Commodity intelligent purchase and sale matching and risk hedging method and system

Abstract

The invention discloses an intelligent commodity purchase and sale matching and risk hedging method and system, and belongs to the technical field of data processing. Multi-source supply chain data and high-frequency financial market quotation data are received, sliding window filtering is performed on the quotation data to generate a low-frequency risk factor sequence, and the low-frequency risk factor sequence and the supply chain data are subjected to time alignment to form a fusion data set; constructing a supply and demand bipartite graph based on the fused data set, comprehensively considering logistics, capital cost and risk cost obtained by quantification of low-frequency risk factors, solving a global optimal matching scheme meeting delivery constraints, and generating a purchase and sale matching order; and further constructing a dynamic risk assessment model, generating a multi-scene price path by utilizing Monte Carlo simulation in combination with order terms and low-frequency risk factors, outputting potential profit and loss distribution, generating bound derivative hedging suggestions according to the potential profit and loss distribution, and finally forming an integrated transaction risk avoiding scheme. According to the method, collaborative optimization of supply chain performance and financial risks is realized, and the matching efficiency and the anti-risk capability are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, and in particular relates to a method and system for intelligent commodity purchase and sale matching and risk hedging. Background Technology

[0002] In the commodity trade sector, information asymmetry and inefficient supply-demand matching have long plagued buyers and suppliers. Traditional purchase and sale matching relies heavily on manual negotiation or simple information dissemination through information platforms, resulting in slow decision-making processes and difficulty in achieving optimal resource allocation in dynamic market environments. Simultaneously, volatile commodity prices expose both parties to significant price risks. While existing technologies include independent inventory warning systems, platforms connecting purchase and sale information, and theories and practices of using financial derivatives for risk hedging, these solutions are often fragmented. Transaction matching systems typically lack real-time risk assessments deeply integrated with specific transactions; and the use of risk management tools (such as futures and options) requires specialized knowledge from trading parties and complex calculations independent of transaction decisions, leading to high operational barriers, significant lag, and difficulty in accurately matching specific transaction risks. Therefore, deeply integrating intelligent matching of physical transactions with real-time management of financial risks at the algorithmic level, and simultaneously generating customized hedging solutions upon transaction completion, has become a key bottleneck in improving the efficiency and robustness of commodity trade. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for intelligent commodity purchase and sale matching and risk hedging. By constructing a real-time sensing network and intelligent analysis model that includes multi-source supply chain and financial market data, it achieves intelligent optimization matching of purchase and sale orders, dynamic risk simulation in multiple scenarios, and automatic generation and binding of customized derivative hedging suggestions. It deeply integrates traditional isolated transaction matching and risk management at the algorithm level to form an integrated transaction hedging solution. This effectively solves the technical problems of low efficiency in purchase and sale matching, high price volatility risk, and disconnect between the two in commodity trading. It significantly improves the level of intelligent decision-making, accuracy of risk control, and operational stability in commodity trading, and provides technical support for the digital and intelligent transformation of supply chain finance.

[0004] To achieve the above objectives, a first aspect of the present invention provides a method for intelligent commodity purchase and sale matching and risk hedging, comprising the following steps: Receive multi-source supply chain data and financial market data; The financial market data is subjected to sliding window filtering to generate a low-frequency risk factor sequence, and the low-frequency risk factor sequence is time-aligned with the multi-source supply chain data to construct a fused dataset; Based on the fused dataset, a supply-demand bipartite graph is constructed and the comprehensive weight of the edges is calculated. The comprehensive weight includes a cost item and a risk cost item determined based on the low-frequency risk factor sequence. Under the constraints of delivery time, location, commodity specifications and quantity, solve the matching scheme that optimizes the overall weight and generate purchase and sales matching order data. A dynamic risk assessment model is constructed. Based on the purchase and sale matching order data and the low-frequency risk factor sequence, the dynamic risk assessment model generates price paths under several market scenarios through Monte Carlo simulation. Combined with the order delivery terms, it calculates the cash flow aggregation output potential profit and loss distribution data under several scenarios. Based on the output potential profit and loss distribution data, derivative hedging suggestions are generated that are bound to the purchase and sale matching order data. Generate an integrated transaction hedging solution that includes the purchase and sale matching order data and the hedging recommendations.

[0005] Among them, financial market data is high-frequency financial market data.

[0006] Furthermore, the fused dataset includes several procurement demand nodes and several supply offer nodes. The procurement demand nodes include information on the demanded goods, demand quantity, expected delivery location, and time window. The supply offer nodes include information on the supplied goods, supply quantity, deliverable location, and time window. Based on the deliverable locations in the supply offer nodes and the expected delivery locations in the procurement demand nodes, combined with the logistics data in the multi-source supply chain data, the expected logistics costs are calculated. Based on the supply price in the supply offer node, the demand in the procurement demand node, and the preset capital interest rate, calculate the expected capital occupation cost; Based on the analysis of the corresponding commodity price series in the current financial market data, the expected price fluctuation risk cost during the expected order fulfillment period is calculated as the expected risk cost.

[0007] Furthermore, it also includes: All feasible supply and demand node connections in the bipartite graph are used as decision variables, and the value of the decision variable indicates whether the corresponding pair of supply and demand nodes is matched. The optimization objective is to minimize the sum of the overall weights of all matched edges. The constraints are that the total matching quantity of each procurement demand node does not exceed its demand quantity, the total matching quantity of each supply offer node does not exceed its supply quantity, and the product specifications, delivery time and location meet the preset tolerance range. Based on the aforementioned decision variables, optimization objectives, and constraints, a mixed-integer linear programming solver is used to obtain the matching results.

[0008] Furthermore, constructing a dynamic risk assessment model specifically includes: Based on the delivery time window of the purchase and sale matching order data, the drift term of the stochastic process model used to describe the evolution of commodity prices is dynamically adjusted to reflect changes in market expectations during the delivery period. Extract the spot price series and futures price series corresponding to the commodities involved in the purchase and sale matching order data from the financial market data, and calibrate the parameters of the stochastic process model based on the series; Using the calibrated stochastic process model, the delivery terms of the purchase and sale matching order data, and the credit status information of the counterparty as inputs, the Monte Carlo simulation method based on importance sampling is adopted. Only the tail risk paths that may lead to significant losses are sampled and simulated, rather than the entire price path is sampled uniformly. The importance sampling, guided by the adjusted drift term, increases the probability of path generation in extreme market scenarios, thereby reducing computational complexity while ensuring the accuracy of risk assessment.

[0009] Furthermore, the process of outputting the potential profit and loss distribution data includes: Construct a calibrated stochastic process model with importance sampling optimization to generate a set of simulated price paths focusing on the tail risk region, with each path covering the time interval from the current moment to the final delivery date of the order; For each simulated price path, the expected cash flow is calculated at each key performance point on the path based on the delivery terms of the purchase and sale matching order data. By combining the credit status information of the counterparty, a credit risk adjustment is applied to the expected cash flow to obtain the adjusted cash flow; The adjusted cash flows obtained under the key sampling path are discounted and statistically aggregated to generate potential profit and loss distribution data of the purchase and sale matching order data. The distribution is then unbiasedly corrected based on the sampling weights to ensure the statistical consistency of the evaluation results.

[0010] Furthermore, the derivative hedging recommendations linked to the purchase and sale matching order data include: Statistical analysis was performed on the potential profit and loss distribution data to extract its risk measurement indicators under a preset confidence level; Based on the aforementioned risk measurement indicators, determine the type, maturity, and strike price range of the target hedging instrument; Within the specified expiration date and strike price range, select multiple standardized derivative contracts to form an initial hedging portfolio; With the goal of minimizing the risk metrics of the hedged portfolio and meeting the preset maximum hedging cost constraint, the positions of each contract in the initial hedging portfolio are optimized to obtain an optimized position configuration scheme, which constitutes the derivatives hedging recommendation.

[0011] Furthermore, it also includes: Receive confirmation from the user regarding the integrated transaction hedging solution; In response to the confirmation instruction, structured electronic contract data is generated according to the terms of the purchase and sale matching order data, and an electronic signing process is triggered; the derivative hedging suggestion is converted into an order instruction message that conforms to the specific financial transaction system interface protocol; When the potential profit and loss distribution data output by the dynamic risk assessment model exceeds the preset threshold, the purchase and sale matching model is automatically recalculated to generate alternative matching schemes and lock the corresponding financial transaction system interface until the hedging instruction is confirmed.

[0012] Furthermore, the multi-source supply chain data also includes credit data of trading counterparties, which includes publicly available financial information and historical transaction performance information of the enterprises.

[0013] Secondly, the present invention also provides a commodity intelligent purchase and sale matching and risk hedging system, which applies the commodity intelligent purchase and sale matching and risk hedging method as described above, including: The data aggregation module is used to aggregate multi-source supply chain data and financial market data in real time. The multi-source supply chain data includes production capacity data, inventory data, and logistics data, and the financial market data includes commodity spot market data and futures market data. The intelligent matching module is used to generate purchase and sales matching order data for buyers and suppliers based on the aggregated multi-source supply chain data and the financial market data. The intelligent matching module includes: The supply and demand modeling unit is used to model multiple procurement demands and multiple supply offers as nodes on both sides of a bipartite graph. The weight calculation unit is used to calculate the comprehensive weight of the edge connecting the supply and demand nodes for each pair of matching product specifications. The comprehensive weight includes the base price cost of the product, the expected logistics cost, the expected capital occupation cost, and the expected risk cost obtained based on a preliminary risk analysis of the current financial market data. The optimization and solution unit is used to construct and solve a constrained multi-objective optimization model based on the bipartite graph, the comprehensive weights, and the preset hard constraints, so as to obtain the optimal matching scheme. An order generation unit is used to generate the purchase and sale matching order data according to the optimal matching scheme. The risk simulation module is used to construct a dynamic risk assessment model for the purchase and sale matching order data, so as to output the potential profit and loss distribution data of the purchase and sale matching order data under various preset market fluctuation scenarios. The hedging suggestion generation module is used to generate derivative hedging suggestions that are intelligently bound to the purchase and sale matching order data based on the potential profit and loss distribution data; The solution output module is used to generate an integrated transaction hedging solution from the purchase and sale matching order data and the derivative hedging suggestions and output it to the user.

[0014] Thirdly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the above-described intelligent commodity purchase and sale matching and risk hedging methods.

[0015] The beneficial technical effects of the present invention are at least as follows: This invention provides a method and system for intelligent commodity purchase and sale matching and risk hedging. Through deep integration and innovation at the algorithm level, it transforms the traditionally separate processes of physical transaction matching and financial risk hedging into an integrated and synchronized intelligent decision-making process, bringing the following significant benefits to bulk commodity trading: Firstly, it achieves the synchronization and intelligentization of trading decisions and risk management, significantly improving overall operational efficiency and risk controllability. In the traditional model, transaction matching and risk management belong to different stages, resulting in problems such as decision-making lag and information disconnect. This invention constructs a real-time sensing and computing network based on multi-source data. During the same algorithm process of generating purchase and sale matching order data, it calls a dynamic risk assessment model to simulate potential profit and loss distribution data under multiple scenarios, and automatically generates linked derivative hedging recommendations accordingly. This integrated "trading + hedging" output shortens the risk assessment and hedging scheme design process, which previously might have taken hours or even days, to near real-time, greatly improving decision-making efficiency. More importantly, it makes risk control endogenous to the trading decision itself, changing the passive situation of ex-post remediation in risk management, and realizing ex-ante identification, quantification, and proactive management of risks, fundamentally enhancing the robustness and security of transactions.

[0016] Secondly, through precise modeling and optimization algorithms, optimal resource allocation and substantial reduction in overall costs have been achieved, thereby improving market liquidity.

[0017] The core of this invention lies in its intelligent matching algorithm, which goes beyond simple matchmaking. Instead, it constructs a comprehensive optimization model based on a supply-demand bipartite graph, incorporating multi-dimensional costs (including price, logistics, capital occupation, and expected risk costs). This ensures that the matching results not only meet basic supply and demand balance but also optimize overall costs and transaction stability from a global perspective. This achieves better matching of resources such as goods, capital, and logistics across a wider range, effectively reducing transaction friction costs and mismatch losses. Simultaneously, the system automatically generates standardized derivative hedging recommendations precisely corresponding to specific transactions, lowering the barrier for non-professional users to utilize complex financial instruments. This allows more traders to manage risks conveniently and accurately, which helps stimulate market participation and indirectly improves overall market liquidity and operational efficiency.

[0018] In summary, through technological innovation, this invention not only achieves automation and intelligence at the operational level, but more importantly, it reconstructs the relationship between transactions and risks at the business logic level. It provides a systematic technical solution to address long-standing pain points in bulk commodity trading, and has significant practical application value and promising prospects for industry promotion. Attached Figure Description

[0019] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the working steps of a smart commodity purchase and sale matching and risk hedging method disclosed in one embodiment of the present invention;

[0021] Figure 2 This is a schematic diagram of the structure of a smart commodity purchase and sale matching and risk hedging system disclosed in one embodiment of the present invention. Detailed Implementation

[0022] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0023] Example 1 refer to Figure 1 This embodiment provides a method for intelligent commodity purchase and sale matching and risk hedging, including the following steps: S1. Receive multi-source supply chain data and financial market data; S2. Perform sliding window filtering on the financial market data to generate a low-frequency risk factor sequence, and align the low-frequency risk factor sequence with the multi-source supply chain data in time to construct a fused dataset; S3. Based on the fused dataset, construct a supply-demand bipartite graph and calculate the comprehensive weight of the edges. The comprehensive weight includes a cost item and a risk cost item determined based on the low-frequency risk factor sequence. S4. Under the constraints of delivery time, location, commodity specifications and quantity, solve for the matching scheme that optimizes the overall weight and generate purchase and sales matching order data. S5. Construct a dynamic risk assessment model. The dynamic risk assessment model is based on the purchase and sale matching order data and the low-frequency risk factor sequence. It generates price paths under several market scenarios through Monte Carlo simulation, calculates cash flow aggregation output potential profit and loss distribution data under several scenarios in combination with order delivery terms, and generates derivative hedging suggestions bound to the purchase and sale matching order data based on the output potential profit and loss distribution data. S6. Generate an integrated transaction hedging solution that includes the purchase and sale matching order data and the hedging recommendations.

[0024] Furthermore, the fused dataset includes several procurement demand nodes and several supply offer nodes. The procurement demand nodes include information on the demanded goods, demand quantity, expected delivery location, and time window. The supply offer nodes include information on the supplied goods, supply quantity, deliverable location, and time window. Based on the deliverable locations in the supply offer nodes and the expected delivery locations in the procurement demand nodes, combined with the logistics data in the multi-source supply chain data, the expected logistics costs are calculated. Based on the supply price in the supply offer node, the demand in the procurement demand node, and the preset capital interest rate, calculate the expected capital occupation cost; Based on the analysis of the corresponding commodity price series in the current financial market data, the expected price fluctuation risk cost during the expected order fulfillment period is calculated as the expected risk cost.

[0025] In the specific implementation of this invention, the collected high-frequency financial market data (including tick-level or minute-level price sequences of commodity spot and futures) is first subjected to sliding window filtering. An exponentially weighted moving average or a bandwidth-adaptive low-pass filter is used to effectively suppress micro-market noise and extract low-frequency risk factor sequences (such as basis, rolling volatility, and forward curve slope) reflecting medium- to long-term trends and volatility characteristics. Subsequently, this low-frequency risk factor sequence is aligned with multi-source supply chain data (such as capacity planning, inventory changes, and logistics status in transit) according to a unified time dimension (such as daily or delivery window level) to construct a spatiotemporally consistent fusion dataset. This processing not only eliminates the interference of high-frequency financial noise on low-frequency supply chain decision-making but also significantly reduces the computational dimension of subsequent matching optimization and risk modeling, improving the overall system operating efficiency and enabling financial risk signals to be accurately embedded into the rhythm of supply chain business.

[0026] In this embodiment, the comprehensive weight The sum of the base cost of the commodity, expected logistics costs, expected capital occupation costs, and expected price fluctuation risk costs serves as the negative indicator of the objective function. The calculation algorithms and formulas for each component are as follows: Commodity base cost The product of the supply price and the procurement demand at the supply offer node is directly used, i.e. ,in For nodes and Match count (variable to be optimized).

[0027] Expected logistics costs Based on deliverable location With expected delivery location straight-line distance (Calculation of latitude and longitude from logistics platform) and unit freight rate (Based on trunk line + short-distance freight rates, yuan / ton·km), and transportation loss rate s (default 0.3% in the rebar industry), the formula is as follows:

[0028] Expected cost of capital Based on supply price Matching volume Weighted average interest rate (Based on the central bank's LPR and corporate financing costs, take 4.8% / year), and the capital occupation time t (the number of days from order signing to delivery completion, based on the time window). , (Confirmed) Calculation, the formula is: Example: =4000 yuan / ton, =1000 tons, t=30 days, then Yuan.

[0029] Expected price volatility risk cost Volatility calculated based on rebar futures price series (last 90 days) Combining the order fulfillment period T (days) and confidence level (Corresponding to Z value 1.645), the formula is: in Daily volatility of futures prices (calculated using historical prices, in the example) =0.012), where 252 is the number of trading days in a year. Example: T=60 days, then Yuan.

[0030] The final formula for the overall weight is:

[0031] Furthermore, the construction and solution of a constrained multi-objective optimization model includes: All feasible supply and demand node connections in the bipartite graph are used as decision variables, and the value of the decision variable indicates whether the corresponding pair of supply and demand nodes is matched. The optimization objective is to minimize the sum of the overall weights of all matched edges. The constraints are that the total matching quantity of each procurement demand node does not exceed its demand quantity, the total matching quantity of each supply offer node does not exceed its supply quantity, and the product specifications, delivery time and location meet the preset tolerance range. Based on the aforementioned decision variables, optimization objectives, and constraints, a mixed-integer linear programming solver is used to obtain the matching results.

[0032] In this embodiment, let For 0-1 variables, Indicates procurement node With supply nodes match, Indicates a mismatch; For matching variables, satisfying The optimization objective is to minimize the sum of the global comprehensive weights, that is: ,

[0033] Constraints include: Demand constraints: (The total matching quantity at each procurement node shall not exceed its demand); Supply constraints: (The total matching quantity of each supply node shall not exceed its supply quantity); Time constraint: (Delivery time tolerance is ±2 days); Location constraints: (Outside the range is considered an unmatchable match).

[0034] The Gurobi mixed-integer linear programming solver was used, with a solution accuracy of 1e-4 and an iteration limit of 1000. The input decision variables, objective function, and constraints were used to obtain the optimal matching scheme. In the example, the solution yielded 6 matching relationships, covering 98% of all procurement needs, and the sum of global comprehensive weights was reduced by 22% compared to random matching. Based on the optimal matching scheme, standardized purchase and sale matching orders were generated, including information such as the buyer, supplier, product specifications, matching quantity, delivery time and location, price, and logistics method.

[0035] Furthermore, the process of outputting the potential profit and loss distribution data includes: Construct a calibrated stochastic process model with importance sampling optimization to generate a set of simulated price paths focusing on the tail risk region, with each path covering the time interval from the current moment to the final delivery date of the order; For each simulated price path, the expected cash flow is calculated at each key performance point on the path based on the delivery terms of the purchase and sale matching order data. By combining the credit status information of the counterparty, a credit risk adjustment is applied to the expected cash flow to obtain the adjusted cash flow; The adjusted cash flows obtained under the key sampling path are discounted and statistically aggregated to generate potential profit and loss distribution data of the purchase and sale matching order data. The distribution is then unbiasedly corrected based on the sampling weights to ensure the statistical consistency of the evaluation results.

[0036] Furthermore, the generation of a set of contractual derivative hedging suggestions intelligently bound to the purchase and sale matching order data includes: Statistical analysis was performed on the potential profit and loss distribution data to extract its risk measurement indicators under a preset confidence level; Based on the aforementioned risk measurement indicators, determine the type, maturity, and strike price range of the target hedging instrument; Within the specified expiration date and strike price range, select multiple standardized derivative contracts to form an initial hedging portfolio; With the goal of minimizing the risk metrics of the hedged portfolio and meeting the preset maximum hedging cost constraint, the positions of each contract in the initial hedging portfolio are optimized to obtain an optimized position configuration scheme, which constitutes the derivatives hedging recommendation.

[0037] In the dynamic risk assessment phase, this invention abandons the traditional uniform sampling Monte Carlo simulation method and proposes a rapid risk assessment algorithm based on importance sampling. Specifically, according to the delivery time window of the purchase and sale matching orders, the drift term of the stochastic process model describing the evolution of commodity prices (such as geometric Brownian motion or mean regression model) is dynamically adjusted to shift it towards potential extreme market conditions. Under this corrected probability measure, only tail-risk paths that may cause significant losses are sampled with high density, greatly reducing the number of simulated invalid intermediate paths. The cash flow of each sampled path is calculated in conjunction with the order delivery terms and unbiased correction is performed through sampling weights, ultimately aggregating to generate a high-precision potential profit and loss distribution.

[0038] Furthermore, it also includes: Receive confirmation from the user regarding the integrated transaction hedging solution; In response to the confirmation instruction, structured electronic contract data is generated according to the terms of the purchase and sale matching order data, and an electronic signing process is triggered; the derivative hedging suggestion is converted into an order instruction message that conforms to the specific financial transaction system interface protocol; The order placement instruction message is sent to the corresponding financial transaction system.

[0039] Furthermore, the real-time aggregated multi-source supply chain data also includes the credit data of the trading counterparties, which includes publicly available financial information and historical transaction performance information of the enterprises.

[0040] refer to Figure 2 The embodiment also provides a smart commodity purchase and sale matching and risk hedging system, which applies the aforementioned smart commodity purchase and sale matching and risk hedging method, including: The data aggregation module is used to aggregate multi-source supply chain data and financial market data in real time. The multi-source supply chain data includes production capacity data, inventory data, and logistics data, and the financial market data includes commodity spot market data and futures market data. The intelligent matching module is used to generate purchase and sales matching order data for buyers and suppliers based on the aggregated multi-source supply chain data and the financial market data. The intelligent matching module includes: The supply and demand modeling unit is used to model multiple procurement demands and multiple supply offers as nodes on both sides of a bipartite graph. The weight calculation unit is used to calculate the comprehensive weight of the edge connecting the supply and demand nodes for each pair of matching product specifications. The comprehensive weight includes the base price cost of the product, the expected logistics cost, the expected capital occupation cost, and the expected risk cost obtained based on a preliminary risk analysis of the current financial market data. The optimization and solution unit is used to construct and solve a constrained multi-objective optimization model based on the bipartite graph, the comprehensive weights, and the preset hard constraints, so as to obtain the optimal matching scheme. An order generation unit is used to generate the purchase and sale matching order data according to the optimal matching scheme. The risk simulation module is used to construct a dynamic risk assessment model for the purchase and sale matching order data, so as to output the potential profit and loss distribution data of the purchase and sale matching order data under various preset market fluctuation scenarios. The hedging suggestion generation module is used to generate a set of contractual derivative hedging suggestions that are intelligently bound to the purchase and sale matching order data based on the potential profit and loss distribution data. The solution output module is used to generate an integrated transaction hedging solution from the purchase and sale matching order data and the derivative hedging suggestions and output it to the user.

[0041] Thirdly, the embodiments also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the intelligent commodity purchase and sale matching and risk hedging methods described above.

[0042] Finally, it should be noted that the intelligent management method and system for waste compression stations based on the Internet of Things disclosed in the embodiments of the present invention are merely preferred embodiments of the present invention, and are only used to illustrate the technical solutions of the present invention, not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent commodity purchase and sale matching and risk hedging, characterized in that, include: Receive multi-source supply chain data and financial market data; The financial market data is subjected to sliding window filtering to generate a low-frequency risk factor sequence, and the low-frequency risk factor sequence is time-aligned with the multi-source supply chain data to construct a fused dataset; Based on the fused dataset, a supply-demand bipartite graph is constructed and the comprehensive weight of the edges is calculated. The comprehensive weight includes a cost item and a risk cost item determined based on the low-frequency risk factor sequence. Under the constraints of delivery time, location, commodity specifications and quantity, solve the matching scheme that optimizes the overall weight and generate purchase and sales matching order data. A dynamic risk assessment model is constructed. Based on the purchase and sale matching order data and the low-frequency risk factor sequence, the dynamic risk assessment model generates price paths under several market scenarios through Monte Carlo simulation. Combined with the order delivery terms, it calculates the cash flow aggregation output potential profit and loss distribution data under several scenarios. Based on the output potential profit and loss distribution data, derivative hedging suggestions are generated that are bound to the purchase and sale matching order data. Generate an integrated transaction hedging solution that includes the purchase and sale matching order data and the hedging recommendations.

2. The intelligent commodity purchase and sale matching and risk hedging method according to claim 1, characterized in that, The fused dataset includes several procurement demand nodes and several supply offer nodes. The procurement demand nodes include information on the required goods, the required quantity, the expected delivery location, and the time window. The supply offer nodes include information on the supplied goods, the supply quantity, the deliverable location, and the time window. Based on the deliverable locations in the supply offer nodes and the expected delivery locations in the procurement demand nodes, combined with the logistics data in the multi-source supply chain data, the expected logistics costs are calculated. Based on the supply price in the supply offer node, the demand in the procurement demand node, and the preset capital interest rate, calculate the expected capital occupation cost; Based on the analysis of the corresponding commodity price series in the current financial market data, the expected price fluctuation risk cost during the expected order fulfillment period is calculated as the expected risk cost.

3. The intelligent commodity purchase and sale matching and risk hedging method according to claim 2, characterized in that, Also includes: All feasible supply and demand node connections in the bipartite graph are used as decision variables, and the value of the decision variable indicates whether the corresponding pair of supply and demand nodes is matched. The optimization objective is to minimize the sum of the overall weights of all matched edges. The constraints are that the total matching quantity of each procurement demand node does not exceed its demand quantity, the total matching quantity of each supply offer node does not exceed its supply quantity, and the product specifications, delivery time and location meet the preset tolerance range. Based on the aforementioned decision variables, optimization objectives, and constraints, a mixed-integer linear programming solver is used to obtain the matching results.

4. The intelligent commodity purchase and sale matching and risk hedging method according to claim 1, characterized in that, The construction of a dynamic risk assessment model specifically includes: Based on the delivery time window of the purchase and sale matching order data, the drift term of the stochastic process model used to describe the evolution of commodity prices is dynamically adjusted to reflect changes in market expectations during the delivery period. Extract the spot price series and futures price series corresponding to the commodities involved in the purchase and sale matching order data from the financial market data, and calibrate the parameters of the stochastic process model based on the series; Using the calibrated stochastic process model, the delivery terms of the purchase and sale matching order data, and the credit status information of the counterparty as inputs, the Monte Carlo simulation method based on importance sampling is adopted. Only the tail risk paths that may lead to significant losses are sampled and simulated, rather than the entire price path is sampled uniformly. The importance sampling, guided by the adjusted drift term, increases the probability of path generation in extreme market scenarios, thereby reducing computational complexity while ensuring the accuracy of risk assessment.

5. The intelligent commodity purchase and sale matching and risk hedging method according to claim 4, characterized in that, The process of outputting potential profit and loss distribution data includes: Construct a calibrated stochastic process model with importance sampling optimization to generate a set of simulated price paths focusing on the tail risk region, with each path covering the time interval from the current moment to the final delivery date of the order; For each simulated price path, the expected cash flow is calculated at each key performance point on the path based on the delivery terms of the purchase and sale matching order data. By combining the credit status information of the counterparty, a credit risk adjustment is applied to the expected cash flow to obtain the adjusted cash flow; The adjusted cash flows obtained under the key sampling path are discounted and statistically aggregated to generate potential profit and loss distribution data of the purchase and sale matching order data. The distribution is then unbiasedly corrected based on the sampling weights to ensure the statistical consistency of the evaluation results.

6. The intelligent commodity purchase and sale matching and risk hedging method according to claim 1, characterized in that, The generation of derivative hedging recommendations linked to the purchase and sale matching order data also includes: Statistical analysis was performed on the potential profit and loss distribution data to extract its risk measurement indicators under a preset confidence level; Based on the aforementioned risk measurement indicators, determine the type, maturity, and strike price range of the target hedging instrument; Within the specified expiration date and strike price range, select multiple standardized derivative contracts to form an initial hedging portfolio; With the goal of minimizing the risk metrics of the hedged portfolio and meeting the preset maximum hedging cost constraint, the positions of each contract in the initial hedging portfolio are optimized to obtain a position configuration scheme, which constitutes the derivative hedging recommendation.

7. The intelligent commodity purchase and sale matching and risk hedging method according to claim 1 or 6, characterized in that, Also includes: Receive confirmation from the user regarding the integrated transaction hedging solution; In response to the confirmation instruction, structured electronic contract data is generated according to the terms of the purchase and sale matching order data, and an electronic signing process is triggered; The derivative hedging recommendations are converted into order placement instruction messages that conform to the specific financial trading system interface protocol; When the potential profit and loss distribution data output by the dynamic risk assessment model exceeds the preset threshold, the purchase and sale matching model is automatically recalculated to generate alternative matching schemes and lock the corresponding financial transaction system interface until the hedging instruction is confirmed.

8. The intelligent commodity purchase and sale matching and risk hedging method according to claim 1, characterized in that, The multi-source supply chain data also includes credit data of counterparties, which includes publicly available financial information and historical transaction performance information of the companies.

9. A smart commodity purchase and sale matching and risk hedging system, applying the smart commodity purchase and sale matching and risk hedging method as described in claims 1-8, characterized in that, include: The data aggregation module is used to aggregate multi-source supply chain data and financial market data in real time. The multi-source supply chain data includes production capacity data, inventory data, and logistics data, and the financial market data includes commodity spot market data and futures market data. The intelligent matching module is used to generate purchase and sales matching order data for buyers and suppliers based on the aggregated multi-source supply chain data and the financial market data. The intelligent matching module includes: The supply and demand modeling unit is used to model multiple procurement demands and multiple supply offers as nodes on both sides of a bipartite graph. The weight calculation unit is used to calculate the comprehensive weight of the edge connecting the supply and demand nodes for each pair of matching product specifications. The comprehensive weight includes the base price cost of the product, the expected logistics cost, the expected capital occupation cost, and the expected risk cost obtained based on a preliminary risk analysis of the current financial market data. The optimization and solution unit is used to construct and solve a constrained multi-objective optimization model based on the bipartite graph, the comprehensive weights, and the preset hard constraints, so as to obtain the optimal matching scheme. An order generation unit is used to generate the purchase and sale matching order data according to the optimal matching scheme. The risk simulation module is used to construct a dynamic risk assessment model for the purchase and sale matching order data, so as to output the potential profit and loss distribution data of the purchase and sale matching order data under various preset market fluctuation scenarios. The hedging suggestion generation module is used to generate derivative hedging suggestions that are intelligently bound to the purchase and sale matching order data based on the potential profit and loss distribution data; The solution output module is used to generate an integrated transaction hedging solution from the purchase and sale matching order data and the derivative hedging suggestions and output it to the user.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent commodity purchase and sale matching and risk hedging method as described in any one of claims 1 to 8.

Images