Multi-currency intelligent settlement method and system for cross-border chemical trade

Abstract

The application provides a multi-currency intelligent settlement method and system for cross-border chemical trade, and belongs to the field of trade data processing. The method comprises the following steps: obtaining the characteristics of chemical products and the characteristics of multiple currencies and settlement constraints of multiple currencies during cross-border chemical trade; calculating a settlement sensitivity coefficient according to the settlement constraints; generating a first settlement scheme, analyzing the first settlement time of the first settlement scheme, performing delay analysis and variation prediction, obtaining a first delay amount and a first predicted variation amount; processing and calculating the first delay amount and the first predicted variation amount to obtain a first settlement risk degree, and iteratively optimizing the settlement scheme to obtain an optimal settlement scheme and perform settlement. The settlement sensitivity coefficient quantifies the settlement difficulty, and the settlement scheme is iteratively optimized by combining delay analysis and exchange rate variation prediction, so that the settlement cycle is shortened, the exchange rate fluctuation risk is reduced, and the multi-currency settlement efficiency and transaction stability of cross-border chemical trade are improved.

Description

Technical Field

[0001] This invention relates to the field of trade data processing, and in particular to a multi-currency intelligent settlement method and system for cross-border chemical trade. Background Technology

[0002] Cross-border chemical trade, as an important component of international trade, involves the exchange of chemical products between multiple countries and regions. Its settlement process typically requires the exchange and allocation of funds across multiple currencies. Due to the diverse types of chemical products, significant batch-to-batch value variations, and the time-sensitive nature of some products' storage, the timeliness and stability of settlement have a crucial impact on the entire transaction process.

[0003] Currently, multi-currency settlement in cross-border chemical trade mainly relies on manual experience or fixed rules for currency selection and amount allocation. In practice, different chemical products have different ranges of currencies that can be used for cross-border settlement due to their different categories. Existing settlement methods, when faced with this constraint, often adopt a relatively single currency combination strategy, lacking the ability to quantitatively assess the complexity of multi-currency settlement. At the same time, the international exchange rate market fluctuates frequently, with different currencies experiencing varying exchange rate changes within the settlement cycle. Existing technology lacks systematic analysis and prediction methods for exchange rate fluctuation risks and fund arrival delays during the settlement process, leading to additional transaction costs during the settlement execution stage due to unfavorable exchange rate fluctuations or fund flow delays. These problems manifest in actual business as passively extended settlement cycles and difficulty in controlling settlement amount deviations caused by exchange rate fluctuations, ultimately reducing the settlement efficiency and transaction stability of cross-border chemical trade. Summary of the Invention

[0004] This invention addresses the technical problems in existing cross-border chemical trade where multi-currency settlement is hampered by a lack of intelligent means for currency selection and amount allocation, leading to long settlement cycles, high exchange rate fluctuation risks, and consequently affecting the stability and efficiency of chemical product transactions. It provides a multi-currency intelligent settlement method and system for cross-border chemical trade to solve these problems.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: In a first aspect, the present invention provides a multi-currency intelligent settlement method for cross-border chemical trade, comprising: acquiring chemical product characteristics during cross-border chemical trade, and capturing multiple currency characteristics and settlement constraints corresponding to the chemical product characteristics through big data; calculating a settlement sensitivity coefficient based on the settlement constraints; generating a first settlement scheme according to the settlement constraints, chemical product characteristics, and multiple currency characteristics; analyzing the first settlement time of the first settlement scheme, performing delay analysis and change prediction to obtain a first delay amount and a first predicted change amount; processing and calculating the first delay amount and the first predicted change amount based on the settlement sensitivity coefficient to obtain a first settlement risk level, and iteratively optimizing the settlement scheme to obtain the optimal settlement scheme and perform settlement.

[0006] Secondly, this invention provides a multi-currency intelligent settlement system for cross-border chemical trade, comprising: a data acquisition module for acquiring chemical product characteristics during cross-border chemical trade, and for capturing multiple currency characteristics and settlement constraints corresponding to the chemical product characteristics through big data; a sensitivity coefficient calculation module for calculating a settlement sensitivity coefficient based on the settlement constraints; a scheme generation and analysis module for generating a first settlement scheme according to the settlement constraints, chemical product characteristics, and multiple currency characteristics, analyzing the first settlement time of the first settlement scheme, performing delay analysis and change prediction, and obtaining a first delay amount and a first predicted change amount; and an iterative optimization settlement module for processing and calculating the first delay amount and the first predicted change amount based on the settlement sensitivity coefficient to obtain a first settlement risk level, and iteratively optimizing the settlement scheme to obtain the optimal settlement scheme and perform settlement.

[0007] The beneficial effects of this invention are: In cross-border chemical trade, the first step is to acquire the characteristics of chemical products and, through big data analysis, extract the characteristics of multiple currencies, along with the settlement constraints corresponding to these characteristics. This clarifies the product value and type information of the current transaction batch, and provides real-time exchange rates for each currency and the range of settlement currencies for the current chemical product category, offering fundamental data support for generating subsequent settlement plans. Next, a settlement sensitivity coefficient is calculated based on the settlement constraints. By quantifying the proportion of settlement currencies among all currencies, the difficulty of multi-currency settlement for the current chemical products is measured. A higher settlement sensitivity coefficient indicates more settlement currencies, higher complexity of multi-currency coordination, and a greater need for stability assurance in settlement processing, providing a quantitative basis for risk assessment of the settlement plan. Finally, a first settlement plan is generated according to the settlement constraints, chemical product characteristics, and multiple currency characteristics. The first settlement time of the first settlement plan is analyzed, and delays are assessed. The process involves analyzing and predicting currency variations to obtain the first delay and the first predicted change. This involves selecting a currency combination within the available settlement currencies and allocating settlement amounts for each currency to meet the transaction value of chemical products. Simultaneously, a systematic analysis and prediction of the fund arrival delay and exchange rate fluctuation trends of the selected currencies during the settlement period are conducted to identify potential time and exchange rate risks in the initial plan. Subsequently, the first delay and the first predicted change are processed and calculated based on the settlement sensitivity coefficient to obtain the first settlement risk level. The settlement plan is then iteratively optimized to obtain the optimal settlement plan. This involves a comprehensive risk assessment of delay and exchange rate fluctuation factors using quantitative indicators of settlement difficulty, and repeated adjustments to the currency combination and amount allocation to reduce the settlement risk level. Ultimately, the optimal settlement plan with a short settlement period and minimal impact from exchange rate fluctuations is obtained, thereby improving the efficiency and stability of multi-currency settlement in cross-border chemical trade.

[0008] By calculating the settlement sensitivity coefficient to quantify the difficulty of multi-currency settlement, and combining delay analysis and exchange rate fluctuation prediction to iteratively optimize the settlement scheme, compared with the existing technology that relies on manual experience or fixed rules for currency selection and amount allocation, this method effectively shortens the settlement cycle, reduces the risk of exchange rate fluctuations, and improves the efficiency and transaction stability of multi-currency settlement in cross-border chemical trade. Attached Figure Description

[0009] Figure 1 This is a flowchart illustrating the multi-currency intelligent settlement method for cross-border chemical trade provided by the present invention. Figure 2 This is a schematic diagram of the multi-currency intelligent settlement system for cross-border chemical trade provided by the present invention.

[0010] In the attached diagram, the components represented by each number are as follows: Data acquisition module 11, sensitivity coefficient calculation module 12, scheme generation and analysis module 13, iterative optimization and settlement module 14. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0013] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0014] Example 1, as Figure 1 As shown, this embodiment of the invention provides a multi-currency smart settlement method for cross-border chemical trade, including: S1. In cross-border chemical trade, obtain the characteristics of chemical products, and through big data, capture the characteristics of multiple currencies, as well as the settlement constraints corresponding to the characteristics of chemical products.

[0015] Specifically, in the context of cross-border chemical trade, when a cross-border chemical product transaction is initiated, the first step is to perform data collection and constraint construction operations to provide basic data support for the generation and optimization of subsequent settlement schemes. Specifically, the characteristics of the chemical products in the current transaction batch are obtained. These characteristics include value characteristics and category characteristics. The value characteristic is the total transaction amount of the current batch of chemical products, expressed in a base currency, which is used as the basis for allocating settlement amounts in various currencies in the subsequent settlement scheme. The category characteristic is the chemical category to which the current batch of chemical products belongs, such as hazardous chemicals, fine chemicals, or basic chemical raw materials. Different categories of chemical products correspond to different cross-border settlement channels and currency applicability in cross-border trade; therefore, the category characteristic directly determines which currencies can be used for settlement of this batch of chemical products and is a key basis for constructing settlement constraints.

[0016] Simultaneously, big data technology is used to capture the currency characteristics of multiple currencies involved in the current cross-border trade environment in real time. Each currency characteristic includes at least the current real-time exchange rate of that currency. Exchange rate data comes from data sources such as publicly available foreign exchange market data interfaces or financial data service platforms. Big data collection methods are used to connect with these data sources to obtain the real-time exchange rate values ​​of multiple currencies relative to a base currency. Acquiring multiple currency characteristics provides the basis for exchange rate conversions in subsequent steps for generating settlement plans, and also provides a data source for subsequent exchange rate fluctuation predictions.

[0017] Next, based on the acquired characteristics of the chemical products, settlement constraints corresponding to these products are constructed. Specifically, category characteristics are extracted from the chemical product features, and information on currently available settlement currencies corresponding to these category characteristics is queried using big data methods. Since different categories of chemical products correspond to different cross-border settlement channels and currency applicability ranges, not all currencies can be used for cross-border settlement of a specific category of chemical products. By connecting to big data sources such as cross-border settlement currency databases and payment channel information platforms, at least one available settlement currency corresponding to the current category characteristics is obtained, and settlement constraints are constructed accordingly. The core content of these settlement constraints is: settlement of the current batch of chemical products can only be conducted using at least one of the aforementioned available settlement currencies, excluding currencies not within the scope of settlement. These settlement constraints will serve as a hard constraint for generating settlement schemes in subsequent steps, ensuring that the generated settlement schemes meet the settlement requirements of cross-border chemical trade.

[0018] Through the above operations, the collection and construction of three types of basic data—chemical product characteristics, multi-currency characteristics, and settlement constraints—were completed. Specifically, the value characteristics of chemical products provided a total benchmark for the allocation of settlement amounts, while the category characteristics provided a basis for constructing settlement constraints. The multi-currency characteristics provided real-time data support for the conversion of amounts in each currency and the prediction of exchange rate fluctuations in subsequent settlement schemes. The settlement constraints defined the compliance boundaries for the generation of subsequent settlement schemes and provided a data foundation for the subsequent calculation of settlement sensitivity coefficients and the generation of settlement schemes.

[0019] S2. Calculate the settlement sensitivity coefficient based on the settlement constraints.

[0020] Specifically, after collecting and constructing the characteristics of chemical products, multi-currency characteristics, and settlement constraints, the settlement sensitivity coefficient is further calculated based on the settlement constraints to quantify the settlement difficulty of the current batch of chemical products in a multi-currency settlement environment.

[0021] The settlement sensitivity coefficient is calculated based on the quantitative relationship between at least one settleable currency determined in the settlement constraints and multiple currencies captured in the current cross-border trade environment. The proportion of the quantity of at least one settleable currency to the total quantity of multiple currencies is calculated as the settlement sensitivity coefficient. That is, the settlement sensitivity coefficient equals the ratio of the quantity of settleable currencies to the total quantity of multiple currencies. A higher proportion of settleable currencies indicates more available settlement currencies for the current chemical products, a greater number of currencies requiring coordination during multi-currency settlement, more complex cross-influence of exchange rate fluctuations between different currencies, a larger combination space for settlement schemes, and higher settlement processing complexity. Correspondingly, a higher settlement sensitivity coefficient indicates greater difficulty in multi-currency settlement for the current batch of chemical products and a need for more stable settlement processing. Conversely, a lower proportion of settleable currencies results in a lower settlement sensitivity coefficient, fewer currencies available for settlement, limited combination space for schemes, and relatively lower settlement processing complexity.

[0022] For example, in the current cross-border trade environment, 10 currencies are captured, and 8 of the current category of chemical products can be settled in any of them. Therefore, the settlement sensitivity coefficient is 8 / 10 = 0.8. If only 2 currencies can be settled, the settlement sensitivity coefficient is 2 / 10 = 0.2. The former has a significantly higher settlement sensitivity coefficient than the latter, reflecting that the former involves more currencies for settlement, has a higher complexity in multi-currency combinations, faces greater exchange rate volatility risks, and requires a more robust settlement solution.

[0023] The settlement sensitivity coefficient is used to quantify the difficulty of multi-currency settlement for the current batch of chemical products. The higher the value, the greater the complexity of multi-currency coordination, and the more robust the subsequent settlement scheme needs to be. This settlement sensitivity coefficient will be used as a weighting parameter in the calculation of settlement risk, enabling the iterative optimization process of the settlement scheme to adaptively adjust the optimization focus according to the actual settlement difficulty.

[0024] S3. Based on the settlement constraints, chemical product characteristics, and multiple currency characteristics, generate a first settlement scheme, analyze the first settlement time of the first settlement scheme, perform delay analysis and change prediction, and obtain the first delay amount and the first predicted change amount.

[0025] Specifically, after calculating the settlement sensitivity coefficient, a first settlement plan is generated based on settlement constraints, chemical product characteristics, and multiple currency characteristics. The settlement time is then predicted, delay analysis is performed, and exchange rate fluctuations are predicted for the first settlement plan to obtain the first delay amount and the first predicted change amount, providing an evaluation basis for subsequent calculation of settlement risk and iterative optimization of the plan.

[0026] First, a first settlement scheme is generated. Specifically, under the constraints of settlement conditions, one or more currencies are randomly selected from at least one settleable currency to obtain at least one first currency. Then, based on the value characteristics of the chemical product and the exchange rates of the corresponding currency characteristics of the selected at least one first currency, the total transaction amount represented by the value characteristics is randomly allocated, such that the sum of the settlement amounts of each first currency converted according to its respective exchange rate equals the total transaction amount corresponding to the value characteristics, thereby obtaining at least one first settlement amount. At least one first currency and its corresponding at least one first settlement amount together constitute the first settlement scheme. That is, the first settlement scheme describes the specific allocation method of which currencies are used to complete the entire transaction in the current transaction batch and how much amount is settled in each currency. The first settlement scheme is generated by randomly selecting currencies and randomly allocating amounts so that the optimal settlement scheme can be gradually searched and converged through the iterative optimization mechanism in the subsequent step S4. As the initial scheme for iterative optimization, the randomness of the first settlement scheme ensures the full exploration of the search space.

[0027] After generating the first settlement scheme, its settlement time is predicted to obtain the first settlement time. Specifically, corresponding settlement time prediction agents are pre-built for multiple currencies. Each settlement time prediction agent is trained using a sample settlement volume set as training data and a sample currency settlement time set as supervision labels. The sample settlement volume set represents the settlement amount data of each historical transaction for that currency, and the sample currency settlement time set represents the time taken to complete the settlement of each corresponding settlement amount. Through training, each settlement time prediction agent learns the mapping relationship between settlement amount and settlement time for that currency. During prediction, the first settlement volume corresponding to at least one first currency in the first settlement scheme is input into the settlement time prediction agent for that currency, and the output is the settlement time for at least one currency. Since the settlement completion time of the entire transaction in a multi-currency settlement scenario depends on the transaction with the longest settlement time among all currencies, the maximum value among the settlement times of at least one currency is selected as the first settlement time. The first settlement time represents the longest waiting time required to complete the settlement of the entire transaction under the first settlement scheme.

[0028] After obtaining the first settlement time, a delay analysis is performed to obtain the first delay amount. Specifically, a delay amount classification table corresponding to the category characteristics of the current chemical products is obtained. This delay amount classification table is pre-established based on historical transaction data of different categories of chemical products and includes the mapping relationship between sample settlement time and sample delay amount. Each sample delay amount represents the proportion of chemical products of that category that fail due to an excessively long settlement cycle at the corresponding sample settlement time. Chemical products have unique characteristics that distinguish them from general trade commodities. Some chemical products have shelf-life limitations, storage requirements, or market timeliness constraints. The longer the settlement cycle, the higher the risk of product deterioration, failure, or missing the market window during transportation and storage. The delay amount classification table quantitatively models this correlation between settlement time and product failure risk. The first settlement time is input into the delay amount classification table, and the corresponding first delay amount is obtained by looking up the table. The larger the first delay amount, the higher the risk of chemical product failure due to settlement time under the first settlement scheme.

[0029] Simultaneously, based on the first settlement time, the exchange rate fluctuation of at least one first currency is predicted to obtain the first predicted change. Specifically, the currency characteristics of at least one first currency within a past preset time range are obtained, resulting in at least one first currency feature sequence. Each first currency feature sequence is the time series data of the exchange rate of that currency within the past preset time range. Then, the currency feature data within the last time window of each first currency feature sequence, which is equal in length to the first settlement time, is extracted as the first currency feature sequence, and the change in exchange rate within that time window is calculated to obtain at least one first currency feature change. This means that the exchange rate fluctuation within the most recent time window of equal length to the first settlement time in history is used as the basis for predicting the possible fluctuation of the exchange rate of that currency within the future first settlement time. Then, based on the at least one first currency feature change, the first predicted change is calculated. When the first settlement scheme involves multiple first currencies, the first predicted change can be obtained by weighting and summing the first currency feature changes according to their respective proportions in the total transaction amount. The larger the first predicted change, the higher the exchange rate fluctuation risk faced by the first settlement scheme within the settlement period.

[0030] Through the above operations, risk assessment indicators were obtained for the first settlement scheme in two dimensions: the first delay and the first predicted change. The first delay assesses the risk of product failure due to an excessively long settlement cycle from the perspective of the timeliness of chemical products, while the first predicted change assesses the risk of economic loss due to currency fluctuations during the settlement cycle from the perspective of exchange rate volatility. The first delay and the first predicted change together constitute the basis for the risk assessment of the first settlement scheme. They will be weighted and calculated in conjunction with a settlement sensitivity coefficient to obtain the first settlement risk level, which will then drive the iterative optimization of the settlement scheme.

[0031] S4. Based on the settlement sensitivity coefficient, process and calculate the first delay amount and the first predicted change amount to obtain the first settlement risk level, and perform iterative optimization of the settlement scheme to obtain the optimal settlement scheme and perform settlement.

[0032] Specifically, after obtaining the first delay amount and the first predicted change amount corresponding to the first settlement plan, the first delay amount and the first predicted change amount are further weighted according to the settlement sensitivity coefficient to obtain the first settlement risk level. New settlement plans are continuously generated and evaluated through an iterative optimization mechanism, and finally the optimal settlement plan is converged and the settlement is executed.

[0033] First, the settlement sensitivity coefficient is configured as a variable weight, and the delay weight is calculated accordingly. Since the settlement sensitivity coefficient is the ratio of the number of settlementable currencies to the total number of currencies, its value ranges from 0 to 1, and it can be directly used as a variable weight. The delay weight equals 1 minus the variable weight, and the sum of the variable weight and the delay weight is always 1, forming a complementary weight set. Because the settlement sensitivity coefficient represents the difficulty of multi-currency settlement for the current batch of chemical products, a larger settlement sensitivity coefficient indicates more settlementable currencies, higher complexity of multi-currency coordination, and greater risk of exchange rate fluctuations. Therefore, the variable weight increases accordingly, making the subsequent calculation of settlement risk more focused on exchange rate fluctuations. During iterative optimization, currency combinations with smaller exchange rate fluctuations are prioritized to ensure settlement stability. When the settlement sensitivity coefficient is small, it indicates fewer settlementable currencies and relatively limited exchange rate fluctuation risks, but limited currency selection may lead to a longer settlement cycle. In this case, the variable weight decreases accordingly, and the delay weight increases relatively, making the optimization process focus more on shortening the settlement cycle and reducing the risk of chemical products failing due to excessive settlement time. Through this weighting mechanism, the settlement sensitivity coefficient effectively transmits the quantified settlement difficulty information to the risk assessment stage, enabling adaptive adjustment of the optimization direction.

[0034] Then, using delay weight and change weight, the first delay amount and the first predicted change amount are weighted and calculated to obtain the first settlement risk level. That is, the first settlement risk level = delay weight × first delay amount + change weight × first predicted change amount. The first settlement risk level comprehensively reflects the overall risk level of the first settlement plan in terms of both product timeliness risk and exchange rate fluctuation risk. The higher the value, the higher the overall risk of the settlement plan.

[0035] After obtaining the initial settlement risk level, the iterative optimization phase begins. Following the same method as in step S3, new settlement schemes are randomly generated under the constraints of the settlement conditions. For each newly generated settlement scheme, settlement time prediction, delay analysis, and exchange rate fluctuation prediction are performed to obtain the corresponding delay and predicted change. These are then weighted using the same weights to obtain the corresponding settlement risk level. In each iteration, a new settlement scheme is generated and its settlement risk level is calculated. This is compared with the currently recorded minimum settlement risk level. If the new scheme has a lower settlement risk level, the optimal record is updated. The iterative process continues until a convergence condition is met. The convergence condition can be that the settlement risk level no longer decreases in a preset number of iterations, or that the total number of iterations reaches a preset upper limit. When the iteration converges, the settlement scheme with the lowest settlement risk level among all iterative schemes is retained as the optimal settlement scheme. The optimal settlement scheme determines the currency combination and settlement amount allocation for the current transaction batch. It is the globally optimal solution obtained through multiple rounds of iterative search, within the constraints of the settlement conditions, comprehensively considering the impact of the settlement cycle on the timeliness of chemical products and the economic loss risk caused by exchange rate fluctuations. Settlement will be executed according to the optimal settlement plan to complete the multi-currency intelligent settlement of the current batch of cross-border chemical products.

[0036] Through the above steps, a fully intelligent settlement process is achieved, encompassing basic data collection, settlement difficulty quantification, settlement scheme generation and risk assessment, and iterative optimization. By adaptively adjusting the weight ratio of delay risk and exchange rate fluctuation risk using a settlement sensitivity coefficient, the iterative optimization process can dynamically adjust its optimization focus based on the actual settlement constraints of different chemical product categories. This achieves a balance between shortening the settlement cycle and reducing exchange rate fluctuation risk, thereby improving the efficiency and transaction stability of multi-currency settlement in cross-border chemical trade.

[0037] Furthermore, in cross-border chemical trade, it is necessary to obtain the characteristics of chemical products and, through big data, capture the characteristics of multiple currencies, as well as the settlement constraints corresponding to the characteristics of chemical products, including: S11. In cross-border chemical trade, obtain the chemical product characteristics of chemical products, including value characteristics and category characteristics; S12. Through big data, capture multiple currency characteristics of multiple currencies, where each currency characteristic includes the exchange rate; S13. Obtain the settlement constraints corresponding to the characteristics of the chemical product.

[0038] In a preferred embodiment, firstly, during cross-border chemical trade, the chemical product characteristics of the current transaction batch are obtained. These characteristics include value characteristics and category characteristics. The value characteristic is the total transaction amount of the current transaction batch, expressed in a base currency, such as the US dollar or RMB, commonly used in cross-border trade. The specific value of the value characteristic is extracted from the trade contract or order system corresponding to the current transaction batch. The value characteristic serves as the total benchmark for allocating settlement amounts in each currency in the subsequent settlement plan, ensuring that the sum of the settlement amounts allocated in each currency, after currency conversion, equals the total transaction amount. The category characteristic is the chemical product category to which the current transaction batch of chemical products belongs. In cross-border chemical trade, chemical products are typically classified according to the internationally accepted HS coding system. Different HS codes correspond to different chemical product categories, such as hazardous chemicals, fine chemicals, basic chemical raw materials, and pharmaceutical intermediates. Different categories of chemical products correspond to different cross-border settlement channels and applicable currencies in cross-border trade. For example, some hazardous chemicals can only be traded and settled through specific channels in specific currencies, resulting in a narrow range of available currencies. In contrast, the settlement restrictions for basic chemical raw materials are relatively lenient, allowing for a wider range of acceptable currencies. Therefore, the category characteristics directly determine which currencies can be used for settlement of a batch of chemical products, serving as a key basis for subsequently constructing settlement constraints.

[0039] Then, using big data technology, the characteristics of multiple currencies involved in the current cross-border trade environment are captured in real time. Each currency characteristic includes at least the current real-time exchange rate of that currency. Exchange rate data comes from data sources such as public data interfaces in the foreign exchange market or financial data service platforms. Big data collection methods are used to connect to these data sources to obtain the real-time exchange rate values ​​of multiple currencies relative to the benchmark currency. The exchange rate data is updated and collected at a preset time frequency, such as on a minute-by-minute or hourly basis, to ensure that the exchange rate data used for subsequent settlement scheme generation and exchange rate change prediction has sufficient timeliness and accuracy. Furthermore, the exchange rate in each currency characteristic can further include the buying rate and the selling rate to more accurately reflect the actual settlement cost in subsequent settlement amount allocation. The acquisition of multiple currency characteristics provides the basis for exchange rate conversion for each currency in the generation of subsequent settlement schemes, and also provides a historical data source for subsequent exchange rate change prediction.

[0040] Next, based on the acquired chemical product characteristics, the settlement constraints corresponding to the chemical product are obtained. The construction of these settlement constraints is based on the category characteristics of the chemical product, using big data analytics to query the available settlement currencies corresponding to the current category characteristics. Since different categories of chemical products correspond to different cross-border settlement channels and currency applicability ranges, not all currencies can be used for cross-border settlement of a specific category of chemical products. For example, certain types of hazardous chemicals can only be settled across borders through specific settlement channels in specific currencies, or some cross-border payment channels only support fund transfers in specific currencies; currencies not supported by these channels cannot be used for settlement. By connecting to big data sources such as cross-border settlement currency databases, payment channel information platforms, and currency support lists provided by payment channel service providers, at least one available settlement currency corresponding to the current category characteristics is obtained, and settlement constraints are constructed accordingly. These settlement constraints limit the settlement of the current batch of chemical products to only at least one of the aforementioned available settlement currencies, excluding currencies not within the settlement range. These settlement constraints serve as a hard constraint for generating settlement schemes in subsequent steps, ensuring that the generated settlement schemes meet the settlement requirements of cross-border chemical trade, and also providing a basis for calculating the settlement sensitivity coefficient.

[0041] Furthermore, obtaining the settlement constraints corresponding to the characteristics of the chemical product includes: S131. Extract the category features within the characteristics of the chemical product; S132. Through big data, obtain the currencies that can be settled for the current category features, obtain at least one settlement currency, and construct settlement constraints that use only the at least one settlement currency for the current chemical product settlement.

[0042] In a preferred embodiment, firstly, category features are extracted from the acquired chemical product characteristics. Category features identify the specific chemical category to which the current batch of chemical products belongs, such as the categories identified above based on the HS coding system, including hazardous chemicals, fine chemicals, basic chemical raw materials, and pharmaceutical intermediates. The purpose of extracting category features is to use them as a query index to retrieve the settlement currencies that can be used for that category of chemical products in the current cross-border trade environment. This ensures that the settlement constraints are built upon specific product categories, rather than applying a uniform currency range to all chemical products.

[0043] Then, using big data analytics, the extracted category features are used to identify the currencies currently available for settlement, resulting in at least one settlement currency. Settlement constraints are then constructed based on this. Specifically, using category features as query criteria, the system connects to big data sources such as cross-border settlement currency databases, payment channel information platforms, and currency support lists provided by payment channel service providers to comprehensively retrieve the settlement currencies available for this category of chemical products in cross-border trade at the current time. The search process considers multiple constraints: first, the product category dimension, i.e., whether there are specific currency restrictions on the application scope of this category of chemical products in cross-border settlement; second, the settlement channel dimension, i.e., whether currently available cross-border payment channels support fund transfers for this category of chemical products in a specific currency; and third, the transacting parties dimension, i.e., whether the cross-border settlement channels corresponding to the countries or regions where the transacting parties are located support fund receipts and payments in the corresponding currencies. After the above multi-dimensional search and filtering, at least one settlement currency corresponding to the current category features is obtained. Based on at least one settlement currency, settlement constraints are constructed. These constraints explicitly stipulate that cross-border settlement of the current batch of chemical products can only be conducted using at least one of the aforementioned settlement currencies, and other currencies not included in the settlement currency range are excluded from the candidate settlement scheme.

[0044] Through the above steps, starting from specific categories of chemical products, the range of acceptable settlement currencies is accurately determined through multi-dimensional big data retrieval, thereby constructing targeted settlement constraints. These constraints ensure that the calculation of settlement sensitivity coefficients and the generation of settlement schemes are conducted within the compliant currency range, guaranteeing the compliance and effectiveness of the entire settlement optimization process.

[0045] Furthermore, based on the aforementioned settlement constraints, a settlement sensitivity coefficient is calculated, including: S21. Calculate the proportion of at least one settlement currency among multiple currencies, and use it as the settlement sensitivity coefficient.

[0046] In a preferred embodiment, the number of at least one settlementable currency is denoted as M, and the total number of multiple currencies captured through big data is denoted as N. The settlement sensitivity coefficient is equal to the ratio of M to N, i.e., settlement sensitivity coefficient = M / N. For example, when N=10 and M=8, the settlement sensitivity coefficient is 8 / 10=0.8; when N=10 and M=2, the settlement sensitivity coefficient is 2 / 10=0.2. The more settlementable currencies there are, the larger their proportion, and the larger the settlement sensitivity coefficient. This means that the range of currencies that can be used for settlement in cross-border settlement of the current batch of chemical products is wider, the complexity of multi-currency combination allocation is higher, the cross-influence of exchange rate fluctuations between different currencies is more complex, and the difficulty of settlement processing is greater. In the extreme case where all currencies are settlementable, i.e., M=N, the settlement sensitivity coefficient is 1, indicating the highest complexity of multi-currency coordination in settlement. When there is only one settlementable currency, the settlement sensitivity coefficient approaches 0, indicating that no multi-currency coordination is required, and the settlement processing complexity is the lowest.

[0047] Through the above steps, the quantitative relationship between settleable currencies and all currencies in the settlement constraints is transformed into a specific numerical indicator, changing the settlement difficulty from a qualitative description to a quantifiable coefficient. This settlement sensitivity coefficient is then used as a weighting parameter in the subsequent calculation of settlement risk, allowing the iterative optimization process to adaptively adjust the weight ratio between delay risk and exchange rate fluctuation risk based on the magnitude of the settlement sensitivity coefficient. This enables differentiated optimization strategies for chemical product categories with varying settlement difficulties.

[0048] Furthermore, based on the settlement constraints, chemical product characteristics, and multiple currency characteristics, a first settlement scheme is generated. The first settlement time of the first settlement scheme is analyzed, and delay analysis and change prediction are performed to obtain a first delay amount and a first predicted change amount, including: S31. According to the settlement constraints and the value characteristics within the chemical product characteristics, randomly select one or more of at least one settlement currency to obtain at least one first currency, and according to the corresponding currency characteristics, randomly allocate at least one first settlement amount equal to the value characteristics as the first settlement scheme. S32. Based on at least one first settlement amount of at least one first currency in the first settlement scheme, predict the settlement time to obtain the first settlement time; S33. Based on the first settlement time, perform a delay analysis to obtain the first delay amount; S34. Based on the first settlement time, predict the change of at least one first currency to obtain the first predicted change amount.

[0049] In a preferred embodiment, firstly, a first settlement scheme is generated according to the settlement constraints and the value characteristics within the chemical product characteristics. Specifically, under the constraints of the settlement constraints, one or more currencies are randomly selected from at least one settleable currency to obtain at least one first currency. The range of random selection is strictly limited to the determined settleable currencies to ensure that the selected currencies meet the requirements of the settlement constraints. After determining at least one first currency, the total transaction amount represented by the value characteristics is randomly allocated according to the exchange rate in the currency characteristics corresponding to each first currency. The constraint of random allocation is that the sum of the settlement amounts of each first currency after conversion according to its respective exchange rate must be equal to the total transaction amount corresponding to the value characteristics, that is, to ensure that the amount of the entire transaction is completely covered without any amount gap or overflow. This yields at least one first settlement amount, and each first settlement amount corresponds to a specific amount to be settled under a first currency. At least one first currency and its corresponding at least one first settlement amount together constitute the first settlement scheme. For example, if the total transaction amount is $1 million, and the currencies are randomly selected as Euros and Japanese Yen, and then randomly allocated according to the current exchange rate, $400,000 worth of Euros will be settled, and $600,000 worth of Japanese Yen will be settled. This allocation scheme is the first settlement scheme. The random selection and allocation method used here is to fully explore the search space of feasible solutions in subsequent iterative optimizations by randomly generating different currency combinations and amount allocation methods in multiple rounds, thus avoiding getting trapped in local optima.

[0050] Then, based on at least one first settlement amount for at least one first currency within the first settlement scheme, the settlement time is predicted to obtain the first settlement time. Specifically, corresponding settlement time prediction agents are pre-built for multiple currencies. Each settlement time prediction agent is constructed using a machine learning model, with the settlement amount data of each currency in historical cross-border transactions as the sample settlement amount set, and the time data of the actual settlement of each settlement amount as the sample currency settlement time set, for supervised training. The cross-border settlement time of different currencies is affected by factors such as the bank clearing mechanism of the country where the currency is located, the processing capacity of the cross-border payment network, and the size of the settlement amount. The larger the settlement amount, the more complex the clearing and review process may be, and the longer the settlement time may be. Through training, each settlement time prediction agent learns the mapping relationship between the settlement amount and the settlement time for that currency. During prediction, the first settlement amount corresponding to each of the at least one first currency in the first settlement scheme is input into the settlement time prediction agent of the corresponding currency, and the settlement time of at least one currency is output. In multi-currency settlement scenarios, the settlement completion time of the entire transaction depends on the longest settlement time among all participating currencies. The entire transaction is considered settled only after all currencies have completed their settlements. Therefore, the maximum value among the settlement times of at least one currency is selected as the first settlement time. The first settlement time represents the longest waiting time required to complete the settlement of the entire transaction under the first settlement scheme.

[0051] Next, based on the first settlement time, a delay analysis is performed to obtain the first delay amount. Specifically, a delay amount classification table corresponding to the category characteristics of the current chemical products is obtained. This delay amount classification table is a lookup table pre-established based on historical transaction data and product characteristics of different categories of chemical products, containing the mapping relationship between sample settlement time and sample delay amount. Each sample delay amount represents the proportion of chemical products of that category that fail due to an excessively long settlement cycle at the corresponding sample settlement time. Different categories of chemical products have different sensitivities to settlement cycles. For example, some active chemical reagents have short shelf lives, and extended settlement cycles may cause degradation and failure during storage; for some chemical raw materials with strong market time sensitivity, excessively long settlement cycles may cause them to miss the production scheduling window of downstream customers, resulting in substantial transaction losses. The delay amount classification table establishes a quantitative correspondence between settlement time and product failure risk precisely for this difference in timeliness among different categories of chemical products. The first settlement time is input into the delay amount classification table, and the corresponding first delay amount is obtained by looking up the table. The larger the first delay amount, the higher the risk of chemical product failure due to excessively long settlement time under the first settlement scheme.

[0052] Subsequently, based on the first settlement time, a change prediction for at least one first currency is made to obtain the first predicted change amount. Specifically, the currency characteristics of at least one first currency within a past preset time range are obtained, resulting in at least one first currency feature sequence. Each first currency feature sequence is the exchange rate time series data of that currency within a past preset time range. The preset time range can be set according to actual needs, such as exchange rate data from the past 30 days or the past 90 days. Then, the currency feature data within the last time window of each first currency feature sequence, which is equal in length to the first settlement time, is extracted as the first currency feature sequence, and the change amplitude of the exchange rate within this time window is calculated to obtain at least one first currency feature change amount. The change amplitude is calculated by first calculating the difference between the maximum and minimum exchange rate values ​​within the time window, then calculating the mean exchange rate within the time window, and finally dividing the difference by the mean. The quotient is the change amplitude of that currency. By dividing by the mean for normalization, the influence of differences in the absolute numerical dimensions of exchange rates between different currencies is eliminated, making the change amplitudes of different currencies comparable. This means that the exchange rate fluctuation range within the most recent time window, which is of equal length to the first settlement period in history, is used as the basis for predicting the potential fluctuation of the exchange rate of that currency within the future first settlement period. Then, based on at least one characteristic change in the first currency, the first predicted change is calculated. The larger the first predicted change, the higher the exchange rate fluctuation risk faced by the first settlement scheme during the settlement period.

[0053] Based on the above, the first settlement plan was quantitatively assessed from two dimensions: the risk of chemical product failure caused by the settlement cycle and the risk of economic loss caused by exchange rate fluctuations. This provides an assessment basis for the subsequent calculation and iterative optimization of settlement risk.

[0054] Furthermore, based on at least one first settlement amount of at least one first currency within the first settlement scheme, a settlement time prediction is performed to obtain the first settlement time, including: S321. Call multiple settlement time prediction agents corresponding to multiple currencies, wherein the training data of each settlement time prediction agent is a set of sample settlement amounts, and the supervision label is a set of sample currency settlement times; S322. Input at least one first settlement amount of at least one first currency in the first settlement scheme into the corresponding settlement time prediction agent, and output the settlement time of at least one currency. S323. Select the maximum value among the settlement times of at least one currency as the first settlement time.

[0055] In a preferred embodiment, firstly, multiple settlement time prediction agents corresponding to multiple currencies are invoked. Settlement time prediction agents are pre-constructed for each of the multiple currencies, with each agent corresponding one-to-one with a specific currency. The training data for each settlement time prediction agent is a set of sample settlement amounts, and the supervision label is a set of sample currency settlement times. For example, in one specific embodiment, the settlement time prediction agent uses a BP neural network model, including an input layer, hidden layers, and an output layer. The input to the input layer is the sample settlement amount, the hidden layer can be set to two layers, each containing several neurons, using ReLU as the activation function, and the output of the output layer is the corresponding predicted settlement time. During training, the set of sample settlement amounts is used as training data, and the set of sample currency settlement times is used as supervision labels. The set of sample settlement amounts represents the settlement amount data for each currency in historical cross-border transactions, and the set of sample currency settlement times represents the time taken from initiating settlement to the actual arrival of funds for each settlement amount. During training, mean squared error is used as the loss function, and the Adam optimizer is used for parameter updates. The network weights are continuously adjusted through backpropagation to gradually reduce the error between the predicted and actual settlement times until the loss function converges to a preset threshold or the training epochs reach a preset upper limit, at which point training stops. The cross-border settlement time for different currencies is affected by various factors, including the bank clearing mechanism of the country where the currency is located, the processing efficiency of the cross-border payment network, the complexity of anti-money laundering review processes, and the size of the settlement amount. Larger settlement amounts may involve more complex clearing and review processes, leading to increased settlement time. Through the above supervised training, each settlement time prediction agent learns the non-linear mapping relationship between the settlement amount and settlement time for that currency, enabling it to accurately predict the time required to complete settlement for that currency based on the input settlement amount.

[0056] Then, at least one first settlement amount of at least one first currency in the first settlement scheme is input into the settlement time prediction agent of the corresponding currency, and the settlement time of at least one currency is output. That is, for each first currency involved in the first settlement scheme, the first settlement amount corresponding to that currency is used as input and sent to the settlement time prediction agent of that currency, and the agent outputs the predicted settlement time required to complete the settlement amount of that currency. When the first settlement scheme involves multiple first currencies, the above prediction operation is performed on each first currency to obtain the settlement time of multiple currencies.

[0057] Next, the maximum value among the settlement times of at least one currency is selected as the first settlement time. In a multi-currency settlement scenario, the settlement processes of each currency proceed in parallel. The completion time of the entire transaction depends on the longest-running transaction among all participating currencies. Only when funds have been received in all currencies is the entire transaction considered settled. Therefore, taking the maximum value among the settlement times of all currencies as the first settlement time accurately represents the actual longest waiting time required to complete the settlement of the entire transaction under the first settlement scheme.

[0058] Through the above steps, a currency-level agent model was used to accurately predict settlement time. By taking the maximum value, the actual settlement cycle under multi-currency parallel settlement was accurately reflected, providing a reliable time benchmark for subsequent delay analysis and exchange rate fluctuation prediction.

[0059] Furthermore, based on the first settlement time, a delay analysis is performed to obtain the first delay amount, including: S331. Obtain a delay classification table corresponding to the category characteristics of chemical products, wherein the delay classification table includes the mapping relationship between sample settlement time and sample delay, and each sample delay includes the proportion of chemical product failure caused by settlement time. S332. Input the first settlement time into the delay amount classification table to obtain the first delay amount.

[0060] In a preferred embodiment, firstly, a delay quantity classification table corresponding to the category characteristics of chemical products is obtained. The delay quantity classification table is a lookup table pre-established for different categories of chemical products, with each category characteristic corresponding to an independent delay quantity classification table. The delay quantity classification table contains the mapping relationship between sample settlement time and sample delay quantity. Each sample delay quantity represents the proportion of chemical products of that category that become invalid due to an excessively long settlement period at the corresponding sample settlement time, with a value ranging from 0 to 1. For example, in the delay quantity classification table for a category of active chemical reagents, when the sample settlement time is 3 days, the corresponding sample delay quantity is 0.02, indicating that under a 3-day settlement period, 2% of the chemical products of this category may become invalid due to degradation, deterioration, or other problems during the waiting period for settlement; when the sample settlement time is 7 days, the corresponding sample delay quantity is 0.08, indicating that the failure rate rises to 8%; and when the sample settlement time is 15 days, the corresponding sample delay quantity is 0.25, indicating that the failure rate reaches 25%. The delay classification table is established based on historical transaction data and product characteristic data of different categories of chemical products. By statistically analyzing the proportion of actual failures of each category of chemical products under different settlement cycles in historical transactions, a correspondence between sample settlement time and sample delay amount is established. Different categories of chemical products exhibit significant differences in their sensitivity to settlement cycles. For example, chemical products with short shelf lives, such as active chemical reagents and biochemical products, show a faster probability of degradation and failure during transportation and storage as the settlement cycle lengthens; their delay amount classification table shows a higher rate of increase with sample settlement time. Conversely, basic chemical raw materials such as soda ash and sulfuric acid, which are chemically stable, are less sensitive to settlement cycles; their delay amount classification table shows a more gradual rate of increase with sample settlement time. By establishing separate delay classification tables for different categories of chemical products, a differentiated and refined assessment of settlement delay risk is achieved.

[0061] Subsequently, the first settlement time is input into the delay amount classification table corresponding to the current chemical product category characteristics to obtain the first delay amount. Specifically, the sample settlement time matching the first settlement time is searched in the delay amount classification table, and the corresponding sample delay amount is read as the first delay amount. When the first settlement time is exactly equal to a sample settlement time in the delay amount classification table, the corresponding sample delay amount is directly taken as the first delay amount. When the first settlement time is between two adjacent sample settlement times in the delay amount classification table, linear interpolation is used to calculate the first delay amount corresponding to the first settlement time based on the two adjacent sample settlement times and their corresponding sample delay amounts, to ensure the continuity and accuracy of the table lookup results. The value range of the first delay amount is 0 to 1. The larger the value, the higher the risk of the current batch of chemical products failing due to excessive settlement time under the first settlement scheme.

[0062] Through the above steps, by using a pre-established categorized delay quantity classification table, the first settlement time is transformed into a first delay quantity with a clear physical meaning, realizing a quantitative mapping from settlement time to the failure risk of chemical products, and providing a risk assessment indicator for the product timeliness dimension for the subsequent calculation of settlement risk.

[0063] Furthermore, based on the first settlement time, a change prediction for at least one first currency is made to obtain a first predicted change amount, including: S341. Obtain at least one currency feature within a preset time range in the past to obtain at least one first currency feature sequence; S342. Extract the first currency feature sequence within the last first settlement time range from the at least one first currency feature sequence, and calculate the change range as the change amount of at least one first currency feature. S343. Calculate the first predicted change based on at least one characteristic change of the first currency.

[0064] In a preferred embodiment, firstly, currency characteristics of at least one first currency within a preset time range are obtained, resulting in at least one first currency feature sequence. Specifically, for each first currency involved in the first settlement scheme, historical exchange rate data of that currency within a preset time range is obtained from a public foreign exchange market data interface or a financial data service platform, forming a first currency feature sequence for that currency. The preset time range can be set according to actual business needs, such as the past 30 days, 60 days, or 90 days. Each first currency feature sequence is a set of exchange rate values ​​arranged in chronological order, recording the trajectory of the exchange rate change of that currency relative to the base currency within the preset time range. The preset time range needs to be set long enough to cover the recent exchange rate fluctuation patterns of that currency, and also needs to include a time window of equal length to the first settlement time for subsequent truncation operations. When the first settlement scheme involves multiple first currencies, historical exchange rate data for each first currency is obtained separately, resulting in multiple first currency feature sequences.

[0065] Then, within at least one first currency characteristic sequence, a first currency characteristic sequence within the last first settlement time range is extracted, and the magnitude of change is calculated as the change amount of at least one first currency characteristic. Specifically, for each first currency characteristic sequence, a time window of exchange rate data equal to the first settlement time is extracted from the end of the sequence as the extracted first currency characteristic sequence. For example, if the first settlement time is 5 days and the preset time range is the past 30 days, then the exchange rate data for the last 5 days is extracted from the exchange rate sequence of the past 30 days for that currency. By extracting the last time window equal to the first settlement time, historical exchange rate data that is closest to the current time and equal to the predicted settlement period is selected as a reference for predicting the exchange rate fluctuation of that currency in the future settlement period, because the exchange rate fluctuation characteristics of the recent period have stronger reference value for the exchange rate trend in the near future. Then, the magnitude of change is calculated for the extracted first currency characteristic sequence to obtain the change amount of the first currency characteristic of that currency. The calculation method for the magnitude of change is as follows: First, calculate the difference between the maximum and minimum exchange rates within the time window. Then, calculate the average exchange rate within the time window. Finally, divide the difference by the average; the quotient is the magnitude of change for that currency. Normalization by dividing by the average eliminates the influence of differences in the absolute numerical dimensions of exchange rates between different currencies, making the magnitudes of change comparable. For example, if the maximum exchange rate of a certain first currency is 7.25, the minimum is 7.10, and the average is 7.18 within a 5-day time window, then the magnitude of change is (7.25-7.10) / 7.18 = 0.0209, meaning the characteristic change of that first currency is 0.0209. When the first settlement scheme involves multiple first currencies, the magnitude of change for each first currency is calculated separately to obtain multiple characteristic changes for each first currency.

[0066] Subsequently, a first predicted change is calculated based on the change in characteristics of at least one first currency. When the first settlement scheme involves only one first currency, the change in characteristics of that first currency is the first predicted change. When the first settlement scheme involves multiple first currencies, the change in characteristics of each first currency is weighted and summed according to its proportion in the total transaction amount to obtain the first predicted change. That is, for each first currency, the ratio of its first settlement amount to its value characteristic is calculated as the amount weight of that currency, the change in characteristics of that currency is multiplied by the corresponding amount weight, and the weighted changes of all first currencies are summed to obtain the first predicted change. By weighting according to the proportion of settlement amount, the exchange rate fluctuations of currencies with larger settlement amounts contribute more to the overall predicted change, and the exchange rate fluctuations of currencies with smaller settlement amounts contribute less to the overall predicted change, thus accurately reflecting the actual economic impact of exchange rate fluctuations of each currency on the entire transaction. The larger the first predicted change, the higher the overall exchange rate fluctuation risk faced by the first settlement scheme during the settlement period.

[0067] Through the above steps, based on historical exchange rate data and settlement time window analysis, a quantitative prediction of the exchange rate fluctuation risk faced by the first settlement plan during the settlement period was achieved, providing a risk assessment indicator for the exchange rate fluctuation dimension for the subsequent calculation of settlement risk.

[0068] Furthermore, based on the settlement sensitivity coefficient, the first delay amount and the first predicted change amount are processed and calculated to obtain the first settlement risk level, and the settlement scheme is iteratively optimized to obtain the optimal settlement scheme, and settlement is performed, including: S41. Configure the settlement sensitivity coefficient as a variable weight and calculate the delay weight; S42. Using the aforementioned delay weight and change weight, the first delay amount and the first predicted change amount are weighted and calculated to obtain the first settlement risk level; S43. Continue to generate settlement schemes and perform iterative optimization until convergence. The settlement scheme with the lowest settlement risk is retained as the optimal settlement scheme.

[0069] In a preferred embodiment, firstly, the settlement sensitivity coefficient is configured as a variable weight, and a delay weight is calculated accordingly. Since the settlement sensitivity coefficient is the ratio of the number of settlementable currencies to the total number of currencies, and its value ranges from 0 to 1, it can be directly used as a variable weight. The delay weight equals 1 minus the variable weight, and the sum of the variable weight and the delay weight is always 1, forming a complementary weight set. Because the settlement sensitivity coefficient characterizes the difficulty of multi-currency settlement for the current batch of chemical products, a larger settlement sensitivity coefficient indicates more settlementable currencies, higher complexity of multi-currency coordination, and greater exchange rate fluctuation risk. Therefore, the variable weight increases accordingly, making the subsequent calculation of settlement risk more focused on exchange rate fluctuation factors. During iterative optimization, currency combinations with smaller exchange rate fluctuations are prioritized to ensure settlement stability. When the settlement sensitivity coefficient is small, it indicates fewer settlementable currencies and relatively limited exchange rate fluctuation risk. However, limited currency selection may lead to a longer settlement cycle. In this case, the variable weight decreases accordingly, and the delay weight increases relatively, making the optimization process focus more on shortening the settlement cycle and reducing the risk of chemical products failing due to excessive settlement time. Through this weighting mechanism, the settlement sensitivity coefficient effectively transmits quantified settlement difficulty information to the risk assessment stage, enabling adaptive adjustment of the optimization direction. For example, when the settlement sensitivity coefficient is 0.8, the change weight is 0.8, the delay weight is 0.2, and the exchange rate change factor accounts for 80% of the weight in the calculation of settlement risk; when the settlement sensitivity coefficient is 0.2, the change weight is 0.2, the delay weight is 0.8, and the delay factor accounts for 80% of the weight in the calculation of settlement risk.

[0070] Subsequently, the first delay amount and the first predicted change amount are weighted and calculated using delay weight and change weight to obtain the first settlement risk level. Specifically, the first settlement risk level is equal to the sum of the delay weight multiplied by the first delay amount and the change weight multiplied by the first predicted change amount. For example, when the delay weight is 0.2, the change weight is 0.8, the first delay amount is 0.08, and the first predicted change amount is 0.0209, the first settlement risk level = 0.2 × 0.08 + 0.8 × 0.0209 = 0.016 + 0.01672 = 0.03272. The first settlement risk level comprehensively reflects the overall risk level of the first settlement scheme in terms of both the timeliness risk of chemical products and the exchange rate fluctuation risk. The larger the value, the higher the comprehensive risk of the settlement scheme. By using a weighted calculation method, the risk indicators of two different dimensions are integrated into a unified risk metric, which facilitates horizontal comparison and selection of different settlement schemes during subsequent iterative optimization.

[0071] Then, settlement schemes are generated and iteratively optimized until convergence. The settlement scheme with the lowest settlement risk is retained as the optimal settlement scheme. Specifically, after obtaining the first settlement risk level, the first settlement scheme and its first settlement risk level are recorded as the current optimal scheme and the current minimum settlement risk level. Then, an iterative loop is entered. In each iteration, following the same method as step S31, under the constraints of the settlement conditions, a new settlement scheme is generated by randomly selecting a settlement currency and randomly allocating a settlement amount. For the newly generated settlement scheme, the settlement time prediction, delay analysis, and exchange rate change prediction are performed sequentially, following the same method as steps S32 to S34, to obtain the corresponding delay amount and predicted change amount. Then, the same delay weight and change weight are used for weighted calculation in step S42 to obtain the settlement risk level corresponding to the settlement scheme. This settlement risk level is compared with the currently recorded minimum settlement risk level. If the settlement risk level of the new scheme is smaller, the new scheme is updated as the current optimal scheme, and the current minimum settlement risk level is also updated; otherwise, the current optimal scheme remains unchanged. The iterative process continues until the convergence condition is met. The convergence condition can be either that the minimum settlement risk no longer decreases within a preset number of iterations (e.g., no lower settlement risk is generated after 50 consecutive iterations, indicating that the search space has been fully explored and further iterations are unlikely to yield a better solution), or that the total number of iterations reaches a preset upper limit (e.g., 1000 iterations). When either convergence condition is met, iteration stops, and the currently recorded optimal solution is output as the optimal settlement solution. The optimal settlement solution determines the currency combination to be used in the current transaction batch and the specific settlement amount for each currency. Settlement is executed according to the optimal settlement solution, completing the multi-currency intelligent settlement of the current batch of cross-border chemical products.

[0072] Through the above steps, the settlement sensitivity coefficient is used to adaptively allocate the weight ratio of delay risk and exchange rate fluctuation risk, and the settlement scheme is globally optimized through a multi-round random iterative search mechanism. This achieves an optimal balance between shortening the settlement cycle and reducing exchange rate fluctuation risk, thereby improving the efficiency and transaction stability of multi-currency settlement in cross-border chemical trade.

[0073] Example 2, as Figure 2 As shown, based on the same inventive concept as the multi-currency smart settlement method for cross-border chemical trade provided in Embodiment 1, this embodiment of the invention also provides a multi-currency smart settlement system for cross-border chemical trade, including: The data acquisition module 11 is used to acquire the characteristics of chemical products in cross-border chemical trade, and to capture the characteristics of multiple currencies and the settlement constraints corresponding to the characteristics of chemical products through big data. Sensitivity coefficient calculation module 12 is used to calculate the settlement sensitivity coefficient according to the settlement constraints. The scheme generation and analysis module 13 is used to generate a first settlement scheme according to the settlement constraints, chemical product characteristics and multiple currency characteristics, analyze the first settlement time of the first settlement scheme, perform delay analysis and change prediction, and obtain the first delay amount and the first predicted change amount. The iterative optimization settlement module 14 is used to process and calculate the first delay amount and the first predicted change amount according to the settlement sensitivity coefficient, obtain the first settlement risk level, and perform iterative optimization of the settlement scheme to obtain the optimal settlement scheme and perform settlement.

[0074] Furthermore, the data acquisition module 11 is also used for: In cross-border chemical trade, it is important to obtain the chemical product characteristics of chemical products, which include value characteristics and category characteristics. By leveraging big data, we can capture multiple currency characteristics for various currencies, including exchange rates for each currency. Obtain the settlement constraints corresponding to the characteristics of the chemical product.

[0075] Furthermore, the data acquisition module 11 is also used for: Extract the category features within the characteristics of the chemical product; By using big data, the currencies that can be settled for the current category characteristics are obtained, at least one settlement currency is obtained, and settlement constraints are constructed to use only the at least one settlement currency for the current chemical product settlement.

[0076] Furthermore, the sensitivity coefficient calculation module 12 is also used for: Calculate the proportion of at least one settlement currency among multiple currencies, and use it as the settlement sensitivity coefficient.

[0077] Furthermore, the solution generation and analysis module 13 is also used for: According to the settlement constraints and the value characteristics within the chemical product characteristics, at least one or more of at least one settlement currency are randomly selected to obtain at least one first currency, and at least one first settlement amount equal to the value characteristics is randomly allocated according to the corresponding currency characteristics, as the first settlement scheme. Based on at least one first settlement amount of at least one first currency in the first settlement scheme, the settlement time is predicted to obtain the first settlement time; Based on the first settlement time, a delay analysis is performed to obtain the first delay amount; Based on the first settlement time, at least one first currency is predicted to change, and a first predicted change amount is obtained.

[0078] Furthermore, the solution generation and analysis module 13 is also used for: Multiple settlement time prediction agents corresponding to multiple currencies are invoked. The training data for each settlement time prediction agent is a set of sample settlement amounts, and the supervision label is a set of sample currency settlement times. Input at least one first settlement amount of at least one first currency in the first settlement scheme into the corresponding settlement time prediction agent, and output the settlement time of at least one currency. Select the maximum value among the settlement times of at least one currency as the first settlement time.

[0079] Furthermore, the solution generation and analysis module 13 is also used for: Obtain a delay classification table corresponding to the category characteristics of chemical products. The delay classification table includes the mapping relationship between sample settlement time and sample delay. Each sample delay includes the proportion of chemical product failure caused by settlement time. Input the first settlement time into the delay amount classification table to obtain the first delay amount.

[0080] Furthermore, the solution generation and analysis module 13 is also used for: Obtain at least one currency feature within a preset time range in the past to obtain at least one first currency feature sequence; Within the at least one first currency feature sequence, extract the first currency feature sequence within the last first settlement time range, and calculate the change range as the change amount of at least one first currency feature; The first predicted change is calculated based on at least one characteristic change of the first currency.

[0081] Furthermore, the iterative optimization settlement module 14 is also used for: Configure the settlement sensitivity coefficient as a variable weight and calculate the delay weight; Using the aforementioned delay weight and change weight, the first delay amount and the first predicted change amount are weighted and calculated to obtain the first settlement risk level; Continue generating settlement plans and iteratively optimizing them until convergence. The settlement plan with the lowest settlement risk is retained as the optimal settlement plan.

[0082] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0083] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0084] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0085] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0086] These computer program instructions may 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, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0087] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0088] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A multi-currency intelligent settlement method for cross-border chemical trade, characterized in that: The method includes: In cross-border chemical trade, it is necessary to obtain the characteristics of chemical products, and through big data, capture the characteristics of multiple currencies, as well as the settlement constraints corresponding to the characteristics of chemical products. Calculate the settlement sensitivity coefficient based on the settlement constraints. Based on the settlement constraints, chemical product characteristics, and multiple currency characteristics, a first settlement scheme is generated. The first settlement time of the first settlement scheme is analyzed, and delay analysis and change prediction are performed to obtain the first delay amount and the first predicted change amount. Based on the settlement sensitivity coefficient, the first delay amount and the first predicted change amount are processed and calculated to obtain the first settlement risk level. The settlement scheme is then iteratively optimized to obtain the optimal settlement scheme, and settlement is carried out.

2. The multi-currency intelligent settlement method for cross-border chemical trade according to claim 1, characterized in that, In cross-border chemical trade, it is necessary to obtain the characteristics of chemical products, and through big data, capture the characteristics of multiple currencies, as well as the settlement constraints corresponding to the characteristics of chemical products, including: In cross-border chemical trade, it is important to obtain the chemical product characteristics of chemical products, which include value characteristics and category characteristics. By leveraging big data, we can capture multiple currency characteristics for various currencies, including exchange rates for each currency. Obtain the settlement constraints corresponding to the characteristics of the chemical product.

3. The multi-currency intelligent settlement method for cross-border chemical trade according to claim 2, characterized in that, Obtaining the settlement constraints corresponding to the characteristics of the chemical product includes: Extract the category features within the characteristics of the chemical product; By using big data, the currencies that can be settled for the current category characteristics are obtained, at least one settlement currency is obtained, and settlement constraints are constructed to use only the at least one settlement currency for the current chemical product settlement.

4. The multi-currency intelligent settlement method for cross-border chemical trade according to claim 1, characterized in that, Based on the aforementioned settlement constraints, calculate the settlement sensitivity coefficient, including: Calculate the proportion of at least one settlement currency among multiple currencies, and use it as the settlement sensitivity coefficient.

5. The multi-currency intelligent settlement method for cross-border chemical trade according to claim 1, characterized in that, Based on the settlement constraints, chemical product characteristics, and multiple currency characteristics, a first settlement scheme is generated. The first settlement time of the first settlement scheme is analyzed, and delay analysis and change prediction are performed to obtain a first delay amount and a first predicted change amount, including: According to the settlement constraints and the value characteristics within the chemical product characteristics, at least one or more of at least one settlement currency are randomly selected to obtain at least one first currency, and at least one first settlement amount equal to the value characteristics is randomly allocated according to the corresponding currency characteristics, as the first settlement scheme. Based on at least one first settlement amount of at least one first currency in the first settlement scheme, the settlement time is predicted to obtain the first settlement time; Based on the first settlement time, a delay analysis is performed to obtain the first delay amount; Based on the first settlement time, at least one first currency is predicted to change, and a first predicted change amount is obtained.

6. The multi-currency intelligent settlement method for cross-border chemical trade according to claim 5, characterized in that, Based on at least one first settlement amount of at least one first currency within the first settlement scheme, a settlement time prediction is performed to obtain the first settlement time, including: Multiple settlement time prediction agents corresponding to multiple currencies are invoked. The training data for each settlement time prediction agent is a set of sample settlement amounts, and the supervision label is a set of sample currency settlement times. Input at least one first settlement amount of at least one first currency in the first settlement scheme into the corresponding settlement time prediction agent, and output the settlement time of at least one currency. Select the maximum value among the settlement times of at least one currency as the first settlement time.

7. The multi-currency intelligent settlement method for cross-border chemical trade according to claim 5, characterized in that, Based on the first settlement time, a delay analysis is performed to obtain the first delay amount, including: Obtain a delay classification table corresponding to the category characteristics of chemical products. The delay classification table includes the mapping relationship between sample settlement time and sample delay. Each sample delay includes the proportion of chemical product failure caused by settlement time. Input the first settlement time into the delay amount classification table to obtain the first delay amount.

8. The multi-currency intelligent settlement method for cross-border chemical trade according to claim 5, characterized in that, Based on the first settlement time, at least one first currency's change is predicted to obtain a first predicted change amount, including: Obtain at least one currency feature within a preset time range in the past to obtain at least one first currency feature sequence; Within the at least one first currency feature sequence, extract the first currency feature sequence within the last first settlement time range, and calculate the change range as the change amount of at least one first currency feature; The first predicted change is calculated based on at least one characteristic change of the first currency.

9. The multi-currency intelligent settlement method for cross-border chemical trade according to claim 1, characterized in that, Based on the settlement sensitivity coefficient, the first delay amount and the first predicted change amount are processed and calculated to obtain the first settlement risk level. The settlement scheme is then iteratively optimized to obtain the optimal settlement scheme, and settlement is performed, including: Configure the settlement sensitivity coefficient as a variable weight and calculate the delay weight; Using the aforementioned delay weight and change weight, the first delay amount and the first predicted change amount are weighted and calculated to obtain the first settlement risk level; Continue generating settlement plans and iteratively optimizing them until convergence. The settlement plan with the lowest settlement risk is retained as the optimal settlement plan.

10. A multi-currency intelligent settlement method for cross-border chemical trade, characterized in that: The system is used to implement the multi-currency smart settlement method for cross-border chemical trade as described in any one of claims 1 to 9, the system comprising: The data acquisition module is used to acquire the characteristics of chemical products in cross-border chemical trade, and to capture the characteristics of multiple currencies and the settlement constraints corresponding to the characteristics of chemical products through big data. The sensitivity coefficient calculation module is used to calculate the settlement sensitivity coefficient based on the settlement constraints. The scheme generation and analysis module is used to generate a first settlement scheme according to the settlement constraints, chemical product characteristics and multiple currency characteristics, analyze the first settlement time of the first settlement scheme, perform delay analysis and change prediction, and obtain the first delay amount and the first predicted change amount. The iterative optimization settlement module is used to process and calculate the first delay amount and the first predicted change amount according to the settlement sensitivity coefficient, obtain the first settlement risk level, and perform iterative optimization of the settlement scheme to obtain the optimal settlement scheme and perform settlement.

Images