A supply chain demand prediction method and system based on a fusion model, equipment and medium

Abstract

The application provides a supply chain demand prediction method and system based on a fusion model, equipment and a medium. First, by collecting historical sales data, inventory conditions and other related supply chain participant operation data, the interaction between the supply chains is determined, and a supply chain relationship model is constructed. Then, using ARIMA and other prediction models, the demand of each link in the supply chain is predicted. Next, the supply chain relationship model and the prediction model are fused to obtain the comprehensive predicted demand of each link. Finally, the prediction results are compared with the actual situation to evaluate the performance of the model. This method can effectively improve the accuracy of supply chain demand prediction, thereby helping enterprises optimize supply chain management and improve operational efficiency.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of supply chain demand prediction, and more particularly relates to a supply chain demand prediction method based on a fusion model. BACKGROUND

[0002] Traditional demand prediction methods mainly include qualitative prediction and quantitative prediction. Qualitative prediction mainly relies on the experience and intuition of experts, such as opinion surveys and expert evaluations. They are useful in new product prediction or in the absence of historical data. However, qualitative methods are often subjective and accurate, and vary from person to person, and cannot handle large amounts of data.

[0003] In contrast, quantitative prediction methods can accurately predict future demand, especially when there is a large amount of historical sales data. These methods include time series analysis, causal models, regression analysis, etc. They can analyze historical data, find patterns and trends hidden in the data, and provide predictions. These methods can handle large amounts of data, but require professional skills and may also be affected by data anomalies.

[0004] In recent years, with the rise of big data and artificial intelligence, new demand prediction methods such as machine learning and deep learning have also begun to be applied to demand prediction. They can handle large and complex data, discover deep association rules and patterns, and make high-precision predictions.

[0005] However, these models often ignore the relationships between participants in the supply chain and their interactions. This may ignore some important factors in demand prediction, leading to prediction errors. In this context, we propose a supply chain demand prediction method based on a fusion model, which improves the accuracy of prediction by considering the relationships in the supply chain. SUMMARY

[0006] The application mainly solves the problem of demand prediction in traditional demand prediction methods, which ignores the interaction between participants in the supply chain and cannot accurately reflect the demand changes and the nature of the supply chain system. By building a fusion of supply chain relationship models and prediction models, the interaction between participants and the interaction strength are taken into account, improving the accuracy of demand prediction and the efficiency of supply chain management.

[0007] To achieve the above purpose, the application adopts the following technical solutions: the method comprises:

[0008] Data collection: collect historical sales data, inventory data, and supply conditions;

[0009] Data processing: ensure data quality and accuracy;

[0010] Building a supply chain relationship model: Identify all major players in the supply chain, including suppliers, manufacturers, distributors, and end customers, and build a supply chain relationship model through their interactions;

[0011] Building a prediction model: Based on the supply chain relationship model, build a prediction model;

[0012] Fusion of supply chain relationship and prediction model: Fuse the prediction model and the supply chain relationship model together;

[0013] Evaluate the fusion model: Use the fusion model to predict future demand and evaluate its performance.

[0014] In one scheme, the data collection is obtained from different information systems such as enterprise resource planning systems, warehouse management systems, or through third-party data service providers.

[0015] In one scheme, the data processing includes: data normalization; data division, dividing the data into training set, validation set and test set.

[0016] In one scheme, the construction of the supply chain relationship model includes:

[0017] (1) Identify the main players in the supply chain, including suppliers, manufacturers, wholesalers, retailers, and end consumers;

[0018] (2) Clarify the relationship between participants: understand the role of each participant in the supply chain and their relationship with other participants;

[0019] (3) Analyze the interaction mode: understand the interaction between participants, including: the flow of goods, the exchange of information, quality control;

[0020] (4) Build a supply chain relationship model using a graph model;

[0021] Treat each participant in the supply chain as a node in the graph, and the interaction between participants as a directed edge; on this basis, add weights to represent the amount of flow or the strength of interaction;

[0022] Set G = (V, E), where V represents a set of nodes, which represent participants in the supply chain, including suppliers, manufacturers, distributors, and retailers; and E represents a set of directed edges, representing the relationship between participants.

[0023] Assume V = {v1, v2, v3,..., vn}, where vi (i = 1, 2,..., n) represents different participants in the supply chain;

[0024] E is a set of edges, defined as E={(vi, vj, w)} | vi, vj belongs to V, w represents the weight, which represents the quantity of goods flow between suppliers and manufacturers, manufacturers and distributors, distributors and retailers;

[0025] Finally, the graph model is represented by an adjacency matrix A (n x n matrix, n represents the number of nodes); if there is an edge from vi to vj, the corresponding matrix element aij is the weight w, otherwise aij=0.

[0026] In one scheme, the construction of the prediction model comprises: using an ARIMA model for demand prediction;

[0027] Specifically, it comprises: ARIMA model construction and prediction.

[0028] The ARIMA model consists of three parts: AR model (autoregressive model), I model (difference integration model) and MA model (moving average model), which correspond to the three parameters p, d and q of the ARIMA model respectively.

[0029] In the ARIMA model, the p-order model of the AR part is expressed as:

[0030]

[0031] X t is the value at the current time, c is a constant, is the autoregressive coefficient, ε t is white noise; the q-order model of the MA part is expressed as:

[0032] X t = μ + ε t + ∑ (from i=1 to q) θ i * ε (t-i)

[0033] Here μ is the mean of the sequence, θ i is the moving average coefficient;

[0034] The difference part includes first-order or multi-order difference of the original time series, so that the time series becomes stationary, which is expressed by the following formula:

[0035] Y t = X t - X (t-1)

[0036] Integrate the influences of the three aspects together, and express the ARIMA (p, d, q) model by the following formula:

[0037]

[0038] where B is the backshift operator, B k X t = X (t-k) , B d is the dth difference, and the left part corresponds to AR and I, and the right part corresponds to MA.

[0039] In one aspect, the fusion supply chain relationship and prediction model comprises:

[0040] (1) Use the prediction model ARIMA(p, d, q) to make demand prediction at each node; that is, use the ARIMA model to predict future demand at participants (nodes) such as suppliers, manufacturers, distributors, and retailers; assume that the demand prediction value of node vi is ARIMA vi (p, d, q), which is expressed as:

[0041] Y vit = ARIMA vi (p, d, q)

[0042] where t represents the prediction time, Y vit represents the predicted demand of node vi at time t;

[0043] (2) Use the supply chain relationship model to adjust the weight at each interaction relationship (edge); assume that the interaction weight from node vi to node vj is wij, then the predicted demand of node vi to vj will be obtained by adjusting the predicted demand value of node vi by weight. It can be expressed as:

[0044] Y' vi_vj_t = wij*ARIMA vi (p, d, q)

[0045] where Y' vi_vj_t represents the predicted demand of node vi to node vj at time t.

[0046] (3) Aggregate the predicted demand at each interaction relationship to obtain the final predicted demand; N is all the interaction relationships in the supply chain, then the final predicted demand Y_t is expressed as:

[0047]

[0048] In one aspect, the evaluation fusion model comprises:

[0049] Collect actual demand values to compare with our predictions; the actual demand value is A t , the predicted demand value Y t , and the prediction error is calculated:

[0050] E t = A t-Y t

[0051] using statistical indicators of errors to comprehensively evaluate the prediction performance,

[0052] The mean value ME of the error represents the average level of the prediction deviation, and is calculated by the following formula:

[0053] ME = ∑E t / T

[0054] Wherein T is the total time of the prediction.

[0055] The variance VE of the error represents the stability of the prediction deviation, and is calculated by the following formula:

[0056]

[0057] In one scheme, 8, the system comprises a data collection module for collecting historical sales data, inventory data, and supply data;

[0058] A data processing module is configured to process the collected data to ensure the quality and accuracy of the data.

[0059] A supply chain relationship model construction module is configured to identify all main participants in the supply chain and construct a supply chain relationship model through the interaction between them.

[0060] A prediction model construction module is configured to construct a prediction model based on the supply chain relationship model. A model fusion module is configured to fuse the prediction model and the supply chain relationship model together.

[0061] An evaluation module is configured to use the fused model to predict future demand and evaluate its performance.

[0062] In another aspect, an electronic device comprises a processor and a memory, wherein the memory stores computer readable instructions, and the computer readable instructions are executed by the processor to implement the supply chain demand prediction method based on the fused model.

[0063] In another aspect, a computer storage medium stores a computer program, and the computer program is executed by a processor to implement the supply chain demand prediction method based on the fused model.

[0064] The present application has the following advantages:

[0065] The application effectively proposes a new demand prediction strategy by fusing the supply chain relationship model and the prediction model, considers the interaction relationship and interaction strength between participants, and significantly improves the accuracy of the prediction result. In addition, the new prediction method can effectively give the supply chain management more predictability and pertinence, so that the production scheduling, inventory strategy and other management work are more efficient, thereby greatly optimizing the operation cost and response speed of the entire supply chain, and improving the economic benefit of the enterprise. At the same time, due to the significant improvement of the accuracy of the prediction result, customer satisfaction and the competitiveness of the enterprise are also greatly enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0066] Figure 1 The method flowchart of the application;

[0067] Figure 2 The supply chain relationship model construction flowchart;

[0068] Figure 3 The flowchart of fusing the supply chain relationship and the prediction model. DETAILED DESCRIPTION

[0069] In order to facilitate the understanding of the present application, the present application will be described more fully below with reference to the accompanying drawings. The drawings show typical embodiments of the present application. However, the present application can be implemented in many different forms and is not limited to the embodiments described in the present application. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive.

[0070] Unless otherwise defined, all technical and scientific terms used in the present application have the same meaning as understood by a person skilled in the art to which the present application belongs. The terms used in the present application in the specification of the present application are only for the purpose of describing the specific embodiments and are not intended to limit the present application. In order to facilitate the understanding of the present application, the present application will be described more fully below with reference to the accompanying drawings. The drawings show typical embodiments of the present application. However, the present application can be implemented in many different forms and is not limited to the embodiments described in the present application. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present application more thorough and comprehensive.

[0071] Embodiment one:

[0072] As shown in the figure, the supply chain demand prediction method based on the fusion model comprises: Figure 1

[0073] S1, data collection: collect historical sales data, inventory data, supply situation.

[0074] Determine the type of data to be collected. In this stage, historical sales data, inventory data, supply situation and other demand prediction related data are required.​

[0075] Identify the data sources from which these data can be obtained. Data sources include internal systems such as ERP, warehouse management systems, etc., as well as external data sources such as supplier data, third-party market reports, etc.

[0076] After obtaining the data, it is necessary to verify the accuracy and completeness of the data. This may include checking for missing values or errors in the data, and comparing with known information or standards.

[0077] Store the collected data in a place where further analysis can be performed, such as a database or data warehouse. At the same time, consider the requirements of data security and privacy, and ensure compliance with relevant laws and industry rules.

[0078] S2, Data processing: Ensure the quality and accuracy of the data.

[0079] Data normalization: Adjust the data to a standard scale, such as between 0-1 (normalized) or to a standard normal distribution (standardized). This ensures that different scales of data do not have a significant impact on the final results in subsequent modeling processes.

[0080] Divide the data into training set, validation set and test set, so as to evaluate the generalization ability of the model and prevent overfitting or underfitting. The ratio of training set, validation set and test set is: 6:2:2.

[0081] S3, Build a supply chain relationship model: Identify all major participants in the supply chain, including suppliers, manufacturers, distributors and end customers, and build a supply chain relationship model through their interactions.

[0082] As shown in Figure 2 , the construction of the supply chain relationship model includes:

[0083] S301, First identify the main participants in the supply chain, including suppliers, manufacturers, wholesalers, retailers and end consumers.

[0084] S302, Clarify the relationship between participants: Understand the role of each participant in the supply chain and their relationship with other participants. For example, the relationship with suppliers may be cooperative, competitive or dependent.

[0085] S303, Analyze the interaction method: Have a clear understanding of the possible interaction methods between participants, such as the flow of goods, exchange of information, quality control, etc.

[0086] S304, Build a supply chain relationship model, build a graph model, which can use mathematical formulas in graph theory to describe the interaction relationship. The following represents the relationship between participants in the supply chain as a mathematical model.

[0087] Each participant in the supply chain is regarded as a node in the graph, and the interaction between participants is regarded as a directed edge; on this basis, weights are added to represent the quantity of flow or the strength of interaction;

[0088] Set the graph G=(V,E), where V represents a set of nodes, which represent participants in the supply chain, including suppliers, manufacturers, distributors and retailers; and E represents a set of directed edges, representing the relationship between participants.

[0089] Assume V={v1,v2,v3,...,vn}, where vi(i=1,2,...,n) represents different participants in the supply chain;

[0090] E is a set of edges, defined as E={(vi,vj,w)}|vi,vj∈V,w represents the weight, which represents the quantity of goods flow between suppliers and manufacturers, manufacturers and distributors, and distributors and retailers;

[0091] Finally, the adjacency matrix A(n x n matrix, n represents the number of nodes) is used to represent the graph model; if there is an edge from vi to vj, then the corresponding matrix element aij is the weight w, otherwise aij=0.

[0092] S4, build a prediction model: based on the supply chain relationship model, build a prediction model.

[0093] The construction of the prediction model includes using the ARIMA model for demand prediction; specifically including the construction and prediction of the ARIMA model. The ARIMA model stands for AutoRegressive Integrated Moving Average Model. It is a prediction model based on the combination of autoregressive process and moving average process for time series data.

[0094] In the ARIMA model, we convert the time series data into a stationary sequence by differencing. Then the stationary sequence is predicted, and the prediction result is de-differentiated to obtain the prediction result of the original sequence.

[0095] The ARIMA model consists of three parts: AR model (autoregressive model), I model (difference integration model) and MA model (moving average model), which correspond to the three parameters p, d and q of the ARIMA model.

[0096] In the ARIMA model, the p-order model of the AR part is represented as:

[0097]

[0098] X t is the value at the current time, c is a constant, is the autoregressive coefficient, ε t is white noise; the q-order model of the MA part is expressed as:

[0099] X t = μ + ε t + ∑ (from i = 1 to q) θ i * ε (t-i)

[0100] Here μ is the mean of the sequence, θ i is the moving average coefficient;

[0101] The difference part includes first-order or multi-order difference of the original time series, so that the time series becomes stationary, which is expressed by the following formula:

[0102] Y t = X t - X (t-1)

[0103] Integrate the influences of the three aspects together, and express the ARIMA (p, d, q) model by the following formula:

[0104]

[0105] Where B is the backshift operator, B k * X t = X (t-k) , the d-th power of B is the d-order difference, and the left part corresponds to AR and I, and the right part corresponds to MA.

[0106] As Figure 3 shown, S5, fusion supply chain relationship and prediction model: fuse the prediction model and the supply chain relationship model together.

[0107] The fusion supply chain relationship and prediction model comprises:

[0108] S501, using the prediction model ARIMA (p, d, q) to predict demand at each node; that is, using the ARIMA model to predict future demand at participants (nodes) such as suppliers, manufacturers, distributors and retailers; assuming that the demand prediction value of node vi is ARIMA vi (p, d, q), expressed as:

[0109] Y vit = ARIMA vi (p, d, q)

[0110] Where t represents the prediction time, Y vitY' represents the predicted demand of node vi at time t.

[0111] S502, applying the supply chain relationship model to each interaction relationship (edge) to adjust the weight; assuming that the interaction weight from node vi to node vj is wij, the predicted demand of node vi to vj will be obtained by adjusting the predicted demand value of node vi. It can be represented as:

[0112] Y' vi_vj_t = wij*ARIMA vi (p,d,q)

[0113] where Y' vi_vj_t represents the predicted demand from node vi to node vj at time t.

[0114] S503, aggregate the predicted demand on each interaction relationship to obtain the final predicted demand; N is all the interaction relationships in the supply chain, then the final predicted demand Y_t is represented as:

[0115]

[0116] S6, evaluate the fusion model: use the fusion model to predict future demand and evaluate its performance.

[0117] The evaluation of the fusion model includes:

[0118] Collect actual demand values to compare with our predictions; actual demand value A t , predicted demand value Y t , calculate the prediction error:

[0119] E t = A t -Y t

[0120] Use statistical indicators of error to comprehensively evaluate the prediction performance,

[0121] The prediction error can reflect the deviation between the predicted value and the actual value, which is an intuitive indicator to evaluate the prediction performance. However, it should be noted that a single error value cannot fully reflect the prediction performance, because the predicted value may be affected by random errors, therefore, statistical indicators of error are used to comprehensively evaluate the prediction performance.

[0122] The mean error ME represents the average level of prediction deviation, which is calculated by the following formula:

[0123] ME = ∑E t / T

[0124] where T is the total prediction time.

[0125] The variance of error VE represents the stability of the prediction bias, which is calculated by the following formula:

[0126]

[0127] Example two:

[0128] A supply chain demand forecasting system based on a fusion model. The system includes a data collection module for collecting historical sales data, inventory data, and supply data;

[0129] A data processing module for processing the collected data to ensure data quality and accuracy;

[0130] A supply chain relationship model construction module for identifying all major participants in the supply chain and constructing a supply chain relationship model through their interactions;

[0131] A prediction model construction module for constructing a prediction model based on the supply chain relationship model; a model fusion module for fusing the prediction model and the supply chain relationship model together;

[0132] An evaluation module for predicting future demand using the fusion model and evaluating its performance.

[0133] Example three:

[0134] Take a manufacturing enterprise as an example, the enterprise's supply chain includes suppliers (Supplier), manufacturing factories (Factory), distributors (Distributor) and retailers (Retailer), which constitute four nodes in the supply chain. In the ERP and warehouse management system of the manufacturing enterprise, we have collected sales data, inventory data, material consumption data and other data in the past two years. And from the supplier, we collected the supply situation of the relevant raw materials, and collected consumer consumption habits and other data from the portal website and industry report.

[0135] In the S2 stage, the collected data is cleaned and processed, abnormal values and missing values are removed, and non-numeric data is digitized and encoded. Finally, the cleaned data is obtained.

[0136] In the S3 stage, set

[0137] V={v1(Supplier),v2(Factory),v3(Distributor),v4(Retailer)} as nodes in the supply chain relationship model, and analyze the interaction between the four, define the weight, such as raw material supply, product manufacturing and sales, etc., respectively, to build the edge from the supplier to the factory v1-v2, from the factory to the distributor v2-v3, from the distributor to the retailer v3-v4, and build a supply chain relationship model.

[0138] In the S4 stage, the ARIMA model is used to independently predict the demand of each node by setting the model parameters p, d, q as 1, 1, 1 respectively, and combining the historical data collected above. Assuming the prediction results are: v1(Supplier) predicted demand: 1000 raw materials / v2(Factory) predicted demand: 900 products / v3(Distributor) predicted demand: 800 products / v4(Retailer) predicted demand: 700 products.

[0139] In the S5 stage, the results of S3 and S4 are fused, and by setting the weight of the edge w_ij and based on the predicted demand of each node obtained above, it can be concluded that the demand prediction value from v1 to v2 after weight adjustment is 0.91000 = 900 products, the demand prediction value from v2 to v3 is 0.89900 = 801 products, and the demand prediction value from v3 to v4 is 0.87*800 = 696 products.

[0140] In the S6 stage, the performance of the model is evaluated by comparing with the actual data, and the evaluation standard is the accuracy of the prediction result and the actual value. Assuming that the actual v4(Retailer) demand is 695 products, which is within the prediction error range, indicating that the model prediction accuracy is high.

[0141] We expect the demand prediction results to guide the company to make more reasonable production decisions, inventory management decisions and supply chain coordination decisions in actual operation, and continuously optimize the model according to the actual results to improve the prediction accuracy. In this way, not only can we reduce unnecessary waste and inventory, optimize supply chain operation, but also improve customer satisfaction and improve the economic benefits of enterprises.

[0142] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The program can include the processes of the above-mentioned embodiments when executed. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0143] It should be understood that the technical solutions of the present application described above with the preferred embodiments are illustrative rather than limiting. Based on the description of the present application, those skilled in the art can modify the technical solutions recorded in each embodiment, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of each embodiment of the present application.

Claims

1. A method for supply chain demand forecasting based on a fusion model, characterized in that: The method comprises: Data collection: collecting historical sales data, inventory data, supply situation; Data processing: ensuring the quality and accuracy of the data; Building a supply chain relationship model: identifying all major players in the supply chain, and building a supply chain relationship model through their interactions; Building a prediction model: based on the supply chain relationship model, building a prediction model; Fusing supply chain relationships and prediction models: fusing the prediction model and the supply chain relationship model together; Evaluating the fusion model: using the fusion model to predict future demand and evaluating its performance; The method of building a supply chain relationship model comprises: (1) identifying the main players in the supply chain, including suppliers, manufacturers, wholesalers, retailers, and end consumers; (2) Clarify the relationship between participants: understand the role of each participant in the supply chain and their relationship with other participants; (3) Analyze the interaction mode: understand the interaction between participants, including: the flow of goods, the exchange of information, quality control; (4) Build a supply chain relationship model using a graph model; Treat each participant in the supply chain as a node in the graph, and the interaction between participants as a directed edge; On this basis, add weights to represent the amount of flow or the strength of interaction; Setting diagram where V represents a set of nodes, which represent the participants in the supply chain, suppliers, manufacturers, distributors, and retailers, and E represents a set of directed edges, which represent the relationships of interactions among the participants. Assume wherein represent different participants in the supply chain; E is a set of edges, defined as , belong to V, w represents weight, which represents the quantity of goods flow between suppliers and manufacturers, manufacturers and distributors, distributors and retailers; Finally, the graph model is represented by the adjacency matrix A n x n of size n x n, where n is the number of nodes; if there is an edge between vi and vj, then the corresponding matrix element aij is the weight w, otherwise ; The method of building a prediction model comprises: using ARIMA model for demand prediction; Specifically, it includes ARIMA model construction and prediction; ARIMA model consists of three parts: AR model, I model and MA model, which correspond to the three parameters p, d and q of ARIMA model respectively; In the ARIMA model, the p-order model of the AR part is represented as: ； is the value at the current time, c is a constant, is the autoregressive coefficient, is the white noise of the AR part; The q-order model of the MA part is represented as: ； Here µ is the mean of the MA part of the sequence, is the moving average coefficient; The difference part includes one or more order differences of the original time series, which makes the time series stationary, represented by the following formula: ； Integrate the three aspects together, and use the following formula to represent the ARIMA(p, d, q) model: ； where B is the backward shift operator, Bd is the dth difference operator. The method of fusing supply chain relationships and prediction models comprises: The demand prediction is performed on each node using a prediction model ARIMA(p, d, q); that is, the future demand is predicted on each node using the ARIMA model; and the demand prediction value of the node vi is denoted as , which is expressed as: ； where t denotes the prediction time, denotes the predicted demand of node vi at time t; Adjust the weights on each edge using the supply chain relationship model; Assuming that the interaction weight from node vi to node vj is wij, the predicted demand from node vi to vj will be obtained by adjusting the predicted demand value of node vi by weight; Represented as: ； wherein, denotes the predicted demand from node vi to node vj at time t; The final forecast demand is obtained by aggregating the forecast demands on each interaction relationship. N is all the interaction relationships in the supply chain, and the final forecast demand is is represented as: 。 2. The method of claim 1, wherein: The data collection is obtained from different information systems such as enterprise resource planning systems, warehouse management systems, or through third-party data service providers.

3. The method of claim 1, wherein: The data processing includes: data normalization; data division, dividing the data into training set, validation set and test set.

4. The method of claim 1, wherein: The method of evaluating the fusion model comprises: Collect actual demand values to compare with our predictions; actual demand values are , predicted demand values , compute prediction errors: ； Use statistical indicators of errors to comprehensively evaluate the prediction performance, The mean error ME represents the average level of prediction bias, which is calculated by the following formula: ； Where T is the total prediction time; The error variance VE represents the stability of the prediction bias, which is calculated by the following formula: 。 5. A fusion model based supply chain demand forecasting system, said system being adapted to the method according to any one of claims 1 to 4, characterized in that: The system comprises a data collection module for collecting historical sales data, inventory data, and supply situation data; A data processing module for processing the collected data to ensure the quality and accuracy of the data; A supply chain relationship model construction module for identifying all major players in the supply chain and constructing a supply chain relationship model through their interactions; A prediction model construction module for constructing a prediction model based on the supply chain relationship model; A model fusion module for fusing the prediction model and the supply chain relationship model together; 6. An electronic device, comprising: An evaluation module for predicting future demand using the fused model and evaluating its performance.

7. A computer storage medium, characterized in that: A computer readable medium having stored thereon computer readable instructions which, when executed by a processor, implement a supply chain demand prediction method based on a fused model according to any one of claims 1 to 4. A computer program stored on a computer readable medium, which, when executed by a processor, implements a supply chain demand prediction method based on a fused model according to any one of claims 1 to 4.

Images