain time series data; The energy concentration, relative entropy characteristics and reconstruction error are weightedly combined to construct a multi-objective optimization function, and the multi-objective optimization function is iteratively optimized using the gradient descent method. When the adjacent iteration difference of the multi-objective optimization function is less than a preset convergence threshold and the function variance meets the stability requirement, the optimal wavelet basis function is selected from the wavelet basis function candidate set.

[0040] Exemplarily, a set of candidate wavelet basis functions is constructed, wherein a wavelet basis function is a mathematical function used for signal processing, which can decompose the original signal into different frequency components for further analysis and processing. A set of candidate wavelet basis functions is selected from multiple known wavelet basis functions, and the standardized supply chain time series data is transformed using the candidate wavelet basis functions to obtain corresponding wavelet coefficients.

[0041] The energy value of the wavelet coefficient corresponding to each wavelet basis function is calculated and the energy value is normalized. The purpose of normalization is to ensure that the wavelet energies of different scales can be compared within the same reference range. Based on the normalized energy value, the distribution probability of energy on different wavelet basis functions is calculated to reflect the adaptability of different wavelet basis functions to time series data.

[0042] The energy concentration of each wavelet basis function is calculated based on the energy distribution probability. The energy concentration is used to measure the concentrated performance of a wavelet basis function in signal analysis. The higher the concentration, the stronger the signal decomposition ability of the wavelet basis function. At the same time, the information entropy corresponding to each wavelet basis function is calculated based on the energy distribution probability. The information entropy is used to measure the complexity of the signal. The information entropy of different wavelet basis functions can reflect their ability to extract the characteristics of supply chain time series data. Based on the ratio of the calculated information entropy value to the theoretical maximum entropy, the relative entropy feature of the wavelet basis function is determined. This feature is used to measure the degree to which the wavelet basis function maintains the characteristics of time series data.

[0043] The calculated wavelet coefficients and the corresponding wavelet basis functions are used to reconstruct the supply chain time series data to obtain the reconstructed signal. The reconstruction error between the reconstructed signal and the original supply chain time series data is calculated. The reconstruction error reflects the information loss of the wavelet basis function in the process of transforming and reconstructing the signal. The smaller the error, the higher the fitting degree of the wavelet basis function to the signal.

[0044] The energy concentration, relative entropy characteristics and reconstruction error are weighted and combined to construct a multi-objective optimization function, which is used to comprehensively evaluate the advantages and disadvantages of different wavelet basis functions in the analysis of supply chain time series data. The multi-objective optimization function is iteratively optimized using the gradient descent method. The gradient descent method is a commonly used optimization algorithm that calculates the function gradient and adjusts the parameters along the gradient direction so that the optimization objectives gradually converge. When the adjacent iteration difference of the multi-objective optimization function is less than the preset convergence threshold and the variance of the optimization function meets the stability requirements, the optimization process is considered to have reached a stable state, and the optimal wavelet basis function is selected from the candidate wavelet basis function set. The optimal wavelet basis function refers to the wavelet basis function with the best performance under the evaluation of the multi-objective optimization function, which can be used for further analysis and prediction of subsequent supply chain time series data.

[0045] The prior art usually uses a fixed wavelet basis function to decompose supply chain time series data, which is difficult to adapt to different data features, resulting in inaccurate feature extraction. In addition, the existing method does not fully consider the reconstruction error when reconstructing the signal, which affects the accuracy of data analysis. At the same time, a single objective strategy is often used in the optimization process, which fails to take into account energy concentration, information entropy characteristics and reconstruction errors, limiting the optimization effect of the wavelet basis function. The present application uses a multi-objective optimization method to dynamically select the optimal wavelet basis function from the wavelet basis function candidate set to adapt to the characteristic changes of supply chain data. By calculating the energy distribution probability, it is ensured that the selected basis function can accurately capture the main features. The relative entropy feature is calculated to reduce the information loss during data conversion. The reconstruction error is further introduced to keep the wavelet basis function at a high fit during signal reconstruction, and the stability and accuracy of the selection are improved through gradient descent optimization iteration. Compared with the prior art, the scheme realizes the adaptive optimization of the wavelet basis function, makes the supply chain data decomposition more accurate, reduces the data reconstruction error, improves the reliability of supply chain prediction, and enhances the adaptability and generalization ability in complex environments.

[0046] In an optional implementation, the node fluctuation characteristic index value is used as input, and the fluctuation characteristic vector is obtained by feature aggregation through the graph attention network enhanced by the knowledge graph, and the correlation strength between different nodes is calculated based on the multi-head self-attention mechanism. The state matrix reflecting the overall fluctuation trend of the supply chain is constructed based on the fluctuation characteristic vector and the correlation strength, including: Obtaining a fluctuation characteristic index value of a supply chain node, embedding the fluctuation characteristic index value and position coding information to obtain a node representation vector; Extracting node association information from the supply chain knowledge graph, building a knowledge-enhanced graph attention network based on the node association information, the knowledge-enhanced graph attention network fuses the node association information with the node representation vector to calculate the attention weights between nodes, and performing feature aggregation on the node representation vector according to the attention weights to generate a node-level fluctuation feature vector; The fluctuation feature vector is input into a multi-head self-attention network, and the multi-head self-attention network performs feature transformation on the fluctuation feature vector through multiple sets of feature mapping matrices to obtain a query vector, a key vector and a value vector, calculates the association strength between different nodes based on the query vector and the key vector, and performs weighted fusion on the value vector using the association strength to generate an association feature vector of the node; A node state vector is constructed based on the fluctuation eigenvector and the associated eigenvector, and the node state vectors are combined to form a state matrix. Multi-scale consistency constraints and temporal consistency constraints are imposed on the state matrix for optimization to obtain a state matrix that reflects the overall fluctuation trend of the supply chain.

[0047] Exemplarily, the fluctuation characteristic index values ​​of the supply chain nodes are first obtained. After obtaining the fluctuation characteristic index values ​​of the supply chain nodes, these fluctuation characteristic index values ​​need to be feature embedded with the corresponding position information to generate a node representation vector. The fluctuation characteristic index value refers to a quantitative parameter used to measure the changes in the supply chain nodes at different time points, such as inventory levels, transportation delays, order volume fluctuations, etc. The position encoding information is used to represent the position of the node in the supply chain network so that the model can use the position information for learning.

[0048] The association information between nodes is extracted from the supply chain knowledge graph, and a knowledge-enhanced graph attention network is constructed based on the association information. In this network, the knowledge-enhanced graph attention network uses the prior information contained in the knowledge graph to calculate the attention weights between nodes. The attention weights reflect the relative importance between different nodes. The node representation vectors are feature aggregated through these weights to obtain the node-level fluctuation feature vector that can reflect the fluctuation of supply chain nodes.

[0049] The node-level fluctuation feature vector is input into the multi-head self-attention network, and the fluctuation feature vector is transformed using multiple sets of feature mapping matrices to generate query vectors, key vectors, and value vectors. The query vector, key vector, and value vector represent the feature representations of different nodes in the supply chain network, where the query vector is used to retrieve information about other nodes, the key vector is used to calculate similarity, and the value vector is used for feature fusion. The query vector and the key vector are used to calculate the association strength between different nodes, which is used to measure the closeness of the relationship between two nodes. The value vector is weighted and fused according to the calculated association strength, and finally the node's association feature vector is generated.

[0050] Combining the fluctuation eigenvector and the correlation eigenvector, a node state vector containing rich information is constructed, and the state vectors of multiple nodes are combined to form a state matrix. The state matrix is ​​used to represent the overall fluctuation trend of the supply chain. In order to optimize the state matrix so that it can more accurately reflect the fluctuation of the supply chain network, multi-scale consistency constraints and temporal consistency constraints are imposed on the state matrix. The multi-scale consistency constraint ensures that the fluctuation pattern at different time scales remains stable, and the temporal consistency constraint ensures that the change trend of the supply chain node is consistent with the actual business logic. These constraints ensure that the model maintains consistency at different time scales and can handle the continuity of node states at different time points. By optimizing these constraints, a state matrix that reflects the overall fluctuation trend of the supply chain can be obtained.

[0051] For example, in a fast-moving consumer goods supply chain management, you can obtain information such as inventory changes, transportation timeliness data, and order demand fluctuation data for each storage center. Input this data into the supply chain knowledge graph, combine it with the existing supply chain link association information in the knowledge graph, and use the graph attention network to calculate the influence relationship between storage centers. Subsequently, use the multi-head self-attention mechanism to calculate the strength of mutual correlation between different storage centers, identify key fluctuation points in the supply chain, and use the optimized state matrix to predict the risks of inventory shortages or logistics delays that may occur in the future, providing companies with reasonable response strategies.

[0052] like Figure 3 As shown, Figure 3 The distribution diagram of the prediction accuracy of the fluctuation trend of this embodiment shows the prediction performance under different fluctuation levels. The size of the bubble indicates the amount of data, and the larger the fluctuation coefficient, the more severe the supply chain fluctuation. Even in the case of high fluctuation (fluctuation coefficient 0.6), this solution can still maintain 86.1% prediction accuracy, while the accuracy of the traditional time series prediction method drops to 73.2%. This proves the robustness of this solution in dealing with complex fluctuation scenarios.

[0053] In this embodiment, by combining the fluctuation characteristic index, knowledge graph and graph attention network, the limitations of existing supply chain management technology in node dynamic fluctuation analysis are solved. The existing technology usually relies on static data or traditional statistical methods, lacks deep mining and dynamic updating of complex relationships between nodes, resulting in the inability of the analysis results to accurately reflect the real-time changes of the supply chain. In contrast, this solution dynamically captures the fluctuation characteristics of nodes through feature embedding and graph attention network, and uses the correlation information between nodes for deep fusion, which can more accurately reflect the interaction and influence between nodes. In addition, with the help of knowledge graph, the solution fully mines the correlation information between nodes and enhances the understanding of the supply chain structure, while traditional methods usually ignore these complex relationships. The multi-head self-attention network further optimizes the feature aggregation of nodes, can more accurately process the correlation features between nodes, and avoids the overly simple aggregation method of traditional methods. Compared with the existing technology, the improvement of this application is to combine multiple feature information with deep learning methods to improve the prediction accuracy of node fluctuation characteristics and the ability to assess supply chain risks. Through this intelligent fusion method, this solution can achieve more efficient and accurate dynamic monitoring and management of the supply chain.

[0054] In an optional implementation, the historical fluctuation pattern is matched with the current state matrix in time series, and the output risk probability distribution includes: The historical fluctuation data is split into a trend component, a period component and a random component by data decomposition, and the trend component, the period component and the random component are respectively calculated to obtain trend features, period features and random features, so as to construct a historical fluctuation pattern feature set; The feature weight is determined by calculating the information entropy of each feature in the historical fluctuation pattern feature set, and the historical fluctuation pattern feature set is weightedly fused using the feature weight to obtain a historical fluctuation pattern matrix; Construct a multi-layer progressive similarity calculation network, calculate the similarity between the current state matrix and the historical fluctuation pattern matrix in each layer of the network, and use the similarity result of each layer as the calculation weight of the next layer to obtain the progressive time series matching weight between layers; Adaptively adjust the time window size according to the distribution characteristics of the timing matching weight, resample the current state matrix based on the time window size, and perform multi-layer progressive similarity calculation at different scales, and stop the calculation when the difference between the timing matching weights obtained by two adjacent calculations is less than a preset weight threshold; The time series matching weights at different scales are nonlinearly combined to obtain a comprehensive matching score, which is used to perform probability mapping on the risk indicators in the historical volatility pattern matrix to output the risk probability distribution.

[0055] Exemplarily, historical fluctuation data is split by data decomposition to obtain trend component, periodic component and random component. These components reflect different changing trends in the data, the trend component indicates the direction of long-term changes, the periodic component indicates periodic fluctuations, and the random component indicates unpredictable noise or random fluctuations. The components obtained by splitting are subjected to feature calculation respectively, thereby extracting trend features, periodic features and random features. These features can effectively represent the different components of historical fluctuation data, thereby constructing a feature set of historical fluctuation patterns.

[0056] Calculate the information entropy of each feature in the historical fluctuation pattern feature set. Information entropy is an indicator to measure feature uncertainty or information volume. By calculating information entropy, a weight can be assigned to each feature. These weights reflect the importance of each feature to the fluctuation pattern. Using these feature weights, perform weighted fusion on the historical fluctuation pattern feature set to generate a historical fluctuation pattern matrix. This matrix contains the comprehensive performance of each feature under different weights and is a comprehensive description of the historical fluctuation pattern.

[0057] Then, a multi-layer progressive similarity calculation network is constructed. Each layer of the network calculates the similarity between the current state matrix and the historical fluctuation pattern matrix. In each layer of calculation, the matching result of the historical fluctuation pattern matrix is ​​used as the weight of the current layer and passed to the next layer. This layer-by-layer transmission method can gradually optimize the similarity calculation and combine the information of each layer to obtain a more accurate time series matching weight.

[0058] Based on the distribution characteristics of the time series matching weight, the system will adaptively adjust the size of the time window. The time window determines the range of historical data used in the calculation. By adjusting the window size, it can adapt to different time scales in the data. When the difference in the time series matching weight obtained from two consecutive calculations is less than the preset threshold, the calculation is stopped to ensure the stability and accuracy of the calculation results.

[0059] The time series matching weights obtained at different scales are nonlinearly combined to obtain a comprehensive matching score. This score reflects the degree of matching between the current state and the historical fluctuation pattern. This score is used to perform probability mapping of risk indicators in the historical fluctuation pattern matrix, thereby outputting a risk probability distribution. This distribution can help predict possible future risks and provide decision support.

[0060] In this embodiment, the trend, cycle and random characteristics of the historical fluctuation pattern are fully characterized through data decomposition and feature extraction, which improves the accuracy and robustness of pattern recognition. The multi-layer progressive similarity calculation network and adaptive time window are adopted to achieve time series matching at different scales, and enhance the adaptability of the algorithm to market fluctuations. The feature weight calculation based on information entropy and the matching score of nonlinear combination make the risk probability distribution more objective and reasonable, providing a reliable basis for risk warning.

[0061] In an optional implementation, the risk probability distribution is input into a graph convolutional network, feature propagation is performed based on the supply chain knowledge graph, a local structural feature matrix in the supply chain network is extracted through multi-layer spatial convolution operations, and the key path of risk transmission is identified by combining expert rules in the supply chain knowledge graph. The risk warning signal containing the transmission probability is generated, including: The risk probability distribution is input into the graph convolutional network, and an adjacency matrix is ​​constructed based on the entity relationship in the supply chain knowledge graph. The adjacency matrix uses bidirectional adaptive weights to dynamically adjust the node relationship. Multi-layer spatial convolution operations are performed on the node features through the adjacency matrix. Each layer of convolution introduces a residual structure and a skip-layer connection with a gating mechanism. The local structure feature matrix is ​​obtained through weight matrix and nonlinear activation function processing; The structural similarity between nodes is calculated using the local structural feature matrix, the importance of historical conduction rules is decayed based on the time series decay function, and the structural similarity and the decayed expert rule score are combined through dynamic weights to obtain rule confidence; Based on the local structural feature matrix, a multi-hop transmission link between nodes is established, a transmission path set is constructed from the risk source node to the target node, and the risk diffusion coefficient of each transmission path is calculated. The risk diffusion coefficient is calculated by the risk probability size of the nodes on the path, the structural similarity between the nodes, and the rule confidence. The transmission path with the largest risk diffusion coefficient is selected as the key risk transmission path; For adjacent nodes on the critical path of risk conduction, the conduction strength is calculated by combining the temporal conduction mode and the spatial conduction mode between the nodes, and the multiplication value of the conduction strength and the risk probability distribution of the node is used as the path conduction probability to generate a risk warning signal including the critical path of risk conduction and the path conduction probability.

[0062] First, the risk probability distribution is input into the graph convolutional network to build an adjacency matrix based on the entity relationships in the supply chain knowledge graph. The adjacency matrix is ​​a matrix that reflects the connection relationship between nodes. In this process, the node relationship in the adjacency matrix is ​​dynamically adjusted using bidirectional adaptive weights. This means that the connection strength between nodes is automatically optimized based on current data and historical information to better reflect the actual impact between nodes.

[0063] Multi-layer spatial convolution operations are performed on node features through the adjacency matrix. Graph convolutional network is a neural network model that can process graph structure data and extract local structural features of nodes through convolution operations. In each layer of convolution, residual structure and skip-layer connection with gating mechanism are introduced. The residual structure helps alleviate the gradient vanishing problem in deep networks and ensures that features are effectively transmitted in multi-layer networks. The gating mechanism controls the transmission of information flow to prevent irrelevant information from interfering with network learning. After convolution, the local structural feature matrix is ​​obtained through weight matrix and nonlinear activation function processing.

[0064] The structural similarity between nodes is calculated through the local structural feature matrix. Structural similarity is an indicator to measure the similarity between two nodes in the graph, reflecting their structural similarity. Based on the time series decay function, the importance of the influence of historical conduction rules is decayed. As time goes by, the influence of historical rules gradually weakens. The structural similarity and the attenuated expert rule score are combined through dynamic weights to obtain the rule confidence, which represents the credibility of the rule at the current moment.

[0065] Establish multi-hop transmission links between nodes. A multi-hop transmission link refers to a multi-level path from one node to another, and the influence of each node on the path on the final target node will be gradually transmitted. Construct a set of transmission paths from the risk source node to the target node, and calculate the risk diffusion coefficient of each path. The risk diffusion coefficient is determined by the risk probability of the nodes on the path, the structural similarity between the nodes, and the rule confidence. Select the path with the largest risk diffusion coefficient as the key path for risk transmission, which has the greatest impact on the entire risk transmission process.

[0066] For adjacent nodes on the critical path of risk transmission, the transmission strength is calculated by combining the temporal transmission mode and spatial transmission mode between nodes. The temporal transmission mode describes the law of risk transmission between nodes over time, while the spatial transmission mode focuses on the geographical or structural relationship between nodes. The calculated transmission strength is combined with the multiplication value of the risk probability distribution of the node as the path transmission probability. The path transmission probability indicates the possibility of risk transmission on a specific path.

[0067] Finally, a risk warning signal containing the key risk transmission path and the path transmission probability is generated. This signal can reflect the potential risk transmission path in the supply chain in real time, provide early warning information for decision makers, and take corresponding preventive measures in advance.

[0068] In this embodiment, by combining graph convolutional networks, knowledge graphs, time-series decay functions and multi-layer conduction calculations, the accuracy of risk prediction and management in the supply chain is significantly improved. The existing technology usually relies on static data or a single rule engine, which makes it difficult to fully capture the complex spatiotemporal relationship between nodes and cannot dynamically adjust the risk conduction path. In contrast, this solution can reflect the actual changes and dependencies between nodes by dynamically adjusting the node relationship in the adjacency matrix. In addition, the multi-layer graph convolutional network combined with the residual structure and gating mechanism enhances the expressiveness of the model, effectively solves the gradient vanishing problem in information transmission, and improves the efficiency of feature extraction. The innovation of the solution is that by adaptively adjusting the weights and the importance decay of historical conduction rules, it effectively integrates time, space and structural information, and enhances the ability to capture multi-dimensional risk conduction paths. The existing technology is mostly static paths or based on single-level information transmission. This technology dynamically identifies and optimizes risk conduction paths through progressive similarity calculations and multi-hop conduction links, further improving the accuracy and flexibility of predictions. In this way, complex node interactions and potential risk diffusion in the supply chain can be better dealt with. In addition, the generation of comprehensive matching scores and path transmission probabilities makes risk warning signals more accurate and real-time, providing decision makers with accurate warning information, thereby effectively avoiding or mitigating the impact of potential risks. Compared with traditional methods, this solution can handle more complex supply chain environments, reduce manual intervention and prediction errors, and improve the overall stability and risk resistance of the supply chain.

[0069] In an optional implementation, for adjacent nodes on the risk conduction critical path, calculating the conduction strength by combining the temporal conduction mode and the spatial conduction mode between the nodes includes: For adjacent nodes on the critical path of risk transmission, the historical fluctuation pattern of the adjacent nodes is extracted based on the sliding time window, and the historical fluctuation pattern is matched with the current state matrix in a multi-dimensional time series through fluctuation frequency matching, fluctuation amplitude matching and fluctuation trend matching. The fluctuation mutation detection is introduced to obtain the fluctuation mutation coefficient, and the multi-dimensional time series matching result and the fluctuation mutation coefficient are weightedly combined to obtain the time series transmission characteristics; Based on the current state matrix, the business association strength, capital flow strength and information transmission strength of the adjacent nodes are extracted, and the conduction impedance coefficient is calculated in combination with the time series conduction characteristics. The conduction impedance coefficient represents the risk resistance of the node. The business association strength, capital flow strength and information transmission strength are corrected by the conduction impedance coefficient to obtain the spatial conduction characteristics; Adaptively weighting and combining the temporal conduction feature and the spatial conduction feature to obtain a basic conduction strength, calculating a conduction threshold value based on the basic conduction strength, and calculating a cascade influence coefficient when the basic conduction strength exceeds the conduction threshold value; The basic conduction strength and the cascade influence coefficient are combined to obtain the conduction strength between adjacent nodes.

[0070] Exemplarily, for adjacent nodes on the critical path of risk transmission, the historical fluctuation pattern of each adjacent node is extracted by means of a sliding time window. The sliding time window is a method for analyzing by selecting data within a time range in chronological order. Through this window, the fluctuation characteristics of the node in different time periods can be extracted. These fluctuation characteristics include fluctuation frequency, fluctuation amplitude and fluctuation trend, which respectively reflect the periodicity, intensity and development direction of node fluctuations. After that, these historical fluctuation patterns are matched with the current state matrix in a multi-dimensional time series. In this process, the fluctuation frequency matching is first performed to compare the periodicity of historical fluctuations with the periodicity of the current state; then the fluctuation amplitude matching is performed to compare the intensity of historical fluctuations with the intensity of the current state; finally, the fluctuation trend matching is performed to compare the direction of historical fluctuations with the direction of change of the current state. In addition, it is necessary to introduce fluctuation mutation detection to identify the mutation points of node fluctuations. These mutation points are often the key points for the occurrence of risks. By calculating the fluctuation mutation coefficient, the information of these mutation points can be effectively integrated into the time series matching results.

[0071] Combining the obtained multi-dimensional time series matching results and volatility mutation coefficients, the time series transmission characteristics are obtained through weighted combination. The time series transmission characteristics reflect the risk transmission characteristics of nodes at different time scales. Through these characteristics, the risk transmission path and intensity can be predicted more accurately.

[0072] Based on the current state matrix, the business association strength, capital flow strength and information transmission strength are extracted from the adjacent nodes. These strengths represent the interactive relationship between nodes at the business level, capital flow and information flow transmission. By combining the timing transmission characteristics, the transmission impedance coefficient is calculated. The transmission impedance coefficient represents the node's ability to resist external risks, similar to the resistance in the circuit. The larger the value, the stronger the node's risk resistance and the smaller the transmission risk.

[0073] By correcting the business association strength, capital flow strength and information transmission strength through the transmission impedance coefficient, the spatial transmission characteristics can be obtained. The spatial transmission characteristics represent the risk transmission characteristics between nodes in space or structure. These characteristics can reflect how the risks between different nodes are affected by factors such as business and capital flow.

[0074] The temporal transmission characteristics and spatial transmission characteristics are adaptively weighted to obtain the basic transmission strength. The basic transmission strength comprehensively considers the impact of the time dimension and the spatial dimension on risk transmission, and can comprehensively evaluate the transmission of risks. The transmission critical value is calculated based on the basic transmission strength. When the basic transmission strength exceeds the transmission critical value, it indicates that the risk transmission has reached a dangerous level, and the cascade impact coefficient needs to be further calculated. The cascade impact coefficient is used to evaluate the degree of mutual influence between nodes in the risk transmission process. When the transmission strength reaches a certain critical value, it may trigger a series of chain reactions, thereby expanding the scope of risk transmission.

[0075] Combining the basic transmission strength with the cascade influence coefficient, we can get the transmission strength between adjacent nodes. This transmission strength reflects the mutual influence and synergy between adjacent nodes under a specific risk transmission path, providing a quantitative basis for risk management and helping to identify potential risk transmission chains.

[0076] In this embodiment, by combining sliding time windows, fluctuation matching, time series transmission characteristics and spatial transmission characteristics of multi-dimensional analysis, the ability to identify and predict supply chain risks is significantly improved. Existing technologies usually rely on static analysis and single-dimensional risk assessment, which is difficult to cope with complex and dynamically changing supply chain environments. Traditional methods often ignore the linkage effect of time and space factors, and this solution effectively overcomes these shortcomings by dynamically matching historical fluctuation patterns, introducing fluctuation mutation detection and transmission impedance coefficients. Through more comprehensive feature extraction and adaptive weighted combination, the identification process of risk transmission paths is optimized. Traditional technologies are often difficult to accurately capture the multi-level and multi-dimensional propagation of risks, and this solution can more accurately evaluate the diffusion process of potential risks by comprehensively considering the interaction between nodes in the time series and space dimensions and dynamically adjusting the transmission intensity. In particular, through the introduction of the cascade impact coefficient, the ability to control the risk transmission chain is further enhanced, thereby avoiding the limitations of single-path risk assessment.

[0077] According to a second aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0078] According to a third aspect of the embodiments of the present invention, A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0079] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A supply chain multi-dimensional fluctuation prediction and inventory optimization method based on artificial intelligence, characterized by: include: Acquire and preprocess multi-dimensional data in the supply chain system, use piecewise linear interpolation to supplement missing data, generate standardized time series data sets, and build a supply chain knowledge graph based on the standardized time series data sets; Based on the standardized time series data set, the wavelet transform method is used to decompose the data at multiple scales, extract the trend component, periodic component and random component in the time series data, and calculate the node fluctuation characteristic index value; The node fluctuation characteristic index value is used as input, and the feature aggregation is performed through the graph attention network enhanced by the knowledge graph to obtain the fluctuation characteristic vector. The correlation strength between different nodes is calculated based on the multi-head self-attention mechanism, and a state matrix reflecting the overall fluctuation trend of the supply chain is constructed based on the fluctuation characteristic vector and the correlation strength. Match the historical fluctuation pattern with the current state matrix in time series and output the risk probability distribution; The risk probability distribution is input into the graph convolutional network, and feature propagation is performed based on the supply chain knowledge graph. The local structural feature matrix in the supply chain network is extracted through multi-layer spatial convolution operations. Combined with the expert rules in the supply chain knowledge graph, the key path of risk transmission is identified, and a risk warning signal containing the transmission probability is generated. According to the risk warning signal, the preset emergency response strategy library is called to generate the corresponding risk prevention and control plan.

2. The method according to claim 1, characterized in that Acquire and preprocess multi-dimensional data in the supply chain system, use piecewise linear interpolation to supplement missing data, generate standardized time series data sets, and build a supply chain knowledge graph based on the standardized time series data sets, including: Acquire multi-dimensional data in the supply chain system, including supplier production data, logistics and transportation data, warehouse inventory data, and market demand data; Establish a feature distribution matrix according to the multi-dimensional data, obtain a density feature vector using a local density calculation method, calculate a local deviation value of a data point based on the density feature vector, and compare it with an adaptive threshold to identify the first round of abnormal data, and record the location information of the first round of abnormal data in an abnormal index list; Extract the time series data segments around the first round of abnormal data according to the abnormal index list, calculate the change trend coefficients between the data points in the time series data segments, construct a data change feature matrix, use the data change feature matrix to perform secondary verification on the first round of abnormal data, and generate a final abnormal data set; Based on the location information of the final abnormal data set, a data interval to be supplemented is divided in the multi-dimensional data, the numerical difference and time span of the data points at both ends of the data interval to be supplemented are calculated, and the supplementary data weight is adaptively generated according to the numerical difference and time span, and the supplementary data weight is used to perform piecewise linear interpolation calculation to obtain a supplementary data set; Filling the supplementary data set into the multi-dimensional data, normalizing the filled data to generate a standardized data matrix; Based on the standardized data matrix, supply chain entity features are extracted, the multidimensional correlation between entities is calculated to obtain a correlation strength matrix, and a supply chain knowledge graph is constructed according to the correlation strength matrix, wherein nodes represent supply chain entities and the weights of connecting edges are determined by the correlation strength matrix.

3. The method according to claim 1, characterized in that Based on the standardized time series data set, the wavelet transform method is used to decompose the data at multiple scales, extract the trend component, periodic component and random component in the time series data, and calculate the node fluctuation characteristic index values ​​including: Calculate the energy concentration and information entropy of different wavelet basis functions based on the standardized supply chain time series data set, construct a multi-objective optimization function according to the energy concentration and information entropy, and obtain the optimal wavelet basis function by solving the multi-objective optimization function; The standardized supply chain time series data set is subjected to an adaptive multi-level decomposition, and the scale coefficient and wavelet coefficient are calculated at each decomposition level. The trend component is reconstructed according to the scale coefficient, and the periodic component and random component are reconstructed according to the wavelet coefficient. The energy distribution characteristics of the trend component, periodic component and random component at different scales are used as fluctuation characteristic indicator values.

4. The method according to claim 3, characterized in that Based on the standardized supply chain time series data set, the energy concentration and information entropy of different wavelet basis functions are calculated, and a multi-objective optimization function is constructed according to the energy concentration and information entropy. The optimal wavelet basis function is obtained by solving the multi-objective optimization function, including: Constructing a wavelet basis function candidate set, and using each wavelet basis function in the wavelet basis function candidate set to perform wavelet transform on the supply chain time series data in the standardized supply chain time series data set to obtain corresponding wavelet coefficients; Calculating the energy value of each wavelet basis function based on the wavelet coefficient and performing normalization processing to obtain an energy distribution probability, and calculating the energy concentration of each wavelet basis function according to the energy distribution probability; Calculating the information entropy of each wavelet basis function based on the energy distribution probability, and obtaining the relative entropy characteristics of each wavelet basis function using the ratio of the information entropy to the theoretical maximum entropy; Reconstructing a signal using the wavelet coefficients and corresponding wavelet basis functions to obtain a reconstructed signal, and calculating a reconstruction error between the reconstructed signal and the supply chain time series data; The energy concentration, relative entropy characteristics and reconstruction error are weightedly combined to construct a multi-objective optimization function, and the multi-objective optimization function is iteratively optimized using the gradient descent method. When the adjacent iteration difference of the multi-objective optimization function is less than a preset convergence threshold and the function variance meets the stability requirement, the optimal wavelet basis function is selected from the wavelet basis function candidate set.

5. The method according to claim 1, characterized in that The node fluctuation characteristic index value is used as input, and the fluctuation characteristic vector is obtained by feature aggregation through the graph attention network enhanced by the knowledge graph. The correlation strength between different nodes is calculated based on the multi-head self-attention mechanism. The state matrix reflecting the overall fluctuation trend of the supply chain is constructed based on the fluctuation characteristic vector and the correlation strength, including: Obtaining a fluctuation characteristic index value of a supply chain node, embedding the fluctuation characteristic index value and position coding information to obtain a node representation vector; Extracting node association information from the supply chain knowledge graph, building a knowledge-enhanced graph attention network based on the node association information, the knowledge-enhanced graph attention network fuses the node association information with the node representation vector to calculate the attention weights between nodes, and performing feature aggregation on the node representation vector according to the attention weights to generate a node-level fluctuation feature vector; The fluctuation feature vector is input into a multi-head self-attention network, and the multi-head self-attention network performs feature transformation on the fluctuation feature vector through multiple sets of feature mapping matrices to obtain a query vector, a key vector and a value vector, calculates the association strength between different nodes based on the query vector and the key vector, and performs weighted fusion on the value vector using the association strength to generate an association feature vector of the node; A node state vector is constructed based on the fluctuation eigenvector and the associated eigenvector, and the node state vectors are combined to form a state matrix. Multi-scale consistency constraints and temporal consistency constraints are imposed on the state matrix for optimization to obtain a state matrix that reflects the overall fluctuation trend of the supply chain.

6. The method according to claim 1, characterized in that Match the historical fluctuation pattern with the current state matrix in time series, and output the risk probability distribution including: The historical fluctuation data is split into a trend component, a period component and a random component by data decomposition, and the trend component, the period component and the random component are respectively calculated to obtain trend features, period features and random features, so as to construct a historical fluctuation pattern feature set; The feature weight is determined by calculating the information entropy of each feature in the historical fluctuation pattern feature set, and the historical fluctuation pattern feature set is weightedly fused using the feature weight to obtain a historical fluctuation pattern matrix; Construct a multi-layer progressive similarity calculation network, calculate the similarity between the current state matrix and the historical fluctuation pattern matrix in each layer of the network, and use the similarity result of each layer as the calculation weight of the next layer to obtain the progressive time series matching weight between layers; Adaptively adjust the time window size according to the distribution characteristics of the timing matching weight, resample the current state matrix based on the time window size, and perform multi-layer progressive similarity calculation at different scales, and stop the calculation when the difference between the timing matching weights obtained by two adjacent calculations is less than a preset weight threshold; The time series matching weights at different scales are nonlinearly combined to obtain a comprehensive matching score, which is used to perform probability mapping on the risk indicators in the historical volatility pattern matrix to output the risk probability distribution.

7. The method according to claim 1, characterized in that The risk probability distribution is input into the graph convolutional network, and feature propagation is performed based on the supply chain knowledge graph. The local structural feature matrix in the supply chain network is extracted through multi-layer spatial convolution operations. Combined with the expert rules in the supply chain knowledge graph, the key path of risk transmission is identified, and risk warning signals containing transmission probability are generated, including: The risk probability distribution is input into the graph convolutional network, and an adjacency matrix is ​​constructed based on the entity relationship in the supply chain knowledge graph. The adjacency matrix uses bidirectional adaptive weights to dynamically adjust the node relationship. Multi-layer spatial convolution operations are performed on the node features through the adjacency matrix. Each layer of convolution introduces a residual structure and a skip-layer connection with a gating mechanism. The local structure feature matrix is ​​obtained through weight matrix and nonlinear activation function processing; The structural similarity between nodes is calculated using the local structural feature matrix, the importance of historical conduction rules is decayed based on the time series decay function, and the structural similarity and the decayed expert rule score are combined through dynamic weights to obtain rule confidence; Based on the local structural feature matrix, a multi-hop transmission link between nodes is established, a transmission path set is constructed from the risk source node to the target node, and the risk diffusion coefficient of each transmission path is calculated. The risk diffusion coefficient is calculated by the risk probability size of the nodes on the path, the structural similarity between the nodes, and the rule confidence. The transmission path with the largest risk diffusion coefficient is selected as the key risk transmission path; For adjacent nodes on the critical path of risk conduction, the conduction strength is calculated by combining the temporal conduction mode and the spatial conduction mode between the nodes, and the multiplication value of the conduction strength and the risk probability distribution of the node is used as the path conduction probability to generate a risk warning signal including the critical path of risk conduction and the path conduction probability.

8. The method according to claim 7, characterized in that For the adjacent nodes on the critical risk transmission path, the transmission intensity is calculated by combining the temporal transmission mode and the spatial transmission mode between the nodes, including: For adjacent nodes on the critical path of risk transmission, the historical fluctuation pattern of the adjacent nodes is extracted based on the sliding time window, and the historical fluctuation pattern is matched with the current state matrix in a multi-dimensional time series through fluctuation frequency matching, fluctuation amplitude matching and fluctuation trend matching. The fluctuation mutation detection is introduced to obtain the fluctuation mutation coefficient, and the multi-dimensional time series matching result and the fluctuation mutation coefficient are weightedly combined to obtain the time series transmission characteristics; Based on the current state matrix, the business association strength, capital flow strength and information transmission strength of the adjacent nodes are extracted, and the conduction impedance coefficient is calculated in combination with the time series conduction characteristics. The conduction impedance coefficient represents the risk resistance of the node. The business association strength, capital flow strength and information transmission strength are corrected by the conduction impedance coefficient to obtain the spatial conduction characteristics; Adaptively weighting and combining the temporal conduction feature and the spatial conduction feature to obtain a basic conduction strength, calculating a conduction threshold value based on the basic conduction strength, and calculating a cascade influence coefficient when the basic conduction strength exceeds the conduction threshold value; The basic conduction strength and the cascade influence coefficient are combined to obtain the conduction strength between adjacent nodes.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method described in any one of claims 1 to 8.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images