Flow control method and device, medium and product

Abstract

The invention provides a flow control method and device, a medium and a product. The method comprises the following steps: in response to a flow control request for a target object, determining a regulation target corresponding to the target object; wherein the regulation and control target is determined based on the traffic proportion of the target object in the total business; acquiring the accumulated real-time service flow of the target object in the current time period; and calculating the flow deviation of the target object according to the regulation and control target and the real-time service flow, and controlling the flow of the target object according to the flow deviation. When traffic control needs to be carried out on the target object, a time period regulation target can be set for the target object and the accumulated real-time service traffic of the target object can be acquired in real time without changing the existing promotion rule and recall rule, so that the traffic deviation is calculated by utilizing a preset algorithm; and the traffic of the target object is dynamically adjusted according to the traffic deviation, so that the service traffic of the target object can be effectively improved, and the service traffic is prevented from being preempted by other services.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and more particularly to flow control methods, devices, media, and products. Background Art

[0002] There are many ways to promote products on product sales platforms, such as sorting by transaction probability.

[0003] In existing technology, platforms promote a wide variety of products, each with numerous products. Different product types have varying levels of popularity. Popular, best-selling products attract more attention, which translates to more traffic. Unpopular products, however, receive less attention and click-through rates, resulting in lower rankings and less traffic. They may even lose traffic to other products. Therefore, a solution is needed to effectively control traffic based on demand. Summary of the Invention

[0004] The present disclosure provides flow control methods, devices, media and products.

[0005] According to one aspect of the present disclosure, a traffic control method is provided, comprising: determining, in response to a traffic control request for a target object, a control target corresponding to the target object; wherein the control target is determined based on the target object's traffic proportion in the total traffic; obtaining the target object's accumulated real-time traffic volume in a current time period; calculating the target object's traffic deviation based on the control target and the real-time traffic volume; and controlling the target object's traffic volume based on the traffic deviation.

[0006] According to the traffic control method of at least one embodiment of the present disclosure, the target traffic allocated to the target object in the total business is determined based on the traffic and traffic proportion of the total business; wherein the target traffic is a target set in units of days; the historical time period traffic of the target object is obtained; based on the target traffic and the historical time period traffic, the time period regulation target is estimated.

[0007] According to the flow control method of at least one embodiment of the present disclosure, the flow of the total business is determined; the total business includes multiple object types, and the flow ratio is determined according to the object type to which the target object belongs; the target flow is calculated using the business flow and the flow ratio.

[0008] According to the flow control method of at least one embodiment of the present disclosure, the time-divided control target is linearly decomposed to obtain the minute control target; the flow deviation of the target object is calculated using the minute control target and the real-time business flow.

[0009] According to the flow control method of at least one embodiment of the present disclosure, the proportional coefficient, integral coefficient and differential coefficient corresponding to the target object are determined; the proportional coefficient, integral coefficient, differential coefficient and flow deviation are input into the proportional-integral-differential formula to calculate the flow control signal; the flow control signal is input into the control formula to obtain the control position offset of the target object; and the flow of the target object is controlled according to the control position offset.

[0010] According to the flow control method of at least one embodiment of the present disclosure, when the flow control signal is greater than zero, the flow control signal, the control target weight and the highest-weighted business weight are input into the control formula to calculate the first control position offset; when the flow control signal is less than zero, the flow control signal and the control target weight are input into the control formula to calculate the second control position offset.

[0011] The control formula is: The calculation formula for the control target weight is: .

[0012] Among them, t is the control target, i is a target topic item, is the model score of target topic item i; is the model score of the item with the highest score; It is the rank position obtained by sorting the model in descending order; It is the control signal output by the proportional integral differential algorithm; To regulate the target weight; γ is used to control tanh(β ) part of the growth rate parameter; β is used to control the regulatory target weight in the tanh function Size parameters; Parameter used to indirectly adjust the sensitivity of ranking downgrade.

[0013] According to the flow control method of at least one embodiment of the present disclosure, whether the sorting rule or recall rule of the target object needs to be adjusted is judged based on the flow difference and the difference threshold; if the flow difference is greater than the difference threshold, the sorting rule or recall rule of the target object is adjusted, and the flow control of the target object is performed according to the adjusted sorting rule or recall rule.

[0014] According to the flow control method of at least one embodiment of the present disclosure, if the flow difference is less than the difference threshold and the flow control response time is less than the time threshold, the proportional coefficient, integral coefficient and differential coefficient in the proportional-integral-differential formula are adjusted.

[0015] According to the traffic control method of at least one embodiment of the present disclosure, the target object is promoted according to the traffic regulation result; or, the target object is mined according to the traffic regulation result; or, the target object is sorted according to the traffic regulation result.

[0016] According to a second aspect of the present disclosure, an electronic device is provided, comprising: a memory storing execution instructions; and a processor executing the execution instructions stored in the memory, so that the processor executes the flow control method of any embodiment of the present disclosure.

[0017] According to a third aspect of the present disclosure, a readable storage medium is provided, in which execution instructions are stored. When the execution instructions are executed by a processor, they are used to implement the flow control method of any embodiment of the present disclosure.

[0018] According to a fourth aspect of the present disclosure, a computer program product is provided, comprising a computer program, wherein when the computer program is executed by a processor, the flow control method of any one embodiment of the present disclosure is implemented.

[0019] Based on the above scheme, in response to the traffic control request for the target object, the time period control target corresponding to the target object is determined; wherein the target object is one of the total businesses; the real-time business traffic accumulated in the current time period of the target object is obtained; based on the time period control target and the real-time business traffic, the traffic deviation of the target object is calculated, and the traffic of the target object is controlled according to the traffic deviation. Through the above method, when it is necessary to control the traffic of the target object, the time period control target can be set for the target object without changing the existing promotion rules and recall rules, and the accumulated real-time business traffic of the target object can be collected in real time, and then the traffic deviation can be calculated using a preset algorithm, and the traffic of the target object can be dynamically adjusted according to the traffic deviation, thereby effectively improving the business traffic of the target object and avoiding the business traffic being preempted by other businesses. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings illustrate exemplary embodiments of the present disclosure and together with the description serve to explain the principles of the present disclosure. These drawings are included to provide a further understanding of the present disclosure and are incorporated in and constitute a part of this specification.

[0021] Figure 1 The figure shows an overall flow chart of a flow control method according to an embodiment of the present application.

[0022] Figure 2 A flow chart illustrating the method for determining the regulatory objectives provided in this application.

[0023] Figure 3A schematic diagram of the flow control process used in this application.

[0024] Figure 4 This is a schematic diagram of the system structure for system flow control as illustrated in this application.

[0025] Figure 5 This is a schematic structural diagram of a flow control device provided in this application.

[0026] Figure 6 This is a schematic block diagram of the structure of an electronic device according to one embodiment of the present application. DETAILED DESCRIPTION

[0027] The present disclosure is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific examples described herein are intended only to illustrate the relevant content and are not intended to limit the present disclosure. It should also be noted that, for ease of description, only the portions relevant to the present disclosure are shown in the accompanying drawings.

[0028] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present disclosure can be combined with each other. The technical solution of the present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0029] In existing product marketing and promotion scenarios, various retrieval and ranking algorithms are often designed to help promote products to consumers. This allows products to be more quickly matched with consumers with purchasing or rental needs. However, in practice, due to the wide variety of product specifications and types, some products are often not very popular. To ensure faster and more effective product sales, conventional promotion algorithms typically increase the exposure of best-selling products, resulting in higher exposure rates for best-selling products. Conversely, slower-selling products receive lower exposure rates.

[0030] To this end, the present disclosure proposes a traffic control method, which increases traffic for certain specific products by adding a traffic control algorithm without changing the existing product promotion rules, thereby giving specific products more exposure opportunities and improving product sales.

[0031] For the convenience of description and to make the technical solutions of the specific embodiments of the present disclosure easier to understand, before describing the flow control method implemented in the present disclosure, the technical terms involved in the specific embodiments of the present disclosure are defined as follows: Traffic refers to the number of users who access a brand's website, content pages, or social media platforms through various promotional channels. It is a key indicator of the effectiveness of brand promotion on the internet. By promoting traffic, brands can expand their exposure, increase conversion rates, and achieve market coverage.

[0032] Exposure refers to the number of times an ad or product is seen by users on an online platform. In online marketing, exposure is a key metric for measuring the effectiveness of ad or product display. When an ad or product is seen by more users, its exposure increases, which helps to increase brand awareness and product attention. The subject of exposure is the ad or product, and the area is the entire online platform. The platform uses rules to control product exposure. When an ad or product closely complies with the platform's rules, exposure increases, and vice versa. These rules include explicit rules and implicit rules. Explicit rules include ad size, pixels, keywords, etc., while implicit rules include high conversion rates and rich content.

[0033] Recall. In online promotion, the term "recall" usually refers to a mechanism in the recommendation system, that is, from a massive library of items (such as products, videos, articles, etc.), based on some characteristics of users and items, a small number of items that users may be interested in are quickly retrieved, and then handed over to the subsequent sorting link for further processing.

[0034] PID, short for Proportional-Integral-Derivative, is a common feedback control algorithm used in closed-loop control systems. The PID control algorithm combines the proportional, integral, and derivative components to maintain the system output at the desired setpoint. The following is a detailed explanation of each component in the PID algorithm: The proportional term (P) generates feedback based on the error between the system output and the setpoint. A larger error results in a larger control signal. The output of the proportional controller is proportional to the input deviation, quickly reflecting the deviation and reducing it, but it cannot eliminate static error. The integral term (I) accumulates the persistent error between the system output and the setpoint, helping to eliminate steady-state error (error caused by the output failing to reach the setpoint). The strength of the integral action is determined by the integral time constant Ti: a larger Ti weakens the integral action, and vice versa. The derivative term (D) generates feedback based on the rate of change of the error, helping to improve system stability and response speed. The derivative component reflects the trend (rate of change) of the error signal and can introduce an effective early correction signal into the system before the error signal becomes excessive, thereby accelerating system action and reducing settling time.

[0035] It should be noted that the target object here can be one type of property (e.g., office buildings, residential buildings, shops), and the corresponding total business can be understood as properties of the same type within a certain area (e.g., a business district). The target object here can also be a type of property, and the corresponding total business can be understood as multiple different types of properties within a certain area (e.g., a business district) (e.g., the total business includes office buildings, residential buildings, and shops).

[0036] Figure 1 FIG. 1 shows a schematic diagram of the overall flow of a flow control method according to an embodiment of the present application. Figure 1 The method shown includes steps 101 to 103. The method can be executed by an electronic device such as a server (a local server or a cloud server).

[0037] Specifically, Figure 1 The method steps shown are as follows.

[0038] Step 101: In response to a traffic control request for a target object, determine a regulation target corresponding to the target object; wherein the regulation target is determined based on a traffic proportion of the target object in the total business.

[0039] For ease of understanding, let's assume the target object refers to a specific type of property (for example, rental housing). While promoting the entire business, including the target object, using the existing backend algorithm, if you discover that the target object's traffic share is too low or sales targets are not meeting expectations, you can consider directly controlling traffic for the target object without changing the existing promotion algorithm. Initiate a traffic control request and determine the traffic control target for the target object.

[0040] When dividing the traffic proportion of each object in the total business, the division can be performed on a daily basis. Furthermore, when performing traffic control, the control is also performed on a daily basis. Of course, if there are special needs, the traffic proportion can also be divided on an hourly basis. For example, in some specific short-term promotion scenarios, the traffic proportion of each object can be divided on an hourly basis. For another example, in some long-term product promotion scenarios, the traffic proportion can be divided on a monthly basis. In the embodiment of the present application, the traffic proportion division and control will be performed on a daily basis.

[0041] Here, we assume that the total business includes three types of properties: offices, retail stores, and residential properties. If statistical analysis reveals that residential sales are not meeting expectations, optimization can be achieved by adjusting the traffic share of residential properties within the total business. In the leasing business, traffic can be segmented. For example, offices can account for 20% of the total traffic, retail stores for 30%, and residential properties for 50%.

[0042] Step 102: Obtain the real-time business traffic accumulated by the target object in the current time period.

[0043] In actual applications, in order to achieve better traffic control and adjustment for the target object, after setting the control target for the target object in the manner described above, it is also necessary to understand the real-time business traffic of the target object in real time.

[0044] Real-time traffic refers to the cumulative traffic from the agreed-upon time to the present. Specifically, since each consumer visit is logged by the server and stored in the database, when calculating traffic statistics for a target object, the target object's real-time traffic can be obtained by reading the log records in the database. The real-time traffic accumulated in the current time period can refer to the access traffic data from a specified time (for example, midnight) to the present, which is a summary of multiple time periods.

[0045] For example, suppose you want to count the cumulative traffic for each object in the total business starting at midnight each day. If the current time is 7:00 AM, the corresponding traffic statistics will be the access traffic for each business from midnight to 7:00 AM, a total of 7 hours, including the traffic for the target object of interest. Find the traffic data for the target object of interest by object type.

[0046] By acquiring business traffic in real time, we can understand the previous traffic change trend of the target object and the gap between the business traffic of the target object and the expected traffic, so as to better regulate the business traffic distribution.

[0047] Step 103: Calculate the traffic deviation of the target object according to the control target and the real-time service traffic, and control the traffic of the target object according to the traffic deviation.

[0048] After obtaining the target object's real-time service traffic and the target object's control target through the aforementioned method, the traffic deviation between the control target and the real-time service traffic can be further calculated. The traffic deviation here can be understood as the difference between the traffic data corresponding to the control target and the traffic data corresponding to the real-time service traffic.

[0049] After calculating the flow deviation, it can be further used to control the flow of the target object. For example, the proportional-integral-differential formula can be used to control the flow of the target object. In addition, other algorithms can be used for flow control, such as the proportional-integral (PI) algorithm and fuzzy control algorithms. Although these algorithms are commonly used in the industrial field, they are used for flow control due to their simple formula parameters and high control efficiency. It can achieve real-time and efficient flow control of the target object without requiring a lot of computing power.

[0050] Based on the method described above, in response to the traffic control request for the target object, the time period control target corresponding to the target object is determined; wherein the target object is a business in the total business; the real-time business traffic accumulated in the current time period of the target object is obtained; based on the time period control target and the real-time business traffic, the traffic deviation of the target object is calculated, and the traffic of the target object is controlled according to the traffic deviation. Through the above method, when it is necessary to control the traffic of the target object, the time period control target can be set for the target object without changing the existing promotion rules and recall rules, and the accumulated real-time business traffic of the target object can be collected in real time, and then the traffic deviation can be calculated using a preset algorithm, and the traffic of the target object can be dynamically adjusted according to the traffic deviation, thereby effectively improving the business traffic of the target object and avoiding the business traffic being preempted by other businesses.

[0051] In one or more embodiments of the present application, Figure 2 The flowchart of the method for determining the control target provided in this application is as follows. Figure 2 As can be seen in the figure, the method for determining the control target specifically includes the following steps: Step 201: Determine the target traffic volume allocated to the target object within the total traffic volume and traffic proportion; the target traffic volume is set on a daily basis. Step 202: Obtain the target object's historical traffic volume for a specific time period. Step 203: Based on the target traffic volume and historical traffic volume for a specific time period, estimate the control target for each time period.

[0052] Determining the target traffic volume allocated to the target object within the total traffic volume and traffic volume percentage as described in step 201 includes the following steps: Step 2011: Determining the total traffic volume. Step 2012: Determining the traffic volume of the target object based on the type of object to which the target object belongs, given that the total traffic volume and traffic volume percentage are included. Step 2013: Calculating the target traffic volume using the traffic volume and traffic volume percentage of the target object.

[0053] In practical applications, in order to make the flow control process more stable, reasonable, and accurate, it is necessary to set reasonable control targets for users. When setting control targets, they can be set by estimation.

[0054] Determine the object types for the various objects included in the overall business, as well as the total traffic corresponding to the total business. Then, based on the actual promotion needs of the objects, allocate different traffic proportions to different types of objects. It's easy to understand that objects that currently require key promotion will receive a higher traffic proportion, while objects with less promotion needs will receive a lower traffic proportion.

[0055] The target traffic here can be understood as calculated using traffic percentage and total business traffic. For example, if the total business traffic is 1000 and the traffic percentage allocated to the target object is 20%, the target traffic corresponding to the target object is calculated as 1000 * 20% = 200. Furthermore, the traffic deviation can be calculated using the target traffic and historical traffic.

[0056] When making an estimate, you can base it on historical data. Obtain the historical traffic flow for the target object at the same time in the historical data. If the historical traffic flow is low, add a certain value to the historical traffic flow and use it as the estimated control target for the time period. If the historical traffic flow is high, subtract a certain value from the historical traffic flow and use it as the estimated control target for the time period.

[0057] After obtaining the estimated control targets for each time period, it is necessary to further evaluate or judge whether the current control targets for each time period are reasonable. If the estimated control targets for each time period are significantly unreasonable, the control targets need to be recalculated or optimized.

[0058] When making an estimate, you can also use the current total ratio. Specifically, obtain the service traffic corresponding to each object in the total service before the current period. Calculate the real-time traffic ratio corresponding to each service flow, and use the calculated real-time traffic ratio of the target object as the estimated control target. This estimation method dynamically adjusts to the real-time changes in the traffic conditions of each service in the total service. This makes the estimated control target more accurate and reasonable.

[0059] In addition, it can also be set in a linear manner. That is, after determining the total target for daily regulation, the total target can be averaged, and the average regulation target for each hour can be used as the regulation target for the current PID formula. For example, if the total flow target of the target object today is 1000, and a day is 24 hours, then the corresponding average value for each hour is 1000 / 24=41.7. Because the accumulated flow in the historical period is used to calculate the flow deviation, here, after obtaining the average target flow per hour, when calculating the regulation target, it is also necessary to make an estimate based on the accumulated time length. In other words, the average value of each hour before the current moment is summed up, and the flow in the corresponding historical period is summed up to determine the regulation target for each time period.

[0060] This approach allows us to specify estimated control targets based on time periods, rather than daily targets. This allows for more timely and accurate adjustments, making the overall control process smoother and more stable, without sudden changes in traffic.

[0061] In one or more embodiments of the present application, the traffic deviation of the target object is calculated based on the control target and the real-time business traffic, including: linearly decomposing the control target of the time period to obtain the minute control target; and using the minute control target and the real-time business traffic to calculate the traffic deviation of the target object.

[0062] In some scenarios, for example, where traffic control sensitivity requires more timely and accurate responses, the hourly control targets can be further decomposed. This decomposition can be done using a linear decomposition method to break down the targets into minute-by-minute control targets. Specifically, the control targets calculated in the aforementioned embodiment are averaged and decomposed into minute-by-minute control targets.

[0063] For example, the previously calculated hourly control target is 600. The historical cumulative value is 540, and the current hourly increment is 60. The corresponding minute control target could be 60 / 60 = 1. Based on this, the corresponding minute control target is also a cumulative value. For example, in the current time period, the minute control target for 10 minutes is 550 = 540 + 10 * 1. Similarly, the minute control target for 30 minutes is 570 = 540 + 30 * 1.

[0064] Through the above method, the adjustment target is further decomposed, and the linear decomposition method is simple and direct. In addition, by setting the minute control target, the control process response is more timely and accurate.

[0065] In one or more embodiments of the present application, flow control of a target object is performed based on the flow deviation, including: determining the proportional coefficient, integral coefficient, and differential coefficient corresponding to the target object; inputting the proportional coefficient, integral coefficient, differential coefficient, and flow deviation into a proportional-integral-differential formula to calculate a flow control signal; inputting the flow control signal into a control formula to obtain a control position offset for the target object; and performing flow control on the target object based on the control position offset.

[0066] In practical applications, after determining the control target (which can be an hourly or minute-by-minute target), a proportional-integral-derivative (PID) calculation can be performed based on this target. The PID calculation formula combines the proportional (P), integral (I), and differential (D) components to generate a control signal and output it. This output is used to adjust the system so that the actual output approaches the control target.

[0067] Simply put: the proportional component (P) directly responds to the current error, providing a correction proportional to the error. The integral component (I) corrects for accumulated historical errors, eliminating long-standing deviations. The differential component (D) predicts the error's changing trend, suppressing rapid error changes and reducing system overshoot. By adjusting the three parameters kp, ki, and kd, the system's response speed, stability, and accuracy can be precisely controlled.

[0068] The global PID is responsible for adjusting the flow input of the entire system so that the total flow is distributed according to the preset flow ratio.

[0069] The global error, or traffic deviation, is calculated as the target value at the current moment after time-based segmentation. Assume the total traffic volume is T, where office traffic = 0.5*T, retail traffic = 0.2*T, and residential traffic = 0.3*T.

[0070] The actual real-time traffic flow for each object is Office Building A, Shop A, and Residential Building A. In the event of traffic flow deviation, E Office Building = T Office Building - Office Building A, E Shop = T Shop - Shop A, and E Residential Building = T Residential Building - Residential Building A. The calculated flow difference is used as feedback input to the PID formula, which then generates a flow control signal.

[0071] After the flow control signal is calculated using the above method, it is used as the input of the control formula. After being processed by the control formula, the control position offset is obtained. Then, the flow of the target object can be controlled based on the control position offset. The specific process of calculating the control position offset using the control formula can be found in the following embodiment and will not be repeated here.

[0072] Through this method, the proportional-integral-derivative (PID) formula and the control formula are used to calculate the control position offset of the target object. This simple and effective calculation method can improve the flow control effect of the target object. At the same time, it does not require tedious calculations or modifications to the previous recall and sorting rules.

[0073] In one or more embodiments of the present application, the flow control signal is input into the control formula to obtain the control position offset of the target object, including: when the flow control signal is greater than zero, the flow control signal, the control target weight and the highest-weighted business weight are input into the control formula to calculate the first control position offset; when the flow control signal is less than zero, the flow control signal and the control target weight are input into the control formula to calculate the second control position offset.

[0074] The control formula is as follows.

[0075] The calculation formula for the control target weight is: .

[0076] Where t is the control target, i is a target topic (item, marked as i in the above formula), is the model score for target topic i; is the model score of the target topic with the highest score; It is the rank position obtained by sorting the model in descending order; It is the flow control signal output by the proportional integral differential algorithm. γ: This parameter controls the growth rate of the tanh (βwt) part. It determines the amplification effect of the adjustment amount Δrank. It can be compared to an amplifier, which is used to control the effect of exposure. If the γ value is large, the control effect will be more obvious; if it is small, the control effect will be relatively mild. β: This parameter is used to control the control target weight in the tanh function ( ), which generally reflects the degree to which the current target subject deviates from the target control plan (such as exposure). β determines the sensitivity of the control. A larger β value means a stronger response to deviations from the control target, while a smaller β means a more gradual response.

[0077] η is different from γ and is more of an indirect control of the sensitivity of ranking downgrades. If η is large, then The adjustment effect on ranking is more obvious. If η is small, the effect of this part is weaker. The relationship between γ and η, these two parameters appear in different parts of the formula, affecting ascending and descending sorting respectively. In ascending sorting (the role of γ): directly amplifies or reduces the impact of the control weight on the improvement of Δrank. γ determines whether the control target (such as business share or tag weight) can quickly show its effect in ascending sorting. In descending sorting (the role of η): controls the sensitivity of the control weight to ranking suppression. η determines the impact of the control signal on ranking adjustment in descending sorting.

[0078] After calculation by PID formula, the calculation result (that is, flow control signal) is obtained As input to the control formula. Calculation results It can be positive or negative. When it is positive, the traffic of the target object is upgraded; otherwise, when the calculation result is When it is negative, the traffic of the target object is degraded.

[0079] When using the above control scheme to control, we should avoid over-control. That is, in order to avoid excessive reduction of traffic for objects that are ranked high in the original ranking (occupying very high traffic), the parameter To ensure that those with high weights are not overly suppressed.

[0080] After calculating the control position offset using the above debugging formula, reordering can be further performed based on the calculation result. The reordering calculation formula is as follows.

[0081] 。

[0082] In actual applications, if the frequency of position offset calculations is high, reordering can be performed based on a certain period. For example, if the calculation frequency is once every minute, the corresponding reordering period is 5 minutes. Therefore, when updating, reordering can be performed based on the above reordering formula. The above formula can update the displacement in a timely manner and at a reasonable period. While increasing the exposure of the target object, it can also avoid causing discomfort to browsing consumers due to overly frequent adjustments.

[0083] In one or more embodiments of the present application, flow control is performed on the target object based on the adjustment position offset, including: judging whether the sorting rule or recall rule of the target object needs to be adjusted based on the flow difference and the difference threshold; if the flow difference is greater than the difference threshold, adjusting the sorting rule or recall rule of the target object, and performing flow control on the target object based on the adjusted sorting rule or recall rule.

[0084] In practical applications, after calculating the control position offset using the above formula, it is necessary to further consider whether to implement flow control for the target object based on the calculated control position offset based on the actual situation. If the flow difference is greater than the difference threshold, it indicates that the deviation is too large. If the proportional integral differential formula and the control formula are used to adjust it, it may be time-consuming. Therefore, other methods for flow control can be used here. For example, the recall rule or sorting rule can be directly modified.

[0085] During the adjustment process, the target object's position in the recalled or sorted set is adjusted according to the modified recall or sorting rules. This directly increases the target object's exposure. Generally speaking, the higher the target object's ranking, the higher its corresponding traffic share; conversely, the lower its ranking, the lower its corresponding traffic share. The more frequently a target object is recalled, the higher its corresponding traffic share; conversely, the lower its ranking, the lower its corresponding traffic share.

[0086] Through the above method, the actual flow difference calculation results are comprehensively considered, and the corresponding method is adopted according to the actual situation to achieve flow control of the target object. This meets the control needs while maximizing the control efficiency.

[0087] In one or more embodiments of the present application, it also includes: if the flow difference is less than the difference threshold and the flow control response time is less than the time threshold, the proportional coefficient, integral coefficient and differential coefficient in the proportional-integral-differential formula are adjusted.

[0088] In actual applications, if the flow difference is less than the difference threshold, it means that the flow difference is not far from the expectation (difference threshold), and the current formula (PID formula, control formula) can meet the control requirements.

[0089] During the control process, the actual control effect of the PID formula must also be considered. For example, gradually increasing the proportional coefficient P. During the adjustment process, it is necessary to pay attention to the PID system to avoid overshoot and oscillation.

[0090] If the flow control response time is less than the time threshold—for example, if the set time threshold is 3 seconds and the actual flow response time is 5 seconds—then the integral coefficient I can be increased to speed up the response. However, fine-tuning is required to avoid system instability and oscillation. If significant oscillation or overshoot occurs after accelerating the response, the differential coefficient can be increased. However, it is important to avoid introducing noise or high-frequency oscillation when using the differential coefficient.

[0091] By adjusting the formula coefficients, the flow control efficiency and accuracy can be improved.

[0092] In one or more embodiments of the present application, it also includes: promoting the target object according to the traffic regulation result; or, mining the target object according to the traffic regulation result; or, sorting the target object according to the traffic regulation result.

[0093] After implementing traffic control through the above solution, you can further execute corresponding tasks based on the control results. Specifically, the traffic control results can be used to promote business targets. For example, if the target target's exposure rate is relatively low during promotion or sales are relatively low within a specified time frame, in order to improve the promotion effect, the target target's business traffic can be increased.

[0094] For example, to better identify which business objects are bestsellers among all business objects, you can use the above methods to increase exposure traffic for each business object. By increasing traffic in a targeted manner, you can discover business objects with greater potential and true bestsellers.

[0095] Furthermore, traffic control results can be used for sorting. Products displayed to consumers are sorted according to certain criteria (e.g., popularity). When performing traffic control, the position of the target object in the sorting list can be changed based on the traffic control results.

[0096] Based on the above content, it can be seen that without modifying the original recall algorithm and sorting rules, by performing flow control, the application needs of target objects in various application scenarios can be met, such as promotion needs, mining needs, etc.

[0097] In order to facilitate understanding of the flow control application scheme proposed in this application. Figure 3 This is a schematic diagram of the flow control process used in this application. Figure 3 It can be seen that in this regulation system, the control target is set through the target configuration platform. In the target configuration platform, the target flow allocated to the target object in the total business can be determined based on the flow of the total business and the proportion of the flow; wherein, the target flow is a target set in units of days; the historical time period flow of the target object is obtained; based on the target flow and the historical time period flow, the control target for each time period is estimated. Among them, the historical time period data comes from the hourly flow report and / or real-time message queue (MessageQueue, mq) in the log storage unit. After the above-mentioned control target is determined, the PID formula and the control formula are used for processing to obtain the flow control signal. On this basis, the recall and sorting tasks are further performed. From Figure 3 As can be seen, the existing sorting and recall algorithms include relevance recall, popular recall, new product recall and other existing algorithms. The existing sorting algorithms include default sorting, diversity sorting, single product weighted sorting and other existing algorithms. When performing recall and sorting tasks, there is no need to modify the existing recall algorithm and sorting algorithm, and add a control recall scheme based on the original recall function, and add a flow control scheme based on the original sorting function. Instead, distinguish and mark the business objects that already exist in the material collection. Specifically, add a control mark to the target object that needs to adjust the sorting, and in subsequent recalls, the target object with the mark will be recalled first. When sorting based on the recall set, the target object is sorted in the front position of the list according to the sorting priority.

[0098] During sorting, the recall and traffic control algorithms are used to replace the content in the original sorted list. Assuming that category B is the target object, when performing the replacement, B1, B2, B3, etc. are pre-placed into the list, and the exposure traffic of category A objects is appropriately reduced. By replacing the target object into the list, the traffic of the target object can be increased in a targeted manner.

[0099] like Figure 4 This is a schematic diagram of the system structure of the system flow control system illustrated in this application. Figure 4As can be seen in the figure, in this system, access information for each service is recorded in logs. These logs, along with other raw data (such as historical traffic), are stored in various data sources, such as Redis, ES, and HBASE databases. To fully utilize and process this underlying data, various supporting tools are required, such as caching, configuration, logging, and real-time messaging. Based on the aforementioned solution, improvements have been made to reordering, recalling, and ranking to effectively control target objects. This means that traffic control has been integrated into these modules. The system diagram shows that the existing sorting and recall algorithms do not need to be modified. Instead, traffic control is added to the existing algorithms. Furthermore, this traffic control allows for multi-level granularity. Traffic proportions can be divided and controlled for the overall service, and further divided for sub-services. This solution enables multi-level and refined traffic control to meet users' diverse service control needs.

[0100] Based on any of the above embodiments, the present disclosure also provides a flow control device. Figure 5 This is a schematic block diagram of the structure of a flow control device provided by this application. Figure 5 As can be seen in the figure, the device includes: The determination module 51 is configured to determine a control target corresponding to the target object in response to a traffic control request for the target object; wherein the control target is determined based on a traffic proportion of the target object in the total business.

[0101] The acquisition module 52 is used to acquire the real-time business traffic accumulated by the target object in the current time period.

[0102] The calculation module 53 is used to calculate the flow deviation of the target object according to the control target and the real-time business flow, and control the flow of the target object according to the flow deviation.

[0103] Optionally, the determination module 51 is also used to determine the target traffic allocated to the target object in the total business based on the traffic and traffic proportion of the total business; wherein the target traffic is a target set in days; obtain the historical time period traffic of the target object; based on the target traffic and historical time period traffic, estimate the time period control target.

[0104] Optionally, the determination module 51 is further used to determine the traffic of the total business; the total business includes multiple object types, and the traffic proportion is determined according to the object type to which the target object belongs; the target traffic is calculated using the business traffic and traffic proportion.

[0105] Optionally, the calculation module 53 is further configured to linearly decompose the time-divided control target to obtain a minute control target; and calculate the flow deviation of the target object using the minute control target and the real-time business flow.

[0106] Optionally, the calculation module 53 is also used to determine the proportional coefficient, integral coefficient and differential coefficient corresponding to the target object; input the proportional coefficient, integral coefficient, differential coefficient and flow deviation into the proportional integral differential formula to calculate the flow control signal; input the flow control signal into the control formula to obtain the control position offset of the target object; and control the flow of the target object according to the control position offset.

[0107] Optionally, the calculation module 53 is also used to input the flow control signal, the control target weight and the highest-weighted business weight into the control formula when the flow control signal is greater than zero, and calculate the first control position offset; when the flow control signal is less than zero, input the flow control signal and the control target weight into the control formula, and calculate the second control position offset.

[0108] The control formula is as follows.

[0109] The calculation formula for the control target weight is: .

[0110] Among them, t is the control target, i is a target topic, is the model score of target topic i; is the model score of the target topic with the highest score; It is the rank position obtained by sorting the model in descending order; It is the control signal output by the proportional integral differential algorithm; To regulate the target weight; γ is used to control tanh(β ) part of the growth rate parameter; β is used to control the regulatory target weight in the tanh function Size parameters; Parameter used to indirectly adjust the sensitivity of ranking downgrade.

[0111] Optionally, the calculation module 53 is also used to determine whether the sorting rules or recall rules of the target object need to be adjusted based on the flow difference and the difference threshold; if the flow difference is greater than the difference threshold, the sorting rules or recall rules of the target object are adjusted, and the flow of the target object is controlled according to the adjusted sorting rules or recall rules.

[0112] Optionally, the calculation module 53 is further configured to adjust the proportional coefficient, integral coefficient and differential coefficient in the proportional-integral-differential formula if the flow difference is less than a difference threshold and the flow control response time is less than a time threshold.

[0113] Optionally, an application module 54 is further included, which is used to promote the target object according to the traffic regulation result; or to mine the target object according to the traffic regulation result; or to sort the target object according to the traffic regulation result.

[0114] Each of the above modules can be implemented through computer program modules. Figure 5 The flow control device 5 shown can be implemented based on a computer program module architecture.

[0115] The execution entity of the flow control method in the specific implementation of the present disclosure may be an electronic device such as a server (a cloud server or a local server).

[0116] Therefore, based on any of the above embodiments, the present disclosure further provides an electronic device, which can execute the flow control method of any of the above embodiments of the present disclosure.

[0117] Figure 6 This is a schematic block diagram of the structure of an electronic device according to an embodiment of the present application. Figure 1 The corresponding flow control method.

[0118] The hardware structure of the electronic device 1000 can be implemented using a bus architecture. The bus architecture can include any number of interconnecting buses and bridges, depending on the specific application and overall design constraints of the hardware. The bus 1100 connects various circuits including one or more processors 1200, memory 1300, and / or hardware modules. The bus 1100 can also connect various other circuits 1400 such as peripheral devices, voltage regulators, power management circuits, external antennas, etc.

[0119] Bus 1100 may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Component Architecture (EISA) bus. Buses can be classified as address buses, data buses, control buses, and the like. For ease of illustration, this figure shows only one connecting line, but this does not imply that there is only one bus or only one type of bus.

[0120] The present disclosure also provides a readable storage medium having a computer program stored therein, which is used to implement the above-mentioned method when the computer program is executed by a processor. "Readable storage medium" can be any device that can contain, store, communicate, propagate or transmit a program for use in an instruction execution system, device or equipment or in combination with these instruction execution systems, devices or equipment. More specific examples of readable storage media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and editable read-only memory (EPROM or flash memory), an optical fiber device, and a portable read-only memory (CDROM), etc.

[0121] The present disclosure also provides a computer program product. The method of the present disclosure can be implemented in whole or in part using software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed, the process or function of the present disclosure is performed in whole or in part.

[0122] A computer program or instruction can be stored in a readable storage medium or transferred from one readable storage medium to another. For example, the computer program or instruction can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The readable storage medium can be any accessible medium or a data storage device such as a server or data center that integrates one or more accessible media. The accessible medium can be a magnetic medium such as a floppy disk, hard disk, or magnetic tape; an optical medium such as a digital video disk; or a semiconductor medium such as a solid-state drive. The computer-readable storage medium can be a volatile or non-volatile storage medium, or can include both volatile and non-volatile types of storage media.

[0123] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present disclosure 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.

[0124] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks 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.

[0125] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0126] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational 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.

[0127] In the description of this specification, the description with reference to the terms "one embodiment / method", "some embodiments / methods", "example", "specific example", or "some examples" means that the specific features, structures, or characteristics described in conjunction with the embodiment / method or example are included in at least one embodiment / method or example of the present disclosure. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment / method or example. Moreover, the specific features, structures, or characteristics described may be combined in a suitable manner in any one or more embodiments / methods or examples. In addition, those skilled in the art may combine and combine different embodiments / methods or examples described in this specification and the features of different embodiments / methods or examples, unless they are contradictory.

[0128] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. Throughout the present disclosure, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0129] Those skilled in the art will appreciate that the above embodiments are merely intended to clearly illustrate the present disclosure and are not intended to limit the scope of the present disclosure. Other changes or modifications may be made based on the above disclosure, and such changes or modifications are still within the scope of the present disclosure.

Claims

1. A flow control method, characterized in that: The method comprises: In response to a traffic control request for a target object, determine a control target corresponding to the target object; wherein the control target is determined based on a traffic proportion of the target object in the total service; Obtain the real-time business traffic accumulated by the target object in the current time period; The flow deviation of the target object is calculated according to the regulation target and the real-time business flow, and the flow of the target object is controlled according to the flow deviation.

2. The method according to claim 1, characterized in that The method for determining the control target includes: Determine the target flow rate allocated to the target object in the total business according to the flow rate of the total business and the flow rate proportion; wherein the target flow rate is a target set in units of days; Obtain the historical time period traffic of the target object; Based on the target flow and the historical time period flow, the control targets for each time period are estimated.

3. The method according to claim 2, characterized in that The determining, according to the traffic of the total service and the traffic proportion, a target traffic allocated to the target object in the total service includes: Determining the flow of the total business; The total business includes multiple object types, and the traffic proportion is determined according to the object type to which the target object belongs; The target flow is calculated using the flow of the business and the flow ratio.

4. The method according to claim 1, characterized in that: The calculating the flow deviation of the target object according to the regulation target and the real-time service flow includes: Linearly decompose the control target in each time period to obtain a minute control target; The flow deviation of the target object is calculated using the minute control target and the real-time business flow.

5. The method according to any one of claims 1 to 4, characterized in that Performing flow control on the target object according to the flow deviation includes: Determine the proportional coefficient, integral coefficient and differential coefficient corresponding to the target object; Input the proportional coefficient, the integral coefficient, the differential coefficient and the flow deviation into a proportional-integral-differential formula to calculate a flow control signal; Inputting the flow control signal into a control formula to obtain a control position offset for the target object; The flow rate of the target object is controlled according to the regulated position offset.

6. The method according to claim 5, characterized in that The step of inputting the flow control signal into a control formula to obtain a control position offset of the target object includes: When the flow control signal is greater than zero, the flow control signal, the control target weight and the highest weighted service weight are input into the control formula to calculate and obtain a first control position offset; When the flow control signal is less than zero, the flow control signal and the control target weight are input into the control formula to calculate the second control position offset; The control formula is: The calculation formula of the control target weight is: Among them, t is the regulation target, i is a target topic item, is the model score of target topic item i; is the model score of the item with the highest score; It is the rank position obtained by sorting the models in descending order; It is the control signal output by the proportional integral differential algorithm; is the target weight for regulation; γ is used to control the growth rate of the tanh (βwt) part; β is used to control the target weight for regulation in the tanh function Size parameters; Parameter used to indirectly adjust the sensitivity of ranking downgrade.

7. The method according to claim 5, characterized in that The performing flow control on the target object according to the adjustment position offset comprises: According to the flow difference and the difference threshold, determining whether the target object needs to adjust the sorting rule or the recall rule; If the flow difference is greater than the difference threshold, the sorting rule or the recalling rule of the target object is adjusted, and the flow of the target object is controlled according to the adjusted sorting rule or the recalling rule.

8. The method according to claim 7, characterized in that Also includes: If the flow difference is smaller than the difference threshold, and the flow control response time is smaller than the time threshold, the proportional coefficient, the integral coefficient and the differential coefficient in the proportional-integral-differential formula are adjusted.

9. An electronic device, characterized in that: include: A memory storing execution instructions; as well as A processor, wherein the processor executes the execution instruction stored in the memory, so that the processor executes the flow control method according to any one of claims 1 to 8.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the flow control method according to any one of claims 1 to 8 is implemented.

Images