Cross-border clearing path generation method and device and electronic equipment

Abstract

The invention provides a cross-border clearing path generation method and device and electronic equipment. The method comprises the following steps: acquiring a cross-border liquidation network, and constructing a weighted directed graph according to the cross-border liquidation network; obtaining a historical cross-border liquidation path, and generating a path generation model based on the historical cross-border liquidation path and the weighted directed graph; after obtaining the first cross-border liquidation requests, inputting the first cross-border liquidation requests to a path generation model, and processing the first cross-border liquidation requests by using the path generation model to obtain a plurality of first liquidation paths corresponding to the first cross-border liquidation requests; and expanding each first clearing path according to a non-dominated sorting genetic algorithm to obtain a plurality of second clearing paths, and generating a cross-border clearing path set according to the first clearing paths and the second clearing paths. The problem that the accuracy of clearing path selection is low due to the fact that existing cross-border clearing path selection is insufficient in the aspects of multi-target adaptability, real-time performance and comprehensive consideration is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cross-border clearing path generation, in particular to a cross-border clearing path generation method and device, a computer readable storage medium and an electronic device. BACKGROUND

[0002] The cross-border clearing system usually needs to complete transaction processing through multiple proxy channels. Unlike the domestic currency clearing system of a single country, domestic and foreign banks need to interact across jurisdictions and different clearing platforms. Such multi-level network structure brings flexibility to cross-border clearing, but also introduces complexity: the path may involve multiple transit lines, multiple currency conversions, different compliance rules, etc., making the path topology loose, unstable and opaque.

[0003] In existing practice, cross-border remittance path selection is often regarded as a shortest path problem in graph theory: bank institutions are regarded as nodes in the graph, proxy relationships are regarded as directed weight edges, and edge weights can be set by indicators such as cost or delay. Through algorithms, the optimal path from the starting point to the end point is found. For example, the current "remittance route optimization" service of the cross-border clearing company uses the classic Dijkstra shortest path algorithm. This algorithm constantly updates the shortest distance from each node to the starting point, and finally determines a path with the minimum cost or time. This method can quickly give a low-cost or low-latency path solution to a certain extent, but it relies on a single weight set in advance and fails to take into account other important factors such as exchange rate risk, business success rate and compliance requirements.

[0004] In addition, actual business personnel often need to combine time zone, compliance requirements and other information to manually search for multiple possible paths before making routing decisions. Dynamic factors such as exchange rate fluctuations and real-time bank availability often change, further increasing the uncertainty of path selection. As the clearing routing field says, the most effective decision should be based on real-time performance data such as historical success rate and current network status to select the optimal path; the current method mainly relies on rule configuration or single-target algorithm, which is difficult to quickly process such multi-dimensional information and cannot guarantee that the selected path is optimal and compliant in a multi-target sense.

[0005] In summary, the network infrastructure for cross-border clearing has matured (SWIFT, CIPS, etc.), and specialized systems have accumulated a large amount of cross-border proxy relationships and business statistics data. Existing technologies mainly rely on graph algorithms for single-target path planning or static rules for routing decisions. However, as business scales expand, globalization and compliance requirements tighten, these methods have shortcomings in multi-target adaptability, real-time performance and comprehensive consideration, and more flexible and intelligent optimization methods are needed. SUMMARY

[0006] The main purpose of the present application is to provide a cross-border clearing path generation method, device, computer readable storage medium and electronic equipment, so as to at least solve the problem that the path selection of the existing cross-border clearing has defects in multi-target adaptability, real-time performance and comprehensive consideration, resulting in low accuracy of clearing path selection.

[0007] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a cross-border clearing path generation method is provided, comprising: acquiring a cross-border clearing network, and constructing a weighted directed graph according to the cross-border clearing network, wherein the weighted directed graph comprises nodes, directed edges and edge weights, the nodes represent business institutions of cross-border clearing, the directed edges represent the clearing relationship between the business institutions, and the edge weights are the weights of the directed edges; acquiring historical cross-border clearing paths, and generating a path generation model based on the historical cross-border clearing paths and the weighted directed graph; after acquiring a first cross-border clearing request, inputting the first cross-border clearing request into the path generation model to process the first cross-border clearing request by using the path generation model to obtain a plurality of first clearing paths corresponding to the first cross-border clearing request; expanding each of the first clearing paths according to a non-dominated sorting genetic algorithm to obtain a plurality of second clearing paths, and generating a cross-border clearing path set according to the first clearing paths and the second clearing paths.

[0008] Optionally, expanding each of the first clearing paths according to the non-dominated sorting genetic algorithm to obtain a plurality of second clearing paths comprises: performing non-dominated sorting on a plurality of the first clearing paths based on the non-dominated sorting genetic algorithm to obtain a path priority sequence, and determining the crowding distance of each of the first clearing paths; selecting a preset number of the first clearing paths to construct an expansion path set according to the crowding distance of each of the first clearing paths and the path priority sequence; and performing genetic operations on the expansion path set based on the non-dominated sorting genetic algorithm to obtain a plurality of the second clearing paths.

[0009] Optionally, performing non-dominated sorting on a plurality of the first clearing paths to obtain a path priority sequence comprises: determining a target vector of each of the first clearing paths according to the formula: wherein f is the target vector, is the historical selection rate of the first clearing path, is the transaction success rate of the first clearing path, is the compliance score of the first clearing path, is the exchange rate volatility index of the first clearing path, is the timeliness index of the first clearing path, ​The total cost index of the first liquidation path is used; the multiple first liquidation paths are non-dominated and sorted according to the target vector of each first liquidation path to obtain the path priority sequence.

[0010] Optionally, performing genetic operations on the extended path set based on the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths includes: a selection step, used to arbitrarily select two first liquidation paths from the extended path set as a first crossover mutation path and a second crossover mutation path based on the non-dominated sorting genetic algorithm; an exchange processing step, used to exchange a first fragment path in the first crossover mutation path with a second fragment path in the second crossover mutation path to obtain two second liquidation paths, wherein the path endpoints of the first fragment path and the second fragment path are the same, and the path endpoints are the business units; executing the selection step and the exchange processing step at least once to obtain multiple second liquidation paths.

[0011] Optionally, after generating a cross-border clearing path set based on the first clearing path and the second clearing path, the method further includes: sending the cross-border clearing path table to the user to obtain the user's path selection result for the cross-border clearing path set, and constructing a preferred path set based on the historical cross-border clearing paths to obtain a target clearing path; and updating the target clearing path to the preferred path set based on the ε-dominance mechanism.

[0012] Optionally, updating the target liquidation path to the preferred path set based on the ε-dominance mechanism includes: determining whether the number of paths in the preferred path set is greater than or equal to a path number threshold; if the number of paths in the preferred path set is greater than or equal to the path number threshold, calculating the difference value between the target liquidation path and each path in the preferred path set based on the ε-dominance mechanism; determining the path with the smallest difference value between the preferred path set and the target liquidation path as a similar path; determining the target vectors of the target liquidation path and the similar path respectively; if the target vector of the target liquidation path is greater than the target vector of the similar path, adding the target liquidation path to the preferred path set and removing the similar path from the preferred path set.

[0013] Optionally, after updating the target clearing path to the preferred path set based on the ε-dominance mechanism, the method further includes: obtaining the user's second cross-border clearing request, and matching the second cross-border clearing request according to the preferred path set to generate a cross-border clearing path corresponding to the second cross-border clearing request.

[0014] According to another aspect of this application, a cross-border clearing path generation apparatus is provided, comprising: a first acquisition unit, configured to acquire a cross-border clearing network, construct a weighted directed graph based on the cross-border clearing network, wherein the weighted directed graph includes nodes, directed edges, and edge weights, the nodes representing cross-border clearing business entities, the directed edges representing clearing relationships between the business entities, and the edge weights being the weights of the directed edges; a second acquisition unit, configured to acquire historical cross-border clearing paths, construct a path generation model based on the historical cross-border clearing paths and the weighted directed graph, acquire a first cross-border clearing request, input the first cross-border clearing request into the path generation model, and obtain multiple first clearing paths corresponding to the first cross-border clearing request; and an extension unit, configured to extend each first clearing path according to the NSGA-II algorithm to obtain multiple second clearing paths, construct a cross-border clearing path table from the first clearing paths and the second clearing paths, and send the cross-border clearing path table to a user.

[0015] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform any of the aforementioned methods for generating cross-border clearing paths.

[0016] According to another aspect of this application, an electronic device is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including a method for performing any of the methods for generating cross-border clearing paths.

[0017] Applying the technical solution of this application, a weighted directed graph is first constructed based on the cross-border clearing network. Nodes and directed edges correspond to the clearing relationships between business institutions and these institutions, while edge weights quantify the weight attributes of the clearing process. Subsequently, a path generation model is established by analyzing historical cross-border clearing paths. Upon receiving the first cross-border clearing request, the model can quickly process and output multiple first clearing paths. Next, based on the Non-Dominated Ranking Genetic Algorithm (NSGA-II), the first clearing paths are expanded and optimized to generate multiple second clearing paths. Finally, all paths are integrated to form a cross-border clearing path set. This series of operations enables the algorithm to capture the optimal solution in real time amidst dynamic information such as exchange rate fluctuations, time zone effects, and real-time bank availability, while ensuring comprehensive consideration of multiple objectives such as compliance, selection rate, and success rate. Automated path recommendation reduces the manual query burden on business personnel and improves decision-making efficiency. This solves the problem of low accuracy in clearing path selection due to shortcomings in multi-objective adaptability, real-time performance, and comprehensive consideration in existing cross-border clearing path selection methods. Attached Figure Description

[0018] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0019] Figure 1 A hardware structure block diagram of a mobile terminal for executing a method for generating a cross-border clearing path is shown in an embodiment of this application.

[0020] Figure 2 A flowchart illustrating a method for generating a cross-border clearing path according to an embodiment of this application is shown.

[0021] Figure 3 A flowchart illustrating a specific method for generating a cross-border clearing path according to an embodiment of this application is shown.

[0022] Figure 4 A structural block diagram of a cross-border clearing path generation apparatus provided according to an embodiment of this application is shown; Detailed Implementation

[0023] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0025] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0026] For ease of description, the following explains some of the nouns or terms used in the embodiments of this application:

[0027] Cross-border clearing: refers to the business process of clearing and settling funds between banks or financial institutions in different regions.

[0028] Multi-objective optimization refers to problems that simultaneously optimize two or more conflicting objective functions. Compared to single-objective optimization, multi-objective optimization seeks a set of Pareto optimal solutions, meaning that no single solution can further improve a particular objective without compromising other objectives.

[0029] Pareto Dominance and Pareto Optimal Solution: For two solutions If for all targets All And there exists at least one objective that is strictly superior to the other objective, that is If the Pareto dominance is achieved, it is called Pareto dominance, also known as... Dominate If no other solution dominates the solution. Then it is called This is the Pareto optimal solution. The boundary of the objective space formed by the set of Pareto optimal solutions is called the Pareto front.

[0030] NSGA-II Algorithm: NSGA-II (Non-dominated sorting genetic algorithm II) is a multi-objective genetic algorithm. It combines fast non-dominated sorting, crowding distance selection, and elitist strategy to effectively maintain the diversity of solutions in the population and find the optimal Pareto solution set.

[0031] Chromosomes, genes, population, and fitness: In genetic algorithms, each candidate solution is called an "individual" or "chromosome," and its internal components are "genes." The population is the set of all chromosomes included in the current iteration. The fitness function is used to evaluate the quality of solutions corresponding to chromosomes, which is a comprehensive consideration of ranking rank and crowding distance. In this scheme, chromosomes can be used to encode the specific transit sequence of a cross-border clearing path.

[0032] Selection rate: This refers to the proportion of a clearing path that has been used in historical transactions, reflecting the path's popularity or maturity. A higher selection rate means that the path is used more frequently and may have higher stability.

[0033] Success rate: refers to the percentage of transactions that successfully arrive after a transaction is initiated using a certain path, used to measure the reliability and feasibility of the path.

[0034] Exchange rate volatility refers to the degree of change in the exchange rate of the currency pair involved. It can be calculated using historical data to determine the standard deviation or range of variation. High volatility indicates greater exchange rate risk.

[0035] Timeliness: refers to the time or delay required for the transfer of funds along the route; the shorter the overall timeliness, the better.

[0036] Costs: This refers to the sum of various handling fees and exchange costs along the route. The lower the total cost, the better.

[0037] Compliance: This refers to the degree to which the path meets compliance and cross-border regulations. Since compliance requirements are typically hard constraints, this solution treats compliance as a prerequisite for decision-making or a high-priority objective.

[0038] Non-dominated sorting: The solutions in the population are sorted hierarchically. The first layer is the Pareto front (not dominated by any solution), the second layer is the Pareto front of the solutions remaining after removing the solutions in the first layer, and so on.

[0039] Crowding distance: Used to measure the density of solutions in the Pareto front. For a given solution, the population is first sorted for each objective function, and the boundary points are assigned a maximum crowding distance. The remaining solutions are calculated with normalized distances based on the differences between the objective values ​​of their two nearest neighbors, and then summed. This invention introduces weights to form a weight-driven crowding reconstruction: that is, the neighbor differences of each objective are multiplied by a preset weight and then summed, thus strengthening the influence of important objectives on the overall crowding.

[0040] Elite preservation strategy: During evolution, the best-performing individuals (Pareto front) in the current population are retained to avoid loss during genetic operations, thereby improving the algorithm's convergence efficiency and solution quality. NSGA-II itself achieves elite preservation by merging parent and child populations and pruning the top few individuals.

[0041] External Elite Archive: In standard NSGA-II, elites are preserved by merging parent and child generations, but there is no separate "historical best solution archive" outside of each generation. To ensure that optimal solutions are not lost, an additional independent set is maintained specifically to store the set of non-dominated solutions discovered so far (i.e., the current best Pareto front). The elite archive does not directly participate in genetic operations; it can be used to provide users with the set of optimal solutions, as a reference for selection operators, and to prevent the loss of optimal solutions.

[0042] As described in the background section, existing cross-border clearing path selection methods are insufficient in terms of multi-objective adaptability, real-time performance, and comprehensive consideration, resulting in low accuracy of clearing path selection. To address the problem of insufficient multi-objective adaptability, real-time performance, and comprehensive consideration in existing cross-border clearing path selection methods, resulting in low accuracy of clearing path selection, embodiments of this application provide a method, apparatus, computer-readable storage medium, and electronic device for generating cross-border clearing paths.

[0043] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0044] The methods and embodiments provided in this application can be executed on a mobile terminal, computer terminal, or similar computing device. Taking running on a mobile terminal as an example, Figure 1 This is a hardware structure block diagram of a mobile terminal for a method of generating a cross-border clearing path according to an embodiment of the present invention. For example... Figure 1 As shown, a mobile terminal may include one or more ( Figure 1 Only one is shown in the diagram. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. The mobile terminal may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the mobile terminal described above. For example, the mobile terminal may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0045] The memory 104 can be used to store computer programs, such as application software programs and modules, like the computer program corresponding to the cross-border clearing path generation method in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, thereby implementing the above-described method. The memory 104 may include high-speed random access memory and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the mobile terminal via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof. The transmission device 106 is used to receive or send data via a network. Specific examples of the aforementioned networks may include wireless networks provided by the mobile terminal's communication provider. In one example, the transmission device 106 includes a network interface controller (NIC), which can be connected to other network devices via a base station to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0046] This embodiment provides a method for generating a cross-border clearing path that runs on a mobile terminal, computer terminal, or similar computing device. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Also, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0047] Figure 2 This is a flowchart of a method for generating a cross-border clearing path according to an embodiment of this application. Figure 2 As shown, the method includes the following steps:

[0048] Step S201: Obtain the cross-border clearing network and construct a weighted directed graph based on the cross-border clearing network. The weighted directed graph includes nodes, directed edges, and edge weights. The nodes represent cross-border clearing business entities, the directed edges represent clearing relationships between the business entities, and the edge weights are the weights of the directed edges.

[0049] Specifically, a cross-border clearing network is a network system used to transfer funds between financial institutions in different countries and regions. A weighted directed graph is constructed based on the clearing relationships between the various financial institutions and business units within the cross-border clearing network.

[0050] Step S202: Obtain historical cross-border clearing paths and generate paths based on the aforementioned historical cross-border clearing paths and the aforementioned weighted directed graph path generation model.

[0051] Specifically, a weighted directed graph has nodes, directed edges, and edge weights. Historical cross-border clearing paths are cross-border clearing paths from the starting point to the end point. The path includes multiple nodes, which are connected by directed edges. Multiple historical cross-border clearing paths are input into the weighted directed graph to construct a path generation model.

[0052] Step S203: After obtaining the first cross-border clearing request, the first cross-border clearing request is input into the path generation model to process the first cross-border clearing request using the path generation model to obtain multiple first clearing paths corresponding to the first cross-border clearing request.

[0053] Specifically, a cross-border clearing request includes the business institution's starting point, business institution's ending point, and business type. Based on the business institution's starting point, business institution's ending point, and business type, the path generation model matches one of multiple historical cross-border clearing paths that match the cross-border clearing request.

[0054] Step S204: Expand each of the first liquidation paths according to the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, and generate a cross-border liquidation path set based on the first liquidation paths and the second liquidation paths.

[0055] Specifically, the non-dominated sorting genetic algorithm is the NSGA-II algorithm. The NSGA-II algorithm combines fast non-dominated sorting, crowding distance selection, and elitist strategy, which can effectively maintain the diversity of solutions in the population and find the optimal Pareto solution set.

[0056] In this embodiment, applying steps S201, S202, S203, and S204, a weighted directed graph is first constructed based on the cross-border clearing network. Nodes and directed edges correspond to the clearing relationships between business institutions and these institutions, while edge weights quantify the weight attributes of the clearing process. Subsequently, a path generation model is established by analyzing historical cross-border clearing paths. Upon receiving a first cross-border clearing request, the model quickly processes and outputs multiple first clearing paths. Next, based on the Non-Dominated Ranking Genetic Algorithm (NSGA-II), the first clearing paths are expanded and optimized to generate multiple second clearing paths. Finally, all paths are integrated to form a cross-border clearing path set. This series of operations enables the algorithm to capture the optimal solution in real time amidst dynamic information such as exchange rate fluctuations, time zone effects, and bank real-time availability, while ensuring comprehensive consideration of multiple objectives such as compliance, selection rate, and success rate. Automated path recommendation reduces the manual query burden on business personnel and improves decision-making efficiency. It solves the problem of low accuracy in clearing path selection due to shortcomings in multi-objective adaptability, real-time performance, and comprehensive consideration in existing cross-border clearing path selection methods.

[0057] In the specific implementation process, the first liquidation paths are expanded according to the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, including: performing non-dominated sorting on the multiple first liquidation paths based on the non-dominated sorting genetic algorithm to obtain a path priority sequence, and determining the congestion distance of each of the first liquidation paths; selecting a preset number of the first liquidation paths to construct an extended path set based on the congestion distance and the path priority sequence of each of the first liquidation paths; and performing genetic operations on the extended path set based on the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths.

[0058] In this embodiment, multiple first liquidation paths are optimized and expanded using a non-dominated sorting genetic algorithm to obtain multiple second liquidation paths. First, the first liquidation paths are prioritized using non-dominated sorting to generate a path priority sequence, while simultaneously calculating the crowding distance of each path to quantify their distribution density in the solution space. Next, based on the crowding distance and priority order, a predetermined number of paths are selected from the first liquidation paths to form an expanded path set. These paths are considered high-quality candidate solutions due to their Pareto front characteristics. Subsequently, genetic operations, including crossover and mutation, are performed on the expanded path set to generate new second liquidation paths. This process allows the algorithm to explore more potential path combinations, enriching the solution set in order to find better Pareto solutions. The genetic operations follow the basic principles of NSGA-II. By introducing a specific multi-layer path encoding scheme and a weight-driven crowding reconstruction strategy, the evolutionary process maintains both the diversity of solutions and highlights the impact of key objectives (such as compliance and selection rate), thereby improving the overall quality and search efficiency of the path solution set.

[0059] Specifically, a non-dominated sort is performed on multiple of the aforementioned first liquidation paths to obtain a path priority sequence, including: according to the formula: Determine the target vector for each of the aforementioned first liquidation paths, where, For the above target vector, The historical selection rate of the aforementioned first liquidation path, The success rate of transactions via the first clearing path mentioned above. The compliance score for the first liquidation path mentioned above, This serves as an indicator of exchange rate volatility for the first clearing path mentioned above. For the timeliness indicators of the aforementioned first liquidation path, The total cost index of the first liquidation path is given; the first liquidation paths are non-dominated and sorted according to their target vectors to obtain the path priority sequence.

[0060] In this embodiment, when optimizing multiple first clearing paths, a target vector F is first defined for each path, consisting of historical selection rate, transaction success rate, compliance score, exchange rate volatility index, timeliness index, and total cost index. A non-dominated ranking algorithm is used to hierarchically divide the first clearing paths according to the target vector, forming a path priority sequence. In specific implementation, the path's historical selection rate, transaction success rate, and compliance are used as positive evaluation indicators, while exchange rate volatility, timeliness, and cost are used as negative evaluation indicators. After normalization, these are transformed into a unified optimization objective. This ranking mechanism ensures that the algorithm can automatically identify and prioritize those cross-border clearing paths that perform best in multi-objective evaluation. The final output path priority sequence can guide business personnel to make efficient and accurate cross-border remittance path selections, considering cost and timeliness, as well as success rate, compliance, and exchange rate risk, achieving optimal path recommendation under multi-objective optimization.

[0061] More specifically, the genetic operation is performed on the extended path set based on the aforementioned non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, including: a selection step, used to arbitrarily select two of the aforementioned first liquidation paths from the extended path set as a first crossover mutation path and a second crossover mutation path based on the aforementioned non-dominated sorting genetic algorithm; an exchange processing step, used to exchange the first segment path in the aforementioned first crossover mutation path with the second segment path in the aforementioned second crossover mutation path to obtain two of the aforementioned second liquidation paths, wherein the path endpoints of the aforementioned first segment path and the aforementioned second segment path are the same, and the path endpoints are the aforementioned business institutions; the selection step and the exchange processing step are executed at least once to obtain multiple of the aforementioned second liquidation paths.

[0062] This embodiment, based on an improved NSGA-II algorithm, performs genetic operations on an expanded path set to explore and generate more second liquidation paths that meet multi-objective optimization criteria. In the selection step, a binary tournament selection mechanism randomly selects two first liquidation paths from the expanded path set as the base individuals for the genetic operation. In the swapping step, a partial mapping crossover (PMX) strategy is employed to ensure the invariance of the path endpoints (i.e., the business initiator and receiver), swapping only the intermediate node sequences, i.e., fragment paths, to generate two new second liquidation paths. This operation guarantees the connectivity and feasibility of the new paths while allowing the algorithm to perform crossovers in multi-level path encoding, exploring a wider range of path combinations that meet business needs. During the algorithm's iteration process, the selection and swapping steps are executed at least once, with each iteration generating multiple second liquidation paths, enriching population diversity and enhancing the algorithm's search capability and the quality of recommended paths. In this way, not only are existing historically optimal paths preserved, but new paths are also generated and optimized. This achieves a multi-objective genetic search for the cross-border clearing path selection problem, improves the overall competitiveness of recommended paths, reduces transaction costs, shortens clearing time, increases business success rate, ensures compliance, and forms a dynamically updated and highly efficient path selection system.

[0063] In addition, the crossover mutation operation in this embodiment may also include random insertion or deletion of a gene (node), which further enhances the diversity of the population while maintaining path coherence, and promotes the algorithm to converge to the Pareto optimal front faster, thereby comprehensively improving the planning efficiency and quality of cross-border remittance paths in a multi-objective sense.

[0064] Furthermore, after generating the cross-border clearing path set based on the first clearing path and the second clearing path, the method further includes: sending the cross-border clearing path table to the user to obtain the user's path selection results for the cross-border clearing path set, constructing a preferred path set based on the historical cross-border clearing paths to obtain the target clearing path; and updating the target clearing path to the preferred path set based on the ε-dominance mechanism.

[0065] This solution sends the generated cross-border clearing path set to users to obtain their path selection results based on business needs and market conditions. User feedback is integrated into a historical cross-border clearing path database to construct a preferred path set, which is then iteratively optimized to obtain the target clearing path. Based on the ε-dominance mechanism, the target clearing path is updated in the preferred path set, ensuring that the solution set within the archive always reflects the optimal multi-objective balance under current conditions. The ε-dominance mechanism, by setting a target difference threshold, allows merging of similar solutions within a small performance difference range, effectively controlling the size of the external elite archive, avoiding excessive redundancy in the solution set, while ensuring the diversity of the solution set and the recording of historical optimality. This dynamic maintenance mechanism reflects the continuous evolution of the algorithm and enhances the real-time adaptability and business guidance value of the recommended paths, enabling this solution to better serve the ever-changing cross-border clearing environment and provide more accurate, efficient, and compliant path optimization services.

[0066] Furthermore, based on the ε-dominance mechanism, the target liquidation path is updated to the preferred path set, including: determining whether the number of paths in the preferred path set is greater than or equal to a path number threshold; if the number of paths in the preferred path set is greater than or equal to the path number threshold, calculating the difference value between the target liquidation path and each path in the preferred path set based on the ε-dominance mechanism; determining the path with the smallest difference value between the preferred path set and the target liquidation path as a similar path; determining the target vectors of the target liquidation path and the similar path respectively; if the target vector of the target liquidation path is greater than the target vector of the similar path, adding the target liquidation path to the preferred path set and removing the similar path from the preferred path set.

[0067] When updating external elite profiles to the preferred path set based on the ε-dominance mechanism, this scheme first determines whether the number of paths in the preferred path set has reached a preset threshold. If the number of paths has reached or exceeded the threshold, the algorithm enters the refinement process of the ε-dominance mechanism. For each target clearing path, the difference between it and each path in the existing preferred path set is calculated, i.e., the difference between the objective function values. Based on the calculation results, the path with the smallest difference from the target clearing path is identified as a similar path. Subsequently, the objective vectors of these two paths are compared. If the target clearing path is superior to the similar path in all objective function values ​​(or has a significant improvement in key indicators), the target clearing path is added to the preferred path set, and the original similar paths are removed accordingly, keeping the size of the preferred path set within the threshold. This mechanism ensures that the preferred path set always contains the best path options, guaranteeing both diversity and maintaining global optimality, providing business personnel with efficient and compliant cross-border clearing path decision support. By dynamically updating the preferred path set, changes in the clearing network, such as exchange rate fluctuations and bank availability, can be reflected in real time, thereby improving the accuracy and applicability of recommended paths and reducing business risks and costs.

[0068] Specifically, after updating the target clearing path to the preferred path set based on the ε-dominance mechanism, the method further includes: obtaining the user's second cross-border clearing request, and matching the second cross-border clearing request according to the preferred path set to generate a cross-border clearing path corresponding to the second cross-border clearing request.

[0069] This solution, after updating the target clearing path to the preferred path set using an ε-dominance mechanism, also includes receiving new cross-border clearing requests from users and quickly matching a suitable cross-border clearing path based on the historically best Pareto front solutions in an external elite database. Specifically, by querying the external elite database, the algorithm can select the most suitable solution from the accumulated high-quality paths in the database as the recommendation result based on the specific conditions of the new request (such as currency, amount, and timeliness requirements). This process greatly accelerates the decision-making cycle because the paths in the external elite database have undergone rigorous multi-objective optimization and verification, ensuring the comprehensive optimality of the recommended path in terms of cost, success rate, and compliance.

[0070] To enable those skilled in the art to better understand the technical solution of this application, the implementation process of the method for generating cross-border clearing paths of this application will be described in detail below with reference to specific embodiments.

[0071] This embodiment relates to a specific method for generating cross-border clearing routes. Based on a cross-border clearing network, an optimization model is established and a complete genetic evolution process adapted to bank clearing route selection scenarios is designed, such as...Figure 3 As shown, it includes the following steps:

[0072] S1: Model Construction and Coding Design: The cross-border clearing network is abstracted as a weighted directed graph. Each node in the weighted directed graph represents a banking institution, and each directed edge represents an agency clearing relationship (including domestic and foreign clearing channels, currency conversion, etc.). Edge weights can comprehensively consider attributes such as timeliness and cost. A multi-layer path coding method is used to represent the clearing path: each chromosome consists of several genes, and each gene can represent a transit node and its corresponding currency / clearing channel type. For example, a gene can be defined as a binary sequence of (bank ID, currency) or (clearing type identifier, bank ID), so that the chromosome simultaneously reflects the transit line sequence and the cross-currency conversion level in the path. This multi-layer coding allows the algorithm to simultaneously search for composite paths across multiple channels such as SWIFT and CIPS, and avoid invalid paths. In addition, the gene hierarchy can also constrain the subsequent cross-mutation of candidate paths to ensure the feasibility of the generated paths.

[0073] S2: Objective Function Definition and Normalization: For each path P, the following objective functions are defined:

[0074] Historical path selection rate (the higher the percentage, the better);

[0075] : Success rate of the path (probability of transaction success);

[0076] : Path compliance score (1 for full compliance, 0 otherwise, and the higher the score, the better).

[0077] Exchange rate volatility indicator;

[0078] Timeliness metrics for the route;

[0079] Total cost of the route metric.

[0080] The optimization objectives are compliance, selection rate, and success rate (the higher the better), while exchange rate fluctuations, latency, and costs are considered negative objectives (converted to positive objectives through normalization). In practice, to eliminate differences in objective scales and avoid optimization bias caused by numerical dominance, the calculation of congestion distance needs to be normalized, and the optimization objective should be uniformly set to maximize all of these.

[0081] For metrics where larger is always better, the following normalization process is performed: ;

[0082] For metrics where smaller is always better, the following normalization process is performed to convert them into metrics where larger is always better: ;

[0083] The target vector of the algorithm is defined as:

[0084] ;

[0085] S3: Population Initialization: Directly select used clearing paths as the initial path set, and randomly generate an initial path set of a certain size (population size N). Merge the two sets as the initialization candidate set. Each individual (chromosome) is encoded as a feasible path from the paying bank to the receiving bank (ensuring path connectivity and satisfying currency conversion rules). Initial valid paths can be generated using breadth-first search (BFS) or existing rules.

[0086] S4: Objective Function Calculation: Calculate the objective function value F for each chromosome in both populations (refer to the definition above). If a solution is found to be compliant... =0 is considered a violation of the constraint, and its fitness is set to the worst, causing it to be eliminated quickly.

[0087] S5: Non-dominated sorting: Perform a fast non-dominated sorting of the population solutions based on the target vector. Specifically, for each solution X, count how many other solutions dominate it and record the solution set it dominates; first, classify individuals not dominated by any solution as the first frontier (level 1), then remove these individuals, continue to find a new first frontier as level 2, and so on.

[0088] S6: Crowding Calculation: Calculate the crowding distance for each individual within the Pareto front. For target j, calculate the crowding distance for the population based on the value. Sort the data and set the crowding degree of the extreme points of the objective function to infinity. For intermediate individuals, their crowding degree is calculated based on the difference in objective values ​​between their neighbors: ;

[0089] in Let the target values ​​of the current individual's neighbors be sorted by the j-th target. , These represent the maximum and minimum values ​​of the target in the population, respectively. Weights are introduced here. Give greater weight to compliance, selection rate, and success rate (e.g.) The highest priority targets are given less weight in the congestion calculation, while exchange rate fluctuations, timeliness, and costs are given less weight.

[0090] S7: Selection Operation: A binary tournament selection based on non-dominance rank and crowding degree is adopted to select individuals from the parent population as crossover parents. Specifically, two individuals are randomly selected each time, with priority given to the one with the lower rank; if the ranks are the same, the individual with the higher crowding degree is selected.

[0091] S8: Crossover and Mutation: In scenarios like banking, where stability and security are paramount, business personnel often prefer previously selected paths. However, when no existing path exists, or when new and better, validated paths emerge, crossover and mutation provide a wider range of choices. Crossover and mutation operations are performed on selected parent paths to generate offspring. Considering the unique characteristics of each path, partial mapping crossover (PMX) is used to ensure the reachability of both the starting and ending points. This involves exchanging only identical gene sequence segments at both ends. For example, two parent paths are selected, and random segments in their gene sequences are swapped to generate two child paths. Simultaneously, mutation operations are performed, such as randomly replacing a transit node or inserting a new one. The newly added node must be at the same level as the original node to ensure the new path remains connected. Through crossover and mutation, new path combinations can be generated, expanding the search space and aligning with the GA (Genetic Algorithm) principle of "providing high-quality solutions to the search through genetic operators (selection, crossover, and mutation)."

[0092] S9: Elite Preservation and Population Update: This ensures that the "best known solutions" are not lost during random operations such as crossover / mutation, while maintaining the diversity of the solution set. That is, existing selected paths are not lost due to crossover or mutation, and new selectable paths are provided for business users to choose from. First, the parent generation... and offspring Merging into a new joint population ,exist Perform a non-dominated sort on the top to obtain the frontier. , , …, prioritizing the loading of cutting-edge technologies into the next generation. Until the next one is loaded This would cause the value to exceed the size N, at which point... The crowding is ranked, and individuals with high crowding are selected until a size N is reached. The choice of cross-border clearing routes for banks has a high priority for compliance, so the ranking rules of traditional elite retention strategies need to be improved. First, the joint population... The pathways are selected based on compliance, followed by non-dominated and crowded ranking. This process ensures elite retention: individuals meeting hard prerequisites, non-dominated frontiers, and high crowding levels are prioritized for preservation in the next generation.

[0093] S10: Iteration Termination: Repeat steps S4–S9 until the maximum algebraic number is reached or the convergence condition is met (Pareto front is stable). The final output is a set of paths in the Pareto front set as the recommendation result.

[0094] S11: Update the external elite profile A using the ε-dominance mechanism: The solution set A guarantees that once a feasible selected path exists, it will be recommended to business personnel. To prevent the elite profile from growing indefinitely, its size M needs to be limited. To ensure diversity, the ε-dominance mechanism needs to be used to deduplicate the elite profile during updates. The target space is "gridized," and a precision threshold is set. "differences less than" Individuals exhibiting ε-dominance are considered equivalent and thus merged. ε-dominance definition:

[0095] Individual x ε dominates y if and only if: ;

[0096] If x is no worse than y on every objective by more than [a certain percentage] If y is considered meaningless, then y is considered meaningless. The definition of depends on the importance of the objective. Therefore, when updating the external elite archive, it only needs to traverse new candidate solutions. If a new candidate solution is dominated by a solution in A, it is removed. If the final number of candidate solutions and solutions in the original archive exceeds the size, crowding distance sorting is performed, and solutions of size M are selected to form the elite archive, ensuring that A is always a sparse, diverse, and high-quality historical optimal solution. The external archive can record historical optimal paths for real-time recommendations (even if the main population is still evolving). The archive can serve as a system output interface, allowing decision-makers to directly query multiple Pareto optimal liquidation paths.

[0097] Key parameters in all the above processes can be defined as follows: typical parameters include population size N=100, maximum number of iterations G=200, and crossover probability. =0.8, Probability of Mutation =0.1. Crossover mutation is performed in S8 each time. Used to control the proportion of offspring generated by crossover, a random number is generated after selecting the operator from the parent population. If r < If the result is a crossover, then the copy will be performed; otherwise, the copy will be directly copied from the parent generation to the child generation. This is used to control the degree of random perturbation in offspring, increasing diversity and preventing premature convergence. A random number is generated during the mutation process. ,if < Then, gene mutation operations are performed. After crossover mutation, the reachability of the path needs to be verified, and a feasible repair plan needs to be implemented.

[0098] Target weight In the allocation process, compliance is given the highest weight, followed by selection rate, then success rate, with exchange rate fluctuations, timeliness, and cost having the lowest weight. Non-dominated ranking within the population follows the Pareto dominance definition. Genetic operators execute according to the standard genetic algorithm procedure. It is important to note that compliance, as a decisive objective, can be treated as a constraint: if a path does not meet compliance requirements, its compliance score is the lowest, inevitably leading to Pareto dominance by other solutions.

[0099] The value is This indicates that a difference of 1% to 5% is allowed to be ignored. Setting the value too high will cause many suitable solutions to be filtered out, resulting in sparse files. Setting it too low will cause approximate solutions to be retained, resulting in overly large files that lose their meaning.

[0100] The specific method for generating cross-border clearing paths proposed in this invention has the following significant advantages over existing technologies:

[0101] 1. Multi-objective comprehensive optimization: By simultaneously optimizing multiple indicators, the generated solution set achieves a trade-off balance in terms of selection rate, success rate, compliance, exchange rate volatility, timeliness, and cost. Compared to single-objective optimization, this solution provides multiple Pareto optimal path options for business personnel to select according to actual needs; its automated approach avoids tedious manual multi-party queries. Practice shows that intelligent routing can significantly improve transaction success rate and revenue, resulting in "fewer false rejections, fewer shopping cart abandonments, and higher actual settlement revenue"; similarly, this embodiment can improve the overall success rate of cross-border clearing.

[0102] 2. Cost and Efficiency Optimization: The recommended path considers both transaction fees and timeliness, effectively reducing transaction costs and overall latency. Similarly, small fee differences can yield significant benefits for large-scale transactions; selecting low-cost channels through routing can save substantial costs. This embodiment utilizes the low-cost path found by the algorithm to reduce overall costs without sacrificing success rate.

[0103] 3. Compliance Assurance: This solution treats compliance as a strong constraint, only outputting paths that meet compliance requirements. Compared to traditional algorithms that may ignore legal risks, this invention ensures that all recommended paths comply with the "legal requirements of the final destination," significantly reducing compliance risks in cross-border clearing.

[0104] 4. Strong dynamic adaptability: Since the algorithm evaluates the path based on real-time data, it can quickly provide new solutions by adjusting the objective function and re-optimizing when exchange rates fluctuate, the status of the correspondent bank changes, or compliance regulations change, thus having better adaptability.

[0105] 5. Easily Integratable Decision-Making Tool: Results are presented as multiple candidate paths, allowing banks and enterprises to directly reference them, making them more intuitive to use. Previous reports have indicated that introducing path planning services can "significantly reduce exchange rate rebates and minimize business friction," achieving positive results. If applied to a real-world system, this invention can also improve cross-border clearing efficiency and reduce transaction failures.

[0106] 6. Innovation and Scalability: This solution optimizes the standard NSGA-II: First, it adopts a multi-layer path encoding design, allowing genetic operations to be applied simultaneously to path combinations at different currency levels, enhancing search flexibility. Second, it introduces weight-driven congestion reconstruction, emphasizing high-priority indicators in congestion calculation to better reflect factors valued by business personnel. Furthermore, it explicitly strengthens the elite retention strategy during the selection process, ensuring that high-quality solutions are not lost. Simultaneously, it allows crossover mutation to generate entirely new path combinations, expanding the search space for candidate sinks. Finally, it incorporates an external elite archive for long-term recording of historical non-dominated solutions, ensuring historical diversity and global optimality of historical solutions, and accelerating business response time. These designs enable the algorithm to automatically generate multiple candidate paths that meet business needs, providing decision-makers with various trade-offs.

[0107] This solution innovates upon the NSGA-II algorithm, improving its performance and applicability. These technological innovations are not limited to cross-border clearing but also offer valuable insights for other path planning problems requiring a balance of multiple factors.

[0108] In summary, this solution achieves intelligent optimization of complex cross-border remittance routes by introducing multi-objective genetic optimization and a series of improvement strategies, making up for the shortcomings of existing single-objective, manual, or experience-based routing methods, and has significant business value and application prospects.

[0109] In addition, the information collected is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant countries and regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation portals are provided for users to choose to authorize or refuse.

[0110] This application also provides an apparatus for generating cross-border clearing paths. It should be noted that this apparatus can be used to execute the method for generating cross-border clearing paths provided in this application. This apparatus is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0111] The following describes the cross-border clearing path generation apparatus provided in the embodiments of this application.

[0112] Figure 4 This is a schematic diagram of a cross-border clearing path generation apparatus according to an embodiment of this application. Figure 4 As shown, the device includes:

[0113] The first acquisition unit 41 is used to acquire the cross-border clearing network and construct a weighted directed graph based on the cross-border clearing network. The weighted directed graph includes nodes, directed edges and edge weights. The nodes represent cross-border clearing business institutions, the directed edges represent clearing relationships between the business institutions, and the edge weights are the weights of the directed edges.

[0114] The second acquisition unit 42 is used to acquire historical cross-border clearing paths, based on the aforementioned historical cross-border clearing paths and the aforementioned weighted directed graph path generation model.

[0115] The third acquisition unit 43 is used to input the first cross-border clearing request into the path generation model after acquiring the first cross-border clearing request, so as to use the path generation model to process the first cross-border clearing request to obtain multiple first clearing paths corresponding to the first cross-border clearing request.

[0116] The extension unit 44 is used to extend each of the first liquidation paths according to the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, and to generate a cross-border liquidation path set based on the first liquidation paths and the second liquidation paths.

[0117] In this embodiment, the first acquisition unit is used to acquire the cross-border clearing network and construct a weighted directed graph based on the cross-border clearing network. The weighted directed graph includes nodes, directed edges, and edge weights. Nodes represent cross-border clearing business entities, directed edges represent clearing relationships between business entities, and edge weights are the weights of the directed edges. The second acquisition unit is used to acquire historical cross-border clearing paths and generate a path generation model based on the historical cross-border clearing paths and the weighted directed graph. The third acquisition unit, after acquiring a first cross-border clearing request, inputs the first cross-border clearing request into the path generation model to process the first cross-border clearing request and obtain multiple first clearing paths corresponding to the first cross-border clearing request. The expansion unit is used to expand each first clearing path according to a non-dominated sorting genetic algorithm to obtain multiple second clearing paths, and generate a cross-border clearing path set based on the first and second clearing paths. First, a weighted directed graph is constructed based on the cross-border clearing network, where nodes and directed edges correspond to the clearing relationships between business entities and them, and edge weights quantify the weight attributes of clearing. Subsequently, a path generation model is established by analyzing historical cross-border clearing paths. Upon receiving the first cross-border clearing request, the model can quickly process and output multiple first clearing paths. Next, based on the Non-Dominated Ranking Genetic Algorithm (NSGA-II), the first clearing paths are expanded and optimized to generate multiple second clearing paths. Finally, all paths are integrated to form a cross-border clearing path set. This series of operations enables the algorithm to capture the optimal solution in real time amidst dynamic information such as exchange rate fluctuations, time zone effects, and real-time bank availability, while ensuring comprehensive consideration of multiple objectives such as compliance, selection rate, and success rate. Automated path recommendation reduces the manual query burden on business personnel and improves decision-making efficiency. It addresses the shortcomings of existing cross-border clearing path selection methods in terms of multi-objective adaptability, real-time performance, and comprehensive consideration, which leads to low accuracy in clearing path selection.

[0118] As an optional scheme, the extension unit includes a first determining module, a first selecting module, and a first executing module; the first determining module is used to perform non-dominated sorting on multiple first liquidation paths based on the above-mentioned non-dominated sorting genetic algorithm to obtain a path priority sequence, and determine the congestion distance of each of the first liquidation paths; the first selecting module is used to select a preset number of the first liquidation paths to construct an extended path set according to the congestion distance and the path priority sequence of each of the first liquidation paths; the executing module is used to perform genetic operations on the extended path set based on the above-mentioned non-dominated sorting genetic algorithm to obtain multiple of the second liquidation paths.

[0119] In one alternative approach, the first determining module includes a determining submodule and a sorting submodule. The determining submodule is used to determine the order based on the formula: Determine the target vector for each of the aforementioned first liquidation paths, where, For the above target vector, The historical selection rate of the aforementioned first liquidation path, The success rate of transactions via the first clearing path mentioned above. The compliance score for the first liquidation path mentioned above, This serves as an indicator of exchange rate volatility for the first clearing path mentioned above. For the timeliness indicators of the aforementioned first liquidation path, The total cost index of the first liquidation path is given; the sorting submodule is used to perform non-dominated sorting of the multiple first liquidation paths according to the target vector of each of the first liquidation paths to obtain the path priority sequence.

[0120] In one optional scheme, the execution module includes a second selection module, an exchange processing module, and a second execution module. The second selection module is used to perform a selection step, which is to arbitrarily select two of the first liquidation paths from the extended path set as the first crossover mutation path and the second crossover mutation path based on the non-dominated sorting genetic algorithm. The exchange processing module is used to perform an exchange processing step, which is to exchange the first fragment path in the first crossover mutation path with the second fragment path in the second crossover mutation path to obtain two of the second liquidation paths, wherein the path endpoints of the first fragment path and the second fragment path are the same, and the path endpoints are the business institutions. The second execution module is used to perform the selection step and the exchange processing step at least once to obtain multiple of the second liquidation paths.

[0121] In one optional embodiment, the apparatus further includes a sending unit and an updating unit; the sending unit is used to send the cross-border clearing path table to the user after generating a cross-border clearing path set based on the first clearing path and the second clearing path, so as to obtain the user's path selection result for the cross-border clearing path set, and construct a preferred path set based on the historical cross-border clearing paths to obtain the target clearing path; the updating unit is used to update the target clearing path to the preferred path set based on the ε-dominance mechanism.

[0122] An optional scheme, the updating unit includes a second determining module, a calculation module, a third determining module, and a removal module; the second determining module is used to determine whether the number of paths in the preferred path set is greater than or equal to a path number threshold; the calculation module is used to calculate the difference value between the target liquidation path and each path in the preferred path set based on the ε-dominance mechanism when the number of paths in the preferred path set is greater than or equal to the path number threshold; the third determining module is used to determine the path in the preferred path set with the smallest difference value from the target liquidation path as a similar path; the removal module is used to determine the target vectors of the target liquidation path and the similar path respectively, and when the target vector of the target liquidation path is greater than the target vector of the similar path, the target liquidation path is added to the preferred path set, and the similar path is removed from the preferred path set.

[0123] In an optional embodiment, the apparatus further includes a matching processing unit, which, after updating the target clearing path to the preferred path set based on the ε-dominance mechanism, obtains the user's second cross-border clearing request, and performs matching processing on the second cross-border clearing request according to the preferred path set to generate a cross-border clearing path corresponding to the second cross-border clearing request.

[0124] The aforementioned cross-border clearing path generation device includes a processor and a memory. The first acquisition unit, the second acquisition unit, the expansion unit, etc., are all stored as program units in the memory, and the processor executes the aforementioned program units stored in the memory to achieve the corresponding functions. All of the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0125] The processor contains a kernel, which retrieves the corresponding program unit from memory. One or more kernels can be configured, and adjusting kernel parameters can address the shortcomings of existing cross-border clearing path selection in terms of multi-target adaptability, real-time performance, and comprehensive considerations, leading to low accuracy in clearing path selection.

[0126] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0127] This invention provides a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device containing the computer-readable storage medium to perform the cross-border clearing path generation method.

[0128] Specifically, the methods for generating cross-border clearing paths include:

[0129] Step S201: Obtain the cross-border clearing network and construct a weighted directed graph based on the cross-border clearing network. The weighted directed graph includes nodes, directed edges, and edge weights. The nodes represent cross-border clearing business entities, the directed edges represent clearing relationships between the business entities, and the edge weights are the weights of the directed edges.

[0130] Step S202: Obtain historical cross-border clearing paths and generate paths based on the aforementioned historical cross-border clearing paths and the aforementioned weighted directed graph path generation model.

[0131] Step S203: After obtaining the first cross-border clearing request, the first cross-border clearing request is input into the path generation model to process the first cross-border clearing request using the path generation model to obtain multiple first clearing paths corresponding to the first cross-border clearing request.

[0132] Step S204: Expand each of the first liquidation paths according to the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, and generate a cross-border liquidation path set based on the first liquidation paths and the second liquidation paths.

[0133] This invention provides a processor for running a program, wherein the program executes the method for generating the cross-border clearing path.

[0134] Specifically, the methods for generating cross-border clearing paths include:

[0135] Step S201: Obtain the cross-border clearing network and construct a weighted directed graph based on the cross-border clearing network. The weighted directed graph includes nodes, directed edges, and edge weights. The nodes represent cross-border clearing business entities, the directed edges represent clearing relationships between the business entities, and the edge weights are the weights of the directed edges.

[0136] Step S202: Obtain historical cross-border clearing paths and generate paths based on the aforementioned historical cross-border clearing paths and the aforementioned weighted directed graph path generation model.

[0137] Step S203: After obtaining the first cross-border clearing request, the first cross-border clearing request is input into the path generation model to process the first cross-border clearing request using the path generation model to obtain multiple first clearing paths corresponding to the first cross-border clearing request.

[0138] Step S204: Expand each of the first liquidation paths according to the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, and generate a cross-border liquidation path set based on the first liquidation paths and the second liquidation paths.

[0139] This invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs at least the following steps:

[0140] Step S201: Obtain the cross-border clearing network and construct a weighted directed graph based on the cross-border clearing network. The weighted directed graph includes nodes, directed edges, and edge weights. The nodes represent cross-border clearing business entities, the directed edges represent clearing relationships between the business entities, and the edge weights are the weights of the directed edges.

[0141] Step S202: Obtain historical cross-border clearing paths and generate paths based on the aforementioned historical cross-border clearing paths and the aforementioned weighted directed graph path generation model.

[0142] Step S203: After obtaining the first cross-border clearing request, the first cross-border clearing request is input into the path generation model to process the first cross-border clearing request using the path generation model to obtain multiple first clearing paths corresponding to the first cross-border clearing request.

[0143] Step S204: Expand each of the first liquidation paths according to the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, and generate a cross-border liquidation path set based on the first liquidation paths and the second liquidation paths.

[0144] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0145] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program having at least the following method steps:

[0146] Step S201: Obtain the cross-border clearing network and construct a weighted directed graph based on the cross-border clearing network. The weighted directed graph includes nodes, directed edges, and edge weights. The nodes represent cross-border clearing business entities, the directed edges represent clearing relationships between the business entities, and the edge weights are the weights of the directed edges.

[0147] Step S202: Obtain historical cross-border clearing paths and generate paths based on the aforementioned historical cross-border clearing paths and the aforementioned weighted directed graph path generation model.

[0148] Step S203: After obtaining the first cross-border clearing request, the first cross-border clearing request is input into the path generation model to process the first cross-border clearing request using the path generation model to obtain multiple first clearing paths corresponding to the first cross-border clearing request.

[0149] Step S204: Expand each of the first liquidation paths according to the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, and generate a cross-border liquidation path set based on the first liquidation paths and the second liquidation paths.

[0150] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0151] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application 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.

[0152] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. 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 processor, 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... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0153] 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.

[0154] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment 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.

[0155] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0156] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0157] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0158] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0159] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0160] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for generating a cross-border clearing path, characterized in that, include: Obtain the cross-border clearing network and construct a weighted directed graph based on the cross-border clearing network. The weighted directed graph includes nodes, directed edges, and edge weights. The nodes represent cross-border clearing business entities, the directed edges represent clearing relationships between the business entities, and the edge weights are the weights of the directed edges. Obtain historical cross-border clearing paths, and generate a path generation model based on the historical cross-border clearing paths and the weighted directed graph; After obtaining the first cross-border clearing request, the first cross-border clearing request is input into the path generation model, so as to use the path generation model to process the first cross-border clearing request to obtain multiple first clearing paths corresponding to the first cross-border clearing request; Each of the first liquidation paths is expanded using a non-dominated sorting genetic algorithm to obtain multiple second liquidation paths. A cross-border liquidation path set is then generated based on the first liquidation paths and the second liquidation paths.

2. The method according to claim 1, characterized in that, By expanding each of the first liquidation paths using a non-dominated sorting genetic algorithm, multiple second liquidation paths are obtained, including: Based on the non-dominated sorting genetic algorithm, multiple first liquidation paths are non-dominatedly sorted to obtain a path priority sequence, and the congestion distance of each first liquidation path is determined. Based on the congestion distance and path priority sequence of each first clearing path, a preset number of first clearing paths are selected to construct an extended path set; Genetic operations are performed on the extended path set based on the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths.

3. The method according to claim 2, characterized in that, A non-dominated sorting of multiple first liquidation paths yields a path priority sequence, including: According to the formula: Determine the target vector for each of the first liquidation paths, where, Let be the target vector. This represents the historical selection rate of the first liquidation path. The transaction success rate of the first clearing path. The compliance score for the first liquidation path is given. This serves as an indicator of exchange rate volatility for the first clearing path. This is the timeliness indicator for the first liquidation path. This represents the total cost indicator for the first liquidation path. The path priority sequence is obtained by performing a non-dominated sorting of the multiple first liquidation paths based on the target vector of each first liquidation path.

4. The method according to claim 2, characterized in that, Genetic operations are performed on the expanded path set based on the non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, including: The selection step is used to arbitrarily select two of the first liquidation paths from the extended path set as the first crossover mutation path and the second crossover mutation path based on the non-dominated sorting genetic algorithm. The interchange processing step is used to interchange the first fragment path in the first cross-mutation path with the second fragment path in the second cross-mutation path to obtain two second clearing paths, wherein the first fragment path and the second fragment path have the same path endpoint, and the path endpoint is the business unit. By performing the selection step and the swapping process at least once, multiple second liquidation paths are obtained.

5. The method according to claim 1, characterized in that, After generating a cross-border clearing path set based on the first clearing path and the second clearing path, the method further includes: The cross-border clearing path table is sent to the user to obtain the user's path selection result for the cross-border clearing path set, and a preferred path set is constructed based on the historical cross-border clearing paths to obtain the target clearing path; Based on the ε-dominance mechanism, the target liquidation path is updated to the preferred path set.

6. The method according to claim 5, characterized in that, Based on the ε-dominance mechanism, updating the target liquidation path to the preferred path set includes: Determine whether the number of paths in the preferred path set is greater than or equal to a path number threshold; If the number of paths in the preferred path set is greater than or equal to the path number threshold, the difference between the target liquidation path and each path in the preferred path set is calculated based on the ε-dominance mechanism. The path with the smallest difference value between the preferred path set and the target liquidation path is identified as the similar path; The target vectors of the target liquidation path and the similar path are determined respectively. If the target vector of the target liquidation path is greater than the target vector of the similar path, the target liquidation path is added to the preferred path set, and the similar path is removed from the preferred path set.

7. The method according to claim 5, characterized in that, After updating the target liquidation path to the preferred path set based on the ε-dominance mechanism, the method further includes: The user's second cross-border clearing request is obtained, and the second cross-border clearing request is matched and processed according to the preferred path set to generate a cross-border clearing path corresponding to the second cross-border clearing request.

8. A device for generating cross-border clearing routes, characterized in that, include: The first acquisition unit is used to acquire the cross-border clearing network and construct a weighted directed graph based on the cross-border clearing network. The weighted directed graph includes nodes, directed edges, and edge weights. The nodes represent cross-border clearing business institutions, the directed edges represent clearing relationships between the business institutions, and the edge weights are the weights of the directed edges. The second acquisition unit is used to acquire historical cross-border clearing paths and generate a model based on the historical cross-border clearing paths and the weighted directed graph path generation model. The third acquisition unit is used to input the first cross-border clearing request into the path generation model after acquiring the first cross-border clearing request, so as to use the path generation model to process the first cross-border clearing request to obtain multiple first clearing paths corresponding to the first cross-border clearing request. An extension unit is used to extend each of the first liquidation paths according to a non-dominated sorting genetic algorithm to obtain multiple second liquidation paths, and to generate a cross-border liquidation path set based on the first liquidation paths and the second liquidation paths.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the computer-readable storage medium to perform the method for generating a cross-border clearing path as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including a method for performing the generation of a cross-border clearing path as described in any one of claims 1 to 7.

Images