A multi-level permission collaborative management method and platform for a performance bond system

Abstract

The application provides a multi-level permission collaborative management method and platform for a guarantee system, relates to the technical field of permission management, and comprises the following steps: after a guarantee task is generated in the guarantee system, multiple collaborative subjects participating in the guarantee task are abstracted as game participants, and the execution process of the guarantee task is constructed as a multi-stage dynamic game process; a strategy space is constructed for each game participant; a multi-subject benefit evaluation model is constructed based on the risk benefit and compliance constraints of the current task state; the optional strategy set of each game participant is cut based on the strategy space and the constraint boundary, and the effective strategy subspace is input into the multi-subject benefit evaluation model; the permission allocation strategy is dynamically configured according to the optimal strategy combination of the stage, and multi-level permission collaborative management is performed. Through the application, the technical problem of poor guarantee system permission collaborative management efficiency in the prior art can be solved, and the technical effects of improving system collaborative efficiency and optimizing the task execution process are achieved.

Description

Technical Field

[0001] This application relates to the field of access control technology, and in particular to a multi-level access control collaborative management method and platform for guarantee systems. Background Technology

[0002] With the rapid development of information technology and automated control, task execution in many fields has gradually become more intelligent and collaborative. Especially in multi-agent systems, the application of game theory and permission allocation strategies has become an important means to optimize task execution and improve efficiency.

[0003] Currently, most existing methods for allocating permissions and adjusting strategies among game participants rely on static rule settings, failing to fully consider the dynamic reactions and strategy changes of participants during task execution. Permission allocation in most systems is typically managed based on fixed permission models, and strategy selection is not adequately adapted to the actual performance of the participants. Since participants' strategies may deviate due to changes in the external environment or uncertainties in internal operations, existing technologies cannot monitor and adjust strategies and permissions in real time. This results in the system struggling to respond and adjust promptly when strategies deviate significantly, thus impacting the overall task execution efficiency.

[0004] In summary, existing technologies suffer from the technical problem that the deviation between the strategies and permissions of game participants cannot be monitored and adjusted in real time, resulting in an inability to effectively respond to dynamic changes in the strategies of the participants during task execution, which further affects the system's collaborative efficiency and the stability of task execution. Summary of the Invention

[0005] The purpose of this application is to provide a multi-level permission collaborative management method and platform for guarantee systems, in order to solve the technical problem in the prior art that the deviation between the strategies and permissions of game participants cannot be monitored and adjusted in real time, which leads to the inability to effectively respond to the dynamic changes of the participants' strategies during task execution, and further affects the collaborative efficiency and task execution stability of the system.

[0006] In view of the above problems, this application provides a multi-level permission collaborative management method and platform for guarantee systems.

[0007] Firstly, this application provides a multi-level permission collaborative management method for a guarantee system, implemented through a multi-level permission collaborative management platform for the guarantee system. The method includes: after a guarantee task is generated within the guarantee system, abstracting multiple collaborating entities participating in the guarantee task as game participants, and constructing the execution process of the guarantee task as a multi-stage dynamic game process, where each stage corresponds to a task state node; constructing a strategy space for each game participant, the strategy space including operational strategies, collaborative strategies, and risk-taking strategies, and representing the permissions possessed by each game participant in the current stage as optional strategies. The system defines the constraint boundaries of a set of strategies; constructs a multi-party benefit evaluation model based on the risk-reward and compliance constraints of the current task state. This model evaluates task completion efficiency, risk exposure, and collaboration costs to calculate the stage benefit values ​​of each game participant under different strategy combinations. Based on the strategy space and the constraint boundaries, the system prunes the optional strategy set of each game participant to establish an effective strategy subspace. This effective strategy subspace is then input into the multi-party benefit evaluation model to output the stage-optimal strategy combination. Finally, the system dynamically configures permission allocation strategies based on the stage-optimal strategy combination to implement multi-level permission collaborative management.

[0008] Preferably, the multi-level permission collaborative management method for a guarantee system further includes: constructing a strategy interaction matrix based on the effective strategy subspace of each game participant, wherein the strategy interaction matrix is ​​used to describe the mutual influence relationship between strategy combinations of different game participants; inputting the strategy interaction matrix into the multi-subject payoff evaluation model to calculate the payoff response value corresponding to each strategy combination under the current permission constraint boundary; performing payoff strategy adjustment iteration based on the payoff response value and the subject characteristics of the mapped game participants; performing iterative update analysis based on the payoff strategy adjustment iteration result, and outputting the optimal strategy combination for the stage.

[0009] Preferably, the multi-level permission collaborative management method for a guarantee system further includes: extracting features of game participants and establishing subject features, wherein the subject features include risk preference features, collaborative dependency features, and permission level features; mapping the subject features to a payoff response adjustment coefficient, wherein the payoff response adjustment coefficient is used to characterize the sensitivity of different game participants to payoff changes; calculating the deviation between the payoff response value of each game participant in the current round and the historical payoff benchmark value, generating a first adjustment constraint; using the payoff response adjustment coefficient as a second adjustment constraint, configuring a strategy adjustment increment according to the first adjustment constraint and the second adjustment constraint, and using the strategy adjustment increment to perform payoff strategy adjustment iterations.

[0010] Preferably, the multi-level permission collaborative management method for a guarantee system further includes: calculating a payoff-risk coupling index based on the payoff response value and stage risk exposure value of each game participant under the current strategy combination, wherein the payoff-risk coupling index is used to characterize the balance between the increase in payoff and the increase in risk brought about by strategy adjustment; mapping the payoff-risk coupling index to a coupling adjustment factor, and adjusting the magnitude and direction of the incremental strategy adjustment based on the coupling adjustment factor; continuously monitoring the payoff-risk coupling index during the iteration process, and determining that the iteration has reached convergence when the incremental change of the strategy adjustment meets a preset stability threshold, and outputting the optimal strategy combination for the stage.

[0011] Preferably, the multi-level permission collaborative management method for a guarantee system further includes: acquiring the historical strategy sequence, payoff response change trend, and stage risk exposure characteristics of each game participant; performing strategy prediction for each game participant in the next round of iteration testing; establishing predictive strategy behavior; using the predicted strategy behavior to perform strategy correction of the strategy adjustment increment; establishing predictive compensation strategy adjustment increment; and using the predictive compensation strategy adjustment increment to perform payoff strategy adjustment iteration.

[0012] Preferably, the multi-level permission collaborative management method for a guarantee system further includes: during the execution of the guarantee task, monitoring the deviation between the actual strategic behavior of each game participant and the preset strategy, and establishing a deviation dataset; determining whether the deviation dataset meets the preset deviation threshold, and if it meets the preset deviation threshold, issuing a deviation anomaly warning.

[0013] Preferably, the multi-level permission collaborative management method for a guarantee system further includes: configuring policy weight feedback based on the deviation dataset; and using the policy weight feedback to perform real-time update management of the permission allocation policy.

[0014] Preferably, the multi-level permission collaborative management method for a guarantee system further includes: configuring a response detection window for each game participant; if the collaborating entity does not have a response in the corresponding response detection window, then sending an activation request to the collaborating entity; and performing multi-level permission collaborative management based on the response result of the collaborating entity.

[0015] Preferably, the multi-level permission collaborative management method for the guarantee system further includes: classifying the risk sensitivity of the task execution status and establishing the classification result; configuring the enhancement or penalty adjustment weight of the stage benefit according to the classification result, and updating the stage benefit value according to the configuration result.

[0016] Secondly, this application also provides a multi-level permission collaborative management platform for a guarantee system, used to execute a multi-level permission collaborative management method for a guarantee system as described in the first aspect, comprising: a multi-stage dynamic game process construction module, used to abstract multiple cooperating entities participating in the guarantee task as game participants after the guarantee task is generated in the guarantee system, and to construct the execution process of the guarantee task as a multi-stage dynamic game process, wherein each stage corresponds to a task state node; and a strategy space construction module, used to construct a strategy space for each game participant, wherein the strategy space includes operation strategies, collaboration strategies, and risk-taking strategies, and represents the permissions possessed by each game participant in the current stage as a set of optional strategies. The system includes: a constraint boundary; a multi-agent benefit evaluation model construction module, used to construct a multi-agent benefit evaluation model based on the risk-reward and compliance constraints of the current task state. The multi-agent benefit evaluation model calculates the stage benefit value of each game participant under different strategy combinations by evaluating task completion efficiency, risk exposure degree, and collaboration cost; a stage-optimal strategy combination output module, used to prune the optional strategy set of each game participant based on the strategy space and the constraint boundary, establish an effective strategy subspace, input the effective strategy subspace into the multi-agent benefit evaluation model, and output the stage-optimal strategy combination; and a multi-level permission collaboration management module, used to dynamically configure permission allocation strategies according to the stage-optimal strategy combination and execute multi-level permission collaboration management.

[0017] The technical solution provided in this application has at least the following technical effects or advantages: by realizing the dynamic monitoring and adjustment of the strategies and permissions of the game participants, it ensures that the strategies and permissions can be optimized in real time according to the actual deviation during the task execution process, thereby improving the system's collaborative efficiency, optimizing the task execution process, and enhancing the stability and accuracy of task execution.

[0018] The above description is merely an overview of the technical solution of this application. To enable a clearer understanding of the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating a multi-level permission collaborative management method for a guarantee system according to this application.

[0021] Figure 2 This is a schematic diagram of the structure of a multi-level permission collaborative management platform for a guarantee system according to this application.

[0022] Figure labeling: Module 1 for constructing the multi-stage dynamic game process, Module 2 for constructing the strategy space, Module 3 for constructing the multi-agent payoff evaluation model, Module 4 for outputting the optimal strategy combination for each stage, and Module 5 for multi-level permission collaborative management. Detailed Implementation

[0023] This application provides a multi-level permission collaborative management method and platform for guarantee systems. It solves the technical problem in existing technologies where the inability to monitor and adjust the deviation between the strategies and permissions of game participants in real time leads to an inability to effectively respond to dynamic changes in participant strategies during task execution, further affecting system collaborative efficiency and task execution stability. The method enables dynamic monitoring and adjustment of game participant strategies and permissions, ensuring that strategies and permissions can be optimized in real time based on actual deviations during task execution. This achieves the technical effects of improving system collaborative efficiency, optimizing the task execution process, and enhancing task execution stability and accuracy.

[0024] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0025] Example 1, please refer to the appendix. Figure 1 This application provides a multi-level permission collaborative management method for a guarantee system, applied to a multi-level permission collaborative management platform for a guarantee system, specifically including the following steps: After a guarantee task is generated within the guarantee system, the multiple collaborating entities participating in the guarantee task are abstracted as game participants, and the execution process of the guarantee task is constructed as a multi-stage dynamic game process, where each stage corresponds to a task state node.

[0026] Specifically, after a guarantee task is generated in the guarantee system, the multiple collaborating entities involved in the task are transformed into game participants. Game participants refer to the various collaborators with different permissions and decision-making abilities during the execution of the guarantee task. Multiple collaborators need to interact and make decisions to achieve a common goal during task execution; therefore, each participant can influence the progress and outcome of the task by formulating strategies. Constructing the execution process of the guarantee task as a multi-stage dynamic game means that the task state at each stage represents a state node in the game process. At each stage, the strategy choices and interactions of the game participants directly affect the task state and subsequent progress of that stage. This allows for precise simulation and management of the interactions among the participants in the guarantee task and their impact on the task progress.

[0027] A strategy space is constructed for each game participant, including operational strategies, collaborative strategies, and risk-bearing strategies. The permissions possessed by each participant at the current stage are represented as the constraint boundaries of the optional strategy set. For operational strategies, the implementation of operational strategies related to multiple tasks involves a pre-defined task state machine. For example, each task state node, such as guarantee issuance, guarantee modification, and guarantee claim settlement, is assigned a unique task identifier. Operational strategies are associated through a task-operation mapping table. For instance, when the game participant is a guarantee institution and the task identifier is "guarantee issuance review," based on the permission constraint boundaries, the optional operational strategy set is matched as {Strategy A: Automatic compliance verification and approval; Strategy B: Transfer to manual review}. For collaborative strategies, coordination and cooperation are implemented as follows: collaborative strategies are configured as strategy entities containing message routing rules and data sharing fields. For example, in the guarantee issuance task, the applicant, i.e., participant 1, selects collaborative strategy S1: {Trigger condition: Submit application; Action: Call API to push 5 core fields such as guarantee amount and term to the bank, i.e., participant 2; Timeout setting: 24 hours}. By parsing the parameters of the coordination strategy, cross-system interface calls are executed to complete the coordination.

[0028] Specifically, each player's strategy space refers to the set of all possible strategies they can choose in a given situation. The construction of the strategy space includes operational strategies, cooperative strategies, and risk-taking strategies. Operational strategies refer to the specific actions taken by players based on task requirements, directly related to the efficiency, quality, and effectiveness of task execution. Cooperative strategies involve how players coordinate and cooperate to achieve common goals, reduce conflict, and optimize collaboration. Risk-taking strategies relate to how players assess and assume potential risks in the task, primarily considering how to balance gains and risks.

[0029] In a game, the authority of a player at a given stage refers to the rights and capabilities they can exercise when performing a task. Authority is represented as the constraint boundary of the set of available strategies, meaning that the strategies each player can adopt are limited by the scope of their authority. In other words, authority defines the feasible range of strategies a player can choose when selecting actions, cooperating, or taking risks; players can only choose within the set of strategies defined by their authority.

[0030] A multi-agent benefit evaluation model is constructed based on the risk-reward and compliance constraints of the current task status. The multi-agent benefit evaluation model calculates the stage benefit value of each game participant under different strategy combinations by evaluating task completion efficiency, risk exposure degree and coordination cost.

[0031] Furthermore, this application also includes: classifying the risk sensitivity of the task execution status and establishing the classification result; configuring the enhancement or penalty adjustment weight of the stage benefit according to the classification result, and updating the stage benefit value according to the configuration result.

[0032] Specifically, constructing a multi-agent benefit evaluation model based on the risk-reward and compliance constraints of the current task status means that during task execution, a multi-agent model is built to evaluate the performance of various aspects of the task, based on the task status at each stage and considering the risks, expected returns, and compliance requirements faced by the task. Specifically, the function for calculating the stage benefit value of game participant i at the current stage is defined as: U i =w1·E i -w2·R i -w3·C i Where w1, w2, and w3 are preset weighting coefficients, and w1 + w2 + w3 = 1. E i R is a metric for task completion efficiency. i C is an indicator of the degree of risk exposure. i This is a collaborative cost indicator. Among them, E... i E i =N done / N total ×[1-(T now -T start ) / (T deadline -T start )]. Where, N done / N represents the number of subtasks completed by the participants. total T represents the total number of subtasks. now T represents the current time. start T is the start time of the phase. deadline The deadline for compliance constraints. R i =P risk ×Amount guarantee Among them, Amountguarantee P represents the business amount involved in the current guarantee task. risk This represents the probability of default, which is directly obtained by mapping the risk score returned by the risk control engine after the risk-taking strategy selected by the participant invokes it. i =α·N msg +β·T wait Among them, N msg T represents the number of cross-entity message interactions generated when executing a collaboration strategy. wait α and β are adjustment coefficients representing the cumulative time spent waiting for responses from other parties. For example, during the guarantee issuance and review stage, participant B, i.e., the guarantor, currently selects the following strategy combination: operational strategy: automatic compliance verification; collaboration strategy: sharing only basic information. This results in the need for secondary communication, N. msg =5,T wait =2, risk-taking strategy selection requires no credit enhancement review, risk control engine returns a high default probability P. risk =0.8. Current guarantee amount (Amount) guarantee The budget is 1 million yuan, the task progress is 50%, and the time schedule is 20%. Substituting into the model, the efficiency is derived as follows: E = 0.5 × (1 0.2) = 0.4; Risk R = 0.8 × 100 = 80. The risk value can be normalized. Assuming the maximum risk exposure is 200, then the normalized R... norm =80 / 200=0.4; Collaboration cost C=0.1×5+0.05×2=0.6 (assuming normalization is 0.6); Then, the stage profit value U of participant B under this strategy combination is... B =0.5×0.4 0.3 × 0.4 0.2 × 0.6 = 0.2 0.12 0.12= 0.04. By traversing other strategy combinations in the effective strategy subspace, such as changing the risk strategy to deep auditing, and recalculating the payoff, the strategy combination that maximizes the stage payoff is finally output as the stage's optimal strategy combination. Furthermore, the different interests and constraints of multiple participants in executing the task are considered. By evaluating task completion efficiency, risk exposure level, and collaboration cost, the payoff obtained by each game participant under different strategy combinations is measured. Task completion efficiency refers to the time and resource consumption required to achieve the predetermined goal, risk exposure level indicates the magnitude of the risk that may be faced during the task, and collaboration cost is related to the costs incurred by multi-party cooperation.

[0033] The risk sensitivity classification of task execution status involves assessing the risks of a task under different states and categorizing these risks into different levels. Each level corresponds to a different level of risk, and the classification results are used for subsequent benefit evaluation. This ensures that risk factors at each stage of task execution can be reasonably identified and managed, thereby providing a basis for benefit assessment.

[0034] The allocation of enhancement or penalty adjustment weights for stage payouts based on the classification results refers to adjusting the payouts of each game participant according to different risk levels. If the risk is high, the penalty adjustment weight may be increased, reducing payouts; conversely, if the risk is low, the payout adjustment weight may be increased. This adjustment mechanism ensures that the payouts at each stage of the task more accurately reflect the actual risk level, thereby ensuring the fairness and rationality of payout evaluation.

[0035] Based on the strategy space and the constraint boundary, the optional strategy set of each game participant is pruned to establish an effective strategy subspace. The effective strategy subspace is then input into a multi-agent payoff evaluation model to output the optimal strategy combination for the stage.

[0036] Furthermore, this application also includes: constructing a strategy interaction matrix based on the effective strategy subspace of each game participant, wherein the strategy interaction matrix is ​​used to describe the mutual influence relationship between strategy combinations of different game participants; inputting the strategy interaction matrix into the multi-agent payoff evaluation model to calculate the payoff response value corresponding to each strategy combination under the current permission constraint boundary; performing payoff strategy adjustment iteration based on the payoff response value and the mapped subject characteristics of the game participants; performing iterative update analysis based on the payoff strategy adjustment iteration result, and outputting the optimal strategy combination for the stage.

[0037] Furthermore, this application also includes: extracting features of the game participants and establishing subject features, the subject features including risk preference features, collaboration dependency features, and authority level features; mapping the subject features to payoff response adjustment coefficients, the payoff response adjustment coefficients being used to characterize the sensitivity of different game participants to payoff changes; calculating the deviation between the payoff response value of each game participant in the current round and the historical payoff benchmark value, generating a first adjustment constraint; using the payoff response adjustment coefficients as a second adjustment constraint, configuring strategy adjustment increments according to the first adjustment constraint and the second adjustment constraint, and using the strategy adjustment increments to perform payoff strategy adjustment iterations.

[0038] Furthermore, this application also includes: calculating a payoff-risk coupling index based on the payoff response value and stage risk exposure value of each game participant under the current strategy combination, wherein the payoff-risk coupling index is used to characterize the balance between the increase in payoff and the increase in risk brought about by strategy adjustment; mapping the payoff-risk coupling index to a coupling adjustment factor, and adjusting the magnitude and direction of the incremental strategy adjustment based on the coupling adjustment factor; continuously monitoring the payoff-risk coupling index during the iteration process, and determining that the iteration has reached convergence when the incremental change of the strategy adjustment meets a preset stability threshold, and outputting the stage optimal strategy combination.

[0039] Furthermore, this application also includes: obtaining the historical strategy sequence, payoff response change trend, and stage risk exposure characteristics of each game participant; performing strategy prediction for each game participant in the next round of iteration testing; establishing predictive strategy behavior; using the predicted strategy behavior to perform strategy correction of the strategy adjustment increment; establishing predictive compensation strategy adjustment increment; and using the predictive compensation strategy adjustment increment to perform payoff strategy adjustment iteration.

[0040] Specifically, based on the strategy space and constraint boundaries, the optional strategy set for each game participant is pruned. This means that when constructing the strategy space for the game participants, the constraint boundaries limit the range of strategy choices for each participant. The pruning process ensures that game participants can only choose effective strategies that meet the actual requirements by removing strategies that do not meet the current task requirements or constraints. The effective strategy subspace refers to the set of all effective strategies that game participants can choose after pruning. Inputting the effective strategy subspace into a multi-agent payoff evaluation model aims to evaluate the payoff of each game participant under different strategy combinations, thereby outputting the optimal strategy combination and determining the most suitable strategy choice for the current stage.

[0041] Specifically, a strategy interaction matrix is ​​constructed based on the effective strategy subspace of each game participant. This matrix describes the mutual influence relationships between different strategy combinations of the game participants. The strategy interaction matrix is ​​a multi-dimensional mathematical tool that clearly reveals the potential interactions and outcomes when different participants choose different strategies. It reflects the dependencies and interactions between strategy choices, providing a basis for subsequent strategy adjustments.

[0042] The strategy interaction matrix is ​​input into the multi-agent benefit evaluation model to calculate the benefit response value corresponding to each strategy combination under the current permission constraints. This means that, given the permission and strategy interaction conditions, the agent benefit evaluation model will calculate the benefit response value for each strategy combination based on its impact on the benefit. This calculation result reflects the effectiveness and benefit level of each strategy combination at the current task stage.

[0043] Feature extraction and principal characterization of game participants involves extracting key characteristics from their behavior and background, and constructing a model reflecting their attributes. Principal characterization includes risk preference, collaboration dependency, and authority level. Risk preference reflects a player's decision-making tendencies when facing risk, quantifying their attitude and tolerance for risk. Collaboration dependency indicates the degree of dependence a player has on other players during task execution; players with higher levels of collaboration tend to choose more cooperative strategies. Authority level describes the scope and level of authority a player possesses within a multi-level authority system, determining the range of strategies they can employ and the freedom of decision-making in the task.

[0044] Mapping subject features to payoff response modifiers means generating a modifier through a quantitative model of subject features. The payoff response modifier characterizes the sensitivity of different game participants to changes in payoffs. Through this mapping relationship, the degree to which different game participants react to payoff changes under different strategies can be quantified; participants with higher sensitivity will react more strongly to payoff changes.

[0045] The first adjustment constraint is generated by calculating the deviation between each player's payoff response in the current round and its historical benchmark payoff. This means comparing each player's payoff response under the current strategy combination with its historical performance, calculating the difference, and generating an adjustment constraint based on this difference. This adjustment constraint is used to adjust the players' behavior and strategies to ensure the reasonableness and stability of their payoff levels.

[0046] Using the payoff response adjustment coefficient as the second adjustment constraint, and configuring the strategy adjustment increment based on the first and second adjustment constraints, means that when adjusting the strategies of game participants, the strategy increment is adjusted in conjunction with both the first and second adjustment constraints. The payoff response adjustment coefficient as the second constraint ensures that the sensitivity of game participants is considered during strategy adjustment, while the first adjustment constraint ensures the rationality of the strategy adjustment. The strategy adjustment increment can be configured more precisely, ensuring optimal payoff during the adjustment process.

[0047] Based on the payoff response values ​​and stage risk exposure values ​​of each player under the current strategy combination, a payoff-risk coupling index is calculated. This means that during the execution of the game participants' strategy combinations, the payoff response value and corresponding risk exposure value of each participant are considered, and a comprehensive evaluation index is calculated to characterize the balance between the increase in payoff and the increase in risk brought about by the strategy adjustment. The increase in payoff refers to the additional gains obtained after the strategy adjustment, while the increase in risk refers to the additional risks brought about by the strategy adjustment. By calculating the payoff-risk coupling index, the balance relationship can be quantified, providing guidance for subsequent strategy adjustments.

[0048] Mapping the return-risk coupling index to a coupling adjustment factor means generating a coupling adjustment factor through analysis of the return-risk coupling index. This factor is used to adjust the magnitude and direction of the strategy's adjustment increment. The coupling adjustment factor determines the degree and direction of the strategy adjustment based on the balance between return and risk, thus ensuring that while increasing returns, risk is not excessively increased. By adjusting the magnitude and direction of the strategy adjustment, the coupling adjustment factor helps achieve strategy optimization.

[0049] During the iteration process, the payoff-risk coupling index is continuously monitored. When the incremental change in strategy adjustment meets a preset stability threshold, the iteration is considered to have converged. This means that during the iterative process of strategy adjustment, the changes in the payoff-risk coupling index are continuously monitored, and when the incremental change in strategy adjustment meets the set stability threshold, the strategy adjustment process is considered to have stabilized and converged to the optimal state. At this point, the optimal strategy combination for the current stage is output, indicating that the strategy combination that best balances payoff and risk has been found and is suitable for the current task stage.

[0050] Furthermore, obtaining the historical strategy sequences, payoff response trends, and stage-specific risk exposure characteristics of each game participant means collecting records of their strategy choices during past task executions and analyzing the trends in payoff response values ​​and the participants' risk exposure during task execution. The historical strategy sequence refers to the chronological set of strategies adopted by the participants in the past; the payoff response trend reflects the changing impact of strategy choices on payoffs; and the stage-specific risk exposure characteristics reflect the degree of risk borne by the participants at different task stages. This provides a basis for strategy prediction in the next iteration, forecasting the possible strategic behaviors of game participants in future task stages.

[0051] Predicting the strategies of each game participant in the next iteration refers to forecasting their strategy choices in the next iteration based on historical data, trends in payoff responses, and risk exposure characteristics. This helps predict the strategies participants might adopt in the future and provides guidance for strategy adjustments.

[0052] Utilizing predicted strategic behavior to execute policy adjustment increments means making corresponding adjustments to the policy adjustment increments based on the prediction results. The policy adjustment increment refers to the fine-tuning of a player's current strategy, and predicting strategic behavior provides the basis for adjusting the increment. By correcting the policy increment, the player's strategy can be adjusted more precisely, thereby optimizing the task execution effect.

[0053] Establishing a predictive compensation strategy adjustment increment refers to setting up a compensation mechanism during the strategy adjustment process by considering the impact of predictive behavior on the adjustment increment. This mechanism compensates for potential negative impacts caused by changes in strategy selection. The compensation strategy adjustment increment ensures that the optimization of the strategy during the iteration process is not adversely affected by prediction errors or sudden changes.

[0054] Utilizing predictive compensation strategy adjustment increments to iterate and adjust the payoff strategy refers to the continuous optimization and adjustment of the game participants' strategies based on predictive compensation strategy adjustment increments. In each iteration, by executing compensatory adjustment increments, the strategy is continuously revised to ensure that the final selected strategy combination maximizes the overall payoff of the task while balancing risk.

[0055] Based on the iterative results of adjusting the profit strategy, iterative update analysis is performed, outputting the optimal strategy combination for each stage. This means that during the iterative process of adjusting the profit strategy, the adjustment results are continuously analyzed, and the strategy is updated based on the analysis results, ultimately yielding the optimal strategy combination for each stage. This ensures that the strategy selection during task execution is always in an optimal state, maximizing the overall profit of the task.

[0056] Based on the optimal strategy combination for the described stage, the permission allocation strategy is dynamically configured to perform multi-level permission collaborative management.

[0057] Furthermore, this application also includes: during the execution of the guarantee task, monitoring the deviation between the actual strategic behavior of each game participant and the preset strategy, and establishing a deviation dataset; determining whether the deviation dataset meets the preset deviation threshold, and if it meets the preset deviation threshold, then issuing a deviation anomaly warning.

[0058] Furthermore, this application also includes: configuring policy weight feedback based on the deviation dataset; and using the policy weight feedback to perform real-time update management of the permission allocation policy.

[0059] Furthermore, this application also includes: configuring a response detection window for each game participant; if the collaborating entity does not have a response in the corresponding response detection window, sending an activation request to the collaborating entity; and performing multi-level permission collaborative management based on the response result of the collaborating entity.

[0060] Specifically, during the execution of a guarantee task, monitoring the deviation between the actual strategic behavior of each game participant and the preset strategy means continuously tracking the actual strategies adopted by the participants during task execution and comparing them with the pre-set ideal strategy. Deviation refers to the degree of difference between the actual strategy and the preset strategy, reflecting the extent to which the game participants have failed to strictly follow the preset plan or rules when executing the task. By monitoring the deviation, strategic deviations during task execution can be detected and adjusted in a timely manner.

[0061] Establishing a deviation dataset refers to collecting and organizing the deviation data from each monitoring session to form a comprehensive dataset. This dataset includes the strategy deviations of all game participants during the execution process, providing necessary data support for subsequent analysis and decision-making. By establishing a deviation dataset, patterns and trends of strategy deviations among game participants can be recorded and analyzed.

[0062] The process involves determining whether the deviation dataset meets a preset deviation threshold. If it does, an anomaly alert is triggered. This means analyzing the collected deviation dataset and determining whether the strategy deviation exceeds an acceptable range based on the set threshold. The preset deviation threshold represents the maximum tolerable strategy deviation during task execution. If the deviation between the actual strategy and the preset strategy exceeds the threshold, an anomaly is detected, triggering an anomaly alert mechanism to notify relevant personnel for timely intervention. Configuring strategy weight feedback based on the deviation dataset means that after analyzing the dataset, each strategy is assigned a weight based on the deviation between the actual strategy and the preset strategy. The weight reflects the severity of the strategy deviation; the greater the deviation, the higher the weight of the corresponding strategy. Configuring strategy weight feedback quantifies the degree of deviation for different strategies, providing a basis for subsequent permission allocation and strategy adjustments.

[0063] Real-time update management of permission allocation strategies using strategy weight feedback refers to adjusting the permission allocation strategy in real time based on the configured strategy weight feedback. The permission allocation strategy determines the permissions and strategy selection range available to game participants during task execution. When the strategy deviates significantly, permissions may be adjusted based on feedback to limit the strategy selection range of participants with significant deviations, or appropriate incentives may be provided to guide them towards the predetermined strategy. Real-time update management ensures that permission allocation during task execution remains consistent with task requirements and objectives, thereby improving the efficiency and stability of task execution.

[0064] Furthermore, configuring a response detection window for each game participant means setting a specific time window for each participant during task execution to monitor their response behavior within that time period. The response detection window is a pre-defined time range within which game participants must respond to relevant task information or requirements. By configuring response detection windows, it is ensured that game participants make necessary responses within the specified time, thereby avoiding task execution problems caused by delays or non-response.

[0065] If no response is received from a collaborating entity within the corresponding response detection window, an activation request is sent to that entity or other relevant collaborating entities. This means that in a collaborative task, when a player fails to respond within a predetermined response detection window, an activation request is sent to that player or other relevant collaborating entities. The activation request is a notification mechanism designed to remind collaborating entities to take action and ensure the smooth progress of the task. If an entity fails to respond in time, the activation request will prompt it to take necessary measures promptly.

[0066] Multi-level permission collaborative management, based on the responses of collaborating entities, means further adjusting and executing multi-level permission collaborative management according to their responses to activation requests. Multi-level permission collaborative management is a dynamic adjustment mechanism that rationally allocates and adjusts the permissions and responsibilities of participating parties during task execution based on their responses. If a collaborating entity responds promptly and executes the corresponding task, the permission configuration is optimized based on its response; if it does not respond, corresponding permission adjustments or incentives are implemented based on the non-response situation to ensure the achievement of task objectives.

[0067] In summary, the multi-level permission collaborative management method for guarantee systems provided in this application has the following technical effects: by realizing the dynamic monitoring and adjustment of the strategies and permissions of game participants, it ensures that the strategies and permissions can be optimized in real time according to the actual deviation during the task execution process, thereby improving the system's collaborative efficiency, optimizing the task execution process, and enhancing the stability and accuracy of task execution.

[0068] Example 2: Based on the same inventive concept as the multi-level permission collaborative management method for a guarantee system described in the foregoing examples, this application also provides a multi-level permission collaborative management platform for a guarantee system. Please refer to the appendix. Figure 2The system includes: a multi-stage dynamic game process construction module 1, used to abstract multiple cooperating entities participating in the guarantee task as game participants after the guarantee task is generated within the guarantee system, and to construct the execution process of the guarantee task as a multi-stage dynamic game process, wherein each stage corresponds to a task state node; a strategy space construction module 2, used to construct a strategy space for each game participant, wherein the strategy space includes operation strategies, cooperation strategies, and risk-taking strategies, and to represent the permissions possessed by each game participant in the current stage as the constraint boundary of the optional strategy set; and a multi-entity benefit evaluation model construction module 3, used to evaluate the risk based on the current task state. A multi-party benefit evaluation model is constructed based on risk returns and compliance constraints. This model calculates the stage benefit value of each game participant under different strategy combinations by evaluating task completion efficiency, risk exposure, and collaboration costs. The stage optimal strategy combination output module 4 is used to prune the optional strategy set of each game participant based on the strategy space and the constraint boundary, establish an effective strategy subspace, input the effective strategy subspace into the multi-party benefit evaluation model, and output the stage optimal strategy combination. The multi-level permission collaboration management module 5 is used to dynamically configure permission allocation strategies according to the stage optimal strategy combination and execute multi-level permission collaboration management.

[0069] Furthermore, the multi-level permission collaborative management platform for the guarantee system is also used for: constructing a strategy interaction matrix based on the effective strategy subspace of each game participant, wherein the strategy interaction matrix is ​​used to describe the mutual influence relationship between strategy combinations of different game participants; inputting the strategy interaction matrix into the multi-subject payoff evaluation model to calculate the payoff response value corresponding to each strategy combination under the current permission constraint boundary; performing payoff strategy adjustment iteration based on the payoff response value and the subject characteristics of the mapped game participants; performing iterative update analysis based on the payoff strategy adjustment iteration result, and outputting the optimal strategy combination for the stage.

[0070] Furthermore, the multi-level permission collaborative management platform for the guarantee system is also used for: extracting features of game participants and establishing subject features, including risk preference features, collaborative dependency features, and permission level features; mapping the subject features to a payoff response adjustment coefficient, which is used to characterize the sensitivity of different game participants to payoff changes; calculating the deviation between the payoff response value of each game participant in the current round and the historical payoff benchmark value, and generating a first adjustment constraint; using the payoff response adjustment coefficient as a second adjustment constraint, configuring a strategy adjustment increment according to the first adjustment constraint and the second adjustment constraint, and using the strategy adjustment increment to execute payoff strategy adjustment iterations.

[0071] Furthermore, the multi-level permission collaborative management platform for the guarantee system is also used to: calculate a payoff-risk coupling index based on the payoff response value and stage risk exposure value of each game participant under the current strategy combination, wherein the payoff-risk coupling index is used to characterize the balance between the increase in payoff and the increase in risk brought about by strategy adjustment; map the payoff-risk coupling index to a coupling adjustment factor, and adjust the magnitude and direction of the incremental strategy adjustment based on the coupling adjustment factor; continuously monitor the payoff-risk coupling index during the iteration process, and when the incremental change magnitude of the strategy adjustment meets the preset stability threshold, determine that the iteration has reached convergence and output the optimal strategy combination for the stage.

[0072] Furthermore, the multi-level permission collaborative management platform for the guarantee system is also used to: obtain the historical strategy sequence, payoff response change trend, and stage risk exposure characteristics of each game participant; execute strategy prediction for each game participant in the next round of iteration testing; establish predictive strategy behavior; use the predicted strategy behavior to execute strategy correction of the strategy adjustment increment; establish predictive compensation strategy adjustment increment; and use the predictive compensation strategy adjustment increment to execute payoff strategy adjustment iteration.

[0073] Furthermore, the multi-level permission collaborative management platform for the guarantee system is also used to: monitor the deviation between the actual strategic behavior of each game participant and the preset strategy during the execution of the guarantee task, and establish a deviation dataset; determine whether the deviation dataset meets the preset deviation threshold, and if it meets the preset deviation threshold, issue a deviation anomaly warning.

[0074] Furthermore, the multi-level permission collaborative management platform for the guarantee system is also used for: configuring policy weight feedback based on the deviation dataset; and using the policy weight feedback to perform real-time update management of the permission allocation policy.

[0075] Furthermore, the multi-level permission collaborative management platform for the guarantee system is also used to: configure a response detection window for each game participant; if the collaborating entity does not have a response in the corresponding response detection window, send an activation request to the collaborating entity; and perform multi-level permission collaborative management based on the response result of the collaborating entity.

[0076] Furthermore, the multi-level permission collaborative management platform for the guarantee system is also used for: classifying the risk sensitivity of the task execution status and establishing the classification results; configuring the enhancement or penalty adjustment weights of the stage benefits according to the classification results, and updating the stage benefit values ​​according to the configuration results.

[0077] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The multi-level permission collaborative management method and specific example for a guarantee system described in the foregoing embodiment one are also applicable to the multi-level permission collaborative management platform for a guarantee system described in this embodiment. Through the foregoing detailed description of the multi-level permission collaborative management method for a guarantee system, those skilled in the art can clearly understand the multi-level permission collaborative management platform for a guarantee system described in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0078] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0079] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. A multi-level permission collaborative management method for a guarantee system, characterized in that, The method includes: After the guarantee task is generated within the guarantee system, the multiple cooperating entities participating in the guarantee task are abstracted as game participants, and the execution process of the guarantee task is constructed as a multi-stage dynamic game process, where each stage corresponds to a task state node. A strategy space is constructed for each game participant, which includes operational strategies, collaborative strategies, and risk-taking strategies. The permissions that each game participant has at the current stage are represented as the constraint boundary of the set of optional strategies. A multi-agent benefit evaluation model is constructed based on the risk-reward and compliance constraints of the current task status. The multi-agent benefit evaluation model calculates the stage benefit value of each game participant under different strategy combinations by evaluating task completion efficiency, risk exposure degree and coordination cost. Based on the strategy space and the constraint boundary, the optional strategy set of each game participant is pruned to establish an effective strategy subspace. The effective strategy subspace is then input into a multi-agent payoff evaluation model to output the optimal strategy combination for the stage. Based on the optimal strategy combination for the described stage, the permission allocation strategy is dynamically configured to perform multi-level permission collaborative management.

2. The multi-level permission collaborative management method for a guarantee system as described in claim 1, characterized in that, The optimal strategy combination in the output phase includes: A strategy interaction matrix is ​​constructed based on the effective strategy subspace of each game participant. The strategy interaction matrix is ​​used to describe the mutual influence relationship between different strategy combinations of game participants. Input the strategy interaction matrix into the multi-agent benefit evaluation model to calculate the benefit response value corresponding to each strategy combination under the current permission constraint boundary; The payoff strategy is adjusted iteratively based on the payoff response value and the main characteristics of the game participants mapped thereon; Based on the results of the iterative adjustment of the profit strategy, perform iterative update analysis and output the optimal strategy combination for the stage.

3. A multi-level permission collaborative management method for a guarantee system as described in claim 2, characterized in that, Based on the payoff response value and the mapped subject characteristics of the game participants, perform payoff strategy adjustment iterations, including: The feature extraction of the game participants is performed to establish subject features, which include risk preference features, collaboration dependency features, and authority level features. The main features are mapped to payoff response adjustment coefficients, which are used to characterize the sensitivity of different game participants to changes in payoffs. Calculate the deviation between the payoff response value of each game participant in the current round and the historical payoff benchmark value, and generate the first adjustment constraint; The revenue response adjustment coefficient is used as the second adjustment constraint. The strategy adjustment increment is configured according to the first adjustment constraint and the second adjustment constraint. The revenue strategy adjustment increment is then used to perform the revenue strategy adjustment iteration.

4. A multi-level permission collaborative management method for a guarantee system as described in claim 3, characterized in that, The strategy of adjusting incremental execution of the benefit strategy adjustment iteration also includes: Based on the payoff response value and stage risk exposure value of each game participant under the current strategy combination, a payoff-risk coupling index is calculated. The payoff-risk coupling index is used to characterize the balance between the increase in payoff and the increase in risk brought about by strategy adjustment. The return-risk coupling index is mapped to a coupling adjustment factor, and the magnitude and direction of the strategy adjustment increment are adjusted based on the coupling adjustment factor. During the iteration process, the return-risk coupling index is continuously monitored. When the incremental change of the strategy adjustment meets the preset stability threshold, the iteration is judged to have reached convergence, and the optimal strategy combination of the stage is output.

5. A multi-level permission collaborative management method for a guarantee system as described in claim 4, characterized in that, The strategy of adjusting incremental execution of the benefit strategy adjustment iteration also includes: Obtain the historical strategy sequence, payoff response trend, and stage risk exposure characteristics of each game participant; perform strategy prediction for each game participant in the next round of iteration testing; and establish predictive strategy behavior. The predicted strategy behavior is used to execute the strategy adjustment increment, and a predictive compensation strategy adjustment increment is established. The predictive compensation strategy is used to adjust the incremental execution revenue strategy iteratively.

6. A multi-level permission collaborative management method for a guarantee system as described in claim 1, characterized in that, Based on the optimal strategy combination for the described stage, a dynamic configuration of the permission allocation strategy is performed to execute multi-level collaborative permission management, including: During the execution of the guarantee task, the deviation between the actual strategic behavior of each game participant and the preset strategy is monitored, and a deviation dataset is established. Determine whether the deviation dataset meets the preset deviation threshold. If it does, issue a deviation anomaly warning.

7. A multi-level permission collaborative management method for a guarantee system as described in claim 6, characterized in that, If the preset deviation threshold is met, the following further applies: Configure strategy weight feedback based on the deviation dataset; The policy weight feedback is used to manage the real-time update of the permission allocation policy.

8. A multi-level permission collaborative management method for a guarantee system as described in claim 1, characterized in that, Based on the optimal strategy combination for the aforementioned stage, the permission allocation strategy is dynamically configured to perform multi-level permission collaborative management, which also includes: Configure a response detection window for each game participant; If no response exists in the corresponding response detection window, an activation request is sent to the collaborating entity. Multi-level permission collaborative management is implemented based on the response results of the collaborating entities.

9. A multi-level permission collaborative management method for a guarantee system as described in claim 1, characterized in that, Calculating the stage payoffs of each player under different strategy combinations also includes: Classify the risk sensitivity of the task execution status and establish the classification results; Based on the classification results, configure the enhancement or penalty adjustment weights for the stage benefits, and update the stage benefit values ​​according to the configuration results.

10. A multi-level permission collaborative management platform for a guarantee system, characterized in that, The steps for implementing a multi-level access control method for a guarantee system according to any one of claims 1 to 9 include: The multi-stage dynamic game process construction module is used to abstract multiple cooperating entities participating in the guarantee task as game participants after the guarantee task is generated in the guarantee system, and to construct the execution process of the guarantee task as a multi-stage dynamic game process, wherein each stage corresponds to a task state node. The strategy space construction module is used to construct a strategy space for each game participant. The strategy space includes operational strategies, collaborative strategies, and risk-taking strategies, and represents the permissions that each game participant has at the current stage as the constraint boundary of the optional strategy set. The multi-agent benefit evaluation model construction module is used to construct a multi-agent benefit evaluation model based on the risk-return and compliance constraints of the current task status. The multi-agent benefit evaluation model calculates the stage benefit value of each game participant under different strategy combinations by evaluating task completion efficiency, risk exposure degree and collaboration cost. The stage-optimal strategy combination output module is used to prune the optional strategy set of each game participant based on the strategy space and the constraint boundary, establish an effective strategy subspace, input the effective strategy subspace into the multi-agent payoff evaluation model, and output the stage-optimal strategy combination. The multi-level permission collaborative management module is used to dynamically configure permission allocation strategies based on the optimal strategy combination of the aforementioned stages, and to perform multi-level permission collaborative management.

Images