Multi-dimensional collection routing dynamic channel scheduling method
Through the dynamic channel scheduling method of multi-dimensional payment routing, routing and monitoring is combined with the payment environment status and risk factors, the problem of single payment channel switching strategy in the existing technology is solved, and efficient and reliable automatic payment channel switching is achieved, improving the improvement of User experience and operational efficiency.
Patent Information
- Application Number
- CN202411475235.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-22
- Publication Date
- 2024-11-19
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
In the prior art, the payment channel switching strategy is single and inflexible, and it is impossible to analyze routing dynamically in multiple dimensions, resulting in low routing switching reliability and efficiency, affecting user experience.
Using the dynamic channel scheduling method of multi-dimensional collection routing, a one-time routing is performed by responding to payment methods and combining payment environment status, secondary routing is performed based on risk factors of payment channels, dynamically monitoring and regulating the availability status of payment channels, so as to achieve user inability automatic payment channel switching.
It significantly improves the efficiency of automatic switching of payment channels and the reliability of selection, enhances user experience, reduces manual operation costs, and improves payment work efficiency and effective utilization of channels.
Smart Images

Figure CN118982350A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of dynamic switching and maintenance of payment routes, and in particular to a multi-dimensional payment route dynamic channel scheduling method. Background Art
[0002] In recent years, the business departments and product lines of insurance companies have focused on expanding Internet-based sales models, and the demand for payment collection functions in new scenarios has expanded rapidly, mainly manifested in: online, mobile business, mobile terminals and other diversified payment collection needs. The rapid development of e-commerce has brought about tremendous changes in customer purchasing habits, and changes in customer consumption habits have put forward higher requirements for the convenience and diversity of online payment functions in insurance companies' online sales scenarios. For example, some users like to use WeChat, some users like to use Alipay, some users like to use Cloud QuickPass, some users like to use digital RMB, etc. At the same time, in order to provide 7 To provide 24-hour uninterrupted payment collection services, multiple channels must be backed up. When a channel has a problem, intelligent switching of channel routing must be implemented, automatically switching from one channel to another, so as not to interrupt the customer's payment actions and affect the operating experience. In addition, it is necessary to monitor the availability of each payment channel in real time to prevent unavailable payment channels from being routed and affecting the smooth progress of the payment process.
[0003] The prior art usually directly judges the timeliness of channel service according to the ratio of the number and duration of channel transactions to the threshold, or determines the channel availability status according to the number of historical failures of the channel, or constructs a prediction model containing the aforementioned channel historical transaction data to predict the channel payment success rate, so as to select the optimal channel; however, judging the timeliness and availability of channel service according to a single ratio in the prior art cannot guarantee the reliability and accuracy of the judgment result. First, the setting of the threshold is absolute, and the impact of the changing trend of user demand in different periods on the channel is not considered, resulting in the inability to adaptively adjust the threshold and objectively evaluate the channel status; secondly, there are many factors for channel failure, and there is an error in determining the channel availability status according to the total number of crashes. For example, if the user inputs wrong information due to his own reasons and the payment task fails, it is unreasonable to judge that the channel is unavailable, which affects the normal transaction operation; at the same time, using historical transaction data to train the prediction model will have the following problems: the division of samples, the quality of sample data and the model parameters all affect the accuracy and efficiency of the prediction model. Data quality generally reduces the accuracy of the model, and improving the data sample will reduce the efficiency of the calculation and cannot guarantee high-quality data, resulting in the inability to balance the efficiency and accuracy of channel selection, affecting the overall operation efficiency and user experience. In summary, the channel switching strategy in the prior art is relatively rigid, single and inflexible. It cannot dynamically analyze the routes in multiple dimensions to achieve free route switching while dynamically adjusting the channel status to ensure routing efficiency and save computing resources. For example, it cannot achieve dynamic route switching based on multiple dimensions such as channel service time, service limit, historical success rate, risk level and congestion level, and prohibit unavailable channels from participating in new routes. In addition, it cannot be done without user perception. For example, when a channel has a problem, the user is usually required to click again or reload and select other payment channels. This results in low payment route switching efficiency, poor reliability of payment channel selection, and poor user experience.
[0004] Chinese patent, publication number: CN114581094A, publication date: June 3, 2022, discloses a payment channel dynamic decision method, server and computer-readable medium, by obtaining a payment request from the payer; in response to the payment request, obtaining static characteristics, and determining a first candidate payment channel from each payment channel between the payer and the payee based on the static characteristics; obtaining the service status of each payment channel, and determining a second candidate payment channel from each payment channel based on the service status; determining a target payment channel based on the first candidate payment channel and the second candidate payment channel, and generating a target transaction instruction corresponding to the payment request. This solution does not take into account the maintenance time, congestion status and peak period of the payment channel at the current time, and simply judges the availability of the payment channel based on whether the historical crash time of the payment channel is within the time period of the payment request and whether the historical crash times exceed the threshold. There are occasional errors, inaccurate judgment conditions and unreliable judgment results. That is, payment failures caused by network signal delays and congestion do not mean that the channel is unavailable. In addition, the solution cannot dynamically adjust the availability status of the payment channel according to the failed payment situation, which will cause the unavailable channel to participate in the next routing, increase routing overhead and occupy additional computing resources. Summary of the invention
[0005] The purpose of the present invention is to address the problem of low reliability and efficiency of route switching due to the lack of multi-dimensional dynamic real-time analysis of routing conditions in the prior art, which leads to the disadvantage that users have a poor experience due to payment obstruction; a multi-dimensional collection route dynamic channel scheduling method is proposed, which realizes multi-dimensional dynamic real-time routing by performing primary route selection and secondary route selection in response to the payment method and the payment environment status, overcomes the problem that the prior art does not select channels in multiple dimensions and perform real-time scheduling and maintenance of channel status, resulting in a single payment channel switching mode, inflexibility, and inability to efficiently perform user-imperceptible and reliable dynamic switching of payment channels and real-time channel status management, realizes dynamic selection of the optimal payment path according to real-time data, and real-time regulation of the availability status of the payment path, significantly improves the automatic switching efficiency of payment channels and the reliability of payment channel selection, and further enhances the user's payment experience.
[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a multi-dimensional payment routing dynamic channel scheduling method, comprising the following steps: S1. Performing a routing selection in response to the payment method combined with the payment environment status to determine a reserve payment channel, wherein the payment environment status at least includes: payment channel availability status, channel upgrade and maintenance status, and payment channel congestion status; S2. Perform secondary routing selection on the prepared payment channel based on the risk factor of the payment channel to determine the target payment channel; S3, the user terminal executes a payment action in response to the target payment channel; S4. Analyze and obtain the payment failure reason based on the payment failure result returned by the target payment channel, and record and update the payment failure statistics of the target payment channel according to the payment failure reason; S5. Determine whether the target payment channel is abnormal based on the payment failure statistical information, make a decision on the disabled state of the target payment channel according to the abnormal result, dynamically manage whether the target payment channel is put into a new round of routing selection, and re-execute S1 to switch to other payment channels to complete the payment task.
[0007] In this solution, through primary routing selection and secondary routing selection, it is possible to dynamically monitor the real-time operation of each payment channel from multiple dimensions such as payment channel status, payment channel congestion, channel maintenance time, and channel risk accidents, so that when a payment channel has a problem, it can quickly and accurately automatically switch to the optimal payment channel. At the same time, according to the failed payment situation of the payment channel, its availability status is dynamically adjusted to avoid the unavailable problem channel being routed again before it returns to normal, thereby affecting the normal transaction of the business, ensuring that user damage is minimized, ensuring the consistency of external user experience, achieving user-free, and enhancing user experience; reducing manual operation costs, significantly improving the efficiency of payment channel switching and the reliability and security of channel selection, and improving payment work efficiency. Among them, the first routing selection can filter out channels that are unavailable and channels that are out of service. Then, by judging whether the filtered channels are congested, it can effectively avoid sending channel requests to congested channels, thereby increasing the channel load and prolonging the channel failure time, thereby improving the effective utilization rate of payment channels; then through the second routing selection, the payment channels with risks are eliminated to achieve the optimal channel selection and determine the final payment channel for this payment task; finally, according to the failed payment results returned by the target payment channel, the availability status of the payment channel is calculated and dynamically adjusted, which helps to invalidate the abnormal channels as soon as possible (here, the availability status of the channel is reset to disabled), ensure the smooth progress of payment business, and protect the rights and interests of third-party channel merchants and users.
[0008] Preferably, S1 comprises the following sub-steps: Obtain all corresponding payment channels according to the payment method selected by the user, determine whether each payment channel is in the channel upgrade and maintenance period, and filter the payment channels that are currently in the channel upgrade and maintenance state; Obtain historical transaction data and real-time payment request volume of the payment channel to determine the time domain load of the payment channel at the current time; Determining the availability of the payment channel status based on the time domain load, and filtering out payment channels whose channel status is available; The congestion rate of the payment channel is calculated in real time to determine whether the payment channel is congested at the current time. If so, S1 is re-executed; if not, the payment channel is used as a reserve payment channel.
[0009] In this solution, by filtering out channels that are unavailable and out of service at the same time through one-time routing, further failed attempts can be effectively prevented, thereby reducing system resource consumption and potential payment risks caused by continuous payment failures; it helps to switch to other stable and available payment channels as soon as possible to ensure the smooth progress of the user payment process and avoid user dissatisfaction and loss due to payment failures. By determining the time domain load of the payment channel at the current time, a more reasonable time threshold can be clearly set according to the peak period, off-peak period or off-peak period of the payment channel, which helps to make subsequent channel status judgment and congestion judgment more accurate and reliable.
[0010] Preferably, the step of obtaining the historical transaction data and the real-time payment request volume of the payment channel and determining the time period load of the payment channel at the current time comprises the following sub-steps: Retrieving historical transaction data based on the log file of the payment channel, and dividing the first time period load of the routing frequency of several payment channels within 24 hours of a natural date according to the historical transaction data, including peak period, off-peak period and low-peak period; Calculate the operation characteristic statistics and characteristic standard deviation of each peak period in the first time period, and determine the peak period threshold of each peak period according to the operation characteristic statistics combined with the characteristic standard deviation; The real-time number of first payment requests in the payment queue of the payment channel is matched with the range of the peak threshold, the peak type of the payment channel within the current time range is determined, and the peak type is used as the time domain load of the payment channel.
[0011] Preferably, the determining the availability of the payment channel status comprises the following sub-steps: Extract the first number of failed records in the statistical channel queue corresponding to the payment channel, and when the recording time of the first number of failed records is in a historical time zone adjacent to the current time, count all the number of failed records in the statistical channel queue to obtain the first failed record statistical value; Wherein, the historical time zone is dynamically adjusted according to the time domain load of the payment channel; The first failure record statistic value is compared with a failure statistic threshold value. If the first failure record statistic value is greater than the failure statistic threshold value, the payment channel is abnormal and the payment channel status is unavailable.
[0012] Preferably, calculating the congestion rate of the payment channel to determine whether the payment channel is congested at the current time includes the following sub-steps: Evaluate and determine the number of pre-processed payment requests per unit time of the payment channel based on historical transaction data corresponding to the time domain load; Counting the number of second payment requests actually processed by the payment channel within a unit time, and calculating the congestion rate according to the number of pre-processed payment requests and the number of second payment requests; When the congestion rate is greater than or equal to the congestion limit and the number of real-time first payment requests in the payment queue increases, congestion occurs in the payment channel.
[0013] Preferably, S2 comprises the following sub-steps: Conduct a channel risk audit on the prepared payment channel based on the risk rule manual of the risk factor, conduct a secondary filtering on the prepared payment channel according to the audit result, obtain the main payment channel, and use the main channel as the target payment channel to complete the secondary routing selection; Among them, the risk factors include at least regulatory penalty events, duplicate collection events, and service events of the payment channel.
[0014] In this solution, secondary routing selection is performed through channel risk events to ensure and improve the security and stability of the target payment channel. If a risky time occurs in the payment channel within the current time range, in order to reduce the risk, the payment channel will be avoided so that it cannot be routed to ensure that user transactions are not affected.
[0015] Preferably, S4 includes the following sub-steps: Analyze the reasons for failed payment results based on user payment information; If the payment information provided by the user does not match or / and the information is entered incorrectly, causing the payment failure of the target payment channel, it is determined to be a user information-related reason; if the payment information provided by the user is accurate, it is determined to be a non-user information-related reason, and the user payment information at least includes: bank card account information, account name information, and ID card information; A statistical channel queue for storing payment failure record information is allocated to the target payment channel. When the target payment channel returns information about a data payment failure, the corresponding payment failure information is written into the statistical channel queue.
[0016] Preferably, S5 comprises the following sub-steps: A1. Periodically scan the payment failure information in the statistical channel queue and extract the first failure record number; A2. judging whether the first number of failed records exceeds the statistical range based on the difference between the time information of the first number of failed records and the current time; if so, removing the first number of failed records and re-executing A1; if not, counting all the payment failure records in the statistical channel queue to obtain a second failure record statistical value; A3. Compare the second failure record statistical value with the failure statistical threshold. If the second failure record statistical value is less than the failure statistical threshold, the target payment channel is in normal state, and the target payment channel is added to a new round of channel routing list. If the second failure record statistical value is greater than or equal to the failure statistical threshold, the available state of the target payment channel is reset to a disabled state, indicating that the target payment channel is abnormal, and clear the corresponding statistical channel queue, and simultaneously issue payment channel abnormality warning information.
[0017] In this solution, statistics on the number of payment failures and the specific time of payment failures can help identify the status of the current payment channel. When an abnormality occurs in a payment channel, it will be set to an unavailable state to ensure that it is not routed before it returns to normal, thereby ensuring the user's payment experience and system stability. At the same time, the channel request that has selected the channel to perform the payment task will be switched to other normal and available payment channels as soon as possible to ensure the user's smooth transaction. Clearing statistical data can prevent old and possibly outdated data from affecting the judgment of the current status of the payment channel, ensuring that when the channel is re-enabled, corresponding decisions are made based on the latest and accurate statistical data.
[0018] Preferably, before executing S3, the method further includes: When there are multiple payment channels that meet the conditions for executing the payment task after the secondary routing selection, the target payment channel is selected three times, and the three selections include at least: Formulate a channel sequence strategy based on the payment channel type and channel cost, and determine the target payment channel through the channel sequence strategy; or, A random selection strategy for the payment channel is formulated based on a random algorithm, and the target payment channel is determined by the random selection strategy.
[0019] Preferably, the method further comprises: Planning a backup payment route for each payment channel, where the execution procedure of the backup payment route for payment requests is the same as that of its corresponding primary payment channel; When congestion occurs in the payment channel, its backup payment route is activated, and the payment requests in the payment queue of the payment channel are sent to the backup payment route to execute the payment task.
[0020] Beneficial effects of the present invention: 1. By analyzing the reasons for failed payment results returned by the target payment channel, recording the failure information according to the reasons and judging whether the payment channel is abnormal, the channel availability status of the abnormal payment channel is updated in real time, which helps the router to quickly identify unavailable channels when other payment tasks are generated, thereby avoiding unavailable channels when selecting payment channels, improving routing selection efficiency, achieving efficient channel switching, and at the same time, improving unified management of various payment channels; 2. Through primary and secondary routing selection, automatic screening and switching based on multi-dimensional conditions, the optimal payment channel is dynamically obtained, which ensures the stability and reliability of user payment tasks and significantly improves the efficiency of payment channel switching and user service quality. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts.
[0022] Figure 1 This is a flow chart of a multi-dimensional payment routing dynamic channel scheduling method of this embodiment.
[0023] Figure 2 This is a schematic diagram of a payment channel type architecture of this embodiment.
[0024] Figure 3 The figure is a schematic diagram of a payment channel calling process of this embodiment.
[0025] Figure 4 This is a traditional payment channel congestion judgment flow chart of this embodiment.
[0026] Figure 5 The figure is a schematic diagram of a routing logic flow of this embodiment.
[0027] Figure 6 This is a flow chart of payment channel abnormality judgment in this embodiment. DETAILED DESCRIPTION
[0028] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0029] Example 1: Figure 1 As shown, a multi-dimensional payment routing dynamic channel scheduling method includes steps S1-S5, wherein: S1. Perform a routing selection in response to the payment method in combination with the payment environment status to determine a reserve payment channel, wherein the payment environment status at least includes: payment channel availability status, channel upgrade and maintenance status, and payment channel congestion status.
[0030] Specifically, Figure 2 and Figure 3 As shown, S1 includes the following sub-steps: Obtain all corresponding payment channels according to the payment method selected by the user, determine whether each payment channel is in the channel upgrade and maintenance period, and filter the payment channels that are currently in the channel upgrade and maintenance status; Obtain historical transaction data and real-time payment request volume of the payment channel to determine the current time domain load of the payment channel; The availability of the payment channel status is determined based on the time domain load, and payment channels with available channel status are selected; The congestion rate of the payment channel is calculated in real time to determine whether the payment channel is congested at the current time. If so, S1 is re-executed; if not, the payment channel is used as a reserve payment channel.
[0031] Furthermore, when a third-party channel operator performs maintenance and upgrades on a payment channel, the payment channel is suspended at this time, that is, the channel cannot be routed to the interval from the start time to the end time of the maintenance, and the channel automatically becomes available only after this period of time has passed; the order of screening the availability status of the payment channel and judging whether the payment channel is in service time can be swapped, and the user can adaptively adjust the order of the two steps according to their own needs; for example, screening the service time of payment channels in available status and shielding payment channels that are currently in the maintenance and upgrade time period can reduce the amount of system calculations, avoid transaction failures caused by suspension of service, affect the user's payment process, and delay the payment completion time; for example, first screen the channel maintenance and upgrade time, and then judge the channel availability status. For example, if a channel that is not in the maintenance and upgrade time happens to be unavailable, this will still incur computing resource overhead, so the order can be dynamically adjusted; however, judging channel congestion requires judging the channel status and channel maintenance and upgrade time.
[0032] In addition, during a route selection, the most basic cost screening of payment routes can also be performed, taking into account the cost of the payment channel. This is suitable for cost-prioritized payment scenarios, and requires the formulation of a cost-prioritized strategy. The optimal solution can be obtained by taking the lowest cost as the goal through relevant optimization algorithms, thereby obtaining the lowest-cost payment channel.
[0033] Specifically, obtaining the historical transaction data and real-time payment request volume of the payment channel and determining the time period load of the payment channel at the current time includes the following sub-steps: Retrieving historical transaction data based on the log files of the payment channels, and dividing the first period load of the routing frequency of several payment channels within 24 hours of a natural date according to the historical transaction data, including peak period, off-peak period and low-peak period; Calculate the operation characteristic statistics and characteristic standard deviation of each peak period in the first period of load, and determine the peak period threshold of each peak period according to the operation characteristic statistics combined with the characteristic standard deviation; The real-time number of first payment requests in the payment queue of the payment channel is matched with the range of the peak threshold, the peak type of the payment channel in the current time range is determined, and the peak type is used as the time domain load of the payment channel.
[0034] As an implementation method, historical transaction data of a payment channel is retrieved, for example, transaction data within the past five days is retrieved, the number of payment requests processed in the same period of each day and the number of payment requests added in the request queue are obtained, and then the payment channel is comprehensively divided into peak periods (peak period, low peak period, and flat peak period) in each period of 24 hours a day; or, the average processing time of the number of payment requests processed in the same period of each day is used, and the average processing time is used as the processing efficiency to distinguish the peak period type; then, the characteristic mean and standard deviation are calculated based on the number of payment requests or the average processing time of the number of payment requests in the same period of the past five days, and the payment request threshold (peak period threshold) of each peak period is set based on the characteristic mean and standard deviation, for example, characteristic mean + standard deviation, or characteristic mean - standard deviation, to obtain a payment request threshold; the number of requests in the payment queue of the payment channel (the number of requests to be processed) is obtained in real time and compared with the peak period threshold. If it is within the range of the corresponding peak period threshold, it means that the payment channel is currently in the corresponding peak period. For example, the peak threshold is 500 payment requests, with a range of 20 payments. If there are 490 payments in the payment queue of the current payment channel, it is considered that the current payment channel is at its peak time. When judging whether the status is available or congested, the time must be shorter.
[0035] Specifically, determining the availability of the payment channel status includes the following sub-steps: Extract the first number of failed records in the statistical channel queue corresponding to the payment channel. When the recording time of the first number of failed records is in the historical time zone adjacent to the current time, count all the number of failed records in the statistical channel queue to obtain the first failed record statistical value. Among them, the historical time zone is dynamically adjusted according to the time domain load of the payment channel; The first failure record statistic value is compared with the failure statistic threshold. If the first failure record statistic value is greater than the failure statistic threshold, the payment channel is abnormal and the payment channel status is unavailable.
[0036] Specifically, calculating the congestion rate of the payment channel to determine whether the payment channel is congested at the current time includes the following sub-steps: Evaluate and determine the number of pre-processed payment requests per unit time of the payment channel based on historical transaction data corresponding to the time domain load; The number of second payment requests actually processed by the payment channel in a unit time is counted, and the congestion rate is calculated based on the number of pre-processed payment requests and the number of second payment requests; When the congestion rate is greater than or equal to the congestion limit and the number of real-time first payment requests in the payment queue increases, the payment channel is congested.
[0037] Furthermore, the congestion rate calculation formula is as follows: Congestion rate = (number of records preprocessed per unit time - number of records actually processed per unit time) / number of records preprocessed per unit time 100%; If the actual number of transactions processed is greater than or equal to the pre-processed number, it means that the channel processes quickly and efficiently without congestion. If the congestion rate continues to increase and the number of records in the payment queue increases, it means that the current channel is very congested.
[0038] By calculating the payment request processing efficiency in the past period of time, the current processing efficiency of the channel can be accurately predicted, and further accurately judge whether the payment channel is likely to be congested. Among them, the number of pre-processed payment requests per unit time is mainly preset according to the peak type of the payment channel at the current time, so that the preset value is more reliable and reasonable, and the calculated congestion rate is more accurate. For example, if the payment channel is in the peak period at the current time, then the amount of real-time payment requests will increase significantly at this time, and the amount of payment requests processed in the same time period will also increase compared to the off-peak period. Therefore, the number of pre-processed requests in the peak period needs to be higher than the number of pre-processed requests in the off-peak period, so as to achieve a balance between the actual processing volume and make the calculated results reliable and accurate; it eliminates the occasional errors in the payment channel and the state misjudgment. In addition, the historical peak period of the payment channel is calculated to predict the peak type of the payment channel in the current time, so as to set the timer to scan the channel statistics queue, and improve the accuracy of the judgment of the status availability of the payment channel in different periods. For example, the transaction volume is large during the peak period, so the channel problem should be judged in a shorter time to control the affected transaction business to the minimum range.
[0039] Traditional methods for determining channel congestion, such as Figure 4 As shown in the figure, the time taken from sending the channel to the user to the time the channel returns the result, if N consecutive transactions exceed a certain time (such as 5 seconds), and the number of requests in the payment queue reaches 50, the channel is considered to be congested. The specific process is as follows: Before sending the channel, the counter is incremented by 1, and then the initiation time (recorded as time A here) is recorded. The counter is used by the payment queue to record the payment requests received by the channel. Each time the channel is sent, it indicates a payment request, and the counter is incremented by 1. After the channel returns the result, the counter is decremented by 1, and then the end time (here recorded as time B) is recorded. The counter is decremented by 1, indicating that a payment request has been completed, and the payment request in the payment queue is decremented by 1 accordingly; According to actual business needs, the standard time threshold is set to 5 seconds, and the maximum request carrying capacity of the payment queue is set to 50. If the BA exceeds 5 seconds and the counter value is greater than 50, the channel is considered to be congested; Since the user's own network connection conditions and the payment channel's network connection conditions may have occasional delays, the traditional channel congestion judgment method, that is, BA for more than 5 seconds to judge channel congestion is not accurate enough and may cause misjudgment due to network factors. Therefore, the calculation of the payment channel congestion rate proposed in this embodiment makes an accurate prediction of the pre-processing situation in the next period of time, and more intuitively understands the payment request load and processing progress of the current channel, taking into account the number of payment requests currently being processed, and the number of payment requests in the detachment queue. If the amount of payment request data being processed tends to increase, and the time for each processing is getting longer, this means that the channel is congested. The congestion situation is judged comprehensively by multiple factors to ensure the accuracy and reliability of the judgment result.
[0040] S2. Perform secondary routing selection on the prepared payment channels based on the risk factors of the payment channels and determine the target payment channel.
[0041] Specifically, S2 includes the following sub-steps: Conduct channel risk audits on the reserve payment channels based on the risk rulebook of risk factors, conduct secondary filtering on the reserve payment channels based on the audit results, obtain the main payment channel, and use the main channel as the target payment channel to complete secondary routing selection; Among them, risk factors include at least regulatory penalty events in payment channels, duplicate collection events, and service events.
[0042] In this embodiment, the regulatory penalty event means that the third-party channel business corresponding to the payment channel has been punished by the regulatory agency for failing to comply with the payment institution supervision and management regulations, such as illegally retaining customer funds, leaking customer payment account information, etc.; the duplicate collection event means that the payment channel has experienced duplicate collection events; the service event means that the third-party channel business corresponding to the payment channel has not responded to and handled the problem in a timely manner, such as slow response caused by the replacement of account managers. The channel risk event factor can be calculated according to the actual business scenario. For example, taking one quarter as the calculation time, within one quarter, once a regulatory penalty event occurs, or there are three duplicate collection events, or five untimely responses to problems, which have an impact on the business, the payment channel will be invalidated or directly abandoned to ensure the security of the payment environment and reduce the risks that may occur during the payment process.
[0043] S3. The user terminal executes a payment action in response to the target payment channel.
[0044] Furthermore, the first routing selection (performing the first-layer routing logic) is necessary, and the second routing selection (performing the second-layer routing logic) is optional. However, from a security perspective, in order to reduce risks, if a channel has recently experienced a major risk event, it is considered that the channel cannot be routed. Therefore, channel risk screening must also be considered; usually, the optimal payment channel is switched in the order of the first-layer routing logic to the second-layer routing logic, such as Figure 5 As shown, the order of the first-layer routing logic and the second-layer routing logic can also be reasonably and adaptively arranged according to the actual business scenario and payment business type.
[0045] S4. Based on the payment failure result returned by the target payment channel, analyze and obtain the payment failure reason, and record and update the payment failure statistics of the target payment channel according to the payment failure reason.
[0046] Specifically, S4 includes the following sub-steps: Analyze the reasons for failed payment results based on user payment information; If the payment information provided by the user does not match or / and the information is entered incorrectly, causing the payment failure of the target payment channel, it is determined to be a user information-related reason; if the payment information provided by the user is accurate, it is determined to be a non-user information-related reason. The user's payment information at least includes: bank card account information, account name information, and ID card information; A statistical channel queue for storing payment failure record information is allocated to the target payment channel. When the target payment channel returns information about a data payment failure, the corresponding payment failure information is written into the statistical channel queue.
[0047] In this embodiment, by analyzing the payment result returned by the target payment channel, the payment failure situation of the payment channel is counted in real time according to the payment result, so as to provide data support for judging whether the payment channel can continue to work normally in the future; when judging the cause of payment failure, the user information type of cause is first checked, for example, according to the comparison of the user's account information, including whether the card number and the account name, the ID number match, whether the account balance is sufficient, whether the payment password entered by the user is correct, etc., it is judged whether the current payment failure belongs to the user information type or the non-user information type. If the payment channel payment failure is caused by the user information type, there is no need to re-route a new payment channel. The user is prompted with an error message to allow the user to re-enter the correct information to complete the current payment request. There is no need to switch the payment channel, because these error reasons will fail no matter which payment channel they are routed to; if it is a non-user information type of cause, such as a network error, a system error, etc., it is necessary to route and switch to other payment channels to ensure the smooth progress of the payment task. For example, if the system error of channel A causes failure, then the request is initiated again by switching to channel B through routing, and B may succeed at this time. Therefore, by returning results through the channel, distinguishing failures due to customer / non-customer reasons, and re-routing, the efficiency of interaction between systems and the success rate of one-time interaction between systems can be improved, and manual operations can be reduced, saving manual workload.
[0048] S5. Determine whether the target payment channel is abnormal based on payment failure statistics, make a decision on disabling the target payment channel based on the abnormal result, dynamically manage whether the target payment channel is put into a new round of routing selection, and re-execute S1 to switch to other payment channels to complete the payment task.
[0049] Specifically, S5 includes the following sub-steps: A1. Regularly scan and count payment failure information in the channel queue and extract the number of the first failure record; A2. Based on the difference between the time information of the first failure record and the current time, determine whether the first failure record exceeds the statistical range. If so, remove the first failure record and re-execute A1. If not, count all payment failure records in the statistical channel queue to obtain the second failure record statistical value. A3. Compare the second failure record statistical value with the failure statistical threshold. If the second failure record statistical value is less than the failure statistical threshold, the target payment channel is in normal state and the target payment channel is added to a new round of channel routing list. If the second failure record statistical value is greater than or equal to the failure statistical threshold, the available state of the target payment channel is reset to a disabled state, indicating that the target payment channel is abnormal, and clear its corresponding statistical channel queue, and simultaneously issue payment channel abnormality warning information.
[0050] As an implementation method, Figure 6 As shown, the specific steps of S5 are as follows: B1. Allocate a queue to each channel (e.g., if there are 4 channels, a, b, c, and d, then allocate 4 channel statistics queues, a1, b1, c1, and d1). When channel a fails due to non-customer reasons, write a message to queue a1 (recording the time of failure, accurate to milliseconds). The same is true for other channels. B2. There is a corresponding timer for each queue, which can regularly scan and count the number of failure records and the time when the failure occurred in the channel queue; the timer can achieve second-level statistics, usually once every 1 minute; B3. When the timer counts the error records in the queue, it first determines whether the difference between the first error time and the current time exceeds the preset value T0. If it exceeds, it is removed from the queue and the next determination is continued until a value that does not exceed is found, and then enters B4; B4. Count the number of failed records in the queue and compare it with the failure statistics threshold to determine whether the payment channel status needs to be reset and whether the queue needs to be cleared; If the number of failed records is less than the failure statistics threshold, there is no need to reset the payment channel status and wait for the next cycle of statistics; If the number of failed records is greater than or equal to the failure statistical threshold, the payment channel will be set to "disabled" status, the queue will be cleared, and an early warning message will be issued to inform relevant staff through the page.
[0051] It is understandable that if the number of failure records is greater than or equal to the failure statistics threshold and the queue is not cleared, the first failure time in the queue will definitely exceed the T0 value when the timer is counting, and they will still be removed one by one. This will increase the system computing load and waste computing resources. Therefore, when the status of the payment channel is set to unavailable, the queue is cleared at one time to prevent old, possibly outdated data from affecting the judgment of the current status of the payment channel, so as to ensure that when the channel is reactivated, the system makes decisions based on the latest and accurate data; at the same time, clearing the failure count means that after the channel is reactivated, it will start from an initial state and will no longer be affected by the previous number of failures. It can start collecting new data again, which will help to fairly evaluate the performance and stability of the channel, thereby more accurately monitoring the operation of the payment channel and ensuring that potential problems are discovered in a timely manner and appropriate measures are taken.
[0052] Furthermore, after the payment channel status is reset to unavailable, it can be restored to available status. The restoration methods include but are not limited to: The first type: pure manual operation; The second type: automatic recovery. The system uses the probe principle to periodically send queries to the channel to verify whether the interface has been restored. You can adaptively select the recovery method based on the actual business scenario. Since the second method can only be through the probe principle, and can only call the query interface (the query interface will not generate new transactions), and the query and collection interfaces on the channel business are separate, a normal query does not mean that the collection is normal. Therefore, the automatic recovery method has a certain inaccuracy. Therefore, you can choose pure manual operation to restart the available status of the payment channel, and then enable it after manual confirmation. This is more rigorous, safer, and has the least impact on the business.
[0053] Furthermore, before executing S3, the following steps are also included: When there are multiple payment channels that meet the conditions for executing the payment task after the secondary routing selection, the target payment channel is selected three times, and the three selections include at least: Formulate a channel sequence strategy based on payment channel type and channel cost, and determine the target payment channel through the channel sequence strategy; or, A random selection strategy for payment channels is formulated based on a random algorithm, and the target payment channel is determined through the random selection strategy.
[0054] Specifically, formulating a channel sequence strategy includes: encoding the payment channels with sequence numbers or other ordered label identifications through pre-set sequence rules, and then preferentially selecting one of the payment channels as the target payment channel according to the order and selection agreement. For example, the order rule is to sort by the first letter of the name of the third-party channel business, and if the first letters are the same, sort by the second letter, and so on, and select the payment channel with the smallest, largest or other position in the order number, so as to select a main channel from multiple payment channels that meet the requirements to perform the payment task.
[0055] Specifically, the random selection strategies include: Step 1: Count the number of channels that meet the conditions, recorded as N; Step 2: Generate a random integer from 1 to N, denoted by R, R = random.nextInt(N) + 1; Step 3: Take the Rth channel.
[0056] Furthermore, a multi-dimensional payment routing dynamic channel scheduling method also includes: Plan a backup payment route for each payment channel. The execution procedure of the backup payment route for payment requests is the same as that of its corresponding primary payment channel. When a payment channel is congested, its backup payment route is activated, and the payment requests in the payment channel payment queue are sent to the backup payment route to execute the payment task.
[0057] When a payment channel is congested, you need to suspend sending payment requests to the payment channel. You need to wait until the channel processes a certain number of payment requests in the payment queue (the number of payment requests in the payment queue is less than the request carrying capacity) before you can continue sending payment requests to the channel.
[0058] The beneficial effects of the embodiment are as follows: through primary routing selection and secondary routing selection, it is possible to dynamically monitor the real-time operation of each payment channel from multiple dimensions such as payment channel status, payment channel congestion, channel maintenance time, and channel risk accidents, so that when a payment channel has a problem, it can automatically switch to the optimal payment channel quickly and accurately. At the same time, the availability status of the payment channel is dynamically adjusted according to the failed payment situation of the payment channel to avoid the unavailable problem channel being routed again before it returns to normal, thereby affecting the normal transaction of the business, ensuring that user damage is minimized, ensuring the consistency of external user experience, achieving user-free, and enhancing user experience; reducing manual operation costs, significantly improving the efficiency of payment channel switching and the reliability and security of channel selection, and improving operational work efficiency. Among them, the first routing selection can filter out the channels that are unavailable and out of service time. Then, by judging whether the filtered channels are congested, it can effectively avoid sending channel requests to the congested channels, thereby increasing the channel load and extending the channel failure time, and improving the effective utilization rate of the payment channel; then through the second routing selection, the payment channels with risks are eliminated to achieve the optimal channel selection and determine the final payment channel for this payment task; finally, according to the failed payment result returned by the target payment channel, the availability status of the payment channel is calculated and dynamically adjusted, which helps to invalidate the abnormal channel in the first time (here, the availability status of the channel is reset to disabled), ensure the smooth progress of the payment business, and protect the rights and interests of third-party channel merchants and users. The processing load of the main payment channel is decomposed through the backup channel, which effectively avoids the long-term congestion of the payment channel and ensures that the main payment channel can maintain normal operation.
[0059] The above specific embodiments are preferred embodiments of the present invention, and are not intended to limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to the present specific embodiments. All equivalent changes made in accordance with the shape, structure, and method of the present invention are within the protection scope of the present invention.
Claims
1. A multi-dimensional payment routing dynamic channel scheduling method, characterized by: The steps include: S1. Performing a routing selection in response to the payment method combined with the payment environment status to determine a reserve payment channel, wherein the payment environment status at least includes: payment channel availability status, channel upgrade and maintenance status, and payment channel congestion status; S2. Perform secondary routing selection on the prepared payment channel based on the risk factor of the payment channel to determine the target payment channel; S3, the user terminal executes a payment action in response to the target payment channel; S4. Analyze and obtain the payment failure reason based on the payment failure result returned by the target payment channel, and record and update the payment failure statistics of the target payment channel according to the payment failure reason; S5. Determine whether the target payment channel is abnormal based on the payment failure statistical information, make a decision on the disabled state of the target payment channel according to the abnormal result, dynamically manage whether the target payment channel is put into a new round of routing selection, and re-execute S1 to switch to other payment channels to complete the payment task.
2. A multi-dimensional payment routing dynamic channel scheduling method according to claim 1, characterized in that: The S1 comprises the following sub-steps: Obtain all corresponding payment channels according to the payment method selected by the user, determine whether each payment channel is in the channel upgrade and maintenance period, and filter the payment channels that are currently in the channel upgrade and maintenance state; Obtain historical transaction data and real-time payment request volume of the payment channel to determine the time domain load of the payment channel at the current time; Determining the availability of the payment channel status based on the time domain load, and filtering out payment channels whose channel status is available; The congestion rate of the payment channel is calculated in real time to determine whether the payment channel is congested at the current time. If so, S1 is re-executed; if not, the payment channel is used as a reserve payment channel.
3. A multi-dimensional payment routing dynamic channel scheduling method according to claim 2, characterized in that: The step of obtaining the historical transaction data and the real-time payment request volume of the payment channel and determining the time period load of the payment channel at the current time includes the following sub-steps: Retrieving historical transaction data based on the log file of the payment channel, and dividing the first period load of the routing frequency of several payment channels within 24 hours of a natural date according to the historical transaction data, including peak period, off-peak period and low-peak period; Calculate the operation characteristic statistics and characteristic standard deviation of each peak period in the first time period, and determine the peak period threshold of each peak period according to the operation characteristic statistics combined with the characteristic standard deviation; The real-time number of first payment requests in the payment queue of the payment channel is matched with the range of the peak threshold, the peak type of the payment channel within the current time range is determined, and the peak type is used as the time domain load of the payment channel.
4. A multi-dimensional payment routing dynamic channel scheduling method according to claim 2, characterized in that: The determining the availability of the payment channel status comprises the following sub-steps: Extract the first number of failed records in the statistical channel queue corresponding to the payment channel, and when the recording time of the first number of failed records is in a historical time zone adjacent to the current time, count all the number of failed records in the statistical channel queue to obtain the first failed record statistical value; Wherein, the historical time zone is dynamically adjusted according to the time domain load of the payment channel; The first failure record statistic value is compared with a failure statistic threshold value. If the first failure record statistic value is greater than the failure statistic threshold value, the payment channel is abnormal and the payment channel status is unavailable.
5. A multi-dimensional payment routing dynamic channel scheduling method according to claim 2, characterized in that: Calculating the congestion rate of the payment channel to determine whether the payment channel is congested at the current time includes the following sub-steps: Evaluate and determine the number of pre-processed payment requests per unit time of the payment channel based on historical transaction data corresponding to the time domain load; Counting the number of second payment requests actually processed by the payment channel within a unit time, and calculating the congestion rate according to the number of pre-processed payment requests and the number of second payment requests; When the congestion rate is greater than or equal to the congestion limit and the number of real-time first payment requests in the payment queue increases, congestion occurs in the payment channel.
6. A multi-dimensional payment routing dynamic channel scheduling method according to claim 1, characterized in that: The S2 comprises the following sub-steps: Conduct a channel risk audit on the prepared payment channel based on the risk rule manual of the risk factor, conduct a secondary filtering on the prepared payment channel according to the audit result, obtain the main payment channel, and use the main channel as the target payment channel to complete the secondary routing selection; Among them, the risk factors include at least regulatory penalty events, duplicate collection events, and service events of the payment channel.
7. A multi-dimensional payment routing dynamic channel scheduling method according to claim 1, characterized in that: The S4 comprises the following sub-steps: Analyze the reasons for failed payment results based on user payment information; If the payment information provided by the user does not match or / and the information is entered incorrectly, causing the payment failure of the target payment channel, it is determined to be a user information-related reason; if the payment information provided by the user is accurate, it is determined to be a non-user information-related reason, and the user payment information at least includes: bank card account information, account name information, and ID card information; A statistical channel queue for storing payment failure record information is allocated to the target payment channel. When the target payment channel returns information about a data payment failure, the corresponding payment failure information is written into the statistical channel queue.
8. A multi-dimensional payment routing dynamic channel scheduling method according to claim 7, characterized in that: The S5 comprises the following sub-steps: A1. Periodically scan the payment failure information in the statistical channel queue and extract the first failure record number; A2. judging whether the first number of failed records exceeds the statistical range based on the difference between the time information of the first number of failed records and the current time; if so, removing the first number of failed records and re-executing A1; if not, counting all the payment failure records in the statistical channel queue to obtain a second failure record statistical value; A3. Compare the second failure record statistic value with the failure statistic threshold. If the second failure record statistic value is less than the failure statistic threshold, the target payment channel is in normal state, and the target payment channel is added to a new round of channel routing list. If the second failure record statistical value is greater than or equal to the failure statistical threshold, the available state of the target payment channel is reset to a disabled state, indicating that the target payment channel is abnormal, and the corresponding statistical channel queue is cleared, and payment channel abnormality warning information is issued simultaneously.
9. A multi-dimensional payment routing dynamic channel scheduling method according to claim 1, characterized in that: Before executing S3, the following steps are also included: When there are multiple payment channels that meet the conditions for executing the payment task after the secondary routing selection, the target payment channel is selected three times, and the three selections include at least: Formulate a channel sequence strategy based on the payment channel type and channel cost, and determine the target payment channel through the channel sequence strategy; or, A random selection strategy for the payment channel is formulated based on a random algorithm, and the target payment channel is determined by the random selection strategy.
10. A multi-dimensional payment routing dynamic channel scheduling method according to claim 1, characterized in that: The method further comprises: Planning a backup payment route for each payment channel, where the execution procedure of the backup payment route for payment requests is the same as that of its corresponding primary payment channel; When congestion occurs in the payment channel, its backup payment route is activated, and the payment requests in the payment queue of the payment channel are sent to the backup payment route to execute the payment task.
Citation Information
Patent Citations
Payment channel abnormity checking method and device, computing equipment and storage medium
CN113435871A
Payment channel dynamic decision-making method, server and computer readable medium
CN114581094A
Payment channel monitoring method and device, parking lot payment management equipment and storage medium
CN118761766A
Real-time order callback method and system
WO2019196244A1
Cited By
Transaction order processing method and system, electronic equipment and storage medium
CN119477483A
Cross-border payment path optimization system and method based on multi-dimensional data analysis
CN120106854A
Multi-channel payment integration method and system for educational scene integrated information service
CN120338798A
Processing method and device for appointment transfer, electronic equipment and storage medium
CN120509891A
A processing method and device for pre-booking a transfer, an electronic device, and a storage medium
CN120509891B
