Multi-target-based quota adjustment optimal allocation method and device, equipment and medium

Abstract

The application discloses a credit adjustment optimal distribution method and device based on multiple targets, equipment and medium, and the method comprises the following steps: obtaining related data to be raised, and pre-processing the related data to form customer wide table information; the related data is product level data and customer level data to be used; according to the customer wide table information, a credit exposure distribution model is constructed based on multiple credit raising targets; the credit exposure distribution model is optimized and solved by using a genetic algorithm to determine the credit raising constraint conditions, and the optimal solution meeting the credit raising target and the credit raising constraint condition is obtained, and the optimal solution is used as the adjustment ratio of the credit exposure, so that the optimal distribution of the credit exposure under multiple targets is realized. The credit adjustment optimal distribution method based on multiple targets can meet the credit exposure distribution under different scenes, the credit exposure adjustment is accurate, and the customer use and customer activation can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a quota adjustment optimal allocation method and device based on multiple targets, equipment and medium. BACKGROUND

[0002] With the rapid development of the financial industry, the cost and difficulty of banks to acquire new customers are becoming higher and higher, so the stock customer operation is becoming more and more important in the daily work of the bank, especially through the promotion of quota to maintain and activate the customer. Usually, for different customer operation scenes, the ultimate goal to be achieved through the promotion of quota is also different.

[0003] The existing quota adjustment method mainly allocates the quota according to the risk target of the customer. The main implementation way is to give the customer rating based on the related model (expert scoring model, risk model), and allocate more quota to the customer with good risk rating and low risk, and allocate less quota to the customer with poor risk rating and high risk.

[0004] The above existing technology only considers a single target, and only adjusts the quota according to the risk rating of the customer, and allocates more quota to the customer with good risk rating and low risk, and allocates less quota to the customer with poor risk rating and high risk, which is not conducive to improving customer use and activating customers. In the fierce market competition, equal attention should be paid to the risk target and the use target, the customer activation target, the improvement of customer experience and the financial service ability.

[0005] Therefore, the present application is proposed. SUMMARY

[0006] The technical problem to be solved by the present application is that the existing quota adjustment method only considers a single target, and only adjusts the quota according to the risk rating of the customer, and allocates more quota to the customer with good risk rating and low risk, and allocates less quota to the customer with poor risk rating and high risk, resulting in inaccurate quota adjustment, which is not conducive to improving customer use and activating customers.

[0007] The present application is based on how to reasonably and effectively allocate the quota under multiple targets. The present application aims to provide a quota adjustment optimal allocation method and device based on multiple targets, equipment and medium, which improves the method of adjusting the quota only according to a single target or expert experience. The present application starts from the multi-dimensional angle of bank customer operation, and the allocation of the quota not only realizes the risk target, but also realizes the customer use target, the customer activation target and the like. The quota adjustment optimal allocation method based on multiple targets of the present application meets the quota exposure allocation under different scenes, and the quota exposure adjustment is accurate, which is conducive to improving customer use and activating customers.

[0008] The present application is realized by the following technical scheme:

[0009] In a first aspect, the present application provides a multi-target-based optimal allocation method for quota adjustment, which comprises:

[0010] Obtaining related data to be raised and pre-processing the related data to form a customer wide table information; the related data is product level data and customer level data to be used for raising;

[0011] Constructing a raising exposure allocation model based on multi-raising targets according to the customer wide table information;

[0012] Determining raising constraint conditions, using a genetic algorithm to optimize the raising exposure allocation model, obtaining an optimal solution satisfying the raising target and the raising constraint conditions, and taking the optimal solution as a raising exposure adjustment ratio to realize optimal allocation of the raising exposure under multi-targets.

[0013] Further, the related data includes product-related information and customer basic information, and the product-related information includes customer risk information, customer credit information and customer life cycle information.

[0014] Further, the customer risk information includes repayment behavior performance information, and the repayment behavior performance information includes historical overdue situation, current overdue situation, early settlement situation, recent application situation and recent credit usage situation, etc.

[0015] The customer credit information includes loan basic information and customer's in-and-out-of-line historical credit usage records, and the loan basic information includes loan product name, loan product type, loan guarantee type, credit time, credit time from now, credit amount and interest rate; the customer's in-and-out-of-line historical credit usage records include credit usage number, settlement and unsettled number, credit usage time and credit usage amount proportion of credit quota, etc.

[0016] The customer life cycle information includes customer behavior information, and the customer behavior information includes borrowing habit, active loss probability, borrowing demand probability, historical raising frequency and historical raising amplitude.

[0017] The customer basic information includes user portrait information.

[0018] Further, the pre-processing of the related data includes:

[0019] By data cleaning and feature mining technology, the product level data is aggregated into the customer level data;

[0020] The customer level data is subjected to data conversion and data anomaly processing to form the customer wide table information;

[0021] The data anomaly processing includes missing value and abnormal value processing.

[0022] Further, according to the customer wide table information, a quota increase opening distribution model is constructed based on multiple quota increase targets, including:

[0023] The product level data is used for customer stratification, customer use rates and overdue rates of different customer groups under different quota increase magnitudes are determined according to historical quota increase data, and multiple quota increase targets are determined; the quota increase targets include maximum customer use rates after quota increase and minimum quota increase risks;

[0024] According to the multiple quota increase targets, a quota increase opening distribution model is constructed; the quota increase opening distribution model is a model formed by maximum customer use rates a 11 x 11 +a 12 x 12 +…+a mn x mn and minimum quota increase risks e 11 x 11 +e 12 x 12 +…+e mn x mn ; wherein, x 11 , x 12 , …, x mn are customer proportions under different customer groups and quota increase magnitudes; a 11 , a 12 , …, a mn are use rates under different customer groups and quota increase magnitudes, e 11 , e 12 , …, e mn are overdue rates under different customer groups and quota increase magnitudes.

[0025] Further, the quota increase constraint conditions include a first constraint condition and a second constraint condition;

[0026] The first constraint condition is a constraint that a customer proportion with a quota increase magnitude in a range g is not less than G, and the formula is: x 11 +x 12 +…+x g1g2 ≥G;

[0027] The second constraint condition is that a proportion of quota increase customers of a certain customer group k is not more than K, and the formula is:

[0028] x 11 +x 12 +…+x k1k2 ≤K

[0029] wherein, g is a quota increase magnitude preset value, G is a customer proportion preset value corresponding to the quota increase magnitude g, k is a customer group category limit preset value, K is a customer proportion preset value corresponding to the customer group category k, x g1g2 is a customer group group under a quota increase magnitude less than or equal to g, and x k1k2Customer grouping group for customer group k.

[0030] Further, the genetic algorithm is used to solve the optimization of the quota opening distribution model, including:

[0031] By setting the initial solution X 11 ,X 12 ,…,X mn , the mutation and crossover of the initial solution are carried out to obtain another group of new candidate solutions X' 11 ,X' 12 ,…,X' mn , the other candidate solutions are screened out by calculating the fitness of the candidate solutions, the mutation and crossover steps are repeatedly carried out, and it is judged whether the quota target and the quota constraint condition are met, so that the optimal solution is obtained.

[0032] In the second aspect, the application further provides a quota adjustment optimal distribution device based on multiple targets, which comprises:

[0033] An acquisition unit is configured to acquire related data to be adjusted, wherein the related data is product level data and customer level data to be used by the data to be adjusted;

[0034] A preprocessing unit is configured to preprocess the related data to form customer wide table information;

[0035] A quota opening distribution model construction unit is configured to construct a quota opening distribution model based on multiple quota targets according to the customer wide table information;

[0036] A quota constraint condition determination unit is configured to determine a quota constraint condition;

[0037] An optimal solution unit is configured to use a genetic algorithm to solve the optimization of the quota opening distribution model, so as to obtain an optimal solution meeting the quota target and the quota constraint condition, and the optimal solution is used as an adjustment ratio of the quota opening.

[0038] In the third aspect, the application further provides a computer device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the optimal distribution method based on multiple targets when executing the computer program.

[0039] In the fourth aspect, the application further provides a computer readable storage medium, which stores a computer program, and the computer program is executable on the processor to implement the optimal distribution method based on multiple targets.

[0040] Compared with the prior art, the application has the following advantages and beneficial effects:

[0041] This invention presents a multi-objective method, apparatus, device, and medium for optimal allocation of credit limit adjustments. It proposes an optimal configuration of credit limit increase exposure based on multiple objectives, overcoming the limitations of a single-objective approach that restricts the scope of credit limit increases. This approach is more applicable and provides precise credit limit adjustment, facilitating increased customer spending and customer engagement. Furthermore, it utilizes a genetic algorithm to solve for the optimal allocation of credit limit increase exposure, overcoming the shortcomings of allocating exposure based on expert experience. This invention addresses a wider range of application scenarios, resulting in superior and more scientific outcomes. Attached Figure Description

[0042] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0043] Figure 1 This is a flowchart of an optimal allocation method for quota adjustment based on multiple objectives according to the present invention.

[0044] Figure 2 This is a schematic diagram of the structure of an optimal allocation device for quota adjustment based on multiple objectives according to the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of this invention are only for explaining this invention and are not intended to limit this invention.

[0046] The existing credit limit adjustment method only considers a single objective and adjusts the credit limit based solely on the customer's risk rating. Through calculation, more credit limits are allocated to customers with good and low risk ratings, while less credit limits are allocated to customers with poor and high risk ratings. This results in inaccurate credit limit adjustments, which is not conducive to improving customer spending and activating customers.

[0047] Given the multi-objective nature of credit limit adjustments, the rational and effective allocation of credit limits is crucial. Therefore, this invention provides a method, apparatus, device, and medium for optimal allocation of credit limit adjustments based on multiple objectives. This improves upon methods that rely solely on a single objective or expert experience for credit limit adjustments. From a multi-dimensional perspective of bank customer operations, the allocation of credit limit increases must not only achieve risk objectives but also customer spending and activation objectives. This invention's multi-objective optimal allocation method for credit limit adjustments satisfies the needs of credit limit allocation in different scenarios, and its precise adjustment of credit limit exposure facilitates increased customer spending and customer activation.

[0048] Example 1

[0049] like Figure 1 As shown, the present invention provides a multi-objective-based optimal allocation method for credit limit adjustment, the method comprising:

[0050] Step 1, obtain the relevant data to be raised and pre-process the relevant data to form a customer wide table information; the relevant data is product level data and customer level data needed for the to-be-raised;

[0051] Among them, the relevant data includes product-related information and customer basic information, and the product-related information includes customer risk information, customer credit information and customer life cycle information.

[0052] Specifically, the customer risk information includes repayment behavior performance information, and the repayment behavior performance information includes historical overdue situation, current overdue situation, early settlement situation, recent application situation and recent loan situation, etc.

[0053] The customer credit information includes loan basic information and customer's in-and-out-of-line historical loan records, and the loan basic information includes loan product name, loan product type, loan guarantee type, credit time, loan credit time, credit amount and interest rate; the customer's in-and-out-of-line historical loan records include the number of loan pens, the number of cleared and unclosed pens, the loan time and the loan amount proportion of the credit limit, etc.

[0054] The customer life cycle information includes customer behavior information, and the customer behavior information includes borrowing habits, active loss probability, borrowing demand probability, historical number of times of raising and historical raising amplitude.

[0055] The customer basic information includes user portrait information.

[0056] Specifically, the pre-processing of the relevant data includes:

[0057] Through data cleaning and feature mining technology, the product level data is aggregated to the customer level data; this is because the obtained relevant data includes customer level data, as well as loan, account, product level, credit habit and customer life cycle data, and since the analysis object is the customer, the loan and other product level data need to be aggregated to the customer level through data cleaning and feature mining;

[0058] The customer level data is subjected to data conversion and data anomaly processing (missing value and abnormal value processing) to form a customer wide table information; and data verification and exploratory analysis are performed to have a preliminary understanding of the data.

[0059] Step 2, based on the customer wide table information, a raising exposure allocation model is constructed based on multiple raising targets; including:

[0060] Step 21, use the product level data for customer stratification, determine the customer use rate and overdue rate (see Tables 2 and 3) of different customer groups under different raising amplitudes (see Table 1) according to historical raising data, and determine multiple raising targets; the raising targets include the maximum customer use rate after raising and the minimum raising risk;

[0061] Table 1: Distribution of the number of people in each group

[0062] Customer type Rate-up band 1 Rate-up band 2 Rate-up band 3 … Rate-up band n Segment 1 x 11 ]]> x 12 ]]> x 13 ]]> … x 1n ]]> Segment 2 x 21 ]]> x 22 ]]> x 23 ]]> … x 2n ]]> Segment 3 x 31 ]]> x 32 ]]> x 33 ]]> … x 3n ]]> … … … … … … Segment m x m1 ]]> x m2 ]]> x m3 ]]> … x mn ]]>

[0063] Table 2: Utilization rate information

[0064] Customer type Rate-up band 1 Rate-up band 2 Rate-up band 3 … Rate-up band n Segment 1 a 11 ]]> a 12 ]]> a 13 ]] … a 1n ]]> Segment 2 a 21 ]]> a 22 ]]> a 23 ]]> … a 2n ]]> Segment 3 a 31 ]]> a 32 ]]> a 33 ]]> … a 3n ]]> … … … … … … Segment m a m1 ]]> a m2 ]]> a m3 ]]> … a mn ]]>

[0065] Table 3: Overdue rate information

[0066]

[0067]

[0068] Step 22: According to multiple up-rating targets, an up-rating exposure allocation model is constructed. The up-rating exposure allocation model is formed by maximizing the customer utilization rate Max: a 11 x 11 +a 12 x 12 +…+a mn x mn and minimizing the up-rating risk Min: e 11 x 11 +e 12 x 12 +…+e mn x mn ; wherein x 11 , x 12 , …, x mn are the proportions of customers under different customer groups and up-rating magnitudes;

[0069] a 11 , a 12 , …, a mn are the utilization rates under different customer groups and up-rating magnitudes, and e 11 , e 12 , …, e mn are the overdue rates under different customer groups and up-rating magnitudes.

[0070] Step 3: Determine the up-rating constraint conditions, use a genetic algorithm to optimize the up-rating exposure allocation model, obtain the optimal solution that satisfies the up-rating targets and the up-rating constraint conditions, and take the optimal solution as the adjustment proportion of the up-rating exposure.

[0071] Specifically, the up-rating constraint conditions include a first constraint condition and a second constraint condition.

[0072] The first constraint condition is that the proportion of customers with up-rating magnitudes within the range g is not less than G, and the formula is: x 11 +x 12 +…+x g1g2 ≥G.

[0073] The second constraint condition is that the proportion of customers of a customer group k is not more than K, and the formula is:

[0074] x 11 +x 12 +…+x k1k2 ≤K

[0075] wherein g is a preset value of the increase range, G is a preset value of the customer proportion corresponding to the increase range g, k is a preset value of the customer group category limit, K is a preset value of the customer proportion corresponding to the customer group category k, x g1g2 is a customer grouping category with an increase range less than or equal to g, and x k1k2 is a customer grouping category of the customer group k.

[0076] In a specific implementation, the constraint condition of the increase is determined according to the actual demand and the effect to be achieved. For example:

[0077] The first constraint condition is that the customers with the increase range of range 1 and range 2 are not less than 60% of the total customer proportion

[0078] x 11 +x 12 +x 21 +x 22 +x 31 +x 32 +…+x m1 +x m2 ≥0.6

[0079] The second constraint condition is that the proportions of customer group 1 and customer group 2 are not more than 50% of the total customer proportion

[0080] x 11 +x 12 +x 13 +…+x 1n +x 21 +x 22 +x 23 +…+x 2n ≤0.5

[0081] x 11 , x 12 , …, x mn are the customer proportions under different customer groups and increase ranges.

[0082] Specifically, the genetic algorithm is used for optimization solution according to the increase data, the usage rate data and the default rate data of the historical customers. The genetic algorithm is a random global search optimization method, which simulates the phenomena of copying, crossing and variation occurring in natural selection and heredity. Starting from any initial population, through random selection, crossing and variation operations, through continuous evolution, the optimal solution of the problem is finally obtained.

[0083] The genetic algorithm is used to solve the optimal solution of the quota opening distribution model, including:

[0084] By setting the initial solution X 11 ,X 12 ,…,X mn , the mutation and crossover of the initial solution are carried out to obtain another group of new candidate solutions X' 11 ,X' 12 ,…,X' mn , the other candidate solutions are screened by calculating the fitness of the candidate solutions, the mutation and crossover steps are repeatedly repeated, and it is judged whether the quota target and the quota constraint condition are met, so as to obtain the optimal solution.

[0085] The main advantage of the present application is to propose an optimal configuration of the quota opening based on multiple targets, which overcomes the single deficiency of the quota scene under only one target, is more applicable, and the quota opening adjustment is accurate, which is beneficial to improve customer use and promote customer activation. In addition, the genetic algorithm is used to solve the optimal distribution of the quota opening, which overcomes the deficiency of the opening distribution according to the expert experience. The use scene of the present application is more relaxed, and the final result is more optimal and scientific.

[0086] Embodiment 2

[0087] As shown in Figure 2 , the difference between the present embodiment and embodiment 1 is that the present embodiment provides an optimal distribution device of quota adjustment based on multiple targets, which is used to realize an optimal distribution method of quota adjustment based on multiple targets; the device comprises:

[0088] An acquisition unit is configured to acquire related data to be raised, wherein the related data comprises product level data and customer level data to be used for raising;

[0089] A preprocessing unit is configured to preprocess the related data to form customer wide table information;

[0090] A quota opening distribution model construction unit is configured to construct a quota opening distribution model based on multiple quota targets according to the customer wide table information;

[0091] A quota constraint condition determination unit is configured to determine a quota constraint condition;

[0092] An optimal solution unit is configured to use a genetic algorithm to solve the optimal solution of the quota opening distribution model, obtain an optimal solution meeting the quota target and the quota constraint condition, and use the optimal solution as an adjustment ratio of the quota opening.

[0093] The execution process of each unit is executed according to the process steps of the optimal distribution method of the quota adjustment based on multiple targets described in embodiment 1, and the detailed description of this embodiment is not repeated.

[0094] Meanwhile, the application further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the multi-target based optimal allocation method of credit adjustment when executing the computer program.

[0095] Meanwhile, the application further provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executable on a processor to implement the multi-target based optimal allocation method of credit adjustment.

[0096] Those skilled in the art should understand that embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0097] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus generate a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the functions specified in the flowcharts and / or block diagrams.

[0098] These computer program instructions can also be stored in a computer-readable memory that can direct the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction means, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for carrying out the functions specified in the flowcharts and / or block diagrams.

[0099] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1one or more processes and / or blocks Figure 1 the steps of the functions specified in the one or more blocks.

[0100] The above detailed description has further explained the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A multi-objective-based optimal allocation method for quota adjustment, characterized in that, The method includes: Obtain relevant data for the credit limit increase and preprocess the relevant data to form a wide table of customer information; the relevant data includes product-level data and customer-level data required for the credit limit increase. Based on the customer wide table information, a credit limit increase exposure allocation model is constructed based on multiple credit limit increase targets; The constraints for increasing the credit limit are determined, and the genetic algorithm is used to optimize the credit limit exposure allocation model to obtain the optimal solution that satisfies the credit limit increase target and the credit limit increase constraints. The optimal solution is then used as the adjustment ratio for the credit limit exposure to achieve the optimal allocation of the credit limit exposure under multiple objectives. The relevant data includes product-related information and customer basic information. The product-related information includes customer risk information, customer credit information, and customer lifecycle information. The customer risk information includes repayment behavior performance information, which includes historical overdue status, current overdue status, early settlement status, recent application status, and recent fund usage status. Customer credit information includes basic loan information and the customer's historical loan usage records both within and outside the bank. The basic loan information includes the loan product name, loan product type, loan guarantee type, credit period, loan credit duration, credit amount, and interest rate. The customer's historical loan usage records both within and outside the bank include the number of loan transactions, the number of settled and outstanding transactions, the loan time, and the proportion of the loan amount to the credit limit. Customer lifecycle information includes customer behavior information, which includes borrowing habits, probability of active churn, probability of borrowing demand, number of historical credit limit increases, and historical credit limit increase magnitude. Customer basic information includes user profile information; Based on the customer broad table information, a credit limit increase exposure allocation model is constructed based on multiple credit limit increase targets, including: The product tier data is used for customer segmentation. Based on historical credit limit increase data, the customer utilization rate and delinquency rate under different credit limit increase ranges for different customer groups are determined. At the same time, multiple credit limit increase targets are determined. The credit limit increase targets include maximizing the customer utilization rate and minimizing the credit limit increase risk. Based on multiple credit limit increase targets, a credit limit increase exposure allocation model is constructed; the credit limit increase exposure allocation model is based on maximizing the customer's spending rate. 11 x 11 +a 12 x 12 +…+a mn x mn And the risk of increasing the credit limit is minimal. 11 x 11 +e 12 x 12 +…+e mn x mn The resulting model; where x 11 x 12 ... x mn The proportion of customers under different customer groups and credit limit increases; a 11 a 12 ... a mn For different customer groups and credit limit increases, e 11 e 12 ..., e mn The delinquency rate for different customer groups and credit limit increases.

2. The optimal allocation method for quota adjustment based on multiple objectives according to claim 1, characterized in that, The relevant data is preprocessed, including: Through data cleaning and feature mining techniques, product-level data is aggregated and merged into customer-level data. The customer tier data is transformed and anomaly handling is performed to form a wide customer table. The data anomaly handling includes handling missing values ​​and outlier values.

3. The optimal allocation method for quota adjustment based on multiple objectives according to claim 1, characterized in that, The constraints on the increase in the quota include a first constraint and a second constraint. The first constraint is that the proportion of customers whose credit limit increase falls within the range g is not less than G, and the formula is: x 11 +x 12 +…+x g1g2 ≥G; The second constraint is that the proportion of customers in a certain customer group k who increase their credit limit does not exceed K, and the formula is: x 11 +x 12 +…+x k1k2 ≤K Where g is the preset value for the credit limit increase, G is the preset value for the customer percentage corresponding to the credit limit increase g, k is the preset value for the customer group category restriction, K is the preset value for the customer percentage corresponding to customer group category k, and x g1g2 For customers whose credit limit increase is less than or equal to g, x k1k2 Group customers of customer group k into different customer groups.

4. The optimal allocation method for quota adjustment based on multiple objectives according to claim 1, characterized in that, The genetic algorithm is used to optimize the credit limit allocation model, including: By setting the initial solution X 11 ,X 12 ,…,X mn Mutation and crossover of the initial solution yields another set of new candidate solutions X'. 11 ,X' 12 ,…,X' mn By calculating the fitness of candidate solutions, other candidate solutions are selected. The mutation and crossover steps are repeated continuously, and it is determined whether the quota increase target and quota increase constraints are met, thereby obtaining the optimal solution.

5. A multi-objective-based optimal allocation device for quota adjustment, characterized in that, The device includes: The acquisition unit is used to acquire relevant data for the credit limit increase, which includes product-level data and customer-level data required for the credit limit increase. The preprocessing unit is used to preprocess the relevant data to form customer wide table information; The credit limit increase exposure allocation model construction unit is used to construct a credit limit increase exposure allocation model based on multiple credit limit increase targets according to the customer wide table information. The unit for determining the constraints for increasing the credit limit is used to determine the constraints for increasing the credit limit. The optimal solution unit is used to perform optimal solution on the credit limit increase allocation model using a genetic algorithm to obtain the optimal solution that satisfies the credit limit increase target and the credit limit increase constraint, and uses the optimal solution as the adjustment ratio of the credit limit increase exposure. The relevant data includes product-related information and customer basic information. The product-related information includes customer risk information, customer credit information, and customer lifecycle information. The customer risk information includes repayment behavior performance information, which includes historical overdue status, current overdue status, early settlement status, recent application status, and recent fund usage status. Customer credit information includes basic loan information and the customer's historical loan usage records both within and outside the bank. The basic loan information includes the loan product name, loan product type, loan guarantee type, credit period, loan credit duration, credit amount, and interest rate. The customer's historical loan usage records both within and outside the bank include the number of loan transactions, the number of settled and outstanding transactions, the loan time, and the proportion of the loan amount to the credit limit. Customer lifecycle information includes customer behavior information, which includes borrowing habits, probability of active churn, probability of borrowing demand, number of historical credit limit increases, and historical credit limit increase magnitude. Customer basic information includes user profile information; Based on the customer broad table information, a credit limit increase exposure allocation model is constructed based on multiple credit limit increase targets, including: The product tier data is used for customer segmentation. Based on historical credit limit increase data, the customer utilization rate and delinquency rate under different credit limit increase ranges for different customer groups are determined. At the same time, multiple credit limit increase targets are determined. The credit limit increase targets include maximizing the customer utilization rate and minimizing the credit limit increase risk. Based on multiple credit limit increase targets, a credit limit increase exposure allocation model is constructed; the credit limit increase exposure allocation model is based on maximizing the customer's spending rate. 11 x 11 +a 12 x 12 +…+a mn x mn And the risk of increasing the credit limit is minimal. 11 x 11 +e 12 x 12 +…+e mn x mn The resulting model; where x 11 x 12 ... x mn The proportion of customers under different customer groups and credit limit increases; a 11 a 12 ... a mn For different customer groups and credit limit increases, e 11 e 12 ..., e mn The delinquency rate for different customer groups and credit limit increases.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a multi-objective-based optimal allocation method for quota adjustment as described in any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements a multi-objective-based optimal allocation method for quota adjustment as described in any one of claims 1 to 4.

Images