Automatic auditing business process control method based on rule optimization

Abstract

The invention discloses an automatic auditing business process control method based on rule optimization. The method comprises the following steps: S1, generating a unified auditing data pool; s2, constructing a multi-layer rule engine, and generating an initial rule set; s3, initializing an improved bee colony optimization algorithm, and dynamically defining a search space of a population; s4, dynamically optimizing the initial rule set by using an improved bee colony optimization algorithm, adjusting a single rule parameter in a search space in a worker bee stage, selecting an optimal rule direction in an observation bee stage, and introducing collaborative bees to optimize conflicts among multiple rules, and re-initializing a low-diversity population in a reconnaissance bee stage, and generating an optimized rule set; s5, generating a rule execution plan, adjusting the node priority of the auditing process and optimizing a task path; and S6, realizing continuous optimization iteration of the rule set and the audit process path based on implementation feedback of the rule execution plan. According to the invention, the multi-layer rule engine and the improved bee colony optimization algorithm realize automatic auditing business process control.

Description

Technical Field

[0001] The present invention relates to the technical field of audit business process control, and in particular to an automated audit business process control method based on rule optimization. Background Art

[0002] With the rapid development of information technology and big data analysis, audit business is gradually transforming towards digitalization and automation, and the traditional audit process that relies on manual work is being gradually replaced by intelligent systems. In existing technologies, automated audit systems usually manage and control audit business processes through rule engines, which process and analyze audit data through predefined rule sets. However, most existing rule engines are based on static rule sets, lack dynamic adaptability and intelligent optimization capabilities, and are difficult to meet the changing business needs in complex audit scenarios.

[0003] The rule engine in the prior art is usually composed of a single rule layer, which is directly applied to all audit scenarios, ignoring the differences between different audit businesses. The scope of application of static rules is fixed, and the logic and applicable conditions of the rules cannot be adjusted according to real-time business scenarios, which leads to problems such as low rule execution efficiency and unreasonable audit node priority allocation. In addition, the prior art has weak conflict handling capabilities between rules, and logical conflicts or redundancy between rules may cause deviations in the execution results of the audit process. For example, when multiple rules are applied to the same node, the allocation of rule priorities usually relies only on predefined weights or simple relevance scores, ignoring the dynamic impact of rule execution history and real-time feedback data on rule effects.

[0004] In terms of audit process optimization, most existing technologies use static path planning to generate task paths based on established rules and node priorities. However, static planning cannot cope with the dynamic changes in node priorities during the audit process, especially when high-priority tasks emerge. It is difficult for existing systems to reallocate resources to ensure the priority completion of key node tasks. In addition, task path planning usually only considers node priority and execution order, while ignoring multi-dimensional factors such as task complexity, resource consumption, and execution time. This not only leads to the overall inefficiency of the audit process, but may also cause serious imbalance in resource allocation.

[0005] In terms of rule optimization and intelligent process control, existing technologies mostly use simple optimization algorithms, such as genetic algorithms and particle swarm algorithms. These algorithms often have problems such as slow convergence and easy to fall into local optimality when facing complex audit scenarios with multiple rules, multiple nodes, and multiple targets, making it difficult to quickly generate high-quality rule optimization solutions. In addition, most existing optimization algorithms ignore real-time feedback data during the rule execution process, and the optimization process lacks dynamic adjustment capabilities. For example, when the scope of application of the rule or the triggering conditions change, the existing system cannot adjust the optimization direction in real time, resulting in insufficient adaptability of the rule set and insufficient accuracy of process control.

[0006] Another significant problem is that existing technologies lack a mechanism for continuous optimization and iteration of rules and processes. The execution of audit processes is usually static and consistent, and the update of rule sets often relies on manual intervention or regular maintenance, which cannot achieve dynamic adaptation of rules and processes. Since the rule execution effect and the real-time status of process nodes are not fully utilized, it is difficult for existing systems to form an efficient closed-loop optimization mechanism, and long-term use may cause rule failure or reduced process efficiency. In addition, existing technologies are also insufficient in the transparency of rule execution plans, and users cannot effectively track the logic and process of rule optimization, which further limits the application of the system in highly complex scenarios.

[0007] Therefore, how to provide an automated audit business process control method based on rule optimization is an issue that technical personnel in this field urgently need to solve. Summary of the invention

[0008] One purpose of the present invention is to propose an automated audit business process control method based on rule optimization. The present invention adopts a multi-layer rule engine and an improved bee colony optimization algorithm to construct an automated audit business process control method. Through dynamic adjustment of rules, node priority optimization and optimal generation of task paths, intelligent control and global optimization of the audit process are achieved. The present invention has the advantages of strong rule flexibility, high dynamic process adaptability, high conflict handling efficiency and fast optimization convergence speed. It can quickly respond to the changing needs of complex audit scenarios and significantly improve the execution efficiency of audit business and the stability of the system.

[0009] According to an embodiment of the present invention, a method for controlling an automated audit business process based on rule optimization includes the following steps: S1. Collect audit data, perform data cleaning, anomaly detection and format standardization, and generate a unified audit data pool; S2. Based on the unified audit data pool, a multi-layer rule engine is constructed, wherein the multi-layer rule engine includes a basic rule layer, a dynamic rule layer, and a personalized rule layer, and an initial rule set is generated; S3, initializing the improved bee colony optimization algorithm, generating a bee population, each bee represents a set of rule parameter combinations, and dynamically defining the search space of the bee population; S4. Use the improved bee colony optimization algorithm to dynamically optimize the initial rule set. In the worker bee stage, adjust the parameters of a single rule in the search space. In the observation bee stage, select the optimal rule direction based on the rule execution effect, and introduce collaborative bees to optimize the conflicts between multiple rules. In the scout bee stage, monitor the population diversity, reinitialize the low-diversity population, and generate the optimized rule set. S5. Input the optimized rule set into the multi-layer rule engine to generate a rule execution plan, dynamically adjust the node priority of the audit process according to the rule execution plan, and optimize the task path; S6. Based on the implementation feedback of the rule execution plan, adjust the rule parameters and priorities, and synchronize them to the basic rule layer and personalized rule layer to achieve continuous optimization and iteration of the rule set and audit process path.

[0010] Optionally, the S2 specifically includes: S21. Based on the unified audit data pool, build the basic rule layer according to industry standards and audit requirements; associate the rules with the audit scenarios through preset trigger conditions and execution logic, and generate a basic rule library; S22. Using real-time data stream analysis technology, dynamically updated audit data is processed, trigger conditions and execution actions of rules are generated according to real-time business scenarios, rule logic is dynamically adjusted, and a dynamic rule layer is constructed; the dynamic rule layer is used to respond to real-time changes in audit data and generate a rule base that matches the current audit scenario; S23. Analyze the historical records in the unified audit data pool through data mining technology, and combine the rule logic set by the enterprise with the audit data to form a personalized rule layer; the personalized rule layer is used to dynamically adapt to the needs of different enterprises; S24. Integrate the rules of the basic rule layer, dynamic rule layer and personalized rule layer in a hierarchical manner, optimize the logical consistency between the rules, remove duplicates and adjust the rules with repeated application scopes or conflicting logics, build a multi-layer rule engine, and generate a multi-layer rule set; S25. Dynamically calculate the weight of each rule based on the trigger frequency, importance score and scope of application score: ; in, Indicates The rule weights, Indicates The weight coefficient of the trigger condition, Indicates The trigger frequency of each trigger condition, Indicates the total number of triggers for all trigger conditions. Indicates the total number of trigger conditions. represents the importance score weight coefficient, Indicates the importance score of the current rule. represents the maximum importance score in the rule set, Indicates the applicable scope scoring weight coefficient, Indicates the number of audit scenarios to which the current rule applies. Indicates the total number of all audit scenarios in the rule set. represents the conflict penalty coefficient, represents the total number of conflicting rules, Indicates The number of conflicting fields for conflicting rules, Indicates The total number of fields in the conflicting rules; S26. Prioritize the rules in the multi-layer rule set according to the rule weights, remove low-priority or logically conflicting rules, and generate an initial rule set, which is stored in the multi-layer rule engine.

[0011] Optionally, the S3 specifically includes: S31. By analyzing the distribution of rule parameters in the unified audit data pool, an improved bee colony optimization algorithm is initialized to generate a bee population, wherein each bee represents a set of rule parameter combinations, and the rule parameter combinations include a precondition set, an execution action set, and an applicable scope; S32. Allocate an initial search space for the bee population, where the initial search space is dynamically defined according to the value range of the rule parameter: ; in, Indicates The initial search range for rule parameters, Indicates A set of rule parameters, Indicates The maximum value of a set of rule parameters, Indicates The minimum value of a set of rule parameters, Indicates the total number of rules, represents the adjustment factor, Indicates The standard deviation of a set of rule parameters; S33. Use the fitness function to evaluate the quality of the rule parameter combination of each bee in the bee population: ; in, Indicates The fitness value of the rule parameter combination, , and represents the weight coefficient, Indicates the accuracy of the rule, represents the efficiency of rule execution, Indicates the coverage of the rule; S34. Dynamically adjust the initial search range of the bee population: ; in, Indicates The search range after the rule parameters are adjusted. represents the adjustment coefficient, Indicates The fitness change value of the rule parameter combination, represents the average fitness; S35, optimizing the rule parameter combination of each bee: ; in, Indicates The rules and The coordination of the rules, Indicates The rules and The similarity of the rules, Indicates The rules and The degree of conflict of rules, represents the smoothing factor; S36. Through multiple iterative adjustments, record the rule parameter combination with the best fitness value in each iteration. When the maximum number of iterations is reached or the fitness change is less than a preset threshold, output the optimized bee population and search space, and dynamically define the search space of the bee population.

[0012] Optionally, the S4 specifically includes: S41. In the worker bee stage, the local attributes of the rules are classified and optimized, and the multi-dimensional perturbation mechanism is used to adjust the parameters of each rule: ; in, Indicates New parameter values ​​for the rule, Indicates The current value of the rule, represents the adjustment factor, represents the disturbance factor, represents a random disturbance, and represents two randomly selected rule parameter values; S42, in the observation bee stage, the priority of the rule is optimized through the global evaluation rule execution effect, the parameter distribution is dynamically adjusted, and the global adaptability optimization value is calculated: ; in, Indicates The global adaptability optimization value of the rule, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates the total number of rules; S43, introduce the global logical consistency of the collaborative bee stage optimization rule set, and construct the rule applicability matrix Analyze conflicts between multiple rules: ; in, Indicates The rules and The conflict cost of a rule, Indicates The rules and The similarity of the rules, Indicates The rules and The degree of conflict of rules, Indicates The complexity of the rules, Indicates The complexity of the rules; S44, the scout bee stage uses a multi-objective weighted mechanism to optimize the rule search distribution and reconstruct the rule weights that are below the preset fitness threshold: ; in, Indicates The update weight of each rule is , and represents the weight coefficient, Indicates The global adaptability optimization value of the rules; S45. Combining the results of the worker bee stage, the observer bee stage, the collaborative bee stage, and the scout bee stage, an optimized rule set is generated.

[0013] Optionally, the S5 specifically includes: S51. Input the optimized rule set into the multi-layer rule engine, perform hierarchical analysis, and generate a rule execution plan according to the priority and applicable scope of the rules: ; in, Indicates The priority of a rule, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates the total number of rules; S52. Dynamically adjust the priority of each node in the audit process according to the output results of the rule execution plan: ; in, Indicates The priority of the node, Indicates The rule for The relevance of nodes, represents the total number of rules associated with the node; S53, optimizing the execution path of the audit task, sorting the task path according to the node priority, and generating the optimal task path with the minimum total time cost as the goal; S54, when executing the task path, monitor the implementation of the rule execution plan in real time, record the execution status of the node, and the monitoring content includes the task completion rate and execution time of the node; S55. Based on the feedback from monitoring data, dynamically adjust the rule execution plan to ensure that the node priority is greater than the set threshold The tasks are completed first.

[0014] Optionally, the S6 specifically includes: S61, based on the implementation feedback of the rule execution plan, collect dynamic data generated during the rule execution process, including rule triggering frequency, execution success rate, task completion rate, response time and resource consumption, to form a rule feedback data set; S62, standardize the rule feedback data set and unify the data formats of different dimensions; S63. Analyze the actual execution effect of the rules based on the standardized rule feedback data set, re-rank the importance of the rules based on the execution success rate and task completion rate, and identify conflicting rules; S64. According to the analysis results of the rule execution effect, the rule parameters and priorities are adjusted, and the applicable scope, triggering conditions and execution logic of the rule are redefined; S65, synchronizing the adjusted rules to the basic rule layer and the personalized rule layer, updating the rule set, and regenerating a new rule execution plan according to the adjusted rule priority; S66. Based on the new rule execution plan, optimize and iterate the audit process path, dynamically adjust the priority and task sequence of process nodes, and achieve continuous optimization and iteration of the rule set and audit process path.

[0015] The beneficial effects of the present invention are: First, the present invention solves the problems of fixed rules and insufficient applicability in the prior art by constructing a multi-layer rule engine architecture, including a basic rule layer, a dynamic rule layer, and a personalized rule layer. The basic rule layer meets general audit requirements, the dynamic rule layer can generate and adjust rules according to real-time business scenarios, and the personalized rule layer builds a customized rule library based on corporate historical data and specific needs. The hierarchical architecture of multi-layer rules not only improves the flexibility and adaptability of the rules, but also ensures the logical consistency and accuracy of the rule set through rule integration and conflict deduplication.

[0016] Secondly, the present invention introduces an improved swarm optimization algorithm, which significantly improves the efficiency and quality of rule optimization. By dynamically adjusting the search space, the algorithm can flexibly optimize the rule parameters according to the real-time feedback of the rules, avoiding the problem that the traditional optimization algorithm falls into the local optimality due to the fixed search range. The collaborative bee mechanism ensures the global consistency of the rule set and the rationality of the execution logic by optimizing the conflicts between multiple rules. At the same time, the multi-objective weighted mechanism comprehensively considers the accuracy, execution efficiency and coverage of the rules, realizes the dynamic balance of the rule optimization goals, and enables the rule set to quickly adapt to complex and changing audit scenarios.

[0017] In addition, the present invention constructs an efficient process control mechanism through the dynamic linkage between the rule execution plan and the audit process. Under the guidance of the rule execution plan, the node priority of the audit process can be dynamically adjusted according to the importance of the rules and real-time business needs. The task path is optimized by combining factors such as node complexity, execution efficiency and resource allocation to generate the optimal path with the lowest time cost. At the same time, by real-time monitoring of node status and rule execution, timely adjustment of inefficient nodes and rule parameters, and reallocation of resources to prioritize key tasks, the execution efficiency and flexibility of the audit process are effectively improved.

[0018] Finally, the present invention establishes a continuous optimization and iteration mechanism for rules and processes. By standardizing and analyzing the real-time feedback data of rule execution, the rule parameters and scope of application are continuously optimized, and the basic rule layer and personalized rule layer are updated synchronously. This closed-loop optimization mechanism ensures that the rule set can dynamically adapt to changes in the business environment and needs, avoiding the problems of rule failure and reduced audit efficiency. At the same time, the continuous iterative optimization of the rule execution plan and the audit process path ensures the long-term stability and efficiency of the system, and provides comprehensive and reliable technical support for complex audit scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 This is an overall flow chart of an automated audit business process control method based on rule optimization proposed by the present invention; Figure 2 This is a schematic diagram of the operation of an improved bee colony optimization algorithm for an automated audit business process control method based on rule optimization proposed in the present invention. DETAILED DESCRIPTION

[0020] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0021] refer to Figure 1 and Figure 2 , an automated audit business process control method based on rule optimization, comprising the following steps: S1. Collect audit data, perform data cleaning, anomaly detection and format standardization, and generate a unified audit data pool; S2. Based on the unified audit data pool, a multi-layer rule engine is constructed, wherein the multi-layer rule engine includes a basic rule layer, a dynamic rule layer, and a personalized rule layer, and an initial rule set is generated; S3, initializing the improved bee colony optimization algorithm, generating a bee population, each bee represents a set of rule parameter combinations, and dynamically defining the search space of the bee population; S4. Use the improved bee colony optimization algorithm to dynamically optimize the initial rule set. In the worker bee stage, adjust the parameters of a single rule in the search space. In the observation bee stage, select the optimal rule direction based on the rule execution effect, and introduce collaborative bees to optimize the conflicts between multiple rules. In the scout bee stage, monitor the population diversity, reinitialize the low-diversity population, and generate the optimized rule set. S5. Input the optimized rule set into the multi-layer rule engine to generate a rule execution plan, dynamically adjust the node priority of the audit process according to the rule execution plan, and optimize the task path; S6. Based on the implementation feedback of the rule execution plan, adjust the rule parameters and priorities, and synchronize them to the basic rule layer and personalized rule layer to achieve continuous optimization and iteration of the rule set and audit process path.

[0022] In this implementation, S2 specifically includes: S21. Based on the unified audit data pool, build the basic rule layer according to industry standards and audit requirements; associate the rules with the audit scenarios through preset trigger conditions and execution logic, and generate a basic rule library; S22. Using real-time data stream analysis technology, dynamically updated audit data is processed, trigger conditions and execution actions of rules are generated according to real-time business scenarios, rule logic is dynamically adjusted, and a dynamic rule layer is constructed; the dynamic rule layer is used to respond to real-time changes in audit data and generate a rule base that matches the current audit scenario; S23. Analyze the historical records in the unified audit data pool through data mining technology, and combine the rule logic set by the enterprise with the audit data to form a personalized rule layer; the personalized rule layer is used to dynamically adapt to the needs of different enterprises; S24. Integrate the rules of the basic rule layer, dynamic rule layer and personalized rule layer in a hierarchical manner, optimize the logical consistency between the rules, remove duplicates and adjust the rules with repeated application scopes or conflicting logics, build a multi-layer rule engine, and generate a multi-layer rule set; S25. Dynamically calculate the weight of each rule based on the trigger frequency, importance score and scope of application score: ; in, Indicates The rule weights, Indicates The weight coefficient of the trigger condition, Indicates The trigger frequency of the trigger condition, Indicates the total number of triggers for all trigger conditions. Indicates the total number of trigger conditions. represents the importance score weight coefficient, Indicates the importance score of the current rule. represents the maximum importance score in the rule set, Indicates the applicable scope scoring weight coefficient, Indicates the number of audit scenarios to which the current rule applies. Indicates the total number of all audit scenarios in the rule set. represents the conflict penalty coefficient, represents the total number of conflicting rules, Indicates The number of conflicting fields for conflicting rules, Indicates The total number of fields in the conflicting rules; S26. Prioritize the rules in the multi-layer rule set according to the rule weights, remove low-priority or logically conflicting rules, and generate an initial rule set, which is stored in the multi-layer rule engine.

[0023] In this implementation, S3 specifically includes: S31. By analyzing the distribution of rule parameters in the unified audit data pool, an improved bee colony optimization algorithm is initialized to generate a bee population, wherein each bee represents a set of rule parameter combinations, and the rule parameter combinations include a precondition set, an execution action set, and an applicable scope; S32. Allocate an initial search space for the bee population, where the initial search space is dynamically defined according to the value range of the rule parameter: ; in, Indicates The initial search range for rule parameters, Indicates A set of rule parameters, Indicates The maximum value of a set of rule parameters, Indicates The minimum value of a set of rule parameters, Indicates the total number of rules, represents the adjustment factor, Indicates The standard deviation of a set of rule parameters; S33. Use the fitness function to evaluate the quality of the rule parameter combination of each bee in the bee population: ; in, Indicates The fitness value of the rule parameter combination, , and represents the weight coefficient, Indicates the accuracy of the rule, represents the efficiency of rule execution, Indicates the coverage of the rule; S34. Dynamically adjust the initial search range of the bee population: ; in, Indicates The search range after the rule parameters are adjusted. represents the adjustment coefficient, Indicates The fitness change value of the rule parameter combination, represents the average fitness; S35, optimizing the rule parameter combination of each bee: ; in, Indicates The rules and The coordination of the rules, Indicates The rules and The similarity of the rules, Indicates The rules and The degree of conflict of rules, represents the smoothing factor; S36. Through multiple iterative adjustments, record the rule parameter combination with the best fitness value in each iteration. When the maximum number of iterations is reached or the fitness change is less than a preset threshold, output the optimized bee population and search space, and dynamically define the search space of the bee population.

[0024] In this implementation manner, the S4 specifically includes: S41. In the worker bee stage, the local attributes of the rules are classified and optimized, and the multi-dimensional perturbation mechanism is used to adjust the parameters of each rule: ; in, Indicates New parameter values ​​for the rule, Indicates The current value of the rule, represents the adjustment factor, represents the disturbance factor, represents a random disturbance, and represents two randomly selected rule parameter values; S42, in the observation bee stage, the priority of the rule is optimized through the global evaluation rule execution effect, the parameter distribution is dynamically adjusted, and the global adaptability optimization value is calculated: ; in, Indicates The global adaptability optimization value of the rule, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates the total number of rules; S43, introduce the global logical consistency of the collaborative bee stage optimization rule set, and construct the rule applicability matrix Analyze conflicts between multiple rules: ; in, Indicates The rules and The conflict cost of a rule, Indicates The rules and The similarity of the rules, Indicates The rules and The degree of conflict of rules, Indicates The complexity of the rules, Indicates The complexity of the rules; S44, the scout bee stage uses a multi-objective weighted mechanism to optimize the rule search distribution and reconstruct the rule weights that are below the preset fitness threshold: ; in, Indicates The update weight of each rule is , and represents the weight coefficient, Indicates The global adaptability optimization value of the rules; S45. Combining the results of the worker bee stage, the observer bee stage, the collaborative bee stage, and the scout bee stage, an optimized rule set is generated.

[0025] In this implementation manner, S5 specifically includes: S51. Input the optimized rule set into the multi-layer rule engine, perform hierarchical analysis, and generate a rule execution plan according to the priority and applicable scope of the rules: ; in, Indicates The priority of a rule, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates the total number of rules; S52. Dynamically adjust the priority of each node in the audit process according to the output results of the rule execution plan: ; in, Indicates The priority of the node, Indicates The rule for The relevance of nodes, represents the total number of rules associated with the node; S53, optimizing the execution path of the audit task, sorting the task path according to the node priority, and generating the optimal task path with the minimum total time cost as the goal; S54, when executing the task path, monitor the implementation of the rule execution plan in real time, record the execution status of the node, and the monitoring content includes the task completion rate and execution time of the node; S55. Based on the feedback from monitoring data, dynamically adjust the rule execution plan to ensure that the node priority is greater than the set threshold The tasks are completed first.

[0026] In this implementation manner, S6 specifically includes: S61, based on the implementation feedback of the rule execution plan, collect dynamic data generated during the rule execution process, including rule triggering frequency, execution success rate, task completion rate, response time and resource consumption, to form a rule feedback data set; S62, standardize the rule feedback data set and unify the data formats of different dimensions; S63. Analyze the actual execution effect of the rules based on the standardized rule feedback data set, re-rank the importance of the rules based on the execution success rate and task completion rate, and identify conflicting rules; S64. According to the analysis results of the rule execution effect, the rule parameters and priorities are adjusted, and the applicable scope, triggering conditions and execution logic of the rule are redefined; S65, synchronizing the adjusted rules to the basic rule layer and the personalized rule layer, updating the rule set, and regenerating a new rule execution plan according to the adjusted rule priority; S66. Based on the new rule execution plan, optimize and iterate the audit process path, dynamically adjust the priority and task sequence of process nodes, and achieve continuous optimization and iteration of the rule set and audit process path.

[0027] Embodiment 1: In order to verify the feasibility of the present invention in implementation, the present invention is applied to the annual financial audit scenario of a large manufacturing enterprise. The enterprise has a complex financial structure and a multi-level business system, including multiple modules such as supply chain management, capital flow and inventory accounting. In the traditional audit process, problems such as fixed rules, unreasonable node priority allocation and insufficient process path optimization lead to a long audit cycle, a low error detection rate and uneven resource allocation. Based on these problems, the automated audit business process control method based on rule optimization proposed in the present invention is put into practical application.

[0028] In this embodiment, multi-source audit data from the enterprise's financial system, business system, and third-party platform are first collected. These data include account records, inventory flow, capital flow, etc. for the past three years, with a data volume of about 12 million. Through the method of the present invention, the data is cleaned, anomaly detected, and formatted to form a unified audit data pool. On this basis, a multi-layer rule engine is constructed, including a basic rule layer, a dynamic rule layer, and a personalized rule layer. The basic rule layer covers the financial regulations and tax standards applicable to the enterprise; the dynamic rule layer generates rules in real time to adapt to abnormal data patterns; and the personalized rule layer generates targeted rule sets for historical problems of the enterprise.

[0029] Subsequently, the rule set was dynamically optimized using an improved bee colony optimization algorithm. The worker bee stage performed a local search on the rule parameters and generated an optimized rule set based on fitness; the observer bee stage selected the optimal rule direction based on the rule execution effect and resolved the logical conflicts between the rules; the collaborative bee stage optimized the adaptability between multiple rules; and the scout bee stage prevented the algorithm from falling into a local optimum by maintaining population diversity. In the optimized rule set, a total of 120 high-priority rules were generated, of which 20 rules were newly added in the dynamic rule layer.

[0030] The optimized rule set is input into the multi-layer rule engine to generate a dynamic rule execution plan. The audit process node priority is adjusted in real time according to the rule execution plan. Taking the high-risk capital flow verification task of the enterprise as an example, it takes an average of 16 hours to complete in the traditional process, but only 10 hours after optimization; the audit time of the inventory verification task is reduced from 8 hours to 5 hours. At the same time, by optimizing the task path, the system can dynamically adjust resource allocation, give priority to completing high-priority node tasks when resources are limited, and significantly improve overall efficiency.

[0031] In the entire audit process, the present invention also monitors the rule execution status and task completion rate of process nodes in real time, continuously adjusts the rule parameters and priorities, generates new rule execution plans, and realizes continuous optimization and iteration of rules and processes. Ultimately, the overall audit cycle was shortened from the traditional 10 days to 6 days, the error detection rate of high-risk transactions was increased from the original 65% to 92%, and the rule applicability rate was increased from 78% to 96%.

[0032] Table 1 Comparison of audit process efficiency and effect between traditional method and the method of the present invention

[0033] According to the data in Table 1 above, the application of the present invention in the annual financial audit scenario of a manufacturing enterprise has significantly improved the efficiency and accuracy of the audit process, while greatly reducing the rule conflict rate and resource waste, reflecting extremely high application value. First, the present invention shortens the audit cycle from 10 days in traditional methods to 6 days through rule optimization and process control, reducing the time cost by 40%. This improvement is of great significance in the annual financial audit of enterprises, especially in time-sensitive audit tasks, as the audit work of the entire process can be completed faster. Secondly, the false detection rate of high-risk transactions has increased from 65% to 92%, significantly improving the accuracy and effectiveness of the audit. Combined with dynamic rule generation and optimization, the present invention can quickly identify hidden financial problems and solve the problems of misjudgment and omission caused by the fixed rules in traditional audit methods.

[0034] In addition, the rule applicability rate increased from 78% to 96%, indicating that the present invention can more efficiently match rules with business scenarios and reduce the execution of invalid rules. This improvement directly reduces the rule conflict rate from 15% to 3%, reducing unnecessary waste of resources while ensuring the logical consistency of rule execution. The time for cash flow verification and inventory verification tasks was shortened from 16 hours and 8 hours to 10 hours and 5 hours, respectively, with an optimization range of 37.5%. Through dynamic adjustment of node priorities and optimization of task paths, the present invention significantly improves the execution efficiency of key tasks and saves enterprises a lot of labor and time costs.

[0035] At the same time, the present invention excels in dynamically generating high-priority rules and resource utilization. 20 high-priority rules were dynamically generated, filling the gap of the lack of dynamic rules in traditional methods; the resource utilization of audit process nodes increased from 68% to 87%, an increase of nearly 28%, fully demonstrating the advantages of the present invention in resource optimization. It can be seen that the present invention solves the problems of fixed rules, insufficient node optimization, and inefficient path planning in traditional audit methods through a multi-layer rule engine and an improved swarm optimization algorithm. It has significant advantages in the intelligence, dynamism, and efficiency of the audit process, and provides a new solution for enterprises in complex audit scenarios.

[0036] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for controlling automated audit business processes based on rule optimization, characterized in that: The steps include: S1. Collect audit data, perform data cleaning, anomaly detection and format standardization, and generate a unified audit data pool; S2. Based on the unified audit data pool, a multi-layer rule engine is constructed, wherein the multi-layer rule engine includes a basic rule layer, a dynamic rule layer, and a personalized rule layer, and an initial rule set is generated; S3, initializing the improved bee colony optimization algorithm, generating a bee population, each bee represents a set of rule parameter combinations, and dynamically defining the search space of the bee population; S4. Use the improved bee colony optimization algorithm to dynamically optimize the initial rule set. In the worker bee stage, adjust the parameters of a single rule in the search space. In the observation bee stage, select the optimal rule direction based on the rule execution effect, and introduce collaborative bees to optimize the conflicts between multiple rules. In the scout bee stage, monitor the population diversity, reinitialize the low-diversity population, and generate the optimized rule set. S5. Input the optimized rule set into the multi-layer rule engine to generate a rule execution plan, dynamically adjust the node priority of the audit process according to the rule execution plan, and optimize the task path; S6. Based on the implementation feedback of the rule execution plan, adjust the rule parameters and priorities, and synchronize them to the basic rule layer and personalized rule layer to achieve continuous optimization and iteration of the rule set and audit process path.

2. According to claim 1, a method for controlling an automated audit business process based on rule optimization is characterized in that: The S2 specifically includes: S21. Based on the unified audit data pool, build the basic rule layer according to industry standards and audit requirements; associate the rules with the audit scenarios through preset trigger conditions and execution logic, and generate a basic rule library; S22. Using real-time data stream analysis technology, dynamically updated audit data is processed, trigger conditions and execution actions of rules are generated according to real-time business scenarios, rule logic is dynamically adjusted, and a dynamic rule layer is constructed; the dynamic rule layer is used to respond to real-time changes in audit data and generate a rule base that matches the current audit scenario; S23. Analyze the historical records in the unified audit data pool through data mining technology, and combine the rule logic set by the enterprise with the audit data to form a personalized rule layer; the personalized rule layer is used to dynamically adapt to the needs of different enterprises; S24. Integrate the rules of the basic rule layer, dynamic rule layer and personalized rule layer in a hierarchical manner, optimize the logical consistency between the rules, remove duplicates and adjust the rules with repeated application scopes or conflicting logics, build a multi-layer rule engine, and generate a multi-layer rule set; S25. Dynamically calculate the weight of each rule based on the trigger frequency, importance score and scope of application score: ； in, Indicates The rule weights, Indicates The weight coefficient of the trigger condition, Indicates The trigger frequency of each trigger condition, Indicates the total number of triggers for all trigger conditions. Indicates the total number of trigger conditions. represents the importance score weight coefficient, Indicates the importance score of the current rule. represents the maximum importance score in the rule set, Indicates the applicable scope scoring weight coefficient, Indicates the number of audit scenarios to which the current rule applies. Indicates the total number of all audit scenarios in the rule set. represents the conflict penalty coefficient, represents the total number of conflicting rules, Indicates The number of conflicting fields for conflicting rules, Indicates The total number of fields in the conflicting rules; S26. Prioritize the rules in the multi-layer rule set according to the rule weights, remove low-priority or logically conflicting rules, and generate an initial rule set, which is stored in the multi-layer rule engine.

3. The method for controlling an automated auditing business process based on rule optimization according to claim 1 is characterized in that: The S3 specifically includes: S31. By analyzing the distribution of rule parameters in the unified audit data pool, an improved bee colony optimization algorithm is initialized to generate a bee population, wherein each bee represents a set of rule parameter combinations, and the rule parameter combinations include a precondition set, an execution action set, and an applicable scope; S32. Allocate an initial search space for the bee population, where the initial search space is dynamically defined according to the value range of the rule parameter: ； in, Indicates The initial search range for rule parameters, Indicates A set of rule parameters, Indicates The maximum value of a set of rule parameters, Indicates The minimum value of a set of rule parameters, Indicates the total number of rules, represents the adjustment factor, Indicates The standard deviation of a set of rule parameters; S33. Use the fitness function to evaluate the quality of the rule parameter combination of each bee in the bee population: ； in, Indicates The fitness value of the rule parameter combination, , and represents the weight coefficient, Indicates the accuracy of the rule, represents the efficiency of rule execution, Indicates the coverage of the rule; S34. Dynamically adjust the initial search range of the bee population: ； in, Indicates The search range after the rule parameters are adjusted. represents the adjustment coefficient, Indicates The fitness change value of the rule parameter combination, represents the average fitness; S35, optimizing the rule parameter combination of each bee: ； in, Indicates The rules and The coordination of the rules, Indicates The rules and The similarity of the rules, Indicates The rules and The degree of conflict of rules, represents the smoothing factor; S36. Through multiple iterative adjustments, record the rule parameter combination with the best fitness value in each iteration. When the maximum number of iterations is reached or the fitness change is less than a preset threshold, output the optimized bee population and search space, and dynamically define the search space of the bee population.

4. The method for controlling an automated auditing business process based on rule optimization according to claim 1 is characterized in that: The S4 specifically includes: S41. In the worker bee stage, the local attributes of the rules are classified and optimized, and the multi-dimensional perturbation mechanism is used to adjust the parameters of each rule: ； in, Indicates New parameter values ​​for the rule, Indicates The current value of the rule, represents the adjustment factor, represents the disturbance factor, represents a random disturbance, and represents two randomly selected rule parameter values; S42, in the observation bee stage, the priority of the rule is optimized through the global evaluation rule execution effect, the parameter distribution is dynamically adjusted, and the global adaptability optimization value is calculated: ； in, Indicates The global adaptability optimization value of the rule, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates the total number of rules; S43, introduce the global logical consistency of the collaborative bee stage optimization rule set, and construct the rule applicability matrix Analyze conflicts between multiple rules: ； in, Indicates The rules and The conflict cost of a rule, Indicates The rules and The similarity of the rules, Indicates The rules and The degree of conflict of rules, Indicates The complexity of the rules, Indicates The complexity of the rules; S44, the scout bee stage uses a multi-objective weighted mechanism to optimize the rule search distribution and reconstruct the rule weights that are below the preset fitness threshold: ； in, Indicates The update weight of each rule is , and represents the weight coefficient, Indicates The global adaptability optimization value of the rules; S45. Combining the results of the worker bee stage, the observer bee stage, the collaborative bee stage, and the scout bee stage, an optimized rule set is generated.

5. The method for controlling an automated auditing business process based on rule optimization according to claim 1 is characterized in that: The S5 specifically includes: S51. Input the optimized rule set into the multi-layer rule engine, perform hierarchical analysis, and generate a rule execution plan according to the priority and applicable scope of the rules: ； in, Indicates The priority of a rule, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates The fitness value of the rule parameter combination, Indicates The scope of application of the rule covers the value, Indicates the total number of rules; S52. Dynamically adjust the priority of each node in the audit process according to the output results of the rule execution plan: ； in, Indicates The priority of the node, Indicates The rule for The relevance of nodes, represents the total number of rules associated with the node; S53, optimizing the execution path of the audit task, sorting the task path according to the node priority, and generating the optimal task path with the minimum total time cost as the goal; S54, when executing the task path, monitor the implementation of the rule execution plan in real time, record the execution status of the node, and the monitoring content includes the task completion rate and execution time of the node; S55. Based on the feedback from monitoring data, dynamically adjust the rule execution plan to ensure that the node priority is greater than the set threshold The tasks are completed first.

6. The method for controlling an automated auditing business process based on rule optimization according to claim 1 is characterized in that: The S6 specifically includes: S61, based on the implementation feedback of the rule execution plan, collect dynamic data generated during the rule execution process, including rule triggering frequency, execution success rate, task completion rate, response time and resource consumption, to form a rule feedback data set; S62, standardize the rule feedback data set and unify the data formats of different dimensions; S63. Analyze the actual execution effect of the rules based on the standardized rule feedback data set, re-rank the importance of the rules based on the execution success rate and task completion rate, and identify conflicting rules; S64. According to the analysis results of the rule execution effect, the rule parameters and priorities are adjusted, and the applicable scope, triggering conditions and execution logic of the rule are redefined; S65, synchronizing the adjusted rules to the basic rule layer and the personalized rule layer, updating the rule set, and regenerating a new rule execution plan according to the adjusted rule priority; S66. Based on the new rule execution plan, optimize and iterate the audit process path, dynamically adjust the priority and task sequence of process nodes, and achieve continuous optimization and iteration of the rule set and audit process path.

Images