A multi-dimensional membership rights distribution system

Abstract

The present invention discloses a multi-dimensional member rights distribution system, which relates to the technical field of member rights distribution. The initial generation module randomly generates a number of first-order distribution strategies based on the multi-dimensional information of the distribution strategy selection. The loop iteration module repeats the steps of the screening module, the vector exchange module, and the vector mutation module to iterate. When the preset conditions are met, the iteration loop is jumped out to obtain the final stage strategy set. The strategy output module generates a proportion value for all the distribution strategies in the final stage strategy set through a fitness function, and the distribution strategy with the largest proportion value is selected and output as the current distribution strategy combination. The distribution system distributes rights to members according to the current distribution strategy combination. The distribution system can generate a corresponding rights distribution strategy combination for members in combination with multi-dimensional information, and generate the optimal rights distribution strategy combination after loop search, which not only effectively improves the rights utilization rate, but also increases the adaptability of the distribution strategy to the environment.

Description

Technical Field

[0001] The present invention relates to the technical field of member rights distribution, and in particular to a multi-dimensional member rights distribution system. Background Art

[0002] Traditional one-size-fits-all services are no longer suitable for modern consumers. People want personalized and customized services, and the member rights distribution system can meet this demand. It provides corresponding rights and benefits based on the personalized needs and behavioral habits of members, and aims to provide members with personalized rights distribution and management services, usually composed of software and platforms.

[0003] Existing distribution systems usually distribute benefits to members based on a single distribution strategy, such as formulating benefits distribution strategies based on member levels, consumption behaviors, or interest preferences. This has the following defects:

[0004] 1. A single distribution strategy often cannot meet the personalized needs of different members. Different members may have different preferences, behavioral habits and needs. A single strategy cannot fully take these factors into account, resulting in a lack of personalization in the distribution of benefits and causing some members to be unable to fully utilize their benefits. If the distribution of benefits is based only on the membership level, low-level members may not be able to enjoy the benefits and privileges that high-level members can obtain, resulting in low utilization of benefits;

[0005] 2. A single strategy often only considers part of the member information, which can easily lead to incomplete information on rights distribution, poor results in the distribution strategy, and difficulty in adapting to changes in the environment and member needs. In a rapidly changing market environment, a single strategy may become outdated and unable to be adjusted in time to adapt to new situations.

[0006] Based on this, the present invention proposes a multi-dimensional member rights distribution system, which can automatically generate an optimal rights distribution strategy by combining multi-dimensional rights distribution strategies to solve the above problems. Summary of the invention

[0007] The purpose of the present invention is to provide a multi-dimensional member benefits distribution system to solve the deficiencies in the background technology.

[0008] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: a multi-dimensional member rights distribution system, including an initial generation module, a calculation module, a screening module, a vector exchange module, a vector mutation module, a loop iteration module, and a strategy output module;

[0009] Initial generation module: The generation tool randomly generates several first-order distribution strategies based on the multi-dimensional information of the distribution strategy selection. Each first-order distribution strategy contains different equity distribution strategy weight vectors, and all first-order distribution strategies are established into the first-stage strategy set;

[0010] Calculation module: Generates proportional values ​​for all first-order distribution strategies in the first-stage strategy set through the fitness function;

[0011] Screening module: Use the screening algorithm to select a corresponding number of first-order distribution strategies from the first-stage strategy set as the second-order distribution strategies, and use all the second-order distribution strategies to establish the second-stage strategy set;

[0012] Vector exchange module: randomly exchange the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set, update the second-order distribution strategies that have completed the strategy weight vector exchange to the third-order distribution strategies, and establish the third-stage strategy set with all the third-order distribution strategies;

[0013] Vector mutation module: Use the modification tool to randomly mutate the strategy weight vectors of all third-order distribution strategies in the third-stage strategy set, and update the third-order distribution strategy that has completed the random mutation of the strategy weight vector to the fourth-order distribution strategy;

[0014] Loop Iteration Module: Repeat the steps of the screening module, vector exchange module, and vector mutation module to iterate. When the preset conditions are met, the iteration loop is jumped out to obtain the final stage strategy set.

[0015] Strategy output module: Generate a proportional value for all distribution strategies in the final stage strategy set through the fitness function, select the distribution strategy with the largest proportional value and output it as the current distribution strategy combination, and distribute rights and interests to members based on the current distribution strategy combination.

[0016] In a preferred embodiment, the initial generation module obtains preset multi-dimensional information, which includes membership level, consumption behavior and interest preference;

[0017] For each first-order distribution strategy, randomly generate a vector containing the weights of different equity distribution strategies. The length of the vector matches the dimension of the equity distribution strategy, and each element represents the weight of the corresponding dimension.

[0018] After generating multiple first-order distribution strategies, the generated first-order distribution strategies are saved and a first-stage strategy set is established.

[0019] In a preferred embodiment, the calculation module obtains the equity utilization rate, member floating amplitude and strategy expenditure cost of the first-order distribution strategy, substitutes the equity utilization rate, member floating amplitude and strategy expenditure cost into the fitness function to calculate the proportion value of the first-order distribution strategy, and the function expression is:

[0020]

[0021] ; In the formula, bl zis the proportional value, qlz is the equity utilization rate, qsz is the strategy expenditure cost, cyd is the member floating range, α, β, γ are the proportional coefficients of equity utilization rate, strategy expenditure cost, and member floating range, and α, β, γ are all greater than 0;

[0022] For each first-order distribution strategy in the first-stage strategy set, the corresponding proportion value is calculated through the fitness function, and the proportion value is associated with the corresponding first-order distribution strategy to form a mapping relationship table of proportion value-strategy.

[0023] In a preferred embodiment, the calculation logic of the member floating amplitude is as follows: the historical member click-through rate and the historical member churn rate of the first-order distribution strategy are obtained, and the period when the historical member click-through rate exceeds the click-through rate threshold is recorded as the member active period, and the period when the historical member churn rate exceeds the churn rate threshold is recorded as the member churn period. After performing integral calculation on the member active period and the member churn period, the member floating amplitude is obtained. The function expression is as follows: H(t) is the change in member retention, [t x , t y ] is the member active period, [t i , t j ] is the period of member loss.

[0024] In a preferred embodiment, the calculation logic of the equity utilization rate is: obtain the historical equity usage amount and the total equity amount of the first-order distribution strategy, and obtain the equity utilization rate by dividing the historical equity usage amount by the total equity amount;

[0025] The logic for obtaining the strategy expenditure cost is: obtain the various equity costs, technology development and maintenance costs, and operating costs provided to members by the first-order distribution strategy, and add up the various equity costs, technology development and maintenance costs, and operating costs provided to members to obtain the strategy expenditure cost.

[0026] In a preferred embodiment, the screening module obtains the proportion values ​​of all first-order distribution strategies in the first-stage strategy set, and adds the proportion values ​​of all first-order distribution strategies to obtain the sum of the proportion values, calculates the selection probability of each first-order distribution strategy, and the selection probability is obtained by dividing the proportion value by the sum of the proportion values, and maps the selection probability to the roulette wheel to construct the selection range;

[0027] Generate a random number through a random generation tool, truncate the random number to one decimal place, select a first-order distribution strategy based on the value range of the random number, and repeatedly generate random numbers to select a first-order distribution strategy;

[0028] When the number of first-order distribution strategies exceeds the quantity threshold, the roulette wheel selection method is stopped, and the selected first-order distribution strategy is used as the second-order distribution strategy.

[0029] In a preferred embodiment, the screening module obtains the ratio values ​​of all first-order distribution strategies in the first-stage strategy set, and compares the ratio values ​​of the first-order distribution strategies with a preset first ratio threshold value, and the first ratio threshold value is used to determine whether the benefits of the first-order distribution strategies are qualified;

[0030] If the ratio value of the first-order distribution strategy is less than the first ratio threshold, the first-order distribution strategy is screened out; if the ratio value of the first-order distribution strategy is greater than or equal to the first ratio threshold, the first-order distribution strategy is selected;

[0031] If the number of selected first-order distribution strategies is less than or equal to the quantity threshold, the selected first-order distribution strategy is used as the second-order distribution strategy;

[0032] If the number of selected first-order distribution strategies is greater than the quantity threshold, the selected first-order distribution strategies are sorted from large to small by the proportion value, and after the first-order distribution strategies after the quantity threshold are screened out, the selected first-order distribution strategies are used as the second-order distribution strategies.

[0033] In a preferred embodiment, the loop iteration module repeats the steps of the screening module, the vector exchange module, and the vector mutation module to iterate, continuously acquires new distribution strategies, and compares the number of iterations with the number threshold. The number threshold is used to determine whether it is necessary to jump out of the iteration loop. When the number of iterations exceeds the preset number threshold, the iteration loop is jumped out, and all new distribution strategies are acquired to establish a final stage strategy set.

[0034] In a preferred embodiment, the vector exchange module randomly exchanges the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set in pairs. The second-order distribution strategies that have exchanged strategy weight vectors no longer exchange strategy weight vectors. If there is a single second-order distribution strategy that has not been exchanged, the second-order distribution strategy that has not been exchanged will be screened out.

[0035] A multi-dimensional member benefits distribution method, the distribution method comprising the following steps:

[0036] S1: The distribution system randomly generates several first-order distribution strategies based on the multi-dimensional information of the distribution strategy selection through the generation tool. Each first-order distribution strategy contains different equity distribution strategy weight vectors. All first-order distribution strategies are established into the first-stage strategy set

[0037] S2: Generate a proportional value for all first-order distribution strategies in the first-stage strategy set through the fitness function, and use the screening algorithm to select a corresponding number of first-order distribution strategies from the first-stage strategy set as second-order distribution strategies, and establish a second-stage strategy set with all second-order distribution strategies;

[0038] S3: Randomly exchange the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set, update the second-order distribution strategies that have completed the strategy weight vector exchange to third-order distribution strategies, and establish the third-stage strategy set with all third-order distribution strategies;

[0039] S4: Perform random mutation of the strategy weight vectors of all third-order distribution strategies in the third-stage strategy set through the modification tool, update the third-order distribution strategy that has completed the random mutation of the strategy weight vector to the fourth-order distribution strategy, and establish the fourth-stage strategy set with all the fourth-order distribution strategies;

[0040] S5: Repeat steps S2, S3, and S4 to iterate. When the preset conditions are met, the iteration loop is exited to obtain the final stage strategy set.

[0041] S6: Generate a proportional value for all distribution strategies in the final stage strategy set through the fitness function, select the distribution strategy with the largest proportional value and output it as the current distribution strategy combination. The distribution system distributes rights and interests to members based on the current distribution strategy combination.

[0042] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0043] 1. The present invention randomly generates several first-order distribution strategies based on the multi-dimensional information of the distribution strategy selection through the initial generation module. Each first-order distribution strategy contains different equity distribution strategy weight vectors. The loop iteration module repeats the steps of the screening module, the vector exchange module, and the vector mutation module to iterate. When the preset conditions are met, the iteration loop is jumped out to obtain the final stage strategy set. The strategy output module generates a proportion value for all distribution strategies in the final stage strategy set through the fitness function, and selects the distribution strategy with the largest proportion value as the output of the current distribution strategy combination. The distribution system distributes equity to members based on the current distribution strategy combination. The distribution system can generate corresponding equity distribution strategy combinations for members in combination with multi-dimensional information, and generate the optimal equity distribution strategy combination after loop search, which not only effectively improves the utilization rate of equity, but also increases the adaptability of the distribution strategy to the environment.

[0044] 2. The present invention uses a vector exchange module to randomly exchange the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set, prioritizes the strategy weight vector exchange for the second-order distribution strategies with large ratio values, and updates the second-order distribution strategies that have completed the strategy weight vector exchange to third-order distribution strategies. The vector mutation module uses a modification tool to randomly mutate the strategy weight vectors of all third-order distribution strategies in the third-stage strategy set, randomly exchanges the strategy weight vectors of the second-order distribution strategies, and randomly mutates the strategy weight vectors of the third-order distribution strategies, which is beneficial to improving the robustness of the present application and its adaptability to different scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0046] Figure 1 It is a system module diagram of the present invention.

[0047] Figure 2 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] Example: See Figure 1 As shown, a multi-dimensional member rights distribution system described in this embodiment includes an initial generation module, a calculation module, a screening module, a vector exchange module, a vector mutation module, a loop iteration module, and a strategy output module;

[0050] Initial generation module: A number of first-order distribution strategies are randomly generated based on the multi-dimensional information of the distribution strategy selection through the generation tool. Each first-order distribution strategy contains different equity distribution strategy weight vectors. For example, [0.3, 0.4, 0.3] is a first-order distribution strategy, which represents the weights based on member level, consumption behavior and interest preference, respectively. That is, in this first-order distribution strategy, the member level accounts for 0.3, the consumption behavior accounts for 0.4, and the interest preference accounts for 0.3. All first-order distribution strategies are used to establish the first-stage strategy set, which is sent to the calculation module;

[0051] Calculation module: Generates a proportional value for all first-order distribution strategies in the first-stage strategy set through the fitness function. The larger the proportional value, the better the historical benefits of the first-order distribution strategy. The proportional value is sent to the screening module and the vector exchange module.

[0052] Screening module: uses a screening algorithm to select a corresponding number of first-order distribution strategies from the first-stage strategy set as second-order distribution strategies, establishes a second-stage strategy set from all second-order distribution strategies, and sends the second-stage strategy set to the vector exchange module;

[0053] Vector exchange module: randomly exchange the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set. The second-order distribution strategies that have exchanged strategy weight vectors will no longer exchange strategy weight vectors. Priority will be given to exchanging strategy weight vectors for the second-order distribution strategies with large proportion values. If there is a single second-order distribution strategy that has not been exchanged, the second-order distribution strategy that has not been exchanged will be screened out, and the second-order distribution strategy that has completed the strategy weight vector exchange will be updated to the third-order distribution strategy. All third-order distribution strategies will establish a third-stage strategy set, and the third-stage strategy set will be sent to the vector mutation module;

[0054] Vector mutation module: The strategy weight vectors of all third-order distribution strategies in the third-stage strategy set are randomly mutated through the modification tool. In this application, the strategy weight vectors of the second-order distribution strategies are randomly exchanged, and the strategy weight vectors of the third-order distribution strategies are randomly mutated, which is conducive to improving the robustness of this application and the adaptability to different scenarios. The third-order distribution strategy that has completed the random mutation of the strategy weight vector is updated to a fourth-order distribution strategy, and all fourth-order distribution strategies are used to establish a fourth-stage strategy set;

[0055] Loop iteration module: repeats the steps of screening module, vector exchange module and vector mutation module to iterate. When the preset conditions are met, the iteration loop is jumped out to obtain the final stage strategy set, which is sent to the strategy output module.

[0056] Strategy output module: Generates a proportional value for all distribution strategies in the final stage strategy set through the fitness function, selects the distribution strategy with the largest proportional value and outputs it as the current distribution strategy combination. The distribution system distributes benefits to members based on the current distribution strategy combination.

[0057] This application randomly generates several first-order distribution strategies based on the multi-dimensional information of the distribution strategy selection through the initial generation module. Each first-order distribution strategy contains different equity distribution strategy weight vectors. The loop iteration module repeats the steps of the screening module, the vector exchange module, and the vector mutation module to iterate. When the preset conditions are met, the iteration loop is jumped out to obtain the final stage strategy set. The strategy output module generates a proportional value for all distribution strategies in the final stage strategy set through the fitness function, and selects the distribution strategy with the largest proportional value as the output of the current distribution strategy combination. The distribution system distributes equity to members based on the current distribution strategy combination. The distribution system can combine multi-dimensional information to generate corresponding equity distribution strategy combinations for members, and generate the optimal equity distribution strategy combination after cyclic search, which not only effectively improves the utilization rate of equity, but also increases the adaptability of the distribution strategy to the environment.

[0058] The present application uses a vector exchange module to randomly exchange the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set in pairs, prioritizes the strategy weight vector exchange for the second-order distribution strategies with large ratio values, and updates the second-order distribution strategies that have completed the strategy weight vector exchange to third-order distribution strategies. The vector mutation module uses a modification tool to randomly mutate the strategy weight vectors of all third-order distribution strategies in the third-stage strategy set, randomly exchanges the strategy weight vectors of the second-order distribution strategies in pairs, and randomly mutates the strategy weight vectors of the third-order distribution strategies, which is beneficial to improving the robustness of the present application and its adaptability to different scenarios.

[0059] See also Figure 2 As shown, the specific workflow is:

[0060] The distribution system randomly generates several first-order distribution strategies based on the multi-dimensional information of the distribution strategy selection through the generation tool. Each first-order distribution strategy contains different equity distribution strategy weight vectors. For example, [0.3, 0.4, 0.3] is used as a first-order distribution strategy, which respectively represents the weights based on membership level, consumption behavior and interest preference. That is, in this first-order distribution strategy, the membership level accounts for 0.3, the consumption behavior accounts for 0.4, and the interest preference accounts for 0.3. All first-order distribution strategies are used to establish a first-stage strategy set. The fitness function is used to generate a proportion value for all first-order distribution strategies in the first-stage strategy set. The larger the proportion value, the better the historical benefit of the first-order distribution strategy. The screening algorithm is used to select the corresponding number of first-order distribution strategies from the first-stage strategy set as the second-order distribution strategy. All second-order distribution strategies are used to establish a second-stage strategy set. The strategy weight vectors of the second-order distribution strategies in the second-stage strategy set are randomly exchanged. The second-order distribution strategies that have exchanged strategy weight vectors will no longer exchange strategy weight vectors. The second-order distribution strategies with large proportion values ​​are prioritized. Weight vector exchange, if there is a single second-order distribution strategy that has not been exchanged, the second-order distribution strategy that has not been exchanged will be screened out, the second-order distribution strategy that has completed the policy weight vector exchange will be updated to the third-order distribution strategy, and all the third-order distribution strategies will establish a third-stage policy set. The policy weight vectors of all the third-order distribution strategies in the third-stage policy set are randomly mutated through the modification tool. In this application, the second-order distribution strategies are randomly exchanged in pairs, and the policy weight vectors of the third-order distribution strategies are randomly mutated, which is beneficial to improving the robustness and adaptability of the present application to different scenarios. The third-order distribution strategy that has completed the random mutation of the policy weight vector is updated to the fourth-order distribution strategy, and all the fourth-order distribution strategies are established. The fourth-stage policy set is established, and the above steps are repeated for iteration. When the preset conditions are met, the iteration loop is exited to obtain the final stage policy set. The fitness function generates a proportional value for all the distribution strategies in the final stage policy set, and the distribution strategy with the largest proportional value is selected and output as the current distribution strategy combination. The distribution system distributes rights and interests to members according to the current distribution strategy combination.

[0061] Initial generation module: A number of first-order distribution strategies are randomly generated based on the multi-dimensional information of the distribution strategy selection through the generation tool. Each first-order distribution strategy contains different equity distribution strategy weight vectors. For example, [0.3, 0.4, 0.3] is a first-order distribution strategy, which represents the weights based on member level, consumption behavior and interest preference, respectively. That is, in this first-order distribution strategy, the member level accounts for 0.3, the consumption behavior accounts for 0.4, and the interest preference accounts for 0.3. All first-order distribution strategies are used to establish the first-stage strategy set, which is sent to the calculation module;

[0062] Get the preset multi-dimensional information, including membership level, consumption behavior and interest preferences, and determine the weight range of each dimension, which is usually a real number between 0 and 1. For example, the weight range can be [0,1], indicating that the weight can take any value between 0 and 1. For each first-order distribution strategy, randomly generate a vector containing the weights of different equity distribution strategies. The length of the vector should match the dimension of the equity distribution strategy, and each element represents the weight of the corresponding dimension. For example, a weight vector similar to [0.3, 0.4, 0.3] can be randomly generated. As needed, repeat the above steps multiple times to generate multiple first-order distribution strategies. The number of generated can be determined according to the specific situation. Usually, enough samples are required for subsequent analysis and comparison. The generated first-order distribution strategy is saved for subsequent use. It can be saved as a data set or file to facilitate subsequent processing and analysis;

[0063] Suppose we choose three dimensions as the equity distribution strategy, which are based on membership level, consumption behavior and interest preference.

[0064] Assume that the weight range of each dimension is [0,1], which means that the weight can take any value between 0 and 1.

[0065] Randomly generate a weight vector:

[0066] The weight vector of the first first-order distribution strategy may be [0.3, 0.4, 0.3], which means that the weight based on membership level is 0.3, the weight based on consumption behavior is 0.4, and the weight based on interest preference is 0.3.

[0067] The weight vector of the second first-order distribution strategy may be [0.2, 0.5, 0.3], which means that the weight based on membership level is 0.2, the weight based on consumption behavior is 0.5, and the weight based on interest preference is 0.3.

[0068] The weight vector of the third first-order distribution strategy may be [0.4, 0.3, 0.3], which means that the weight based on membership level is 0.4, the weight based on consumption behavior is 0.3, and the weight based on interest preference is 0.3.

[0069] The above steps may be repeated multiple times as needed to generate more first-order distribution strategies. For example, a total of 10 first-order distribution strategies may be generated.

[0070] Save distribution strategy: Save the generated first-order distribution strategy, which can be saved as a data set or file for subsequent use.

[0071] Calculation module: Generates a proportional value for all first-order distribution strategies in the first-stage strategy set through the fitness function. The larger the proportional value, the better the historical benefits of the first-order distribution strategy. The proportional value is sent to the screening module and the vector exchange module.

[0072] Obtain the equity utilization rate, member floating range and strategy expenditure cost of the first-order distribution strategy, substitute the equity utilization rate, member floating range and strategy expenditure cost into the fitness function to calculate the proportion value of the first-order distribution strategy. The function expression is:

[0073]

[0074] ; In the formula, bl z is the proportional value, qlz is the equity utilization rate, qsz is the strategy expenditure cost, cyd is the member floating range, α, β, γ are the proportional coefficients of equity utilization rate, strategy expenditure cost, and member floating range, and α, β, γ are all greater than 0;

[0075] For each first-order distribution strategy in the first-stage strategy set, apply the defined fitness function to calculate its corresponding ratio value. The higher the ratio value, the better the distribution strategy, and the lower the ratio value, the worse the distribution strategy. The calculated ratio value is recorded for subsequent analysis and comparison. The ratio value can be associated with the corresponding first-order distribution strategy to form a ratio value-strategy mapping relationship table. For each first-order distribution strategy in the first-stage strategy set, the ratio value is calculated through the fitness function until the ratio values ​​of all first-order distribution strategies are calculated.

[0076] The calculation logic of the equity utilization rate is as follows: obtain the historical equity usage amount and the total equity amount of the first-order distribution strategy, and then obtain the equity utilization rate by dividing the historical equity usage amount by the total equity amount. The historical equity usage amount can be the equity amount actually used by the member in a certain period of time, and the total equity amount is the total equity amount provided to the member by the system. The larger the equity utilization rate of the first-order distribution strategy, the higher the use efficiency of the first-order distribution strategy, specifically:

[0077] Improved member satisfaction: If members use the benefits provided more frequently, it means that they are more satisfied with the benefits, believe that these benefits are valuable, and their loyalty to the company will also increase.

[0078] Increased member engagement: By increasing the utilization of benefits, members are more actively involved in benefit activities and their engagement increases, which means that they pay more attention to and participate in the company and may participate more in other activities or services related to the company.

[0079] Improved brand loyalty: Members’ high utilization of benefits is likely to enhance their trust and loyalty to the brand because they believe that the benefits they enjoy are part of the brand value, thereby deepening their emotional connection with the brand.

[0080] Increase sales and revenue: By increasing the utilization of benefits, members are more likely to purchase the company's products or services, thereby increasing sales and revenue.

[0081] Reduced churn rate: Increased member utilization may reduce member churn rate because they feel that membership benefits have real value to them and are more motivated to continue their relationship with the company.

[0082] The calculation logic of the member floating range is as follows: obtain the historical member click-through rate and historical member churn rate of the first-order distribution strategy, record the period when the historical member click-through rate exceeds the click-through rate threshold as the member active period, and record the period when the historical member churn rate exceeds the churn rate threshold as the member churn period. After performing integral calculations on the member active period and the member churn period, obtain the member floating range. The function expression is: H(t) is the change in member retention, [t x , t y ] is the member active period, [t i , t j ] is the member loss period. The larger the member floating amplitude of the first-order distribution strategy, the higher the member retention rate of the first-order distribution strategy, that is, the higher the use efficiency, specifically:

[0083] Enhance member loyalty: A high retention rate means that members are more inclined to maintain a long-term relationship with the company, indicating that they are more loyal to the company. Through an effective equity distribution strategy, members feel the care and value of the company, and are more willing to continue to maintain a cooperative relationship with the company.

[0084] Reduce customer churn costs: A high retention rate reduces the cost of customer churn, because the cost of retaining an existing customer is usually lower than the cost of attracting a new customer. Therefore, improving member retention rates can reduce a company's customer acquisition costs and improve efficiency.

[0085] Stable sales revenue: A high member retention rate means a stable customer base who will continue to purchase the company's products or services, thereby stabilizing the company's sales revenue.

[0086] Word-of-mouth and brand image: Loyal members tend to become loyal supporters of the brand and are likely to recommend the company’s products or services to their friends, family, and social networks, thereby increasing the company’s word-of-mouth and brand image.

[0087] Increase cross-selling and up-selling opportunities: Members with high retention rates have more trust in the company and are more willing to accept recommendations for other products or services from the company, thereby increasing cross-selling and up-selling opportunities.

[0088] The logic for obtaining the strategy expenditure cost is as follows: obtain the various equity costs, technology development and maintenance costs, and operating costs provided to members by the first-order distribution strategy, and add up the various equity costs, technology development and maintenance costs, and operating costs provided to members to obtain the strategy expenditure cost. The larger the strategy expenditure cost of the first-order distribution strategy, the higher the expenditure cost of the first-order distribution strategy, that is, the lower the use benefit, specifically:

[0089] Reduced ROI: If the first-order distribution strategy has high expenditure costs but fails to bring corresponding revenue growth, the return on investment (ROI) of the strategy may be reduced. This means that the company needs to invest more money to obtain the same benefits, thus reducing the effectiveness of the strategy.

[0090] Reduced effectiveness: A high-cost distribution strategy may result in reduced effectiveness because the company needs to spend more resources to implement the strategy but may fail to achieve the expected results, or the results may not reach the expected level of cost.

[0091] Reduced profitability: If the expenditure costs of the first-order distribution strategy are too high, the company may face the risk of reduced profit margins. High-cost strategies may squeeze the company's profit margins, thereby reducing the company's profitability.

[0092] Reduced ROI: A high-cost distribution strategy may extend the ROI because it takes longer to recoup the investment. This may reduce the company's willingness to invest in the strategy, thus affecting the effectiveness of the strategy.

[0093] Waste of resources: If the cost of a strategy is too high, it may lead to a waste of resources, including manpower, money, time, etc. These resources could have been used for more effective strategies or other business development areas.

[0094] Screening module: uses a screening algorithm to select a corresponding number of first-order distribution strategies from the first-stage strategy set as second-order distribution strategies, establishes a second-stage strategy set from all second-order distribution strategies, and sends the second-stage strategy set to the vector exchange module;

[0095] Among them, the screening algorithms include roulette wheel selection method and threshold comparison selection method;

[0096] Roulette selection method:

[0097] Get the proportion values ​​of all first-order distribution strategies in the first-stage strategy set, add up the proportion values ​​of all first-order distribution strategies to get the total proportion value, calculate the selection probability of each first-order distribution strategy, and get the selection probability by dividing the proportion value by the total proportion value. Map the selection probability to the roulette wheel to construct a selection range. For example, if the selection probability of the first first-order distribution strategy is 0.2, then the corresponding selection range of the first first-order distribution strategy on the roulette wheel is 0 to 0.2. If the selection probability of the second first-order distribution strategy is 0.8, then the corresponding selection range of the second first-order distribution strategy on the roulette wheel is 0.2 to 1.0. By analogy, the greater the selection probability of the first-order distribution strategy, the larger the range of the first-order distribution strategy on the roulette wheel. A random number is generated by a random generation tool, and the random number is truncated to one decimal place. For example, if the generated random number is 0.7, the second first-order distribution strategy is selected. By analogy, random numbers are repeatedly generated to select first-order distribution strategies. When the number of first-order distribution strategies exceeds the quantity threshold, the roulette selection method is stopped, and the selected first-order distribution strategy is used as the second-order distribution strategy. The roulette selection method increases randomness, which can effectively improve the adaptability and robustness of the present application.

[0098] Threshold comparison selection method:

[0099] The proportion values ​​of all first-order distribution strategies in the first-stage strategy set are obtained, and the proportion values ​​of the first-order distribution strategies are compared with the preset first proportion threshold value. The first proportion threshold value is used to determine whether the benefits of the first-order distribution strategies are qualified. If the proportion value of the first-order distribution strategy is less than the first proportion threshold value, the first-order distribution strategy is screened out. If the proportion value of the first-order distribution strategy is greater than or equal to the first proportion threshold value, the first-order distribution strategy is selected. If the number of selected first-order distribution strategies is less than or equal to the quantity threshold value, the selected first-order distribution strategy is used as the second-order distribution strategy. If the number of selected first-order distribution strategies is greater than the quantity threshold value, the selected first-order distribution strategies are sorted from large to small by proportion value. After screening out the first-order distribution strategies after the quantity threshold value, the selected first-order distribution strategy is used as the second-order distribution strategy. The threshold comparison selection method is beneficial to improving accuracy.

[0100] Vector exchange module: randomly exchange the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set. The second-order distribution strategies that have exchanged strategy weight vectors will no longer exchange strategy weight vectors. Priority will be given to exchanging strategy weight vectors for the second-order distribution strategies with large proportion values. If there is a single second-order distribution strategy that has not been exchanged, the second-order distribution strategy that has not been exchanged will be screened out, and the second-order distribution strategy that has completed the strategy weight vector exchange will be updated to the third-order distribution strategy. All third-order distribution strategies will establish a third-stage strategy set, and the third-stage strategy set will be sent to the vector mutation module;

[0101] Randomly exchange strategy weight vectors: For each pair of second-order distribution strategies in the second-stage strategy set, randomly select a strategy and exchange its weight vector. The exchanged weight vector will be used as the new distribution strategy.

[0102] Record the exchange status: For the second-order distribution strategies that have been exchanged, record them so that you can determine whether to continue exchanging them later.

[0103] Prioritize exchanging the second-order distribution strategies with larger proportion values: If there is a single second-order distribution strategy that has not been exchanged, then prioritize exchanging the weight vectors of the second-order distribution strategies with larger proportion values.

[0104] Update the exchanged strategy: For the second-order distribution strategy that has undergone weight vector exchange, update it to a third-order distribution strategy.

[0105] Screen out unexchanged strategies: If there is a single second-order distribution strategy that is not exchanged, screen out the strategy and no longer consider it as a third-order distribution strategy.

[0106] Establish the third-stage strategy set: exchange all completed strategy weight vectors and update them to the strategy of the third-order distribution strategy as the third-stage strategy set.

[0107] To better illustrate the above solution, an example is given below:

[0108] Assume that there is a second-stage policy set that includes the following second-order distribution policies:

[0109] Initial Phase 2 policy set:

[0110] Strategy A: [0.2, 0.5, 0.3];

[0111] Strategy B: [0.5, 0.2, 0.3];

[0112] Strategy C: [0.4, 0.3, 0.3];

[0113] Strategy D: [0.1, 0.6, 0.3];

[0114] Randomly swap the policy weight vector:

[0115] First, we randomly swap the strategy weight vectors of the second-order distribution strategies with large ratio values. Strategies with large ratio values ​​are more likely to be better strategies, so we want to swap them first.

[0116] Suppose we choose to randomly swap the weight vectors of strategies A and B. The strategies after the swap are as follows:

[0117] Strategy A: [0.2, 0.3, 0.5];

[0118] Strategy B: [0.3, 0.5, 0.2];

[0119] Updated to a three-stage distribution strategy:

[0120] The second-order distribution strategy that completes the exchange of the strategy weight vector is updated to the third-order distribution strategy. The updated third-order distribution strategy forms the third-stage strategy set.

[0121] Vector mutation module: The strategy weight vectors of all third-order distribution strategies in the third-stage strategy set are randomly mutated through the modification tool. In this application, the strategy weight vectors of the second-order distribution strategies are randomly exchanged, and the strategy weight vectors of the third-order distribution strategies are randomly mutated, which is conducive to improving the robustness of this application and the adaptability to different scenarios. The third-order distribution strategy that has completed the random mutation of the strategy weight vector is updated to a fourth-order distribution strategy, and all fourth-order distribution strategies are used to establish a fourth-stage strategy set;

[0122] Random mutation of the third-order distribution strategy: For each third-order distribution strategy in the third-stage strategy set, the strategy weight vector is randomly mutated through the modification tool. The mutation method can include increasing or decreasing the weight value, randomly replacing the weight value, and randomly perturbing the weight value.

[0123] Update the mutated strategy: Update the third-order distribution strategy that has undergone random mutation of the strategy weight vector to a fourth-order distribution strategy. This can introduce a certain degree of randomness and variation, increase the diversity of the algorithm search space, and help avoid falling into a local optimal solution.

[0124] Establish the fourth-stage strategy set: All strategies that have undergone random mutation of the strategy weight vector and updated to the fourth-order distribution strategy are used as the fourth-stage strategy set.

[0125] To better illustrate the above solution, an example is given below:

[0126] There is a third-stage strategy set, which contains three-order distribution strategies, each of which consists of a weight vector. Now we need to randomly mutate these strategies through the vector mutation module and update the mutated strategies to the fourth-order distribution strategy.

[0127] Initial Phase 3 policy set:

[0128] Suppose we have the following three three-stage distribution strategies:

[0129] Strategy 1: [0.2, 0.5, 0.3];

[0130] Strategy 2: [0.3, 0.4, 0.3];

[0131] Strategy 3: [0.4, 0.3, 0.3];

[0132] Random mutation:

[0133] Through the vector mutation module, we randomly mutate each third-order distribution strategy. For example, we can choose to randomly perturb or replace each weight value.

[0134] Assume that after mutation, we get the following mutated strategy:

[0135] Mutated strategy 1: [0.2, 0.55, 0.25];

[0136] Mutated strategy 2: [0.35, 0.45, 0.2];

[0137] Mutated strategy 3: [0.45, 0.25, 0.3];

[0138] Establish the fourth-stage strategy set: The updated four-stage distribution strategies are combined into the fourth-stage strategy set.

[0139] Loop iteration module: repeats the steps of screening module, vector exchange module and vector mutation module to iterate. When the preset conditions are met, the iteration loop is jumped out to obtain the final stage strategy set, which is sent to the strategy output module.

[0140] Repeat the steps of screening module, vector exchange module and vector mutation module to iterate, continuously obtain new distribution strategies, and compare the number of iterations with the number threshold. The number threshold is used to determine whether it is necessary to jump out of the iteration loop to avoid too many iterations of the control system and increase the operating burden of the control system. When the number of iterations exceeds the preset number threshold, jump out of the iteration loop and obtain all new distribution strategies to establish the final stage strategy set.

[0141] Strategy output module: Generates a proportional value for all distribution strategies in the final stage strategy set through the fitness function, selects the distribution strategy with the largest proportional value and outputs it as the current distribution strategy combination. The distribution system distributes benefits to members based on the current distribution strategy combination.

[0142] Select the maximum proportion value strategy: Select the largest one from the generated proportion values ​​as the current distribution strategy combination. This means selecting the distribution strategy with the best fitness as the final equity distribution strategy.

[0143] Output the current distribution strategy combination: The distribution strategy corresponding to the selected maximum ratio value is used as the current distribution strategy combination and output to the distribution system for use.

[0144] Benefit distribution: The distribution system distributes benefits to members according to the current distribution strategy combination. The specific distribution process is carried out according to the selected distribution strategy combination, which may involve the analysis of member level, consumption behavior, interest preferences and other information and the distribution of benefits.

[0145] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0146] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0147] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A multi-dimensional membership rights distribution system, characterized by: It includes initial generation module, calculation module, screening module, vector exchange module, vector mutation module, loop iteration module and strategy output module; Initial generation module: The generation tool randomly generates several first-order distribution strategies based on the multi-dimensional information of the distribution strategy selection. Each first-order distribution strategy contains different equity distribution strategy weight vectors, and all first-order distribution strategies are established into the first-stage strategy set; Calculation module: Generates proportional values ​​for all first-order distribution strategies in the first-stage strategy set through the fitness function; Screening module: Use the screening algorithm to select a corresponding number of first-order distribution strategies from the first-stage strategy set as the second-order distribution strategies, and use all the second-order distribution strategies to establish the second-stage strategy set; Vector exchange module: randomly exchange the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set, update the second-order distribution strategies that have completed the strategy weight vector exchange to the third-order distribution strategies, and establish the third-stage strategy set with all the third-order distribution strategies; Vector mutation module: Use the modification tool to randomly mutate the strategy weight vectors of all third-order distribution strategies in the third-stage strategy set, and update the third-order distribution strategy that has completed the random mutation of the strategy weight vector to the fourth-order distribution strategy; Loop Iteration Module: Repeat the steps of the screening module, vector exchange module, and vector mutation module to iterate. When the preset conditions are met, the iteration loop is jumped out to obtain the final stage strategy set. Strategy output module: Generates a proportional value for all distribution strategies in the final stage strategy set through the fitness function, selects the distribution strategy with the largest proportional value and outputs it as the current distribution strategy combination, and distributes benefits to members based on the current distribution strategy combination; The initial generation module obtains preset multi-dimensional information, which includes membership level, consumption behavior and interest preference; For each first-order distribution strategy, randomly generate a vector containing the weights of different equity distribution strategies. The length of the vector matches the dimension of the equity distribution strategy, and each element represents the weight of the corresponding dimension. After generating multiple first-order distribution strategies, the generated first-order distribution strategies are saved and a first-stage strategy set is established; The calculation module obtains the equity utilization rate, member floating amplitude and strategy expenditure cost of the first-order distribution strategy, substitutes the equity utilization rate, member floating amplitude and strategy expenditure cost into the fitness function to calculate and obtain the proportion value of the first-order distribution strategy. The function expression is: ; In the formula, is the ratio value, is the equity utilization rate, Cost of strategy, The floating range for members, , , are the proportional coefficients of equity utilization rate, strategy expenditure cost, and member floating amplitude, respectively, and , , All are greater than 0; For each first-order distribution strategy in the first-stage strategy set, the corresponding proportion value is calculated through the fitness function, and the proportion value is associated with the corresponding first-order distribution strategy to form a mapping relationship table of proportion value-strategy.

2. A multi-dimensional member benefits distribution system according to claim 1, characterized in that: The calculation logic of the member floating amplitude is as follows: obtain the historical member click-through rate and historical member churn rate of the first-order distribution strategy, record the period when the historical member click-through rate exceeds the click-through rate threshold as the member active period, and record the period when the historical member churn rate exceeds the churn rate threshold as the member churn period, and after performing integral calculation on the member active period and the member churn period, obtain the member floating amplitude. The function expression is: , Retain changes for members, This is the member active period. This is the period of member loss.

3. A multi-dimensional member benefits distribution system according to claim 2, characterized in that: The calculation logic of the equity utilization rate is as follows: obtain the historical equity usage amount and the total equity amount of the first-order distribution strategy, and obtain the equity utilization rate by dividing the historical equity usage amount by the total equity amount; The logic for obtaining the strategy expenditure cost is: obtain the various equity costs, technology development and maintenance costs, and operating costs provided to members by the first-order distribution strategy, and add up the various equity costs, technology development and maintenance costs, and operating costs provided to members to obtain the strategy expenditure cost.

4. A multi-dimensional member benefits distribution system according to claim 3, characterized in that: The screening module obtains the proportion values ​​of all first-order distribution strategies in the first-stage strategy set, and adds the proportion values ​​of all first-order distribution strategies to obtain the sum of the proportion values, calculates the selection probability of each first-order distribution strategy, and obtains the selection probability by dividing the proportion value by the sum of the proportion values, maps the selection probability to the roulette wheel, and constructs the selection range; Generate a random number through a random generation tool, truncate the random number to one decimal place, select a first-order distribution strategy based on the value range of the random number, and repeatedly generate random numbers to select a first-order distribution strategy; When the number of first-order distribution strategies exceeds the quantity threshold, the roulette wheel selection method is stopped, and the selected first-order distribution strategy is used as the second-order distribution strategy.

5. A multi-dimensional member benefits distribution system according to claim 4, characterized in that: The screening module obtains the ratio values ​​of all first-order distribution strategies in the first-stage strategy set, and compares the ratio values ​​of the first-order distribution strategies with a preset first ratio threshold value, where the first ratio threshold value is used to determine whether the benefits of the first-order distribution strategies are qualified; If the ratio value of the first-order distribution strategy is less than the first ratio threshold, the first-order distribution strategy is screened out; if the ratio value of the first-order distribution strategy is greater than or equal to the first ratio threshold, the first-order distribution strategy is selected; If the number of selected first-order distribution strategies is less than or equal to the quantity threshold, the selected first-order distribution strategy is used as the second-order distribution strategy; If the number of selected first-order distribution strategies is greater than the quantity threshold, the selected first-order distribution strategies are sorted from large to small by the proportion value, and after the first-order distribution strategies after the quantity threshold are screened out, the selected first-order distribution strategies are used as the second-order distribution strategies.

6. A multi-dimensional member rights distribution system according to claim 1, characterized in that: The loop iteration module repeats the steps of the screening module, the vector exchange module, and the vector mutation module to iterate, continuously obtains new distribution strategies, and compares the number of iterations with the number threshold. The number threshold is used to determine whether it is necessary to jump out of the iteration loop. When the number of iterations exceeds the preset number threshold, the iteration loop is jumped out, and all new distribution strategies are obtained to establish a final stage strategy set.

7. A multi-dimensional member benefits distribution system according to claim 6, characterized in that: The vector exchange module randomly exchanges the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set in pairs. The second-order distribution strategies that have exchanged strategy weight vectors no longer exchange strategy weight vectors. If there is a single second-order distribution strategy that has not been exchanged, the second-order distribution strategy that has not been exchanged will be screened out.

8. A multi-dimensional member benefits distribution method, used to implement the distribution system according to any one of claims 1 to 7, characterized in that: The distribution method comprises the following steps: S1: The distribution system randomly generates several first-order distribution strategies based on the multi-dimensional information of the distribution strategy selection through the generation tool. Each first-order distribution strategy contains different equity distribution strategy weight vectors. All first-order distribution strategies are established into the first-stage strategy set S2: Generate a proportional value for all first-order distribution strategies in the first-stage strategy set through the fitness function, and use the screening algorithm to select a corresponding number of first-order distribution strategies from the first-stage strategy set as second-order distribution strategies, and establish a second-stage strategy set with all second-order distribution strategies; S3: Randomly exchange the strategy weight vectors of the second-order distribution strategies in the second-stage strategy set, update the second-order distribution strategies that have completed the strategy weight vector exchange to third-order distribution strategies, and establish the third-stage strategy set with all third-order distribution strategies; S4: Perform random mutation of the strategy weight vectors of all third-order distribution strategies in the third-stage strategy set through the modification tool, update the third-order distribution strategy that has completed the random mutation of the strategy weight vector to the fourth-order distribution strategy, and establish the fourth-stage strategy set with all the fourth-order distribution strategies; S5: Repeat steps S2, S3, and S4 to iterate. When the preset conditions are met, the iteration loop is exited to obtain the final stage strategy set. S6: Generate a proportional value for all distribution strategies in the final stage strategy set through the fitness function, select the distribution strategy with the largest proportional value and output it as the current distribution strategy combination. The distribution system distributes rights and interests to members based on the current distribution strategy combination.

Images