A risk control method and device based on the supply chain
By using LSTM network model-based supply chain analysis and OSP front-end display interface, the problem of supplier risk identification and management in the construction industry has been solved, achieving accurate risk identification and control, ensuring the smooth implementation of projects and the steady development of enterprises.
Patent Information
- Application Number
- CN202510155958.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-12
- Publication Date
- 2026-03-06
- Estimated Expiration
- 2045-02-12
AI Technical Summary
The construction industry faces a variety of risk factors, including poor customer credit, unstable supplier service quality, limited fund supervision, unhealthy financial structure, and irregular joint venture and affiliation management, which can lead to problems such as tight cash flow, project quality and safety hazards.
By analyzing historical data of construction projects based on the LSTM network model, a weighted supply chain graph structure is established to identify and assess supplier risks. Risk monitoring and early warning are then conducted using the OSP front-end display interface, thus constructing a multi-dimensional risk assessment system.
It enables accurate identification and control of potential risks in the construction project supply chain, ensures smooth project implementation, provides intuitive risk warnings and management, and enhances the company's ability to withstand risks.
Smart Images

Figure CN120146555B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing technology, and more specifically to a risk control method and apparatus based on the supply chain. Background Technology
[0002] In the current operating environment of the construction industry, a variety of risk factors are intertwined, which brings many obstacles to the stable development of construction companies.
[0003] Firstly, from the customer's perspective, some customers have poor credit, leading to high accounts receivable for the company. Large amounts of funds remain tied up in accounts for extended periods, making recovery extremely difficult. Due to concentrated capital investment in construction projects, slow cash flow results in a tight cash flow for the company, severely hindering subsequent project progress and the payment of daily operating expenses. Issues such as insufficient funds for construction material procurement and delayed wages for construction workers occur frequently.
[0004] Secondly, observing the supplier level reveals significant issues. The quality of supplier services varies greatly. On one hand, the timeliness of material supply is poor. At critical construction stages, such as the main structure construction phase, essential building materials like steel and cement cannot be supplied on time, disrupting the planned construction schedule, causing idle work, and resulting in project delays. On the other hand, product quality control is lax, with substandard building materials mixed into the construction site, creating hidden dangers for project quality. For example, substandard concrete strength may lead to insufficient structural load-bearing capacity, and poor-quality wall bricks are prone to cracking, affecting the building's safety and functionality.
[0005] Third, from the perspective of capital supervision, construction companies are subject to strict constraints from various policies and regulations, limiting the flexibility of capital allocation and resulting in persistently high financial risks. At the same time, frequent legal proceedings further complicate their situation. Once involved in litigation, the company's bank accounts are often frozen by the judiciary, leaving them with virtually no available funds. This exposes them to the risk of a broken cash flow, with crises such as the interruption of daily operating funds and the suspension of construction projects potentially erupting at any time.
[0006] Fourth, focusing on the financial structure, the phenomenon of high levels of both deposits and loans is quite common in the industry. High debt places a heavy burden on enterprise development, and high interest expenses severely erode enterprise profits. The high interest payments each period greatly limit the profitability of enterprises, hindering their plans to expand new businesses and purchase advanced construction equipment to increase production. It is also difficult to raise the funds needed for technological upgrades, leading to difficulties in enterprise development.
[0007] Finally, joint ventures and affiliations are a prominent problem in the construction industry. While this model may help expand business in the short term, its lax management and unclear definition of responsibilities among all parties easily lead to a series of secondary problems, such as loss of control over project quality, frequent safety accidents at construction sites, and constant economic disputes.
[0008] In conclusion, the most urgent task for construction companies is to face these risks squarely, scientifically plan and build a comprehensive risk management system, and promote this in a coordinated manner from multiple aspects such as strict customer credit assessment, standardized supplier management, ensuring compliant fund operation, and strengthening internal control, so as to comprehensively improve their risk resistance capabilities and move forward steadily in the complex and ever-changing construction market. Summary of the Invention
[0009] In view of this, the present invention provides a risk prevention and control method and apparatus based on the supply chain, which can solve the technical problems of identifying potential risks of suppliers in the supply chain, thereby preventing and controlling risks in construction projects.
[0010] To solve the above-mentioned technical problems, the present invention is implemented as follows.
[0011] A supply chain-based risk control method includes:
[0012] Step S1: Obtain historical data of the construction project, extract feature vectors from the historical data, the feature vectors include the start time, planned end time, actual end time, project content, product category, quantity, parameters, and production time of each supplier; input the feature vectors into the trained LSTM network model, the output of the LSTM network model is the time series corresponding to each supply chain, the time series reflects the relationship between each supplier in the supply chain in the time dimension and its impact on the construction project;
[0013] Step S2: Establish a weighted supply chain graph structure for the construction company. The weighted supply chain graph structure includes all suppliers of the construction company, with each supplier as a graph node. Establish connections between graph nodes based on the entire supply chain, and the connections are represented by directed edges. Determine the weights between graph nodes with direct connections based on the entire time series.
[0014] Step S3: Obtain project information of the construction project to be controlled, and determine the risk of the construction project to be controlled based on the weighted supply chain diagram structure and the project information.
[0015] Preferably, step S2, determining the weights between graph nodes with direct connections based on the entire time series, includes:
[0016] For each graph node: obtain all supply chains corresponding to the graph node; obtain the proportion of the time segment corresponding to the graph node in the time series of each supply chain of the graph node relative to the sum of all time segments; sum the proportions of all supply chains and use them as the weight of the graph node.
[0017] The weights of two graph nodes that have a direct connection are added together to form the weight between the graph nodes that have a direct connection.
[0018] Preferably, step S2 further includes: determining the risk factor value of each graph node in the weighted supply chain graph structure, wherein the risk factor value of a graph node is the sum of the weights of the directed edges corresponding to that graph node.
[0019] Preferably, in step S3, the project information includes the start time of the construction project, the planned end time, the project content, and the product categories, quantities, parameters, and production times of the products supplied by each supplier corresponding to the construction project.
[0020] Preferably, step S3, determining the risk of the construction project to be controlled based on the weighted supply chain graph structure and the project information, includes:
[0021] Step S31: Based on the weighted supply chain graph structure and the project information, determine the weighted critical path of the supplier corresponding to the construction project; obtain the risk factor value of each node in the weighted critical path;
[0022] Step S32: Based on the number of products delivered by suppliers corresponding to each node in the weighted critical path and the funds allocated by the construction company to suppliers corresponding to each node, determine the risk level of the construction project to be controlled.
[0023] Preferably, for the suppliers corresponding to each node in the weighted critical path, risk monitoring indicators involving four aspects are also set: fund fraud risk, fund compliance risk, fund liquidity risk, and financial market risk.
[0024] Preferably, the risks are displayed through the OSP front-end interface.
[0025] A supply chain-based risk control device includes:
[0026] Time series acquisition module: configured to acquire historical data of construction projects, extract feature vectors from the historical data, the feature vectors include the start time, planned end time, actual end time, project content, product category, quantity, parameters, and production time of products supplied by each supplier; input the feature vectors into a trained LSTM network model, the output of which is the time series corresponding to each supply chain, the time series reflecting the relationship between suppliers in the supply chain in the time dimension and their impact on the construction project;
[0027] Graph structure building module: configured to build a weighted supply chain graph structure for the construction company. The weighted supply chain graph structure includes all suppliers of the construction company, with each supplier as a graph node. Connections between graph nodes are established based on the entire supply chain, and these connections are represented by directed edges. The weights between graph nodes with direct connections are determined based on the entire time series.
[0028] Risk determination module: configured to acquire project information of the construction project to be controlled, and determine the risk of the construction project to be controlled based on the weighted supply chain diagram structure and the project information.
[0029] The present invention provides a computer-readable storage medium storing a plurality of instructions; the plurality of instructions are used by a processor to load and execute the method as described above.
[0030] The present invention provides an electronic device, characterized in that the electronic device comprises:
[0031] A processor is used to execute multiple instructions;
[0032] Memory, used to store multiple instructions;
[0033] The plurality of instructions are to be stored in the memory and loaded and executed by the processor as described above.
[0034] Beneficial effects:
[0035] (1) This invention uses historical data of construction projects and deep learning models to analyze the potential relationships between various suppliers in the construction project supply chain, which can more objectively determine the risks.
[0036] (2) The present invention combines graph structure and dynamically adjusts the weights between supplier nodes in the graph structure, making risk identification more accurate.
[0037] (3) This invention can effectively control the risks in the construction project implementation process, provide timely risk warnings, and ensure the smooth implementation of the project.
[0038] (4) The present invention displays risks through the OSP front-end display interface, which is intuitive and easy to read.
[0039] (5) This invention can effectively identify various risks of suppliers. Attached Figure Description
[0040] Figure 1 This is a schematic diagram of the supply chain-based risk control method of the present invention;
[0041] Figure 2 This is a schematic diagram of the risk control device based on the supply chain of the present invention. Detailed Implementation
[0042] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.
[0043] like Figure 1 As shown, this invention proposes a risk control method based on the supply chain, the method comprising:
[0044] Step S1: Obtain historical data of the construction project, extract feature vectors from the historical data, the feature vectors include the start time, planned end time, actual end time, project content, product category, quantity, parameters, and production time of each supplier; input the feature vectors into the trained LSTM network model, the output of the LSTM network model is the time series corresponding to each supply chain, the time series reflects the relationship between each supplier in the supply chain in the time dimension and its impact on the construction project;
[0045] Step S2: Establish a weighted supply chain graph structure for the construction company. The weighted supply chain graph structure includes all suppliers of the construction company, with each supplier as a graph node. Establish connections between graph nodes based on the entire supply chain, and the connections are represented by directed edges. Determine the weights between graph nodes with direct connections based on the entire time series.
[0046] Step S3: Obtain project information of the construction project to be controlled, and determine the risk of the construction project to be controlled based on the weighted supply chain diagram structure and the project information.
[0047] In this invention, since the quality and texture of the products supplied by the supplier are affected by the place of origin and the mining season, the product category, quantity, parameters, and production time of the products supplied by the supplier are taken as factors affecting the construction project, thereby extracting relevant features.
[0048] Each element in the time series is a time segment, which is the sum of the logistics time and production time of its corresponding supplier in the supply chain.
[0049] Step S2, which determines the weights between graph nodes with direct connections based on the entire time series, includes:
[0050] For each graph node: obtain all supply chains corresponding to the graph node; obtain the proportion of the time segment corresponding to the graph node in the time series of each supply chain of the graph node relative to the sum of all time segments; sum the proportions of all supply chains and use them as the weight of the graph node.
[0051] The weights of two graph nodes that have a direct connection are added together to form the weight between the graph nodes that have a direct connection.
[0052] Step S2 further includes: determining the risk factor value of each graph node in the weighted supply chain graph structure, wherein the risk factor value of a graph node is the sum of the weights of the directed edges corresponding to that graph node.
[0053] In step S3, the project information includes the start time, planned end time, project content, and the product categories, quantities, parameters, and production times of the products supplied by each supplier for the construction project.
[0054] Step S3, determining the risk of the construction project to be controlled based on the weighted supply chain graph structure and the project information, includes:
[0055] Step S31: Based on the weighted supply chain graph structure and the project information, determine the weighted critical path of the supplier corresponding to the construction project; obtain the risk factor value of each node in the weighted critical path;
[0056] Step S32: Based on the number of products delivered by suppliers corresponding to each node in the weighted critical path and the funds allocated by the construction company to suppliers corresponding to each node, determine the risk level of the construction project to be controlled.
[0057] Furthermore, for the suppliers corresponding to each node in the weighted critical path, 39 risk monitoring indicators were set up, covering four aspects: fund fraud risk, fund compliance risk, fund liquidity risk, and financial market risk, as shown in Table 1.
[0058] Table 1 Risk Monitoring Indicators
[0059]
[0060] This invention further constructs a multi-dimensional risk assessment system, establishing a comprehensive multi-dimensional risk assessment system that accurately identifies various potential risks. It enables quantitative assessment of construction projects in advance, providing enterprises with comprehensive, accurate, and forward-looking risk assessment results. This invention automatically collects data on multiple risk indicators daily through task statistics, displaying the latest risk data daily, covering the quantity and amount for each business type. Risk items are displayed intuitively. For all risk functions, detailed information can be accessed through linked searches, displaying row-level data, and enabling filtering and deeper data mining visualization.
[0061] This invention relies on the OSP development platform to construct an intuitive and detailed front-end display page. It utilizes the SVR (Service Verification and Registration) service defined by the OSP platform; the back-end organizes data, with data items including risk category, business category, risk control function, risk level, quantity, and amount information, returning the data in a tabular format to the front-end, which displays the information on the OSP front-end interface. It also allows switching between query units to display risk information for different units.
[0062] This invention uses the OSP front-end framework to enable extended data settings for cells. For construction projects with a high risk level, the risk level cell is rendered in red; for construction projects with a medium risk level, the risk level cell is rendered in orange; and for construction projects with a low risk level, the risk level cell is rendered in yellow.
[0063] This invention provides a user interaction interface, where users can click on the risk control function or the quantity to jump to the details page via hyperlinks.
[0064] Furthermore, this invention provides early warnings for incompatibility checks of payment instrument custodians, fund verification status, long-dormant bank accounts, inefficient and invalid accounts, automatic bank account collection limits, external loan bidding execution, complaint leads, overdue margin deposits, overdue accounts receivable, soon-to-expire checks, large-amount private transactions, large-amount payments, guarantee expiration reminders (30 days), credit contract expiration reminders (30 days), letter of guarantee expiration reminders (20 days), comprehensive analysis of deposit and loan status, suspected joint venture / affiliation risk monitoring, suspected fraudulent trade risk monitoring, overseas commission-related business risk warnings, frozen account balance reminders, and available account balance reminders. The system includes 39 risk control functions, such as bond principal repayment reminder (20 days), entrusted loan maturity warning (20 days), reverse factoring maturity repayment reminder (20 days), accounts receivable factoring maturity warning (20 days), letter of credit maturity reminder (10 days), external loan repayment reminder (20 days), external loan interest repayment reminder (20 days), asset securitization maturity warning (20 days), internal loan repayment reminder (10 days), cross-border fund maturity reminder (20 days), liquidity risk reminder, key subsidiary credit monitoring list, bill payment warning, bill maturity reminder (6 days), interest rate change reminder, exchange rate change reminder, supplier risk information monitoring, and customer risk information monitoring.
[0065] Taking the inefficient and invalid account early warning risk control function as an example, it has 100 relevant records. Clicking on the number of records will redirect to the "Inefficient and Invalid Account Early Warning - Joint Inquiry" interface, which displays detailed information based on the OSP platform's GWT component framework. The "General Search" function, also based on the OSP platform framework, can automatically filter data based on column information. The "Export" function, also based on the OSP platform framework, can export system data to a local computer in EXCEL file format. The "Refresh" button can reload the data.
[0066] This invention performs daily data processing at the underlying level, automatically invokes the SVR service, and achieves efficient interaction with the Oracle database. It employs efficient SQL query statements and indexing strategies, significantly improving the speed and accuracy of data retrieval. This service processes early warning information according to the business logic of each risk control function, involving risk category, business category, risk control function, risk level, quantity, and amount information.
[0067] The financial fraud risks of this invention are categorized according to the relevant business categories of the supplier, such as account management, fund concentration, and debt financing; and further categorized according to risk control functions, including payment instrument custodian incompatibility verification warnings, fund verification status warnings, long-dormant bank account reminders, inefficient and invalid account warnings, automatic bank account aggregation limit warnings, and external loan bidding execution status. A query function interface is created through the OSP platform, and data is organized and displayed according to the querying entity through the system's SVR service.
[0068] The financial compliance risks of this invention are categorized according to the relevant business categories of suppliers, such as debt collection, fund settlement, debt financing, and prevention of abnormal business. They are further categorized by risk control functions, including monitoring of complaint leads, monitoring of overdue security deposits, monitoring of overdue accounts receivable, early warning of soon-to-be-due checks, early warning of large-amount private transactions, monitoring of large-amount fund payments, reminders of guarantee expiration (30 days), reminders of credit contract expiration (30 days), reminders of letter of guarantee expiration (20 days), comprehensive analysis of deposit and loan status, monitoring of suspected joint venture / affiliation risks, monitoring of suspected fraudulent trade risks, and early warning of risks related to overseas commission-based businesses. A query function interface is created through the OSP platform, and data is organized and displayed according to the querying entity through the system's SVR service.
[0069] The liquidity risk of this invention is categorized according to business type, such as account management, debt financing, and credit management; and further categorized according to risk control function, including: account frozen balance reminder, account available balance reminder, bond principal repayment reminder (20 days), entrusted loan maturity warning (20 days), reverse factoring maturity repayment reminder (20 days), accounts receivable factoring maturity warning (20 days), letter of credit maturity reminder (10 days), external loan repayment reminder (20 days), external loan interest repayment reminder (20 days), asset securitization maturity warning (20 days), internal loan repayment reminder (10 days), cross-border fund maturity reminder (20 days), and liquidity risk warning. A query function interface is created through the OSP platform, and data is organized and displayed according to the querying unit through the system's SVR service.
[0070] The financial market risks in this invention are categorized by business type, such as bill management, financial markets, and credit management; and further categorized by risk control function, including bill redemption early warning, bill maturity reminder (6 days), interest rate change reminder, exchange rate change reminder, supplier risk information monitoring, and customer risk information monitoring. A query function interface is created through the OSP platform, and data is organized and displayed according to the querying entity through the system's SVR service.
[0071] The OSP platform architecture of this invention provides rich and diverse front-end display effects and flexibility. When handling large-scale tasks or tasks requiring multi-node collaboration, this invention offers powerful distributed task processing advantages and possesses comprehensive task monitoring and management functions. The integration scheme between the OSP platform and Oracle database, along with the data interaction protocol of the SVR service, ensures the accuracy and efficiency of data exchange between systems. This invention features efficient Oracle query algorithms and indexing strategies, providing SQL query algorithms and optimized indexing strategies for fast and accurate data retrieval.
[0072] The present invention also has the following advantages:
[0073] 1. High efficiency and stability: Through the seamless integration of the OSP platform with the Oracle database, as well as optimized data retrieval algorithms and indexing strategies, fast and accurate data processing is achieved, ensuring the system's high efficiency and stability.
[0074] 2. Flexibility: Supports queries from different units to meet users' diverse data query needs.
[0075] 3. Accuracy: Daily automated tasks filter, clean, and organize data, greatly improving the efficiency and accuracy of risk assessment results.
[0076] 4. Data security assurance: Design a comprehensive data collection and preprocessing process to ensure the legality, privacy protection and quality of data, and provide a reliable and secure data source for risk prevention and control.
[0077] 5. Easy to expand and maintain: The system architecture built on the OSP development platform has good scalability and maintainability, which facilitates the expansion of subsequent functions and system upgrades.
[0078] The present invention also provides a supply chain-based risk control device, the device comprising:
[0079] Time series acquisition module: configured to acquire historical data of construction projects, extract feature vectors from the historical data, the feature vectors include the start time, planned end time, actual end time, project content, product category, quantity, parameters, and production time of products supplied by each supplier; input the feature vectors into a trained LSTM network model, the output of which is the time series corresponding to each supply chain, the time series reflecting the relationship between suppliers in the supply chain in the time dimension and their impact on the construction project;
[0080] Graph structure building module: configured to build a weighted supply chain graph structure for the construction company. The weighted supply chain graph structure includes all suppliers of the construction company, with each supplier as a graph node. Connections between graph nodes are established based on the entire supply chain, and these connections are represented by directed edges. The weights between graph nodes with direct connections are determined based on the entire time series.
[0081] Risk determination module: configured to acquire project information of the construction project to be controlled, and determine the risk of the construction project to be controlled based on the weighted supply chain diagram structure and the project information.
[0082] The specific embodiments described above only illustrate the design principles of the present invention. The shapes and names of the components in this description may differ and are not limited. Therefore, those skilled in the art can modify or make equivalent substitutions to the technical solutions described in the foregoing embodiments; and these modifications and substitutions do not depart from the inventive spirit and technical solutions of the present invention, and should all fall within the protection scope of the present invention.
Claims
1. A supply chain-based risk prevention and control method, characterized in that, The method comprises: Step S1: obtaining historical data of a construction project, extracting a feature vector from the historical data, the feature vector comprising a start time, a planned end time, an actual end time, a project content, product categories, product quantities, product parameters and product production times of products supplied by respective suppliers of the construction project; inputting the feature vector into a trained LSTM network model, an output result of the LSTM network model being time series corresponding to respective supply chains, the time series reflecting relationships of the respective suppliers of the supply chains in a time dimension and influences of the respective suppliers on the construction project; Step S2: establishing a supply chain graph structure with weights of a construction enterprise, the supply chain graph structure with weights comprising all suppliers of the construction enterprise, each supplier being taken as a graph node, a connection relationship between the graph nodes being established based on all supply chains, the connection relationship being represented by a directed edge; determining weights between graph nodes having a direct connection relationship based on all time series; Step S3: obtaining project information of a construction project to be prevented and controlled, determining a risk of the construction project to be prevented and controlled based on the supply chain graph structure with weights and the project information.
2. The method of claim 1, wherein, In the step S2, the weights between the graph nodes having the direct connection relationship are determined based on all time series, and the method comprises: for each graph node: obtaining all supply chains corresponding to the graph node; obtaining a proportion of a time segment corresponding to the graph node in a sum of all time segments in a time series corresponding to each supply chain of the graph node; adding the proportions corresponding to all supply chains to obtain a weight of the graph node; adding the respective weights of two graph nodes having the direct connection relationship to obtain a weight between the graph nodes having the direct connection relationship.
3. The method as claimed in claim 2, characterized in that, The step S2 further comprises: determining a risk factor value of each graph node in the supply chain graph structure with weights, wherein the risk factor value of the graph node is a sum of weights of directed edges corresponding to the graph node.
4. The method of any one of claims 1-3, wherein, In the step S3, the project information comprises a start time, a planned end time, a project content, product categories, product quantities, product parameters and product production times of products supplied by respective suppliers corresponding to the construction project.
5. The method of any one of claims 1-3, wherein, In the step S3, the risk of the construction project to be prevented and controlled is determined based on the supply chain graph structure with weights and the project information, and the method comprises: Step S31: determining a critical path with weights of a supplier corresponding to the construction project based on the supply chain graph structure with weights and the project information; obtaining risk factor values of respective nodes in the critical path with weights; Step S32: determining a risk level of the construction project to be prevented and controlled based on product quantities delivered by the respective nodes corresponding to the suppliers in the critical path with weights and funds allocated by the construction company to the respective nodes corresponding to the suppliers.
6. The method of claim 5, wherein, For the respective nodes corresponding to the suppliers in the critical path with weights, risk monitoring indexes in four aspects of fund fraud risk, fund compliance risk, fund liquidity risk and financial market risk are further set.
7. The method of any one of claims 1-3, wherein, The risk is displayed through an OSP front-end display interface.
8. A supply chain-based risk prevention and control device, characterized in that, The device comprises: The time sequence acquisition module is configured to acquire historical data of the construction project, extract a feature vector from the historical data, the feature vector including a start time, a planned end time, an actual end time, a project content, product categories, product quantities, product parameters, and product production times of products supplied by each supplier of the construction project, and input the feature vector into the trained LSTM network model, an output result of the LSTM network model being a time sequence corresponding to each supply chain, the time sequence reflecting relationships of the suppliers of the supply chain in a time dimension and influences of the suppliers on the construction project. The graph structure establishment module is configured to establish a weighted supply chain graph structure of the construction enterprise, the weighted supply chain graph structure including all suppliers of the construction enterprise, each supplier being taken as a graph node, a connection relationship between the graph nodes being established based on all the supply chains, the connection relationship being represented by a directed edge, and weights between the graph nodes having a direct connection relationship being determined based on all the time sequences. The risk determination module is configured to acquire project information of a construction project to be prevented and controlled, and determine a risk of the construction project to be prevented and controlled based on the weighted supply chain graph structure and the project information.
9. A computer-readable storage medium, characterized in that, The storage medium stores a plurality of instructions, and the plurality of instructions are used to load and execute the method in any one of claims 1-7 by a processor.
10. An electronic device, comprising: The electronic device comprises: a processor configured to execute the plurality of instructions; a memory configured to store the plurality of instructions; wherein the plurality of instructions are used to store in the memory, load and execute the method in any one of claims 1-7 by the processor.
Citation Information
Patent Citations
Risk probability assessment for cargo shipment operations and methods of use thereof
CA3219376A1
Supply chain financial risk monitoring method and system based on neural network
CN115907937A