Supply chain dynamic risk control method and system based on smart contract

Abstract

The invention belongs to the cross technical field of block chain technology and supply chain finance, and provides a supply chain dynamic risk control method and system based on a smart contract, and the technical scheme is as follows: obtaining node data of each participant in a supply chain; dynamically evaluating the risk of the supply chain in combination with the obtained node data of each participant and the constructed smart contract; wherein in the basic contract, the logistics data and the invoice data are verified respectively, and a verification result is sent to the business contract after verification is passed; in a business contract, according to received enterprise data, introducing a timestamp mechanism, recording creditor's right generation time in a block chain, and realizing multi-level creditor's right splitting and merging through a Merkle tree; in the risk control contract, calculating a risk assessment index based on the on-chain data, the off-chain data and the risk assessment model; and adjusting the credit granting strategy in real time according to the change of the risk assessment index to obtain an adjusted credit granting strategy. Risk control is embedded into contract execution logic, so that the risk loss of the supply chain is reduced.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of cross technology of blockchain and supply chain finance, and particularly relates to a supply chain dynamic risk control method and system based on a smart contract. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute prior art.

[0003] The current supply chain business mainly relies on traditional centralized systems, which has the following defects: First, information silos lead to ineffective credit transmission, and core enterprise credit is difficult to effectively penetrate to multiple suppliers. Trade data, logistics information, and fund flow of upstream and downstream enterprises are scattered in isolated systems such as banks, core enterprise ERP, and third-party logistics, and data cannot be verified in real time across agencies (such as invoice authenticity, order status). Financial institutions need to invest a lot of manpower in cross-verification, and there are still information lag and authenticity risks.

[0004] Second, manual processes cause efficiency bottlenecks. Financing applications involve contract signing, invoice auditing, mortgage registration, and fund transfer, and are highly dependent on offline manual operations, resulting in low efficiency of small and medium-sized enterprise capital turnover and inability to meet the demand for instant financing.

[0005] Third, fraud risk is high. In the traditional mode, it is difficult to eliminate fraudulent behaviors such as fake trade background (such as fake orders / invoices) and repeated mortgage (the same invoice is financed in multiple institutions).

[0006] Although there are currently corresponding solutions, single-institution data integration solutions (such as bank connection to core enterprise ERP) can only cover part of the supply chain and cannot solve the problem of multi-subject data silos; third-party platform centralized verification has data tampering risks and needs to bear high system operation and maintenance costs; traditional blockchain storage applications only realize data chaining and are not deeply coupled with credit logic, and still rely on offline approval. SUMMARY

[0007] In order to solve at least one of the technical problems in the background art, the present application provides a supply chain dynamic risk control method and system based on a smart contract, which establishes a cross-agency trusted data exchange mechanism, constructs an end-to-end automated credit process combined with a smart contract, dynamically identifies fraud patterns, and embeds risk control into the contract execution logic to reduce the risk loss of the supply chain.

[0008] In order to achieve the above purpose, the present application adopts the following technical solutions: The first aspect of the present application provides a supply chain dynamic risk control method based on a smart contract, comprising the following steps: acquiring node data of each participant in the supply chain; combining the acquired node data of each participant and the constructed smart contract to dynamically evaluate the risk of the supply chain; wherein, in the basic contract, the logistics data and the invoice data are verified respectively, and the verification result is sent to the business contract after verification; in the business contract, according to the received enterprise data, a timestamp mechanism is introduced, the generation time of the right is recorded in the blockchain, and multi-level right splitting and merging is realized through Merkle tree; in the risk control contract, the risk evaluation index is calculated based on on-chain data, off-chain data and risk evaluation model; According to the change of the risk evaluation index, the credit strategy is adjusted in real time to obtain the adjusted credit strategy.

[0009] Further, the method further comprises encrypting the enterprise node data based on the zero-knowledge proof of Groth16 algorithm.

[0010] Further, in the risk control contract, the risk evaluation index is calculated based on on-chain data, off-chain data and risk evaluation model, specifically comprising: extracting time series features and derived features based on the processed on-chain data and off-chain data; training XGBoost in combination with the extracted time series features and derived features to obtain trained XGBoost; Based on the trained XGBoost, the prediction is performed on the to-be-predicted object, and the continuous value output by the XGBoost is converted into a risk evaluation index.

[0011] Further, the time series features include volatility of historical repayment records and moving average of transaction frequency; wherein, the volatility of historical repayment records is used to reflect the stability of repayment, and the moving average of transaction frequency is used to eliminate the influence of short-term fluctuations; the derived features include the ratio index of transaction amount and enterprise asset scale and the standard deviation of associated enterprise credit score; wherein, the standard deviation of associated enterprise credit score is used to measure the dispersion degree of associated risk.

[0012] Further, the combination of the extracted time series features and derived features to train XGBoost to obtain trained XGBoost, comprising: for each parameter in the XGBoost model, the XGBoost model learns the non-linear relationship between the time series features, the derived features and the default rate by integrating multiple decision trees, each tree is recursively split according to the time series features and the derived features, and finally the leaf node corresponds to a default probability, the prediction results of all trees are weighted and summed and converted into the final default rate through the sigmoid function.

[0013] Further, the adjusting the credit policy according to the change of the risk index obtains an adjusted credit policy, including: initializing the Markov decision process model, including defining a state space, an action space and designing a reward function; adjusting the interest rate according to the current risk state and the optimal strategy in the Markov decision, judging whether the adjustment range exceeds the limit, if exceeding the limit, adjusting the interest rate to the maximum limit range, otherwise adjusting the interest rate based on the calculated interest rate; judging whether the adjustment frequency exceeds the limit, if not exceeding the limit, restoring the original interest rate and not performing the current adjustment, otherwise performing the interest rate adjustment and recording the adjustment time.

[0014] The second aspect of the present application provides a supply chain dynamic risk control system based on a smart contract, including: a data acquisition module for acquiring node data of each participant in the supply chain; a risk assessment module for combining the acquired node data of each participant and the constructed smart contract to dynamically assess the risk of the supply chain; wherein in the basic contract, the logistics data and the invoice data are verified respectively, and the verification result is sent to the business contract after verification; in the business contract, according to the received enterprise data, a timestamp mechanism is introduced, the generation time of the right is recorded in the blockchain, and multi-level right splitting and merging is realized through the Merkle tree; in the risk control contract, the risk assessment index is calculated based on the on-chain data, off-chain data and risk assessment model; a risk control module for adjusting the credit policy in real time according to the change of the risk assessment index to obtain an adjusted credit policy.

[0015] The third aspect of the present application provides a computer readable storage medium.

[0016] A computer readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the steps in the supply chain dynamic risk control method based on a smart contract as described above.

[0017] The fourth aspect of the present application provides a computer device.

[0018] A computer device includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor executes the program to implement the steps in the supply chain dynamic risk control method based on a smart contract as described above.

[0019] Compared with the prior art, the present application has the following advantages: 1、The application combines the constructed smart contract, combines the contract after trade background verification, in the risk control contract, combines the risk of supply chain through trade background verification and real-time risk monitoring dynamic assessment, and adjusts the credit strategy in real time, effectively prevents fraudulent behaviors such as false trade background and repeated pledge, and reduces the fraud loss rate of supply chain finance.

[0020] 2、The application uses zero-knowledge proof (ZKP) technology to realize the encryption verification of enterprise sensitive data, and protects the privacy security of enterprises.

[0021] 3、The application realizes the sharing and interaction of data through blockchain technology, breaks the information barrier between each participant, and improves the authenticity and reliability of data.

[0022] The advantages of the additional aspects of the application will be partially given in the following description, partially will become obvious from the following description, or will be known by the practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0023] The drawings accompanying the specification of the application are used to provide further understanding of the application, the illustrative embodiments of the application and the description thereof are used to explain the application, and do not constitute improper limitation on the application.

[0024] Figure 1 is a flow chart of a supply chain dynamic risk control method based on a smart contract provided by the embodiment of the application; Figure 2 is an encryption process based on zero-knowledge proof of Groth16 algorithm provided by the embodiment of the application; Figure 3 is a risk assessment process based on XGBoost provided by the embodiment of the application; Figure 4 is a dynamic interest rate adjustment process based on Markov decision process provided by the embodiment of the application. DETAILED DESCRIPTION

[0025] The application will be further described below in combination with the drawings and embodiments.

[0026] It should be pointed out that the following detailed description is all exemplary, and is intended to provide further description of the application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as generally understood by those skilled in the art to which the application belongs.

[0027] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0028] Example 1 like Figure 1 As shown, this embodiment provides a supply chain dynamic risk control method based on smart contracts, including the following steps: Step 1: Construct nodes for each participant in the supply chain and obtain node data for each participant; In this embodiment, the various participants in the supply chain include financial institutions, core enterprises, logistics companies, and regulatory agencies. Each participant's node is constructed using a Fabric consortium blockchain. The Fabric consortium blockchain offers high scalability and flexibility, capable of meeting the needs of different participants in supply chain finance. Through the consortium blockchain, nodes can share and interact with data, ensuring the authenticity and reliability of the data. In this embodiment, the node data of each participant specifically includes: obtaining relevant logistics data based on the logistics company's nodes, including logistics tracking numbers, logistics trajectory data, and logistics transportation nodes, such as loading, transshipment, and unloading; In this embodiment, a channel mechanism is used to achieve data isolation between nodes, establishing dedicated communication channels for financial institutions, core enterprises, logistics companies, and regulatory agencies. An endorsement policy is used to set multi-node joint signature rules; for example, financing operations require dual signature confirmation from both the core enterprise and the financial institution. The PBFT (Practical Byzantine Fault Tolerance) algorithm is employed, in 3 f +1 ( f With a node configuration that allows for a maximum number of Byzantine nodes, block confirmation is achieved within 2 seconds, ensuring real-time transaction performance.

[0029] Step 2: Encrypt sensitive corporate data; This embodiment uses zero-knowledge proofs based on the Groth16 algorithm for encryption. Groth16 typically uses specific elliptic curves, such as the Bayer-Neven (BN) curve, which is a type of elliptic curve that satisfies the bilinear pairwise friendly property. Common parameters for generating the elliptic curve system include the base point G of the elliptic curve. These parameters will serve as the basis for subsequent calculations, and both the prover and the verifier need to know these common parameters.

[0030] Please refer to Figure 2, specifically comprising the following steps: Step 201: Key generation: Generate the proof key pk and the verification key vk through a trusted setup process. The proof key pk is held by the prover and used to generate zero-knowledge proofs; the verification key vk is held by the verifier and used to verify the validity of the proof. Commitment value generation: The prover wants to mask the real amount x, and uses the commitment scheme to generate the commitment value C.

[0031] The commitment scheme is generally based on the elliptic curve discrete logarithm problem, for example, using Pedersen commitment C = rG + xH, where r is a random number, G and H are generators of the elliptic curve. The commitment value C generated in this way hides the real amount x, while having binding and hiding properties. Auxiliary value calculation: The prover uses elliptic curve parameters to perform related calculations to obtain some auxiliary calculation results. These calculations may involve elliptic curve point addition, multiplication, etc., for subsequent proof construction.

[0032] Step 202: Generate proof π: The prover uses the proof key pk, combines the auxiliary values calculated above and the commitment value C, and generates a zero-knowledge proof π according to the specific rules of the Groth16 algorithm. This process encodes the statements to be proven (such as the transaction amount meeting certain ranges, etc.) into the proof. Proof verification receives the commitment value C and the proof π sent by the prover. Elliptic curve point operation: The verifier performs elliptic curve point operations, and through elliptic curve bilinear pairing operations (such as using the Miller algorithm and the final exponential operation to calculate the bilinear pairing), the relevant values in the proof π and the commitment value C are substituted into the verification equation.

[0033] Step 203: Bilinear pairing operation is the key to the Groth16 verification process, which can efficiently verify whether the proof satisfies the pre-set relationship. Verification result return: According to the result of the bilinear pairing operation, it is judged whether the proof is valid. If the verification equation is established, the verification result is returned, indicating that the statement proved by the prover is true and no sensitive information is leaked; if the verification equation is not established, the verification fails.

[0034] Step 3: Combine the node data of each participant obtained and the smart contract constructed to dynamically assess the risk of the supply chain, and obtain the corresponding transaction strategy according to the risk level; Specifically comprising the following steps: Step 301, in the base contract, respectively verify the logistics data and invoice data, and send the verification result to the business contract after verification; Specifically, when verifying invoice data, a double-hash verification mechanism is adopted, SHA256 hash value of the original invoice file is calculated, and cross verification is performed with the RIPEMD160 hash value stored in the blockchain. If they are the same, the verification is passed, otherwise an automatic early warning is issued. When verifying logistics data, the logistics verification protocol is combined, the third-party logistics API is interfaced, the logistics track is obtained through the OpenAPI call of the smart contract, the state machine model is used to verify the transportation node sequence (loading → transfer → unloading), and if they are consistent, the verification is passed, otherwise the abnormal node triggers an automatic early warning.

[0035] In the basic contract, the authenticity of the trade background can be automatically verified, such as invoice hash value comparison and logistics number verification. Through the comparison of invoice hash value, the authenticity and integrity of the invoice can be ensured; through the verification of logistics number, the actual transportation of goods can be verified, thereby effectively preventing the risk of false trade background.

[0036] Step 302, in the business contract, according to the received enterprise data, the timestamp mechanism is introduced, the time of generating the right is recorded in the blockchain, and the multi-level right splitting and merging is realized through the Merkle tree; The business contract mainly includes the following data: Debt subject information, accounts receivable core data, financing progress data and fund flow data; Among them, the debt subject information includes the identity of the creditor (enterprise / individual) and the debtor (the party in arrears) (such as blockchain address, enterprise code processed by hash), ownership proof (digital signature public key, etc.; clearly financing participants, through smart contract verification "right attribution", to ensure that only the legal creditor initiates financing.

[0037] The financing progress data includes right confirmation status (to be confirmed, confirmed), loan status (to be loaned, loaned), and repayment status (to be repaid, partially repaid, cleared; The fund flow data includes loan amount, repayment amount, and fund flow address; Specifically, the following steps are included: Step 3021, accounts receivable right confirmation stage: Merkle tree splitting evidence, associate sub-debt with total debt, combine timestamp hash and zero-knowledge proof, blockchain automatically obtains on-chain timestamp (verify debt time limit), historical transaction hash (prevent duplicate right confirmation); Step 3022, financing loan stage: creditor submits loan request (sub-debt ID, collection address), zero-knowledge proof (verify debt compliance again); fund provider submits zero-knowledge proof (prove "sufficient funds"); smart contract associates confirmed debt data and fund provider pre-authorization amount.

[0038] For example, 1 million receivables can be split into 10 sub-debt of 100,000 respectively for financing.

[0039] Atomic Swap technology is adopted to ensure the atomicity of loan and right confirmation operations, avoiding the risk of inconsistency between funds and receivables.

[0040] The receivables right confirmation, financing loan, and repayment clearing operations are executed through the business contract. The financing process can be automatically processed, reducing manual intervention and improving financing efficiency. At the same time, through the execution of the smart contract, the transparency and traceability of the financing process can be ensured.

[0041] Step 303, in the risk control contract, calculate the risk indicators based on on-chain data, off-chain data, and risk assessment model; As shown in Figure 3 The specific steps include: Step 3031, obtain on-chain data and off-chain data; In this embodiment, the on-chain data includes historical repayment records, transaction frequency, and associated enterprise credit stored in the on-chain database; The off-chain data can be obtained through the Oracle (Oracle) to obtain enterprise credit investigation and judicial litigation data; Step 3032, preprocess the obtained on-chain data and off-chain data; In this embodiment, preprocessing the obtained on-chain data and off-chain data includes on-chain data cleaning and off-chain data fusion; The on-chain data cleaning specifically includes: Historical repayment records: convert the repayment timestamp to the number of overdue days, for example, if the actual repayment time is 10 days later than the agreed time, the number of overdue days is 10; Transaction frequency: count the number of transactions in the past 30 days, and if the number of transactions is less than 50% of the industry average, normalize the data; Associated enterprise credit: extract the on-chain credit score of the associated enterprise, and if there are multiple associated enterprises, take the lowest score; The off-chain data fusion specifically includes: After obtaining the enterprise credit report through the Oracle, extract core indicators such as asset-liability ratio and liquidity ratio; The judicial litigation data is binarized, with a record of 1 for litigation and 0 otherwise; Missing values are handled using the multiple imputation by chained equations (MICE) method, which iteratively predicts missing values using chain equations; Step 3033, extract time series features and derived features based on the processed on-chain data and off-chain data; In this embodiment, the time series features include the volatility of historical repayment records and the moving average of transaction frequency; Wherein, the volatility of historical repayment records is used to reflect the stability of repayment, and the moving average of transaction frequency is used to eliminate the influence of short-term fluctuations; The derived features include the ratio of transaction amount to enterprise asset size and the standard deviation of associated enterprise credit scores; wherein the standard deviation of associated enterprise credit scores is used to measure the degree of diversification of associated risks; Specifically, the calculation of the volatility of historical repayment records: the historical repayment records can be regarded as a time series. When calculating the volatility, first calculate the adjacent repayment time intervals, and then calculate the standard deviation of these time intervals as the volatility of the repayment records. The larger the standard deviation, the greater the fluctuation of the repayment time, the worse the stability of the repayment, and the higher the credit risk. For example, assuming the repayment time interval sequence is [5, 8, 10, 6, 9], first calculate the mean value as (5 + 8 + 10 + 6 + 9) / 5 = 7.6, and then calculate the standard deviation as sqrt (((5 - 7.6)^2 + (8 - 7.6)^2 + (10 - 7.6)^2 + (6 - 7.6)^2 + (9 - 7.6)^2) / 5) ≈ 2.07.

[0042] Moving average of transaction frequency calculation: the moving average of transaction frequency can be calculated using the sliding window method. For example, set the window size to 5, for the transaction frequency sequence [3, 5, 4, 6, 7, 8], the first moving average is (3 + 5 + 4) / 3 = 4, the second moving average is (5 + 4 + 6) / 3 = 5, and so on. The moving average can smooth out short-term fluctuations and more clearly reflect the long-term trend of transaction frequency.

[0043] Derived features: Construction of transaction amount to enterprise asset size ratio indicator: obtain enterprise asset size data from off-chain data and calculate transaction amount from on-chain transaction records. Then calculate the ratio of the two, the formula is: transaction amount to asset size ratio = transaction amount / enterprise asset size. This indicator can reflect the relationship between the transaction activity of the enterprise and the asset size, the higher the ratio, the higher the operating risk of the enterprise. Calculation of standard deviation of associated enterprise credit scores: after obtaining the credit scores of multiple associated enterprises, calculate the standard deviation of these scores. The larger the standard deviation, the greater the difference in credit scores of associated enterprises, the higher the degree of diversification of associated risks, and it may also imply that the enterprise faces more complex associated risks.

[0044] Step 3034, training of XGBoost combined with extracted time series features and derived features is performed to obtain the trained XGBoost; In this embodiment, during training, the Bayesian optimization algorithm is used to search for the best parameter combination of XGBoost, including n_estimators (number of trees), max_depth (maximum depth of tree), learning_rate (learning rate), in combination with time series features and derived features. Different parameters indirectly affect the calculation accuracy of default rates by affecting the model structure and training process: Among them, max_depth: controls the depth of the tree, too deep may lead to overfitting (capturing noise), too shallow may underfit (ignoring key feature relationships). subsample and colsample_bytree: enhance generalization ability through random sampling, especially for high-dimensional derived features (such as volatility, standard deviation) have regularization effect. min_child_weight: limit the minimum sample weight of leaf node, avoid excessive sensitivity to a small number of samples, improve stability.

[0045] Finally, 5-fold cross-validation is used to evaluate the performance of the model, and the evaluation indicators include mean squared error (MSE), mean absolute error (MAE), and coefficient of determination (R-squared) ); Specifically, during training, for each parameter, the XGBoost model learns the non-linear relationship between time series features, derived features, and default rates by integrating multiple decision trees. Each tree recursively splits based on time series features and derived features, and the final leaf node corresponds to a default probability. The prediction results of all trees are weighted and summed (the weight is controlled by learning_rate), and then converted to the final default rate through the sigmoid function.

[0046] Step 3035, based on the trained XGBoost, the prediction of the to-be-predicted object is performed to obtain the risk prediction result; The output of XGBoost is a continuous value, and the continuous value is converted to a probability through logistic regression. The converted default rate is compared with the set threshold value, and if it is greater than the set threshold value, it is determined as high risk. In this embodiment, the risk threshold is set to 0.3; The model converts the output through the sigmoid function to a probability value between 0 and 1, which directly represents the probability of the object defaulting within a certain period of time (such as 1 year). For example: an output of 0.85 indicates an 85% probability of default.

[0047] When the financial institution issues a loan, it analyzes the distribution of "default samples" in historical data to preset a risk threshold. If the delinquency rate exceeds 5%, it will result in a loss, and the default probability ≥ 5% can be set as a high-risk threshold.

[0048] Step 4: Real-time adjustment of credit strategy according to the change of risk indicators to obtain an adjusted credit strategy; As shown in the following steps: Figure 4 Step 401, initialize the Markov decision process model; Specifically, it includes: Step 4011, define the state space; In this embodiment, the state space includes different levels of risk, specifically, the risk levels include 5 levels: A (low risk), B (low-medium risk), C (medium risk), D (high-medium risk), and E (high risk). Each state contains risk score, historical interest rate, repayment record, and other information.

[0049] Step 4012, define the action space, including the definition of interest rate adjustment action and credit strategy action; In this embodiment, the interest rate adjustment action is defined as: [-300BP, -150BP, 0, 150BP, 300BP]; The credit strategy action is defined as: [maintain the limit, increase the limit, decrease the limit, freeze the limit]; Step 4013, design the reward function including positive reward and negative reward; Wherein, the positive reward is +10 points reward for on-time repayment; the negative reward is -20 points penalty for overdue repayment; the interest rate adjustment reward is given according to the adjusted income; Step 4014, combine the defined state space, action space, and designed reward function to build the Markov decision process model; Step 402, trigger model recalculation according to each repayment data of the user, and dynamically adjust to obtain the final credit strategy; Specifically, it includes the following steps: Step 4021, adjust the interest rate according to the current risk state and the optimal strategy, judge whether the adjustment amplitude exceeds the limit, if it exceeds the limit, adjust the interest rate to the maximum limit amplitude, otherwise adjust based on the calculated interest rate; ​In this embodiment, the optimal policy adjustment interest rate is the optimal policy in Markov decision process (MDP), which is the policy that can maximize the expected cumulative reward of the decision maker in the long run among all possible policies. It is the core goal of MDP solution, which embodies the "optimal action selection rule" in each state.

[0050] The optimal policy of MDP is solved by value iteration algorithm, that is, the action combination that can maximize the cumulative reward in each state: The core logic of value iteration algorithm: value iteration updates the state value function V (s) by iteration until convergence, and finally derives the optimal policy ; Example: When the state s is B level (medium-low risk), the optimal action may be "-150BP (interest rate reduction) + increase the quota", because this combination can bring the maximum cumulative reward of customer retention and revenue growth; When the state s is D level (medium-high risk), the optimal action may be "+300BP (interest rate increase) + reduce the quota" to cover the risk and control the loss.

[0051] Step 4022, judge whether the adjustment frequency exceeds the limit, if not, restore the original interest rate, do not adjust this time, otherwise, perform interest rate adjustment and record the adjustment time.

[0052] Embodiment two The embodiment provides a supply chain dynamic risk control device based on a smart contract, comprising: A data acquisition module is configured to acquire node data of each participant in the supply chain; A risk assessment module is configured to combine the acquired node data of each participant and the constructed smart contract to dynamically assess the risk of the supply chain; wherein in the basic contract, the logistics data and the invoice data are verified, and the verification result is sent to the business contract after verification; in the business contract, according to the received enterprise data, a timestamp mechanism is introduced, the generation time of the right is recorded in the blockchain, and multi-level right splitting and merging is realized through Merkle tree; in the risk control contract, the risk assessment index is calculated based on on-chain data, off-chain data and a risk assessment model; A risk control module is configured to adjust the credit strategy in real time according to the change of the risk assessment index to obtain an adjusted credit strategy.

[0053] It should be noted that the specific implementation mode of the supply chain dynamic risk control device based on the smart contract in the embodiment of the present application is similar to the specific implementation mode of the supply chain dynamic risk control method based on the smart contract in the embodiment of the present application, and specific reference can be made to the description in the method part. In order to reduce redundancy, this part will not be repeated here.

[0054] Embodiment three The embodiment provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the steps in the supply chain dynamic risk control method based on the smart contract.

[0055] Embodiment four The embodiment provides a computer device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor realizes the steps in the supply chain dynamic risk control method based on the smart contract when executing the program.

[0056] Embodiment five The embodiment provides a program product, which is a computer program product, comprising a computer program, and the computer program is executed by a processor to realize the steps in the supply chain dynamic risk control method based on the smart contract.

[0057] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can be in the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to magnetic disk storage and optical storage, etc.) containing computer usable program code.

[0058] The present application is described with reference to flowcharts and / or block diagrams of the method, device (system), and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be realized by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to produce a machine, so that the instructions executed by the computer or other programmable data processing devices produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The function of one flow or multiple flows and / or blocks Figure 1 The device for realizing the function specified in one block or multiple blocks.

[0059] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the flow Figure 1 one or more flow or block Figure 1 one or more blocks or blocks specified in the flow.

[0060] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions executed on the computer or other programmable apparatus provide steps for implementing the flow Figure 1 one or more flow or block Figure 1 one or more blocks or blocks specified in the flow.

[0061] Those of ordinary skill in the art can understand that all or part of the flow of the above-mentioned embodiment method can be completed by computer program instructions to instruct the relevant hardware, and the program can be stored in a computer readable storage medium. When the program is executed, it can include the flow of the above-mentioned embodiment of each method. Among them, the storage medium can be a magnetic disc, an optical disc, a read-only memory (ROM) or a random access memory (RAM) and the like.

[0062] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A supply chain dynamic risk control method based on smart contracts, characterized in that, include: Acquire node data from each participant in the supply chain; By combining the node data acquired from each participant with the constructed smart contracts, the risks of the supply chain are dynamically assessed. Specifically, in the basic contract, logistics data and invoice data are verified separately, and the verification results are sent to the business contract after successful verification. In the business contract, based on the received enterprise data, a timestamp mechanism is introduced to record the time of debt generation on the blockchain, and multi-level debt splitting and merging are realized through Merkle trees. In the risk control contract, risk assessment indicators are calculated based on on-chain data, off-chain data, and risk assessment models. The adjusted credit strategy is obtained by adjusting the credit granting strategy in real time based on changes in risk assessment indicators.

2. The supply chain dynamic risk control method based on smart contracts as described in claim 1, characterized in that, The method also includes encrypting enterprise node data using zero-knowledge proofs based on the Groth16 algorithm.

3. The supply chain dynamic risk control method based on smart contracts as described in claim 1, characterized in that, In the risk control contract, risk assessment indicators are calculated based on on-chain data, off-chain data, and a risk assessment model, specifically including: Time series features and derived features are extracted based on the processed on-chain and off-chain data; XGBoost is trained by combining the extracted time-series features and derived features to obtain the trained XGBoost; Based on the trained XGBoost, predictions are made on the target object, and the continuous values ​​output by XGBoost are transformed into probabilistic values ​​to obtain risk assessment indicators.

4. The supply chain dynamic risk control method based on smart contracts as described in claim 3, characterized in that, Time series features include the volatility of historical repayment records and the moving average of transaction frequency; the volatility of historical repayment records is used to reflect the stability of repayments, and the moving average of transaction frequency is used to eliminate the impact of short-term fluctuations; derived features include the ratio of transaction amount to corporate asset size and the standard deviation of related corporate credit scores; the standard deviation of related corporate credit scores is used to measure the degree of diversification of related risks.

5. The supply chain dynamic risk control method based on smart contracts as described in claim 3, characterized in that, The process of training XGBoost by combining extracted time-series features and derived features to obtain the trained XGBoost includes: for each parameter in the XGBoost model, the XGBoost model learns the non-linear relationship between time-series features, derived features and default rate by integrating multiple decision trees. Each tree is recursively split according to the time-series features and derived features, and the final leaf node corresponds to a default probability. The prediction results of all trees are weighted and summed and converted into the final default rate through the sigmoid function.

6. The supply chain dynamic risk control method based on smart contracts as described in claim 1, characterized in that, The process of adjusting the credit granting strategy in real time based on changes in risk indicators to obtain the adjusted credit granting strategy includes: Initialize the Markov decision process model, including defining the state space, action space, and designing the reward function; The interest rate is adjusted based on the current risk status and the optimal strategy in Markov decision-making. It is then determined whether the adjustment exceeds the limit. If it does, the interest rate is adjusted to the maximum limit; otherwise, it is adjusted based on the calculated interest rate. Determine if the adjustment frequency exceeds the limit. If it does not exceed the limit, restore the original interest rate and do not make this adjustment. Otherwise, execute the interest rate adjustment and record the adjustment time.

7. A supply chain dynamic risk control device based on smart contracts, characterized in that, include: The data acquisition module is used to acquire node data from various participants in the supply chain. The risk assessment module dynamically assesses supply chain risks by combining node data from various participants with the constructed smart contracts. Specifically, the basic contract verifies logistics and invoice data, sending the verification results to the business contracts upon successful verification. The business contracts, based on received enterprise data, introduce a timestamp mechanism to record the debt generation time on the blockchain and use a Merkle tree to achieve multi-level debt splitting and merging. The risk control contract calculates risk assessment indicators based on on-chain and off-chain data and the risk assessment model. The risk control module is used to adjust the credit granting strategy in real time based on changes in risk assessment indicators to obtain the adjusted credit granting strategy.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the smart contract-based supply chain dynamic risk control method as described in any one of claims 1-6.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the supply chain dynamic risk control method based on smart contracts as described in any one of claims 1-6.

10. A program product, said program product being a computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps in the supply chain dynamic risk control method based on smart contracts as described in any one of claims 1-6.

Images