Recovery allocation method and system based on game coordination mechanism

Abstract

The application discloses a collection distribution method and system based on a game cooperation mechanism, comprising the steps of automatically generating a collection task queue according to overdue customer data, constructing a cooperative game model for the collection task queue, regarding a plurality of collectors as game participants, and distributing all collection tasks to each candidate alliance, wherein each candidate alliance comprises at least one collector, obtaining the marginal contribution value of each collector in different candidate alliances, dynamically distributing the collection task weight of each collector in different candidate alliances based on the marginal contribution value, triggering an intelligent contract to execute the latest performance of each collector in different candidate alliances based on the collection results of each collector in different candidate alliances, and updating the current performance based on the latest performance. The application realizes dynamic distribution of task weights, automatically settles performance in combination with the intelligent contract, and solves the problems of unfair task distribution, non-transparent performance settlement and low cooperation efficiency in the prior art.

Description

Technical Field

[0001] This invention relates to the field of information processing technology, specifically to a collection and distribution method and system based on a game-theoretic collaborative mechanism. Background Technology

[0002] With the proliferation of cash loans and consumer loans, collection efforts are essential to recover these loans. Currently, for ease of management, most financial institutions use collection systems to manage the collection of cash loans and consumer loans. However, existing collection systems have the following technical shortcomings:

[0003] 1. Unreasonable collection task allocation mechanism: Traditional collection systems usually use fixed rules or simple priorities to allocate collection tasks, which cannot dynamically reflect the actual contributions and cooperation relationships of collection agents, resulting in waste of resources and low efficiency.

[0004] 2. Lack of transparency in performance settlement: Manual statistics and settlement of performance are easily affected by subjective factors, there is a risk of data tampering, and the collection effect cannot be fed back in real time.

[0005] 3. Insufficient distributed collaboration capabilities: Most existing collection systems are based on a centralized architecture, which makes it difficult to handle large-scale concurrent collection tasks and lacks elastic scaling and fault tolerance mechanisms. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the purpose of this invention is to provide a collection and allocation method and system based on a game-theoretic collaborative mechanism, which enables dynamic allocation of task weights and combines smart contracts for automatic performance settlement, thereby solving problems such as unfair task allocation, opaque performance settlement, and low collaborative efficiency in existing systems.

[0007] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part from practice of this application.

[0008] According to a first aspect of this application, a collection and distribution method based on a game-theoretic collaborative mechanism is provided, comprising:

[0009] A collection task queue is automatically generated based on overdue customer data, and the collection task queue includes several collection tasks.

[0010] For the collection task queue, a cooperative game model is constructed. The model is used to treat several collectors as game participants. Based on the collection task queue, all collection tasks are assigned to each candidate alliance. Each candidate alliance includes at least one collector. The marginal contribution value of each collector in different candidate alliances is obtained. Based on the marginal contribution value, the collection task weight of each collector in different candidate alliances is dynamically allocated.

[0011] Based on each collector's collection results in different candidate alliances, the smart contract is triggered to execute the latest performance of each collector in different candidate alliances, and the current performance is updated based on the latest performance.

[0012] In some embodiments of this application, based on the foregoing scheme, the automatic generation of a collection task queue based on overdue customer data is as follows:

[0013] Several collection tasks are established based on the customer's basic information;

[0014] For each collection task, obtain standardized results such as overdue amount, overdue days, historical recovery rate, customer risk level, last collection time and result, and customer response tendency analysis.

[0015] Establish a collection scoring model. For each collection task, weight the standardized results of overdue amount, overdue days, historical repayment rate, customer risk level, last collection time and result, and customer response tendency analysis to obtain a collection score for each collection task. Based on the collection scores, sort all collection tasks in descending order to establish a collection task queue.

[0016] In some embodiments of this application, based on the foregoing scheme, the step of constructing a cooperative game model for the collection task queue, wherein the model is used to treat several collection agents as game participants, and to allocate all collection tasks to various candidate alliances based on the collection task queue, wherein each candidate alliance includes at least one collection agent, obtaining the marginal contribution value of each collection agent in different candidate alliances, and dynamically allocating the collection task weight of each collection agent in different candidate alliances based on the marginal contribution value, further includes:

[0017] Define the candidate alliance payoff function, and calculate it using the following formula:

[0018]

[0019] in, Indicating candidate alliance The probability of recovery, Indicating candidate alliance The possible amount to be recovered, Indicating candidate alliance Collection costs, Indicating candidate alliance The benefits;

[0020] A hierarchical dynamic sampling strategy is used to generate the candidate coalition set, specifically as follows:

[0021] Candidate alliances are categorized based on their size. Hierarchy, among which, ;

[0022] Based on the hypermodel of the candidate alliance's profit function, sampling weights are assigned to candidate alliances at each level. The calculation formula is:

[0023]

[0024] in, For scale The expected revenue of the candidate alliance under the optimal collection task allocation;

[0025] Total number of samples By weight Candidate alliances are assigned to each level, and a pool is generated for each level. Each candidate alliance is a collection of candidate alliances at various levels.

[0026] All collection tasks are assigned to the candidate alliance set based on the collection task queue.

[0027] In some embodiments of this application, based on the foregoing scheme, the allocation of all collection tasks to the candidate consortium set based on the collection task queue includes:

[0028] Encode the current collection task allocation status into a collection allocation vector. The candidate alliance set is The collection task set is Set actions and rewards ,action This means assigning a collection task to a candidate alliance and receiving a reward. To increase the incremental revenue of candidate alliances after allocation, a Monte Carlo tree search is used to simulate multiple allocation processes. The value function of actions is learned through a deep Q-network. All value function values ​​are sorted in descending order, and the action with the highest value function value is selected first and executed. The collection allocation vector is updated until all collection tasks are allocated, and the collection tasks and their corresponding candidate alliances are output.

[0029] In some embodiments of this application, based on the foregoing scheme, the step of constructing a cooperative game model for the collection task queue, wherein the model is used to treat several collection agents as game participants, and to allocate all collection tasks to various candidate alliances based on the collection task queue, wherein each candidate alliance includes at least one collection agent, obtaining the marginal contribution value of each collection agent in different candidate alliances, and dynamically allocating the collection task weight of each collection agent in different candidate alliances based on the marginal contribution value, further includes:

[0030] Get a debt collector Marginal contribution value in the candidate alliance The calculation formula is:

[0031]

[0032] in, For including debt collectors candidate alliances, For the size of the candidate alliance, The total number of debt collectors. This indicates a gathering of debt collectors.

[0033] In some embodiments of this application, based on the foregoing scheme, the step of triggering a smart contract to execute the latest performance of each collector in different candidate alliances based on the collection results of each collector in different candidate alliances, and updating the current performance based on the latest performance, includes:

[0034] Construct a contribution model to obtain the contribution of each collector in the candidate alliance. The calculation formula is as follows:

[0035]

[0036] in, debt collector In the Candidate Alliances The amount of recovery processed in the middle, debt collector In the Candidate Alliances The amount of recovery processed in the middle, As a priority index, , All are weighting coefficients;

[0037] Upon successful collection by each collector in the candidate consortium, the smart contract is triggered to obtain each collector's individual contribution and latest performance based on their contribution level in the candidate consortium. The calculation formula is as follows:

[0038]

[0039] The current performance is updated based on the latest performance data. The calculation formula is as follows:

[0040]

[0041] in, debt collector Current performance, debt collector Historical performance.

[0042] In some embodiments of this application, based on the foregoing scheme, the following further methods are also included:

[0043] The smart contract is deployed on a blockchain platform.

[0044] According to a second aspect of this application, a collection and distribution system based on a game-theoretic collaborative mechanism is provided, the system comprising:

[0045] The collection task generation module is used to automatically generate a collection task queue based on overdue customer data;

[0046] The collection task allocation module is used to construct a cooperative game model for the collection task queue. The model is used to regard several collectors as game participants, allocate all collection tasks to each candidate alliance based on the collection task queue, and each candidate alliance includes at least one collector. The marginal contribution value of each collector in different candidate alliances is obtained, and the collection task weight of each collector in different candidate alliances is dynamically allocated based on the marginal contribution value.

[0047] The performance allocation module is used to trigger the smart contract to execute the latest performance of each collector in different candidate alliances based on the collection results of each collector in different candidate alliances, and update the current performance based on the latest performance.

[0048] According to a third aspect of this application, a computer-readable storage medium is provided that stores a computer program thereon, the computer program including executable instructions that, when executed by a processor, implement the method described above.

[0049] According to a fourth aspect of this application, an electronic device is provided, comprising:

[0050] One or more processors;

[0051] A memory for storing executable instructions of the processor, which, when executed by the one or more processors, cause the one or more processors to implement the method described above.

[0052] The beneficial effects of this application are as follows:

[0053] The collection allocation method and system based on a game-theoretic collaborative mechanism provided in this application automatically generates a collection task queue based on overdue customer data, ensuring that resources are prioritized for tasks with high recovery potential. It treats collection agents as game participants and allocates all collection tasks to various candidate alliances through cooperative game theory. The task weight of collection agents in different candidate alliances is dynamically adjusted according to their marginal contribution value to avoid resource misallocation. Furthermore, smart contracts automatically trigger performance calculations to avoid human intervention errors. It can collaboratively handle complex overdue cases and improve the overall recovery rate, achieving intelligent, fair, and efficient management of collection tasks.

[0054] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0055] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and are intended to explain the invention, but do not constitute an undue limitation thereof. In the drawings:

[0056] Figure 1 This is a flowchart of a collection and distribution method based on a game-theoretic collaborative mechanism in this embodiment;

[0057] Figure 2 This is a schematic diagram of a collection and distribution system based on a game-theoretic collaborative mechanism in this embodiment;

[0058] Figure 3 This is a schematic diagram of the electronic device in this embodiment. Detailed Implementation

[0059] Specific embodiments of the invention will now be described in detail with reference to the accompanying drawings, which illustrate examples of the invention. Although the invention will be described in conjunction with specific embodiments, it will be understood that it is not intended to limit the invention to the embodiments described herein. Rather, it is intended to cover variations, modifications, and equivalents included within the spirit and scope of the invention as defined by the appended claims. It should be noted that the method steps described herein can be implemented by any functional block or functional arrangement, and any functional block or functional arrangement can be implemented as a physical entity or a logical entity, or a combination of both.

[0060] To enable those skilled in the art to better understand the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0061] Note: The examples described below are merely specific examples and are not intended to limit the embodiments of the present invention to the specific steps, values, conditions, data, order, etc. Those skilled in the art can utilize the concept of the present invention to construct more embodiments not mentioned herein by reading this specification.

[0062] Figure 1 A flowchart of a debt collection and allocation method based on a game-theoretic collaborative mechanism is shown in this embodiment. This embodiment provides a debt collection and allocation method based on a game-theoretic collaborative mechanism, including:

[0063] Step 101: Automatically generate a collection task queue based on overdue customer data.

[0064] In some implementations of this embodiment, overdue customer data is acquired. The overdue customer data includes basic customer information, overdue amount, overdue days, historical repayment rate, customer risk level, last collection time and result, customer response tendency analysis, and other data. A collection task queue is established based on the overdue customer data.

[0065] In some embodiments of this example, establishing a collection task queue based on overdue customer data includes:

[0066] Several collection tasks are established based on the customer's basic information;

[0067] All collection tasks are sorted based on overdue customer data to create a collection task queue.

[0068] In some embodiments of this example, the step of sorting all collection tasks based on overdue customer data and establishing a collection task queue includes:

[0069] For each collection task, obtain standardized results such as overdue amount, overdue days, historical recovery rate, customer risk level, last collection time and result, and customer response tendency analysis.

[0070] Establish a collection scoring model. For each collection task, weight the standardized results of overdue amount, overdue days, historical repayment rate, customer risk level, last collection time and result, and customer response tendency analysis to obtain a collection score for each collection task. Based on the collection scores, sort all collection tasks in descending order to establish a collection task queue.

[0071] In some embodiments of this example, the collection task queue is... ,in, This represents the total number of collection tasks.

[0072] Step 102: For the collection task queue, construct a cooperative game model. The model is used to treat several collectors as game participants. Based on the collection task queue, all collection tasks are assigned to each candidate alliance. Each candidate alliance includes at least one collector. Obtain the marginal contribution value of each collector in different candidate alliances. Based on the marginal contribution value, dynamically allocate the collection task weight of each collector in different candidate alliances.

[0073] In this way, several candidate alliances are generated, each of which includes at least one debt collector. Debt collection tasks are reasonably allocated to each debt collector based on the marginal contribution value. This can identify the true value of each debt collector in a specific candidate alliance, avoid high-skilled personnel being occupied by inefficient debt collection tasks, and the marginal contribution value of each debt collector objectively reflects their global contribution, thus solving the subjectivity problem of traditional allocation based on experience.

[0074] In this embodiment, common forms of candidate alliance collaboration are as follows:

[0075] (1) Small-scale combination: 2-3 collectors temporarily form a team to try to cooperate with a certain type of difficult customers. This is similar to senior collectors mentoring newcomers, and is also a candidate alliance combination.

[0076] (2) Business-oriented combination: group customers by region (such as the candidate alliance of customers in the same city) and type (candidate alliance of high-delinquency enterprise customers) and focus on collaboration in specific business scenarios;

[0077] (3) Dynamically adjust the combination: Reorganize at any time according to the collection progress, staff availability, etc. For example, the collection staff who were scattered in the early stage can be reorganized into a team in the later stage to tackle a large number of overdue customers.

[0078] In some embodiments of this example, the debt collectors are grouped together. Candidate Alliance Indicated by the candidate alliance Gathered by debt collectors A collaborative team composed of several debt collectors.

[0079] In some embodiments of this example, for the collection task queue, a cooperative game model is constructed. This model treats several collection agents as game participants, allocates all collection tasks to various candidate alliances based on the collection task queue, and each candidate alliance includes at least one collection agent. The marginal contribution value of each collection agent in different candidate alliances is obtained, and the collection task weight of each collection agent in different candidate alliances is dynamically allocated based on the marginal contribution value. The model also includes:

[0080] Define the candidate alliance payoff function;

[0081] A hierarchical dynamic sampling strategy is employed to generate a candidate consortium set, ensuring that high-yield potential candidate consortia are fully covered. The candidate consortium set is as follows: ,in, The total number of candidate alliances, with each candidate alliance including at least one debt collector;

[0082] All collection tasks are assigned to the candidate alliance set based on the collection task queue.

[0083] In some implementations of this embodiment, a candidate alliance revenue function is defined, including:

[0084]

[0085] in, Indicating candidate alliance The probability of recovery, Indicating candidate alliance The possible amount to be recovered, Indicating candidate alliance Collection costs, Indicating candidate alliance The benefits.

[0086] In some implementations of this embodiment, candidate alliance The formula for calculating the recovery probability is:

[0087]

[0088] in, For the intercept term; These are the weighting coefficients for each influencing factor; The various influencing factors include overdue days, historical repayment rate, and customer risk level.

[0089] In some implementations of this embodiment, candidate alliance The formula for calculating the potential recoverable amount is:

[0090]

[0091] in, The amount that needs to be recovered; This refers to the dynamic recovery ratio.

[0092] In some implementations of this embodiment, different dynamic recovery ratios are preset according to the customer's risk level (such as A / B / C / D). For example, if the customer's risk level is A, the dynamic recovery ratio is 80%~90%; if the customer's risk level is B, the dynamic recovery ratio is 60%~70%; if the customer's risk level is C, the dynamic recovery ratio is 40%~50%; and if the customer's risk level is D, the dynamic recovery ratio is 10%~20%.

[0093] In some implementations of this embodiment, candidate alliance The formula for calculating collection costs is:

[0094]

[0095] in, These are all cost-influencing factors, such as labor costs, communication costs, technology costs, and travel costs. This embodiment does not limit these factors.

[0096] In some embodiments of this example, a hierarchical dynamic sampling strategy is used to generate a candidate alliance set, specifically including:

[0097] Candidate alliances are categorized based on their size. Hierarchy, among which, The first tier indicates that there is 1 debt collector in the candidate alliance, the second tier indicates that there are 2 debt collectors in the candidate alliance, ..., the tier ... The tier indicates the number of debt collectors in the candidate consortium. indivual.

[0098] Based on the hypermodel of the candidate alliance's profit function, sampling weights are assigned to candidate alliances at each level. The calculation formula is:

[0099]

[0100] in, For scale The expected return of candidate alliances under optimal collection task allocation. Higher-level candidate alliances are assigned higher sampling weights, while lower-level candidate alliances are assigned lower sampling weights.

[0101] Total number of samples By weight Candidate alliances are assigned to each level, and a pool is generated for each level. Each candidate alliance is integrated into a candidate alliance set, which combines candidate alliances at various levels.

[0102] Specifically, the scale is The expected returns of the candidate alliance under the optimal collection task allocation are estimated based on pre-experimentation or historical data.

[0103] In some implementations of this embodiment, a collection task allocation strategy is implemented, which allocates all collection tasks to the candidate consortium set based on the collection task queue, and uses a combination of Monte Carlo Tree Search (MCTS) and Deep Q Network (DQN) to dynamically adjust the collection task allocation strategy.

[0104] Specifically, a method combining Monte Carlo Tree Search (MCTS) and Deep Q-Network (DQN) is used to dynamically adjust the collection task allocation strategy, including:

[0105] Encode the current collection task allocation status into a collection allocation vector. The candidate alliance set is The collection task set is Set actions and rewards ,action This means assigning a collection task to a candidate alliance and receiving a reward. To increase the incremental revenue of candidate alliances after allocation, a Monte Carlo tree search is used to simulate multiple allocation processes. The value function of actions is learned through a deep Q-network. All value function values ​​are sorted in descending order, and the action with the highest value function value is selected first and executed. The collection allocation vector is updated until all collection tasks are allocated, and the collection tasks and their corresponding candidate alliances are output.

[0106] Thus, by reducing the number of candidate alliance combinations generated from exponential to polynomial through Monte Carlo simulation, and by combining reinforcement learning to optimize the allocation of collection tasks, large-scale collection tasks and candidate alliances can be processed within a reasonable time. Furthermore, the collaborative effect of candidate alliances is incorporated into the candidate alliance payoff function, which solves the problem of traditional allocation methods ignoring team collaboration. Through a matching model of collection task characteristics and candidate alliance capabilities, the optimal fit between collection tasks and candidate alliances is achieved, thereby improving the total recovery revenue.

[0107] In some embodiments of this example, for the collection task queue, a cooperative game model is constructed. This model treats several collection agents as game participants, allocates all collection tasks to various candidate alliances based on the collection task queue, and each candidate alliance includes at least one collection agent. The marginal contribution value of each collection agent in different candidate alliances is obtained, and the collection task weight of each collection agent in different candidate alliances is dynamically allocated based on the marginal contribution value. The model also includes:

[0108] Get a debt collector Marginal contribution value in the candidate alliance The calculation formula is:

[0109]

[0110] in, For including debt collectors candidate alliances, For the size of the candidate alliance, The total number of debt collectors. This indicates a gathering of debt collectors.

[0111] The collection workload is allocated based on the marginal contribution value, so that collectors can get more high-value collection tasks.

[0112] Step 103: Based on the collection results of each collector in different candidate alliances, trigger the smart contract to execute the latest performance of each collector in different candidate alliances, and update the current performance based on the latest performance.

[0113] In some implementations of this embodiment, based on the collection results of each collector in different candidate alliances, the smart contract is triggered to execute the latest performance of each collector in different candidate alliances, and the current performance is updated based on the latest performance. This also includes:

[0114] Construct a contribution model to obtain the contribution of each collector in the candidate alliance. The calculation formula is as follows:

[0115]

[0116] in, debt collector In the Candidate Alliances The amount of recovery processed in the middle, debt collector In the Candidate Alliances The amount of recovery processed in the middle, As a priority index, , All are weighting coefficients.

[0117] Upon successful collection by each collector in the candidate consortium, the smart contract is triggered to obtain each collector's individual contribution and latest performance based on their contribution level in the candidate consortium. The calculation formula is as follows:

[0118]

[0119] The current performance is updated based on the latest performance data. The calculation formula is as follows:

[0120]

[0121] in, debt collector Current performance, debt collector Historical performance.

[0122] In some embodiments of this example, the debt collector... The stronger the collection ability, the higher the priority index. The higher the score, the more the latest performance results are updated in real time to the collector's file, and the data is fed back into the contribution model to adjust the priority index. If a debt collector consistently exceeds their targets, their priority level will be increased. When a debt collector If a debt collector fails to complete a task multiple times consecutively, their priority level will be lowered. .

[0123] In some implementations of this embodiment, the smart contract is deployed on a blockchain platform (such as Ethereum or Hyperledger) and triggers performance updates through the following rules.

[0124] In this way, performance calculation and updates are automatically triggered by smart contracts, reducing manual intervention, improving efficiency, and making the contribution and the latest performance calculation rules public on the blockchain to ensure fairness. The method provided in this embodiment can support real-time updates of performance by period or collection task completion, enhance the incentive effect, and quantify the collaborative gains of candidate alliances.

[0125] According to the second aspect of this application, such as Figure 2 As shown, this embodiment provides a collection and distribution system based on a game-theoretic collaborative mechanism, including:

[0126] The collection task generation module 201 is used to automatically generate a collection task queue based on overdue customer data.

[0127] The collection task allocation module 202 is used to construct a cooperative game model for the collection task queue. The model is used to regard several collectors as game participants, allocate all collection tasks to each candidate alliance based on the collection task queue, each candidate alliance includes at least one collector, obtain the marginal contribution value of each collector in different candidate alliances, and dynamically allocate the collection task weight of each collector in different candidate alliances based on the marginal contribution value.

[0128] The performance allocation module 203 is used to trigger the smart contract to execute the latest performance of each collector in different candidate alliances based on the collection results of each collector in different candidate alliances, and update the current performance based on the latest performance.

[0129] According to a third aspect of this application, this embodiment provides a computer-readable storage medium having a computer program stored thereon, the computer program including executable instructions that, when executed by a processor 301, implement the method described above.

[0130] The present invention can implement all or part of the processes in the above methods, or it can be accomplished by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor 301, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable medium can include: any entity or system capable of carrying computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory 302, read-only memory 302 (ROM), random access memory 302 (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0131] According to a fourth aspect of this application, an electronic device is provided, such as... Figure 3 As shown, it includes:

[0132] One or more processors 301;

[0133] The memory 302 is used to store the executable instructions of the processor 301, which, when executed by one or more processors 301, cause one or more processors 301 to implement the above-described method.

[0134] The electronic device is manifested in the form of a general-purpose computing device. The components of the electronic device may include, but are not limited to: at least one processor 301, at least one memory 302, and a bus 303 connecting different system components (including memory 302 and processor 301).

[0135] The processor 301 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor 301, or any conventional processor 301. The processor 301 is the control center of the computer system, connecting various parts of the computer system through various interfaces and lines.

[0136] The memory 302 can be used to store computer programs and / or modules. The processor 301 implements various functions of the computer system by running or executing the computer programs and / or modules stored in the memory 302 and calling the data stored in the memory 302. The memory 302 may mainly include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function (such as sound playback function, image playback function, etc.). The data storage area may store data created according to the use of the mobile phone (such as audio data, video data, etc.). In addition, the memory 302 may include high-speed random access memory 302, and may also include non-volatile memory 302, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device 302, flash memory device, or other volatile solid-state memory 302.

[0137] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, servers, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention 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 302 and optical storage 302, etc.) containing computer-usable program code.

[0138] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), servers, and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor 301 of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor 301 of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A system that specifies functions in one or more boxes.

[0139] These computer program instructions may also be stored in a computer-readable storage medium 302 that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium 302 produce an article of manufacture including an instruction system implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0140] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0141] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A debt collection and distribution method based on a game-theoretic collaborative mechanism, characterized in that, include: A collection task queue is automatically generated based on overdue customer data, and the collection task queue includes several collection tasks. For the collection task queue, a cooperative game model is constructed. This model treats several collection collectors as game participants and assigns all collection tasks to various candidate alliances based on the collection task queue. Each candidate alliance includes at least one collection collector. The marginal contribution value of each collection collector in different candidate alliances is obtained. Based on the marginal contribution value, the collection task weights of each collection collector in different candidate alliances are dynamically allocated, including: Define the candidate alliance payoff function, and calculate it using the following formula: in, Indicating candidate alliance The probability of recovery, Indicating candidate alliance The possible amount to be recovered, Indicating candidate alliance Collection costs, Indicating candidate alliance The benefits; A hierarchical dynamic sampling strategy is used to generate the candidate coalition set, specifically as follows: Candidate alliances are categorized based on their size. Hierarchy, among which, ; Based on the hypermodel of the candidate alliance's profit function, sampling weights are assigned to candidate alliances at each level. The calculation formula is: in, For scale The expected revenue of the candidate alliance under the optimal collection task allocation; Total number of samples By weight Candidate alliances are assigned to each level, and a pool is generated for each level. Each candidate alliance is a collection of candidate alliances at various levels. All collection tasks are assigned to the candidate alliance set based on the collection task queue: Encode the current collection task allocation status into a collection allocation vector. The candidate alliance set is The collection task set is Set actions and rewards ,action This means assigning a collection task to a candidate alliance and receiving a reward. To increase the incremental revenue of candidate alliances after allocation, Monte Carlo tree search is used to simulate multiple allocation processes. The value function value of actions is learned through a deep Q-network. All value function values ​​are sorted in descending order, and the action with the highest value function value is selected first and executed. The collection allocation vector is updated until all collection tasks are allocated, and the collection tasks and their corresponding candidate alliances are output. Based on each collector's collection results in different candidate alliances, the smart contract is triggered to execute the latest performance of each collector in different candidate alliances, and the current performance is updated based on the latest performance.

2. The method according to claim 1, characterized in that, The collection task queue is automatically generated based on overdue customer data. Several collection tasks are established based on the customer's basic information; For each collection task, obtain standardized results such as overdue amount, overdue days, historical recovery rate, customer risk level, last collection time and result, and customer response tendency analysis. Establish a collection scoring model. For each collection task, weight the standardized results of overdue amount, overdue days, historical repayment rate, customer risk level, last collection time and result, and customer response tendency analysis to obtain a collection score for each collection task. Based on the collection scores, sort all collection tasks in descending order to establish a collection task queue.

3. The method according to claim 1, characterized in that, For the collection task queue, a cooperative game model is constructed. This model treats several collection collectors as game participants, assigns all collection tasks to various candidate alliances based on the collection task queue, and each candidate alliance includes at least one collection collector. The marginal contribution value of each collection collector in different candidate alliances is obtained, and the collection task weight of each collection collector in different candidate alliances is dynamically allocated based on the marginal contribution value. The model also includes: Get a debt collector Marginal contribution value in the candidate alliance The calculation formula is: in, For including debt collectors candidate alliances, For the size of the candidate alliance, The total number of debt collectors. This indicates a gathering of debt collectors.

4. The method according to claim 1, characterized in that, The process of triggering a smart contract to execute the latest performance data for each collector across different candidate alliances, based on their collection results, and updating the current performance data accordingly, includes: Construct a contribution model to obtain the contribution of each collector in the candidate alliance. The calculation formula is as follows: in, debt collector In the Candidate Alliances The amount of recovery processed in the middle, debt collector In the Candidate Alliances The amount of recovery processed in the middle, As a priority index, , All are weighting coefficients; Upon successful collection by each collector in the candidate consortium, the smart contract is triggered to obtain each collector's individual contribution and latest performance based on their contribution level in the candidate consortium. The calculation formula is as follows: The current performance is updated based on the latest performance data. The calculation formula is as follows: in, debt collector Current performance, debt collector Historical performance.

5. The method according to claim 4, characterized in that, Also includes: The smart contract is deployed on a blockchain platform.

6. A debt collection and distribution system based on a game-theoretic collaborative mechanism, applied to the debt collection and distribution method based on a game-theoretic collaborative mechanism as described in any one of claims 1-5, characterized in that, The system includes: The collection task generation module is used to automatically generate a collection task queue based on overdue customer data; The collection task allocation module is used to construct a cooperative game model for the collection task queue. The model is used to regard several collectors as game participants, allocate all collection tasks to each candidate alliance based on the collection task queue, and each candidate alliance includes at least one collector. The marginal contribution value of each collector in different candidate alliances is obtained, and the collection task weight of each collector in different candidate alliances is dynamically allocated based on the marginal contribution value. The performance allocation module is used to trigger the smart contract to execute the latest performance of each collector in different candidate alliances based on the collection results of each collector in different candidate alliances, and update the current performance based on the latest performance.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program includes executable instructions that, when executed by a processor, implement the method of any one of claims 1-5.

8. An electronic device, characterized in that, include: One or more processors; A memory for storing executable instructions of the processor, which, when executed by the one or more processors, cause the one or more processors to perform the method according to any one of claims 1-5.

Images