Aggregate payment system based on intelligent routing

Abstract

The invention relates to the technical field of intelligent payment, and discloses an aggregation payment system based on intelligent routing, and the system comprises an acquisition module which is configured to acquire multi-dimensional data; the payment channel management module is configured to access a payment channel and monitor the payment channel in real time; the intelligent routing decision module is configured to collect payment data, obtain a comprehensive score of a payment channel based on the payment data, generate a candidate channel based on the comprehensive score, determine transaction cost based on a three-dimensional rate state model, and trigger a fusing mechanism based on the transaction cost; the equipment compatibility module is configured to construct an equipment fingerprint, collect an equipment operation index, calculate a health degree score and trigger a grading response mechanism based on the health degree score; and the profit sharing optimization module is configured to perform automatic account sharing based on the intelligent contract and verify the routing strategy based on gray level distribution. The dynamic adaptability is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent payment technology, and in particular to an aggregated payment system based on intelligent routing. Background Art

[0002] Aggregated payment (also known as "fourth-party payment") involves integrating multiple third-party payment channels (such as Alipay, WeChat Pay, UnionPay QuickPass, and bank card payments) through technical means to provide merchants with a unified payment interface and management platform. Its core goal is to simplify the merchant payment integration process, reduce operating costs, and enhance the convenience of the consumer payment experience.

[0003] However, existing aggregated payment systems generally use static rules or single indicators to select payment channels and rely on manual intervention to adjust strategies. These systems also suffer from the following problems: 1. Poor dynamic adaptability and inability to perceive the health status of channels in real time, resulting in increased failure rates in peak transaction scenarios; 2. Insufficient multi-objective coordination: The weights of indicators such as success rate, cost, and user experience are rigid, making it difficult to balance the optimization needs of complex transaction scenarios; 3. Device compatibility defects: Traditional device identification relies on a single parameter, resulting in a high mismatch rate and leading to payment terminal driver conflicts; 4. Existing technologies lack a dynamic weight adjustment mechanism and the ability to integrate heterogeneous data, and fail to incorporate real-time transaction characteristics, device status, and user behavior into a unified decision-making framework.

[0004] Therefore, it is necessary to design an aggregate payment system based on intelligent routing to solve the problems existing in current technology. Summary of the Invention

[0005] In view of this, the present invention proposes an aggregate payment system based on intelligent routing, aiming to solve the problem of low efficiency of current aggregate payment.

[0006] The present invention proposes an aggregate payment system based on intelligent routing, comprising:

[0007] A collection module configured to collect multi-dimensional data of the aggregated payment system;

[0008] a payment channel management module configured to access the payment channel based on a plug-in architecture, and further configured to monitor the payment channel in real time and trigger a weight adjustment mechanism when the health value of the payment channel falls below a threshold;

[0009] an intelligent routing decision module configured to collect payment data, obtain a comprehensive score of the payment channel based on the payment data, and generate candidate channels based on the comprehensive score; the intelligent routing decision module is further configured to determine a transaction cost based on a three-dimensional rate state model and trigger a circuit breaker mechanism based on the transaction cost;

[0010] a device compatibility module configured to obtain the multi-dimensional data and construct a device fingerprint, the device compatibility module further configured to collect device operating indicators and calculate a health score, and trigger a graded response mechanism based on the health score;

[0011] The profit-sharing optimization module is configured to automatically split accounts based on smart contracts. The profit-sharing optimization module is also configured to verify the routing strategy based on grayscale distribution.

[0012] Furthermore, when accessing payment channels based on a plug-in architecture, it includes:

[0013] Build an adapter for each newly added payment channel based on the system interface and define its interface specification;

[0014] Implement hot deployment of the adapter based on Java dynamic loading mechanism and register the adapter through the service registration center;

[0015] Building a configuration library based on a distributed database and storing static parameters of each of the payment channels;

[0016] The static parameters include rate tiers, transaction limits, and priority weights.

[0017] Furthermore, the payment channel is monitored in real time. When the health value of the payment channel is lower than a threshold, the circuit breaker mechanism is triggered, including:

[0018] Build a computing pipeline based on the Flink streaming computing engine and obtain transaction data from the payment channel;

[0019] Obtaining a health indicator of the payment channel based on a sliding time window;

[0020] The health indicators include transaction success rate, average latency and number of transactions;

[0021] Performing weighted calculation based on the transaction success rate, the average delay, and the number of transactions to obtain health value data;

[0022] When the health value data is lower than the threshold, the weight degradation is triggered and the weight parameters are updated through the configuration library;

[0023] When the health value data is lower than the threshold for three consecutive times, the circuit breaker mechanism is triggered and the traffic allocation of the payment channel is suspended.

[0024] Furthermore, obtaining a comprehensive score of the payment channel based on the multi-dimensional data and generating candidate channels based on the comprehensive score include:

[0025] Normalize the multi-dimensional data and calculate the degree of discreteness of the entropy value of each data;

[0026] Assigning an indicator weight to each data based on the degree of discreteness of the entropy value, and constructing a decision matrix for the data after assigning the indicator weight;

[0027] The ideal solution is formed based on the optimal value of each data, and the negative ideal solution is formed based on the worst value of each data;

[0028] Calculating the Euclidean distance between the payment channel and the ideal solution and the negative ideal solution based on a chaotic game optimization algorithm, and obtaining the comprehensive score;

[0029] The comprehensive scores are arranged in descending order, and the top three payment channels are selected as candidates.

[0030] Furthermore, when inputting the payment channel into the three-dimensional rate state model to determine the transaction cost, it includes:

[0031] Combine transaction amount tiers, time period coefficients, and channel types into a three-dimensional matrix, and store the corresponding rate rules in each cell;

[0032] The three-dimensional matrix is traversed by the simplex method to screen out the payment channel with the lowest transaction cost.

[0033] Furthermore, when the circuit breaker mechanism is triggered based on the transaction cost, it includes:

[0034] Determine the transaction cost threshold based on historical data, and trigger a circuit breaker mechanism when the transaction cost exceeds the transaction cost threshold;

[0035] The circuit breaker mechanism includes a first-level circuit breaker and a second-level circuit breaker. When the transaction cost exceeds the threshold for three consecutive times, the first-level circuit breaker is triggered and the candidate payment channel is switched. When the transaction cost fluctuation is greater than twice the threshold, the second-level circuit breaker is triggered and all transactions are frozen.

[0036] Furthermore, when constructing a device fingerprint based on the multi-dimensional data, it includes:

[0037] Constructing a fingerprint feature vector based on the multi-dimensional data, and calculating the fingerprint feature vector based on a hash algorithm to obtain the device fingerprint code;

[0038] Upload the device fingerprint code to the cloud engine and bind it;

[0039] A candidate driver set in the cloud engine is screened based on the device fingerprint code, and an optimal driver is screened based on the firmware version.

[0040] Furthermore, the device operation indicators are collected and a health score is calculated. When a hierarchical response mechanism is triggered based on the health score, the following steps are included:

[0041] Based on the device fingerprint code, the device operation index is obtained, and a comprehensive health score is calculated based on a weighted formula, wherein the weight ratios in the weighted formula are 30% for online rate, 25% for transaction success rate, 20% for delay coefficient, 15% for security monitoring index, and 10% for resource utilization rate;

[0042] When the comprehensive health score is lower than 80 points, a yellow warning is triggered and performance log tracking is started;

[0043] When the comprehensive health score is lower than 60 points, an orange warning is triggered and traffic degradation and driver version rollback are initiated;

[0044] When the comprehensive health score is lower than 40 points, a red alert is triggered, the device connection is forcibly disconnected, and the driver matching is frozen.

[0045] Furthermore, when automatic account distribution is based on smart contracts, it includes:

[0046] Define a profit-sharing rule model and build a smart contract based on the profit-sharing rule model;

[0047] When the transaction is completed, the contract call mechanism is triggered based on the smart contract. The contract call mechanism includes calling the chain code and passing in the transaction details. The smart contract determines the amount to be distributed based on the profit-sharing rule model, distributes the amount to be distributed to the account, writes the account splitting execution result into the blockchain ledger, and generates a transaction snapshot.

[0048] Furthermore, when verifying the routing strategy based on grayscale distribution, it includes:

[0049] Conduct a phased release of the profit-sharing rule model and compare profit-sharing data under different profit-sharing rules;

[0050] The contract incentive parameters are adjusted based on the profit sharing data. When the profit sharing execution fails or the smart contract is abnormal, the account is rolled back and the original amount is stored in the pending queue.

[0051] Compared with existing technologies, the present invention offers the following advantages: The payment channel management module, through its plug-in architecture, enables access to new payment methods without being constrained by the underlying framework. This loosely coupled structure ensures both scalability across different payment tools and the independence of payment channels, thus avoiding the chain reaction caused by a single channel failure. A real-time health monitoring mechanism, combined with a dynamic weight adjustment strategy, balances the operational load of payment channels, maintaining service continuity while mitigating the risk of potential service interruptions. The device compatibility module, through the construction of a multi-dimensional device fingerprint library, accurately identifies the differentiated characteristics of various terminal devices, reducing transaction anomalies caused by device compatibility issues. The intelligent routing decision module, through the introduction of a multi-dimensional evaluation system, upgrades traditional single-dimensional rate-based decision-making to a three-dimensional assessment model that integrates service quality, response time, and cost structure. The three-dimensional rate status model enables cost calculation to transcend the limitations of static pricing and dynamically reflect market changes and channel operating conditions. Combined with a circuit breaker mechanism, this allows for rapid switching to alternative channels in the event of sudden cost fluctuations or a surge in abnormal transactions, ensuring both transaction success rates and avoiding unexpected cost losses. The profit-sharing optimization module automates fund allocation through smart contracts, reducing the lag of traditional manual reconciliation and ensuring the immediacy and accuracy of multi-party interest distribution. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0053] Figure 1 This is a structural block diagram of the intelligent routing-based aggregate payment system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0054] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, unless there is a conflict, the embodiments of the present disclosure and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0055] In some embodiments of the present application, see Figure 1 As shown, an aggregate payment system based on intelligent routing includes:

[0056] The collection module is configured to collect multi-dimensional data of the aggregated payment system.

[0057] The payment channel management module is configured to access the payment channel based on a plug-in architecture. The payment channel management module is also configured to monitor the payment channel in real time. When the health value of the payment channel is lower than a threshold, a weight adjustment mechanism is triggered.

[0058] The intelligent routing decision module is configured to collect payment data, obtain a comprehensive score of the payment channel based on the payment data, and generate candidate channels based on the comprehensive score. The intelligent routing decision module is also configured to determine the transaction cost based on the three-dimensional rate status model and trigger the circuit breaker mechanism based on the transaction cost.

[0059] The device compatibility module is configured to obtain multi-dimensional data and build device fingerprints. The device compatibility module is also configured to collect device operating indicators and calculate health scores, and trigger a graded response mechanism based on the health scores.

[0060] The profit sharing optimization module is configured to automatically split accounts based on smart contracts. The profit sharing optimization module is also configured to verify the routing strategy based on grayscale distribution.

[0061] Specifically, the multi-dimensional data collected by the collection module includes, but is not limited to: MAC address, firmware version, hardware configuration parameters, operating system characteristics, sensor data and network environment, TPS (transactions per second), success rate (successful transaction ratio), average response time, network latency, current fund balance, available balance ratio, abnormal transaction rate, error code frequency, volatility index, channel congestion rate, risk control interception ratio, and availability status code. First, the payment channel management module serves as a frontier, responsible for monitoring payment channels. All payment requests are first allocated routing resources through this module. By adopting a plug-in architecture, adding new payment channels only requires developing standardized adapters, and hot deployment technology enables zero-downtime updates. The configuration library stores parameters such as rate tiers and transaction limits for each channel, supports setting priority weights by merchant type, and then, based on the Flink streaming computing engine, updates 12 indicators such as channel TPS, success rate, and fund balance every 5 seconds, and calculates the health index using a sliding time window algorithm. When the health of a channel falls below a threshold, routing weight degradation is automatically triggered. The intelligent routing decision module then makes decisions based on channel monitoring data, acting as the intelligent hub of the overall process. This module uses the TOPSIS algorithm, entropy weighting, and chaos game optimization to select multi-objective channels. It optimizes transaction costs by constructing a three-dimensional rate matrix. The device compatibility module constructs a unique device fingerprint using six of these multi-dimensional data dimensions: MAC address, firmware version, hardware configuration parameters, operating system characteristics, sensor data, and network environment. It then uses a SHA-256 hash to generate an identifier and match the driver version. It also performs health monitoring every three seconds and initiates a driver rollback mechanism when issues arise. The device compatibility module interacts with the front-end to ensure device driver compatibility and provide device-specific references for routing strategies. The profit-sharing optimization module dynamically adjusts the profit-sharing ratio based on transaction type, ensuring traceability through the use of smart contracts and blockchain. It also integrates grayscale testing to evaluate the effectiveness of strategy optimization. The profit-sharing optimization module is the final settlement component of the transaction process, ensuring transparent and flexible fund allocation.

[0062] The plug-in architecture of the payment channel breaks through the constraints of traditional integration models, enabling modular scalability for the integration of new payment tools. By combining standardized adapters with hot deployment technology, channel updates can be performed without disrupting existing services. The establishment of a dynamic parameter configuration library enables merchants to implement differentiated fee policies and transaction limit management in real time, meeting the personalized needs of different business scenarios. The streaming data processing engine collects payment channel operational metrics and, combined with a sliding time window algorithm, captures instantaneous fluctuations in channel status. When health anomalies are detected, a weight adjustment mechanism ensures that payment requests are dynamically diverted to backup channels, avoiding chain reactions caused by service interruptions. Intelligent routing decisions integrate multi-objective optimization algorithms to construct a three-dimensional evaluation model that balances cost, efficiency, and stability. The application of entropy weighting ensures objective weighting of each evaluation metric, avoiding decision-making bias caused by subjective factors. Chaos game optimization enhances strategy optimization capabilities in complex scenarios, ensuring that channel selection not only meets current transaction characteristics but also remains adaptable to market fluctuations. The introduction of a three-dimensional rate model upgrades traditional static rate calculation to a dynamic pricing system that incorporates time, transaction size, and channel status, making cost estimates more closely aligned with real-time market conditions. A circuit breaker mechanism automatically switches to a backup plan upon identifying abnormal cost fluctuations, ensuring transaction success rates while preventing the continued escalation of unexpected cost losses. Device fingerprinting technology builds a precise device identification system by cross-validating multi-dimensional hardware characteristics and software environment data. A hashing algorithm ensures data security while enabling intelligent matching and version control of device drivers. Real-time health monitoring collects device operating indicators and triggers graded responses based on pre-set threshold gradients, maximizing service continuity while avoiding resource waste caused by excessive responses. Device-dimensional data provides decision-making basis for routing policies at the terminal adaptation level, reducing transaction failure rates due to device compatibility issues. The integration of smart contracts and blockchain transforms profit-sharing rules into automatically executed digital agreements. A dynamic transaction type matching mechanism ensures that profit-sharing ratios adapt to changing business scenarios in real time, meeting the needs of complex business collaboration models while avoiding errors caused by manual intervention. The introduction of a grayscale verification mechanism creates a risk buffer during routing policy updates: through small-scale pilots, real business data is collected to verify the effectiveness of policy optimization before gradually expanding its scope. This incremental deployment approach reduces systemic risks potentially associated with policy changes while providing a closed-loop data feedback loop for continuous optimization. The traceability of capital flows strengthens trust among multiple participants and provides an immutable audit basis for dispute resolution.

[0063] In some embodiments of the present application, accessing a payment channel based on a plug-in architecture includes:

[0064] Build an adapter based on the system interface for each new payment channel and define its interface specification.

[0065] The hot deployment of the adapter is realized based on the Java dynamic loading mechanism, and the adapter is registered through the service registration center.

[0066] Build a configuration library based on a distributed database and store the static parameters of each payment channel.

[0067] Static parameters include fee tiers, transaction limits, and priority weights.

[0068] Understandably, for each new payment channel, an adapter that complies with the system interface specifications (such as implementing a unified RESTful API or SDK) must first be developed. Standard transaction request, response, and exception handling interfaces must also be defined. The adapter must encapsulate channel-specific logic such as protocol conversion, signature verification, and parameter mapping. Furthermore, a Java dynamic class loading mechanism (such as the OSGi framework or Spring Boot DevTools) is used to implement hot deployment of the adapter, allowing new plug-ins to be loaded without restarting the service. Once deployed, the adapter is registered with a service registry (such as ZooKeeper or Nacos). The system automatically identifies and includes it in the list of available channels. The configuration library uses a distributed database (such as MySQL or MongoDB) to store static parameters for each channel: fee tiers (setting fees based on transaction amount, e.g., 0.3% for 0-1,000 yuan and 0.25% for amounts above 1,000 yuan); transaction limits (including single transaction limits, daily cumulative limits, and merchant-specific limits (e.g., single transaction limits for small and micro merchants ≤ 50,000 yuan); and priority weights (dynamically assigned weights based on merchant type, such as cross-border merchants and VIP merchants).

[0069] In some embodiments of the present application, the payment channel is monitored in real time. When the health value of the payment channel is lower than a threshold, the circuit breaker mechanism is triggered, including:

[0070] Build a computing pipeline based on the Flink streaming computing engine and obtain transaction data from the payment channel.

[0071] Obtain payment channel health indicators based on a sliding time window.

[0072] Health metrics include transaction success rate, average latency, and number of transactions.

[0073] Health value data is obtained by weighted calculation based on transaction success rate, average latency and number of transactions.

[0074] When the health value data is lower than the threshold, the weight degradation is triggered and the weight parameters are updated through the configuration library.

[0075] When the health value data is below the threshold for three consecutive times, the circuit breaker mechanism is triggered and the traffic allocation of the payment channel is suspended.

[0076] It is understandable that the real-time transaction data of the channel is collected through the embedding point. The real-time transaction data includes TPS (transactions per second), success rate (successful transaction ratio), average response time, network latency, current fund balance, available balance ratio, abnormal transaction rate, error code frequency, volatility index, channel congestion rate, risk control interception ratio and availability status code, etc. By using Apache Flink (distributed data flow engine) to build a real-time computing pipeline: Data access layer: Receive the original transaction log through the Kafka message queue, and computing layer: define Flink jobs, trigger a window calculation every 5 seconds, aggregate the indicator data of each channel, and then ensure the timeliness of the data through the sliding window, and calculate the statistical values ​​of the data in the window: success rate = Number of successful transactions / total number of transactions, average delay = Σ response time / number of transactions, TPS = total number of transactions within the window period / window duration. The health index is calculated using a weighted formula: the health index is composed of 40% success rate + 30% average delay + 20% capital adequacy ratio + 10% TPS. The values of each indicator are normalized and then weighted and summed. When the health of a channel is lower than the threshold (0.7), the weight downgrade is triggered. The routing decision module takes effect by updating the weight parameters in real time. When the health value index is lower than the threshold for three consecutive times, the circuit breaker mechanism is triggered and the traffic allocation of the payment channel is suspended. At the same time, when the channel recovers, a gradual traffic increase strategy is adopted (increasing the traffic by 10% every 10 minutes) until the weight returns to normal.

[0077] In some embodiments of the present application, obtaining a comprehensive score of a payment channel based on multi-dimensional data and generating candidate channels based on the comprehensive score include:

[0078] Normalize the multi-dimensional data and calculate the degree of discreteness of the entropy value of each data.

[0079] Based on the degree of discreteness of entropy values, indicator weights are assigned to each data, and a decision matrix is constructed using the data after the indicator weights are assigned.

[0080] The optimal value of each data constitutes an ideal solution, and the worst value of each data constitutes a negative ideal solution.

[0081] Based on the chaotic game optimization algorithm, the Euclidean distance between the payment channel and the ideal solution and the negative ideal solution is calculated, and a comprehensive score is obtained.

[0082] Arrange the comprehensive scores in descending order and select the top three payment channels as candidates.

[0083] Specifically, the multi-dimensional data collected in this embodiment include transaction performance data: TPS (transactions per second), success rate, average response delay, and maximum response delay. Fund status data: real-time channel balance, fund adequacy ratio (balance / preset threshold). Network and security data: network jitter rate, channel API availability, signature verification failure rate. User behavior data: transaction amount distribution, time period distribution, payment method preference (from user portrait), by implanting points at the payment gateway layer, recording the original log of each transaction (including success / failure status, time consumption, error code, etc.), and then using a lightweight log collection tool (such as Filebeat or Fluentd) to push the log to the Kafka message queue in real time, and then using streaming data Apache Flink as the stream processing engine, subscribing to the original data stream through Kafka Consumer, partitioning by payment channel ID, and normalizing each indicator to eliminate dimensional differences, and calculating the entropy value of each indicator: Among them, p ij is the proportion of the i-th channel in the j-th indicator, m is the total number of payment channels, x′ ij is the normalized indicator set, j is the jth indicator, i is the i-th channel, and then the entropy value of the jth indicator is calculated: Among them, E j is the entropy value of the jth indicator, k = 1 / In(m) is the normalization constant, ensuring that the entropy value is between 0 and 1, and then assigning indicator weights to the indicators. The weight formula is: Among them, w j is the weight of the jth indicator, n is the total number of indicators, and then the weighted matrix is constructed: v ij =w j *x′ ij , where v ij is the weighted value of the i-th channel on the j-th indicator. Finally, the ideal solution is calculated for each indicator. For positive indicators (such as success rate), the maximum value of all channels is taken. For negative indicators (such as delay), the minimum value is taken. By quantifying the closeness of each channel to the ideal solution and the negative ideal solution, the distance to the ideal solution is calculated using the Euclidean distance: Distance to the negative ideal solution: in, is the distance from the ith channel to the ideal solution, is the distance from the ith channel to the negative ideal solution, is the reference value of the negative ideal solution for each indicator, The reference value of the ideal solution for each indicator is used to calculate the comprehensive score of the ideal solution and the negative ideal solution: Among them, S iis the comprehensive score of the i-th channel. The closer the value is to 1, the better it is. Sort by the comprehensive score from high to low, and select the top three as candidate channels. At the same time, introduce the chaos game optimization algorithm to improve the calculation efficiency and reduce the number of algorithm iterations: k+1 =μx k (1-x k ), where x k is the chaotic variable of the kth iteration, and μ is the control parameter (taken as 4).

[0084] In some embodiments of the present application, inputting the payment channel into the three-dimensional rate state model to determine the transaction cost includes:

[0085] The transaction amount tiers, time period coefficients, and channel types are combined into a three-dimensional matrix, and each cell stores the corresponding rate rule.

[0086] The simplex method is used to traverse the three-dimensional matrix and select the payment channel with the lowest transaction cost.

[0087] It is understandable that the transaction amount is first divided into grades, such as small amount: 0-1000 yuan, medium amount: 1001-5000, large amount 5001 and above, and the time period coefficient is divided, for example: the rate during peak hours (10:00-12:00) is increased by 5%, the rate during off-peak hours (14:00-17:00) remains at the base rate, and the rate during off-peak hours (1:00-5:00) is reduced by 3%. Then the channel type is further classified, such as bank direct connection, third-party payment, cross-border payment, and the transaction amount grades, time period coefficient and channel type are combined into a three-dimensional matrix (three-dimensional matrix). Rate matrix), convert the three-dimensional rate matrix into a linear programming problem, with the goal of minimizing transaction costs, that is, each transaction must and can only select one payment channel combination, and some channels may be unavailable due to limits or circuit breakers. The selected time period coefficient needs to be consistent with the actual transaction time, and the transaction amount must meet the tiering standards. The three-dimensional matrix is traversed by the simplex method to screen out the payment channels that meet the constraints and have the lowest cost. First, an initial solution that meets all constraints is selected (such as randomly selecting an available channel). The test number (Reduced Cost) is calculated to determine whether it can be optimized, and the basis variables are gradually replaced. When all test numbers are non-negative, the optimal solution is reached, and then the payment channel combination corresponding to the optimal solution is output, that is, the channel with the lowest cost.

[0088] In some embodiments of the present application, when a circuit breaker mechanism is triggered based on transaction costs, it includes:

[0089] The transaction cost threshold is determined based on historical data. When the transaction cost exceeds the transaction cost threshold, the circuit breaker mechanism is triggered.

[0090] The circuit breaker mechanism includes level 1 and level 2 circuit breakers. When the transaction cost exceeds the threshold for three consecutive times, the level 1 circuit breaker is triggered and the payment channel is switched to the candidate payment channel. When the transaction cost fluctuation is greater than twice the threshold, the level 2 circuit breaker is triggered and all transactions are frozen.

[0091] It is understandable that the transaction cost data (such as handling fees, exchange rate losses, etc.) of each payment channel in the historical data are collected. The data items include single cost, average daily cost and cost volatility. The threshold is determined by taking the moving average of the historical data, and the fluctuation is determined according to the time of the preset historical data. The threshold is updated periodically. When the real-time transaction cost of a payment channel exceeds the threshold for three consecutive times, the exceeding event is recorded and the first-level circuit breaker is triggered to reduce the traffic allocation ratio of the channel and switch to the candidate payment channel. If the transaction cost fluctuation of a payment channel exceeds twice the threshold, the second-level circuit breaker is triggered, and all new transaction allocations of the channel are stopped, and it switches to the backup channel, and manual inspection is triggered at the same time.

[0092] In some embodiments of the present application, when constructing a device fingerprint based on multi-dimensional data, the process includes:

[0093] Construct a fingerprint feature vector based on multi-dimensional data, and calculate the fingerprint feature vector based on the hash algorithm to obtain the device fingerprint code.

[0094] Upload the device fingerprint code to the cloud engine and bind it.

[0095] The candidate driver set in the cloud engine is filtered based on the device fingerprint code, and the optimal driver is selected based on the firmware version.

[0096] Specifically, the MAC address OUI code (accurate to the manufacturer model) and device serial number are obtained in real time through the embedded SDK, the User-Agent field in the HTTP request header is parsed to extract the operating system version and browser fingerprint features, and finally the 12 heterogeneous data are standardized (such as unifying the MAC address separator to a colon format), and a 128-bit device unique identifier is generated through the SHA-256 hash operation, which is then double-matched with the 6 types of device models in the cloud driver library - first quickly filter the candidate driver set according to the device type code, and then compare the firmware version compatibility matrix to select the optimal driver version.

[0097] In some embodiments of the present application, when collecting device operating indicators and calculating health scores, triggering a hierarchical response mechanism based on the health scores includes:

[0098] Based on the device fingerprint code, the device operation indicators are obtained, and the comprehensive health score is calculated based on the weighted formula. The weight ratios in the weighted formula are 30% for online rate, 25% for transaction success rate, 20% for delay coefficient, 15% for security monitoring indicators and 10% for resource utilization.

[0099] When the overall health score is lower than 80 points, a yellow warning is triggered and performance log tracking is started.

[0100] When the overall health score is lower than 60 points, an orange alert is triggered and traffic degradation and driver version rollback are initiated.

[0101] When the overall health score is lower than 40 points, a red alert is triggered, the device connection is forcibly disconnected, and driver matching is frozen.

[0102] It can be understood that the health score is calculated by a weighted formula, health score = online rate 30% + transaction success rate 25% + delay coefficient 20% + security monitoring index 15% + resource utilization 10% to obtain the final health score, and then trigger different levels of response strategies based on the health score. When the health score is lower than or equal to 80 points but greater than 60 points, a yellow warning is triggered and detailed logging is started, and the log is stored. When the health score is lower than or equal to 60 points but greater than 40 points, the transaction traffic allocation ratio of the device is reduced, and the remaining traffic is switched to the backup device or channel, and it is automatically rolled back to the last stable driver version. The health score is recalculated after the rollback. When the health score is lower than 40 points, all active sessions of the device are terminated, and new connection requests are rejected. The reason for the disconnection is recorded, and the device is marked as frozen in the cloud, prohibiting the allocation of new drivers.

[0103] In some embodiments of the present application, when automatically splitting accounts based on smart contracts, the following steps are included:

[0104] Define the profit-sharing rule model and build a smart contract based on the profit-sharing rule model.

[0105] When the transaction is completed, the contract call mechanism is triggered based on the smart contract. The contract call mechanism includes calling the chain code and passing in the transaction details. The smart contract determines the amount to be distributed based on the profit-sharing rule model, distributes the amount to be distributed to the account, writes the account execution results into the blockchain ledger, and generates a transaction snapshot.

[0106] Specifically, a profit-sharing rule model is established for different transaction types, merchant attributes and agency levels, including but not limited to the merchant level and agency relationship map, transaction type, transaction amount and channel type, dynamic weight factors (such as rate return ratio, incentive policy), etc., and a corresponding smart contract template is deployed for each type of profit-sharing rule based on the blockchain. Each contract has the function of accepting transaction events (including amount, channel, time, etc.), querying profit-sharing parameters and current strategies, executing fund allocation calculations and account details. After the transaction is completed, the account execution results are written into the blockchain ledger, which includes the transaction number and time, the account participants and the corresponding amount, the profit-sharing rules used, the execution smart contract version number, the status code and the exception mark (such as processing failure or pending review).

[0107] In some embodiments of the present application, when verifying a routing strategy based on grayscale distribution, the following steps are included:

[0108] The profit-sharing rule model is released in phased mode, and the profit-sharing data under different profit-sharing rules are compared.

[0109] Adjust the contract incentive parameters based on the profit sharing data. When the profit sharing execution fails or the smart contract is abnormal, the account will be rolled back and the original amount will be stored in the pending queue.

[0110] It is understandable that in order to verify the impact of the profit-sharing strategy on trading behavior, grayscale releases are carried out among multiple account-sharing strategies, and user activity, changes in agent income, and channel costs under different strategies are dynamically compared to evaluate the effect and adjust the parameters of the contract incentives. When the contract execution fails or the record on the blockchain is abnormal, the account-sharing status is automatically rolled back, and the original amount is placed in the pending queue, triggering an early warning at the same time. It will not be reissued until the exception is handled.

[0111] In summary, the beneficial effects of the present invention are as follows: the payment channel management module, through its plug-in architecture, enables access to new payment methods without being constrained by the underlying framework. This loosely coupled structure ensures both the scalability of different payment tools and the independence of each payment channel, thus avoiding the chain reaction caused by a single channel failure. A real-time health monitoring mechanism, combined with a dynamic weight adjustment strategy, balances the operating load of payment channels, maintaining service continuity while reducing the risk of potential service interruptions. The device compatibility module, through the construction of a multi-dimensional device fingerprint library, accurately identifies the differentiated characteristics of various terminal devices, reducing transaction anomalies caused by device compatibility issues. The intelligent routing decision module, through the introduction of a multi-dimensional evaluation system, upgrades traditional single-dimensional rate-based decision-making to a three-dimensional evaluation model that integrates service quality, response time, and cost structure. The three-dimensional rate status model enables cost calculation to break through the limitations of static pricing and dynamically reflect changes in the market environment and channel operating conditions. Furthermore, combined with a circuit breaker mechanism, it can quickly switch to a backup channel in the event of sudden cost fluctuations or a surge in abnormal transactions, thus ensuring transaction success rates while avoiding unexpected cost losses. The profit-sharing optimization module automates fund allocation through smart contracts, reducing the lag of traditional manual reconciliation and ensuring the immediacy and accuracy of multi-party interest distribution.

[0112] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0113] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program instructions may also be stored in a computer-readable storage device that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable storage device produce an article of manufacture comprising an instruction device that implements the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. An aggregate payment system based on intelligent routing, characterized in that: include: A collection module configured to collect multi-dimensional data of the aggregated payment system; a payment channel management module configured to access the payment channel based on a plug-in architecture, and further configured to monitor the payment channel in real time and trigger a weight adjustment mechanism when the health value of the payment channel falls below a threshold; an intelligent routing decision module configured to collect payment data, obtain a comprehensive score of the payment channel based on the payment data, and generate candidate channels based on the comprehensive score; the intelligent routing decision module is further configured to determine a transaction cost based on a three-dimensional rate state model and trigger a circuit breaker mechanism based on the transaction cost; a device compatibility module configured to obtain the multi-dimensional data and construct a device fingerprint, the device compatibility module further configured to collect device operating indicators and calculate a health score, and trigger a graded response mechanism based on the health score; The profit-sharing optimization module is configured to automatically split accounts based on smart contracts. The profit-sharing optimization module is also configured to verify the routing strategy based on grayscale distribution.

2. The intelligent routing-based aggregate payment system according to claim 1, characterized in that: When accessing payment channels based on a plug-in architecture, it includes: Build an adapter for each newly added payment channel based on the system interface and define its interface specification; Implement hot deployment of the adapter based on Java dynamic loading mechanism and register the adapter through the service registration center; Building a configuration library based on a distributed database and storing static parameters of each of the payment channels; The static parameters include rate tiers, transaction limits, and priority weights.

3. The intelligent routing-based aggregate payment system according to claim 2, characterized in that: Real-time monitoring of the payment channel. When the health value of the payment channel falls below a threshold, the circuit breaker mechanism is triggered, including: Build a computing pipeline based on the Flink streaming computing engine and obtain transaction data from the payment channel; Obtaining a health indicator of the payment channel based on a sliding time window; The health indicators include transaction success rate, average latency and number of transactions; Performing weighted calculation based on the transaction success rate, the average delay, and the number of transactions to obtain health value data; When the health value data is lower than the threshold, the weight degradation is triggered and the weight parameters are updated through the configuration library; When the health value data is lower than the threshold for three consecutive times, the circuit breaker mechanism is triggered and the traffic allocation of the payment channel is suspended.

4. The intelligent routing-based aggregate payment system according to claim 3, characterized in that: Obtaining a comprehensive score of the payment channel based on the multi-dimensional data, and generating candidate channels based on the comprehensive score, including: Normalize the multi-dimensional data and calculate the degree of discreteness of the entropy value of each data; Assigning an indicator weight to each data based on the degree of discreteness of the entropy value, and constructing a decision matrix for the data after assigning the indicator weight; The ideal solution is formed based on the optimal value of each data, and the negative ideal solution is formed based on the worst value of each data; Calculating the Euclidean distance between the payment channel and the ideal solution and the negative ideal solution based on a chaotic game optimization algorithm, and obtaining the comprehensive score; The comprehensive scores are arranged in descending order, and the top three payment channels are selected as candidates.

5. The intelligent routing-based aggregate payment system according to claim 4, characterized in that: Inputting the payment channel into the three-dimensional rate state model to determine the transaction cost includes: Combine transaction amount tiers, time period coefficients, and channel types into a three-dimensional matrix, and store the corresponding rate rules in each cell; The three-dimensional matrix is traversed by the simplex method to screen out the payment channel with the lowest transaction cost.

6. The intelligent routing-based aggregate payment system according to claim 5, characterized in that: When the circuit breaker mechanism is triggered based on the transaction cost, it includes: Determine the transaction cost threshold based on historical data, and trigger a circuit breaker mechanism when the transaction cost exceeds the transaction cost threshold; The circuit breaker mechanism includes a first-level circuit breaker and a second-level circuit breaker. When the transaction cost exceeds the transaction cost threshold for three consecutive times, the first-level circuit breaker is triggered and the candidate payment channel is switched. When the transaction cost fluctuation is greater than twice the transaction cost threshold, the second-level circuit breaker is triggered and all transactions are frozen.

7. The intelligent routing-based aggregate payment system according to claim 6, characterized in that: When constructing a device fingerprint based on the multi-dimensional data, it includes: Constructing a fingerprint feature vector based on the multi-dimensional data, and calculating the fingerprint feature vector based on a hash algorithm to obtain the device fingerprint code; Upload the device fingerprint code to the cloud engine and bind it; A candidate driver set in the cloud engine is screened based on the device fingerprint code, and an optimal driver is screened based on the firmware version.

8. The intelligent routing-based aggregate payment system according to claim 7, characterized in that: Collect device operating indicators and calculate health scores. When triggering a graded response mechanism based on the health scores, it includes: Based on the device fingerprint code, the device operation index is obtained, and a comprehensive health score is calculated based on a weighted formula, wherein the weight ratios in the weighted formula are 30% for online rate, 25% for transaction success rate, 20% for delay coefficient, 15% for security monitoring index, and 10% for resource utilization rate; When the comprehensive health score is lower than 80 points, a yellow warning is triggered and performance log tracking is started; When the comprehensive health score is lower than 60 points, an orange warning is triggered and traffic degradation and driver version rollback are initiated; When the comprehensive health score is lower than 40 points, a red alert is triggered, the device connection is forcibly disconnected, and the driver matching is frozen.

9. The intelligent routing-based aggregate payment system according to claim 8, characterized in that: Automatic account splitting based on smart contracts includes: Define a profit-sharing rule model and build a smart contract based on the profit-sharing rule model; When the transaction is completed, the contract call mechanism is triggered based on the smart contract. The contract call mechanism includes calling the chain code and passing in the transaction details, the smart contract determining the amount to be distributed based on the profit-sharing rule model, distributing the amount to be distributed to the account, and writing the account splitting execution result into the blockchain ledger and generating a transaction snapshot.

10. The intelligent routing-based aggregate payment system according to claim 9, characterized in that: When verifying the routing policy based on grayscale distribution, it includes: Conduct a phased release of the profit-sharing rule model and compare profit-sharing data under different profit-sharing rules; The contract incentive parameters are adjusted based on the profit sharing data. When the profit sharing execution fails or the smart contract is abnormal, the account is rolled back and the original amount is stored in the pending queue.

Images