A management system maturity advancement system and method thereof

Abstract

The application relates to the technical field of management system promotion, and discloses a management system maturity promotion system and a method thereof, which comprises the following steps: firstly, collecting business data generated by the operation of underlying business systems in real time, and recombining and mapping the business data into a multi-dimensional business activity tensor; secondly, listening to the interface calling links between each underlying business system in real time to build a system topology graph, and dynamically analyzing and generating a state transition matrix representing the internal coupling relationship of the management system based on the system topology graph; and thirdly, defining the management system maturity as a continuous maturity hidden state vector. Through real-time collection of the business data generated by the operation of the underlying business systems, and recombination and mapping of the business data into a multi-dimensional business activity tensor containing time series, process nodes and data features, the underlying business flow state is extracted in a non-sensitization, objectification and full-quantization manner, and the deviation caused by artificial subjective judgment is effectively avoided.

Description

Technical Field

[0001] This invention relates to the field of management system improvement technology, specifically to a management system maturity improvement system and method. Background Technology

[0002] As enterprises deepen their digital transformation, establishing a scientific and efficient management system has become crucial for enhancing their core competitiveness. To measure and continuously optimize the operational level of management systems, management system maturity models (such as CMMI and ISO standards) have been widely adopted. Corresponding management system maturity improvement systems have also gradually become important auxiliary tools for enterprises in compliance auditing, process optimization, and internal governance.

[0003] Currently, conventional maturity assessment and improvement techniques mainly rely on periodic online questionnaires, document reviews, or static rule mapping methods that combine data from underlying business systems. These methods typically score the compliance of various indicators across preset management dimensions and use weighted summation to arrive at an overall maturity score. If an indicator is found to be low, a fixed rectification recommendation report must be generated by relying on an expert knowledge base or a human consulting team.

[0004] However, the aforementioned technical solutions suffer from significant delays in data acquisition and evaluation due to their reliance on periodic, post-event assessment models. This makes it difficult to reflect the true state of business changes in a real-time and dynamic manner. Furthermore, the heavy reliance on subjective experience in data entry reduces the objectivity of the final evaluation results. When characterizing the coupling relationships between elements within the management system, traditional methods often treat different management dimensions in isolation or rely solely on manually setting static correlation weights. This rigid evaluation model struggles to perceive the complex microservice call chains within the underlying IT business systems and cannot adaptively adjust to changes in organizational structure and business processes, easily leading to a disconnect between the evaluation logic and the actual physical operating environment. Moreover, when directly acquiring underlying business data, existing systems lack effective data noise filtering mechanisms, easily misinterpreting occasional random fluctuations in daily operations (such as sudden network latency or single operational errors) as a substantial decline in management maturity.

[0005] On the other hand, after diagnosing a low level of maturity, existing technologies struggle to comprehensively consider the enterprise's resource consumption constraints and time costs, and cannot automatically plan a reasonable improvement implementation path. Current rectification guidance mostly remains at the level of outputting macro-level written suggestions, unable to reverse-engineer abstract management optimization intentions into machine-readable configuration scripts. Consequently, it cannot directly intervene in the underlying heterogeneous business systems through automation (e.g., dynamically adding or deleting process nodes, automatically adjusting permission distribution, etc.). This situation leads to a disconnect between the assessment and diagnosis process and the actual rectification actions at the underlying level, making it difficult to truly form a dynamic control closed loop that requires no manual intervention and can adaptively iterate and optimize. Therefore, this invention designs a management system maturity improvement system and method based on the proposed problems. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a management system maturity improvement system and method, which solves the problems of lagging and subjective data collection, as well as the lack of automated closed-loop control intervention based on mathematical optimization in existing management system maturity assessment and improvement methods.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for improving the maturity of a management system, comprising the following steps:

[0008] S1. First, collect business data generated by the underlying business system in real time, and then reorganize and map the business data into a multi-dimensional business activity tensor.

[0009] S2. Monitor the interface call links between the underlying business systems in real time to construct a system topology diagram, and dynamically parse and generate a state transition matrix representing the internal coupling relationship of the management system based on the system topology diagram;

[0010] S3. Define the maturity of the management system as a continuous maturity latent state vector, construct a state space model that includes observation equations and state transition equations, use the multidimensional business activity tensor as the input of the observation equations, and use the state transition matrix as the parameter of the state transition equations, and continuously solve the current maturity latent state vector at the current time node through the state space model.

[0011] S4. Set a target maturity state vector to improve the maturity of the management system, construct a target cost function based on the deviation between the current maturity hidden state vector and the target maturity state vector, and solve for the optimal control input sequence that minimizes the target cost function;

[0012] S5. The optimal control input sequence is reverse-parsed into a configuration execution script for adjusting the underlying business system, and the configuration execution script is automatically sent to the corresponding underlying business system for change execution. The changed underlying business system generates new business data and new interface call links to trigger a closed loop for continuous dynamic improvement of the maturity of the management system.

[0013] Preferably, the process of reorganizing and mapping business data into a multi-dimensional business activity tensor in step S1 specifically includes:

[0014] Extract the flow information from the operation logs of the underlying business system, and construct the third-order multidimensional business activity tensor. The dimensions of the multidimensional business activity tensor include a discretized time series dimension, a global business process node dimension covering cross-system flow, and a data feature dimension containing timeliness characteristics and resource consumption attributes.

[0015] Preferably, the process of constructing the system topology diagram in step S2 specifically includes:

[0016] The captured inter-system interface call chains are constructed into a dynamic directed weighted graph, which serves as the system topology graph.

[0017] When a new interface is added to the underlying business system or when a data flow is redirected, the node and edge weights of the directed weighted graph are updated in real time.

[0018] Preferably, the process of dynamically generating the state transition matrix representing the internal coupling relationships of the management system based on the system topology diagram specifically includes:

[0019] The influence propagation attenuation coefficient and time delay of each node in the directed weighted graph are analyzed using a graph representation learning algorithm.

[0020] Based on the extracted influence propagation attenuation coefficient and time delay, the state transition matrix that dynamically changes over time is generated.

[0021] Preferably, the state-space model constructed in step S3 specifically includes:

[0022] The observation equation is used to map the multidimensional business activity tensor to maturity status using the observation matrix and to filter out business fluctuation noise.

[0023] The state transition equation is used to combine the state transition matrix, the preset control input matrix, the current system intervention action, and the system evolution noise to characterize the dynamic evolution of the maturity hidden state vector.

[0024] Preferably, the step of continuously calculating the current maturity latent state vector at the current time point through the state-space model includes:

[0025] A recursive filtering algorithm is used to continuously solve the system of equations consisting of the observation equation and the state transition equation, filter out noise in the business data, and output the smoothed true maturity state as the current maturity hidden state vector.

[0026] Preferably, the process of constructing the target cost function in step S4 specifically includes:

[0027] The objective cost function is composed of a weighted sum of a state deviation penalty term and a control input resource consumption term.

[0028] The state deviation penalty term is determined by the degree to which the current maturity hidden state vector deviates from the target maturity state vector and by the state penalty weight matrix; the control input resource consumption term is determined by the control input weight matrix characterizing resource consumption cost and the control input vector to be solved.

[0029] Preferably, the process of finding the optimal control input sequence that minimizes the objective cost function specifically includes:

[0030] The optimal control input sequence is obtained by using discrete-time algebraic equation solving or model predictive control optimization algorithms to solve in reverse the calculation sequence that minimizes the objective cost function. The optimal control input sequence represents the set of optimal management intervention actions under the current resource constraints.

[0031] Preferably, step S5, which involves reverse-parsening the optimal control input sequence into a configuration execution script for adjusting the underlying business system, specifically includes:

[0032] The optimal control input sequence is passed to the reverse proxy gateway, and the policy translation engine is used to map the mathematical solution vector in the optimal control input sequence to the configuration execution script.

[0033] The configuration execution script is issued by calling the open application programming interface of the underlying business system. The specific change execution actions include at least one of dynamically adding or deleting compliance verification nodes in the business process engine and adjusting the system operation permission threshold in the identity management system.

[0034] Preferably, a management system maturity improvement system includes:

[0035] The tensor construction module is used to collect business data generated by the operation of the underlying business system in real time, and reorganize and map the business data into multi-dimensional business activity tensors.

[0036] The topology parsing and matrix reconstruction module is used to monitor the interface call links between the underlying business systems in real time to construct a system topology diagram, and dynamically parse and generate a state transition matrix representing the internal coupling relationship of the management system based on the system topology diagram.

[0037] The hidden state continuous observation module is used to define the maturity of the management system as a continuous maturity hidden state vector, construct a state space model containing observation equations and state transition equations, take the multi-dimensional business activity tensor as the input of the observation equations, and take the state transition matrix as the parameter of the state transition equations, and continuously solve the current maturity hidden state vector at the current time node through the state space model.

[0038] The cost assessment and optimization module is used to set a target maturity state vector for improving the maturity of the management system, construct a target cost function based on the deviation between the current maturity hidden state vector and the target maturity state vector, and solve for the optimal control input sequence that minimizes the target cost function.

[0039] The reverse proxy and distribution module is used to reverse-parse the optimal control input sequence into a configuration execution script for adjusting the underlying business system, and automatically distribute the configuration execution script to the corresponding underlying business system for change execution. After the underlying business system is modified, it generates new business data and new interface call chains to trigger a closed loop for continuous dynamic improvement of the maturity of the management system.

[0040] This invention provides a management system maturity improvement system and method. It has the following beneficial effects:

[0041] 1. This invention collects business data generated by the operation of the underlying business system in real time, and reconstructs and maps the business data into a multi-dimensional business activity tensor containing time series, process nodes and data features. This changes the limitations of traditional management system maturity assessment, which relies heavily on manual questionnaires and post-audit. It achieves seamless, objective and fully quantitative extraction of the underlying business flow status, thereby effectively avoiding the bias caused by subjective human judgment and fundamentally eliminating the technical pain point of serious data lag in assessment.

[0042] 2. This invention transmits the solved abstract optimal control input sequence to the reverse proxy gateway, which then reverse-parses it into a configuration execution script for adjusting the underlying business system and automatically distributes the changes. This forms a complete system dynamic closed loop of "environmental perception, model optimization, physical execution, and effect feedback" without human intervention, thereby enhancing the maturity of the management system and enabling it to have a fully automatic execution capability for continuous adaptive evolution.

[0043] 3. This invention defines the true maturity of the management system, which cannot be directly observed, as a continuous maturity latent state vector. It also constructs a state-space model that integrates multi-dimensional business activity tensor inputs and dynamic state transition matrix parameters. This ingeniously abstracts the management system into a continuous dynamic system with inertia. Combined with a recursive filtering algorithm, it filters out the accidental random fluctuations and evolutionary noise in daily business operations during continuous solution. This solves the problem of "state misjudgment" caused by directly mapping maturity using fluctuations of a single business indicator. It stably outputs a true maturity latent state that is extremely smooth and objectively reflects the essential change trend of the overall management capability of the system.

[0044] 4. This invention constructs a target cost function by setting a target maturity state vector, which includes both the penalty term for deviation from the current maturity hidden state and the resource consumption term for system intervention actions. It then uses an optimization algorithm to solve in reverse the optimal control input sequence that minimizes the target cost function. This upgrades the traditional "static identification of management defects" to "automatic planning of the optimal improvement path". Under the premise of ensuring that the predetermined maturity target can be achieved, it automatically deduces and outputs the set of management intervention actions that minimize the resistance to the normal operation of the enterprise and the lowest resource consumption cost, thereby enhancing the system's global intelligent planning and resource coordination capabilities in the management improvement process.

[0045] 5. This invention constructs a dynamically updated system topology diagram by real-time monitoring of the interface call links between various underlying business systems. It then uses a graph representation learning algorithm to dynamically parse and generate a state transition matrix representing the internal coupling relationships of the management system based on the system topology diagram. This breaks through the technical bottleneck of existing technologies that rely on expert experience and rigidly preset evaluation rules. This enables the system to adaptively perceive changes in the enterprise's underlying IT physical architecture and business flow, accurately quantify and capture the impact attenuation and time delay relationship of cross-system propagation between different management elements, thereby giving the evaluation model the ability to dynamically evolve with changes in the enterprise's organizational structure. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the method architecture of the present invention;

[0047] Figure 2 This is one of the schematic diagrams of the method flow of the present invention;

[0048] Figure 3 This is a second schematic diagram of the method flow of the present invention;

[0049] Figure 4 This is the third schematic diagram of the method flow of the present invention;

[0050] Figure 5 This is the fourth schematic diagram of the method flow of the present invention;

[0051] Figure 6 This is the fifth schematic diagram of the method flow of the present invention. Detailed Implementation

[0052] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Please see the appendix Figure 1 -Appendix Figure 6 This invention provides a method for improving the maturity of a management system, comprising the following steps:

[0054] S1. First, collect business data generated by the operation of the underlying business system in real time, and reorganize and map the business data into a multi-dimensional business activity tensor. The process of reorganizing and mapping the business data into a multi-dimensional business activity tensor in this step specifically includes: extracting the flow information in the operation log of the underlying business system, constructing a third-order multi-dimensional business activity tensor. The dimensions of the multi-dimensional business activity tensor include a discretized time series dimension, a global business process node dimension covering cross-system flow, and a data feature dimension containing timeliness characteristics and resource consumption attributes.

[0055] Specifically, during implementation, enterprises often have multiple heterogeneous underlying business systems, such as enterprise resource planning systems. Traditional management system assessments, such as those for Office Automation (OA), Manufacturing Execution System (MES), and Human Resource Management (HR), rely on manual data entry, resulting in data lag. In this step, the system deploys log collection probes or connects to the enterprise's data bus (such as Kafka) to collect operation logs and business flow data generated by the aforementioned underlying business systems in real time. To facilitate subsequent mathematical modeling, this invention abandons the traditional two-dimensional relational database storage method and maps and reconstructs the extracted flow information into a higher-order mathematical model, namely a third-order multidimensional business activity tensor. ,in:

[0056] For discretized time series dimensions, it represents the data collection time slice (e.g., time step in hours or days).

[0057] To cover the global business process nodes that flow across systems, such as a reimbursement process that spans the OA application node, the department manager approval node, and the ERP finance payment node, these nodes are all coded in the dimension. middle;

[0058] The data feature dimension includes specific business behavior attributes for each node, such as timeliness characteristics (node ​​flow time), compliance verification results (e.g., whether form fields are complete), and resource consumption attributes (e.g., approver level, amount of funds consumed, etc.), forming a multidimensional business activity tensor. This forms the real-time, objective underlying observation base for the subsequent operation of the entire system.

[0059] S2. Real-time monitoring of interface call chains between underlying business systems to construct a system topology graph, and dynamic parsing and generation of a state transition matrix representing the internal coupling relationship of the management system based on the system topology graph. The process of constructing the system topology graph specifically includes: constructing the captured interface call chains between systems into a dynamic directed weighted graph as the system topology graph; when a new interface is added to the underlying business system or a data flow is redirected, the node and edge weights of the directed weighted graph are updated in real time. The process of dynamically parsing and generating a state transition matrix representing the internal coupling relationship of the management system based on the system topology graph specifically includes: using graph representation learning algorithms to analyze the influence propagation attenuation coefficient and time delay of each node in the directed weighted graph; generating a state transition matrix that dynamically changes over time based on the extracted influence propagation attenuation coefficient and time delay.

[0060] Specifically, the various elements within the management system (such as human resource management and production quality control) are not isolated but deeply coupled. To capture this coupling, this step involves monitoring the call chain and data packet flow of the microservice application programming interfaces (APIs) between the underlying business systems.

[0061] First, the captured inter-system interface call links are constructed into a dynamic directed weighted graph as a system topology graph. The nodes in the graph represent the underlying microservices or system modules, the directed edges represent the call flow, and the edge weights represent the call frequency and response latency. When the underlying business system adds interfaces or the data flow is redirected due to business adjustments, the node and edge weights of the directed weighted graph will be automatically updated in real time.

[0062] Subsequently, graph representation learning algorithms (such as Graph Convolutional Networks, GCN, etc.) are used to perform graph embedding representation on the directed weighted graph, and the influence propagation attenuation coefficient and time delay between nodes in the graph are parsed. Based on this, the system automatically extracts and generates a state transition matrix that changes dynamically over time. ;

[0063] The physical significance of this matrix lies in its mathematical quantitative representation of the extent and speed at which a change in a certain management link (such as congestion at a certain process node) will affect other management links under the current physical IT network topology of an enterprise.

[0064] S3. Define the maturity of the management system as a continuous implicit maturity state vector. Construct a state-space model containing observation equations and state transition equations. Use the multidimensional business activity tensor as the input to the observation equations and the state transition matrix as the parameter of the state transition equations. Continuously solve the current implicit maturity state vector at the current time point through the state-space model. Specifically, the state-space model constructed in this step includes: the observation equations, which use the observation matrix to map the multidimensional business activity tensor to maturity states and filter out business fluctuation noise; and the state transition equations, which combine the state transition matrix, the preset control input matrix, the current system intervention actions, and system evolution noise to characterize the dynamic evolution of the implicit maturity state vector. Continuously solving the current implicit maturity state vector at the current time point through the state-space model includes: using a recursive filtering algorithm to continuously solve the equation set composed of the observation equations and the state transition equations, filtering noise in the business data, and outputting the smoothed true maturity state as the current implicit maturity state vector.

[0065] Specifically, this invention views management system maturity as a continuously changing, inertial dynamic system state, rather than a static report score. True maturity is defined as a continuous, implicit state vector of maturity that cannot be directly observed. ,in For the preset number of management system dimensions (such as maturity indicators for quality, safety, R&D, etc.), the system constructs a state-space model consisting of observation equations and state transition equations:

[0066] The observation equation is expressed as: ,in, The business activity tensor input in the above steps, The observation matrix maps multidimensional business data to macro-management status. This refers to occasional fluctuations and noise in business operations (such as timeouts caused by network latency on a certain day, which do not necessarily indicate a decline in management maturity).

[0067] The state transition equation is expressed as: .in, The state transition matrix generated by the above steps; The control input matrix represents the effect of system interventions on maturity status. This refers to the sequence of management intervention actions currently being executed by the system. This is system evolution noise;

[0068] Under this model architecture, the system uses a recursive filtering algorithm (such as the extended Kalman filter) to continuously solve the above equations. By performing recursive Bayesian estimations on historical data and current observations, noise in the business data is filtered out. and The final output is a smooth and accurate true state, namely the current maturity hidden state vector. .

[0069] S4. Define a target maturity state vector to improve the maturity of the management system. Construct a target cost function based on the deviation between the current maturity hidden state vector and the target maturity state vector. Solve for the optimal control input sequence that minimizes the target cost function. The process of constructing the target cost function in this step specifically includes: the target cost function is composed of a weighted sum of a state deviation penalty term and a control input resource consumption term; wherein, the state deviation penalty term is determined by the degree to which the current maturity hidden state vector deviates from the target maturity state vector and the state penalty weight matrix; the control input resource consumption term is determined by the control input weight matrix representing the resource consumption cost and the control input vector to be solved. Solving for the optimal control input sequence that minimizes the target cost function specifically includes: using discrete-time algebraic equations or model predictive control optimization algorithms to inversely solve for the calculation sequence that minimizes the target cost function, which serves as the optimal control input sequence. The optimal control input sequence represents the optimal set of management intervention actions under the current resource constraints.

[0070] Specifically, after obtaining the accurate current state, this step focuses on solving the problem of how to improve maturity from the optimal path. First, it sets the expected target maturity state vector. Next, a target cost function based on limited resources and time constraints is constructed. This function consists of a weighted sum of a state deviation penalty term and a control input resource consumption term, and its specific expression is designed as follows:

[0071] ;

[0072] in: Represents the field of view for predicting the time step; This represents the state deviation penalty term, i.e., the current predicted state. With the target state The degree of deviation, where the matrix This is the state penalty weight matrix, used to set the tolerance for deviations from the target in each management dimension; The matrix represents the cost of controlling input resource consumption items, i.e., the cost required to implement management interventions (such as increasing approval personnel, purchasing new systems, etc.). To control the input weight matrix; The sequence of control intervention actions for the system to be solved;

[0073] To make the objective cost function To minimize the optimal control input sequence, the system employs either a discrete-time algebraic Riccati equation solving algorithm or a model predictive control (MPC) optimization algorithm for inverse optimization. This sequence represents the optimal (least resistant, lowest cost) set of management interventions that the system should take to achieve the set maturity goals under the current organizational structure and IT resource constraints of the enterprise.

[0074] S5. The optimal control input sequence is reverse-parsed into a configuration execution script for adjusting the underlying business system, and the configuration execution script is automatically distributed to the corresponding underlying business system for change execution. After the underlying business system is modified, new business data and new interface call chains are generated to trigger a closed loop for continuous dynamic improvement of the maturity of the management system. The process of reverse-parseting the optimal control input sequence into a configuration execution script for adjusting the underlying business system specifically includes: passing the optimal control input sequence to the reverse proxy gateway, using the policy translation engine to map the mathematical solution vector in the optimal control input sequence into the configuration execution script; and distributing the configuration execution script by calling the open application programming interface of the underlying business system. Specific change execution actions include at least one of dynamically adding or deleting compliance verification nodes in the business process engine and adjusting the system operation permission threshold in the identity management system.

[0075] Specifically, this step represents a leap from mathematical theory guidance to practical IT system control. The optimal control input sequence obtained in the above steps... This step involves deploying a reverse proxy gateway, which acts as the translation engine, instead of simply providing an abstract mathematical solution vector. The reverse proxy gateway's internal policy translation engine translates the mathematical requirements of the optimal control input sequence into configuration execution scripts (such as JSON or XML format) that the underlying business system can recognize, based on a mapping table. These scripts are then automatically distributed and implemented by calling the application programming interfaces (APIs) exposed by the corresponding underlying business system. Specific execution actions may include, but are not limited to:

[0076] Dynamically insert additional compliance verification node interfaces into the routing branches of the Business Process Management (BPM) engine to improve compliance;

[0077] Alternatively, the threshold for critical system operation permissions for certain positions can be dynamically lowered in the Identity Management (IAM) system to reduce operational risks.

[0078] Once these configuration execution scripts are issued and executed, the underlying business system's operating logic is substantially changed, resulting in entirely new business operation behaviors (new business data) and entirely new microservice call relationships (new interface call chains). These newly generated data flows will be immediately captured by the tensor construction module in step S1 and the topology parsing module in step S2, thereby triggering the next round of state assessment and path optimization, thus completely forming a continuous dynamic improvement closed loop of "perception-diagnosis-optimization-control".

[0079] A management system maturity improvement system includes:

[0080] The tensor construction module is used to collect business data generated by the underlying business system in real time and reorganize and map the business data into multi-dimensional business activity tensors.

[0081] The topology parsing and matrix reconstruction module is used to monitor the interface call links between various underlying business systems in real time to build a system topology diagram, and dynamically parse and generate a state transition matrix representing the internal coupling relationship of the management system based on the system topology diagram.

[0082] The hidden state continuous observation module is used to define the maturity of the management system as a continuous maturity hidden state vector, construct a state space model containing observation equations and state transition equations, take the multi-dimensional business activity tensor as the input of the observation equations, and take the state transition matrix as the parameter of the state transition equations. The current maturity hidden state vector at the current time node is continuously calculated through the state space model.

[0083] The cost assessment and optimization module is used to set the target maturity state vector for improving the maturity of the management system, construct the target cost function based on the deviation between the current maturity hidden state vector and the target maturity state vector, and solve for the optimal control input sequence that minimizes the target cost function.

[0084] The reverse proxy and distribution module is used to reverse-parse the optimal control input sequence into a configuration execution script for adjusting the underlying business system, and automatically distribute the configuration execution script to the corresponding underlying business system for change execution. After the underlying business system is changed, it generates new business data and new interface call chains to trigger a closed loop of continuous dynamic improvement of the maturity of the management system.

[0085] Specifically, in response to the above method, this invention also provides a management system maturity improvement system, which includes corresponding hardware and software operation modules. The internal operation logic and method steps of each module correspond one-to-one. The automated calculation and closed-loop control of S1 to S5 are achieved by the server processor calling preset program instructions stored in the memory. Each module includes:

[0086] The tensor building module is used to execute S1;

[0087] The topology parsing and matrix reconstruction module is used to execute S2;

[0088] The hidden state continuous observation module is used to execute S3;

[0089] The cost assessment and optimization module is used to execute S4;

[0090] The reverse proxy and distribution module is used to execute S5 to ensure the reliable deployment and stable operation of this improvement method in an industrial-grade information system environment.

[0091] In summary, this invention provides a management system maturity improvement system and method. By collecting business data generated by the underlying business systems in real time and reconstructing and mapping this data into a multi-dimensional business activity tensor containing time series, process nodes, and data characteristics, it completely changes the limitations of traditional management system maturity assessments, which heavily rely on manually filled questionnaires and post-audit. It achieves seamless, objective, and fully quantitative extraction of the underlying business flow status, effectively avoiding biases caused by subjective human judgment and fundamentally eliminating the technical pain point of severely delayed assessment data. This provides a solid data perspective with high-dimensional characteristics and real-time timeliness for subsequent maturity analysis. The system measures the base and, by passing the solved abstract optimal control input sequence to the reverse proxy gateway, reverse-parses it into a configuration execution script for adjusting the underlying business system and automatically distributes the changes. This achieves an automated implementation leap from macro-management system optimization strategies to the underlying micro-IT system parameter control. Furthermore, the new business data and new call chains generated after the underlying system executes the changes can immediately trigger the data collection and topology reconstruction at the front end. Without human intervention, a complete system dynamic closed loop of "environmental perception, model optimization, physical execution, and effect feedback" is formed, enabling the management system to continuously improve its maturity and possess fully automatic execution capabilities for adaptive evolution.

[0092] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for improving the maturity of a management system, characterized in that, Includes the following steps: S1. First, collect business data generated by the underlying business system in real time, and then reorganize and map the business data into a multi-dimensional business activity tensor. S2. Monitor the interface call links between the underlying business systems in real time to construct a system topology diagram, and dynamically parse and generate a state transition matrix representing the internal coupling relationship of the management system based on the system topology diagram; S3. Define the maturity of the management system as a continuous maturity latent state vector, construct a state space model that includes observation equations and state transition equations, use the multidimensional business activity tensor as the input of the observation equations, and use the state transition matrix as the parameter of the state transition equations, and continuously solve the current maturity latent state vector at the current time node through the state space model. S4. Set a target maturity state vector to improve the maturity of the management system, construct a target cost function based on the deviation between the current maturity hidden state vector and the target maturity state vector, and solve for the optimal control input sequence that minimizes the target cost function; S5. The optimal control input sequence is reverse-parsed into a configuration execution script for adjusting the underlying business system, and the configuration execution script is automatically sent to the corresponding underlying business system for change execution. The changed underlying business system generates new business data and new interface call links to trigger a closed loop for continuous dynamic improvement of the maturity of the management system.

2. The method for improving the maturity of a management system according to claim 1, characterized in that, The process of reorganizing and mapping business data into multidimensional business activity tensors in step S1 specifically includes: Extract the flow information from the operation logs of the underlying business system, and construct the third-order multidimensional business activity tensor. The dimensions of the multidimensional business activity tensor include a discretized time series dimension, a global business process node dimension covering cross-system flow, and a data feature dimension containing timeliness characteristics and resource consumption attributes.

3. The method for improving the maturity of a management system according to claim 1, characterized in that, The process of constructing the system topology diagram in step S2 specifically includes: The captured inter-system interface call chains are constructed into a dynamic directed weighted graph, which serves as the system topology graph. When a new interface is added to the underlying business system or when a data flow is redirected, the node and edge weights of the directed weighted graph are updated in real time.

4. The method for improving the maturity of a management system according to claim 1, characterized in that, The process of dynamically generating the state transition matrix representing the internal coupling relationships of the management system based on the system topology diagram specifically includes: The influence propagation attenuation coefficient and time delay of each node in the directed weighted graph are analyzed using a graph representation learning algorithm. Based on the extracted influence propagation attenuation coefficient and time delay, the state transition matrix that dynamically changes over time is generated.

5. The method for improving the maturity of a management system according to claim 1, characterized in that, The state-space model constructed in step S3 specifically includes: The observation equation is used to map the multidimensional business activity tensor to maturity status using the observation matrix and to filter out business fluctuation noise. The state transition equation is used to combine the state transition matrix, the preset control input matrix, the current system intervention action, and the system evolution noise to characterize the dynamic evolution of the maturity hidden state vector.

6. The method for improving the maturity of a management system according to claim 1, characterized in that, The step of continuously solving the current maturity hidden state vector at the current time point through the state-space model includes: A recursive filtering algorithm is used to continuously solve the system of equations consisting of the observation equation and the state transition equation, filter out noise in the business data, and output the smoothed true maturity state as the current maturity hidden state vector.

7. The method for improving the maturity of a management system according to claim 1, characterized in that, The process of constructing the objective cost function in step S4 specifically includes: The objective cost function is composed of a weighted sum of a state deviation penalty term and a control input resource consumption term. The state deviation penalty term is determined by the degree to which the current maturity hidden state vector deviates from the target maturity state vector and by the state penalty weight matrix; the control input resource consumption term is determined by the control input weight matrix characterizing resource consumption cost and the control input vector to be solved.

8. The method for improving the maturity of a management system according to claim 1, characterized in that, The process of finding the optimal control input sequence that minimizes the objective cost function specifically includes: The optimal control input sequence is obtained by using discrete-time algebraic equation solving or model predictive control optimization algorithms to solve in reverse the calculation sequence that minimizes the objective cost function. The optimal control input sequence represents the set of optimal management intervention actions under the current resource constraints.

9. The method for improving the maturity of a management system according to claim 1, characterized in that, Step S5, which involves reverse-parsening the optimal control input sequence into a configuration execution script for adjusting the underlying business system, specifically includes: The optimal control input sequence is passed to the reverse proxy gateway, and the policy translation engine is used to map the mathematical solution vector in the optimal control input sequence to the configuration execution script. The configuration execution script is issued by calling the open application programming interface of the underlying business system. The specific change execution actions include at least one of dynamically adding or deleting compliance verification nodes in the business process engine and adjusting the system operation permission threshold in the identity management system.

10. A management system maturity improvement system, characterized in that, A method for improving the maturity of a management system according to any one of claims 1-9, comprising: The tensor construction module is used to collect business data generated by the operation of the underlying business system in real time, and reorganize and map the business data into multi-dimensional business activity tensors. The topology parsing and matrix reconstruction module is used to monitor the interface call links between the underlying business systems in real time to construct a system topology diagram, and dynamically parse and generate a state transition matrix representing the internal coupling relationship of the management system based on the system topology diagram. The hidden state continuous observation module is used to define the maturity of the management system as a continuous maturity hidden state vector, construct a state space model containing observation equations and state transition equations, take the multi-dimensional business activity tensor as the input of the observation equations, and take the state transition matrix as the parameter of the state transition equations, and continuously solve the current maturity hidden state vector at the current time node through the state space model. The cost assessment and optimization module is used to set a target maturity state vector for improving the maturity of the management system, construct a target cost function based on the deviation between the current maturity hidden state vector and the target maturity state vector, and solve for the optimal control input sequence that minimizes the target cost function. The reverse proxy and distribution module is used to reverse-parse the optimal control input sequence into a configuration execution script for adjusting the underlying business system, and automatically distribute the configuration execution script to the corresponding underlying business system for change execution. After the underlying business system is modified, it generates new business data and new interface call chains to trigger a closed loop for continuous dynamic improvement of the maturity of the management system.

Images