Electric power digital information system operation parameter regulation and control method based on heuristic algorithm

Abstract

The invention discloses an electric power digital information system operation parameter regulation and control method based on a heuristic algorithm, and the method comprises the steps: constructing a multi-target heuristic algorithm model based on a dynamic weight adjustment mechanism and an improved adaptive search strategy, and enabling the multi-target heuristic algorithm model to adapt to a complex electric power digital information system; a multi-target heuristic algorithm is utilized to process a plurality of optimization targets in parallel, and algorithm parameters and search areas are adjusted in combination with real-time state data of the electric power digital information system and prediction environment changes. A multi-level verification mechanism is adopted to verify the system stability after the operation parameters of the power digital information system are adjusted on line, and configuration parameters are automatically adjusted through feedback information to reduce manual intervention; according to the electric power digital information system operation parameter regulation and control method based on the heuristic algorithm, the expression stability of the algorithm is improved, the problem that an existing method is poor in adaptability in a complex scene is solved, the overall processing efficiency is improved, the optimization result is more reliable, repeatability is higher, and the stability and safety of system operation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems and automation thereof, and in particular to a method for controlling operating parameters of a power digital information system based on a heuristic algorithm. Background Art

[0002] In the operation and optimization of modern power digital information systems, heuristic algorithms are widely used to optimize the configuration of power digital information system operation parameters. Heuristic algorithms are a type of optimization algorithm that simulates natural phenomena to perform global search, such as genetic algorithms and particle swarm optimization algorithms. These algorithms simulate natural evolution or group behavior to find solutions close to the global optimal solution to improve the economy and reliability of power digital information systems. However, current heuristic optimization methods still have many challenges and limitations in practical applications.

[0003] First, when faced with large and complex power digital information system networks, heuristic algorithms often perform less well than expected and have weak adaptability. The scale of complex power and the changing operating environment make it difficult for these algorithms to maintain high efficiency and stability, resulting in optimization results deviating from the optimal solution. Second, in the process of configuring operating parameters, heuristic algorithms usually consume a lot of computing time and resources, and have low processing efficiency, which affects the real-time and practicality of the optimization process. In addition, this type of optimization method is highly sensitive to the initial parameter settings. Improper selection of initial values ​​may lead to unreliable or unrepeatable optimization results. This sensitivity not only increases the complexity of the algorithm, but also affects the robustness of the algorithm. When applied to actual power digital information systems, the tuned parameters may affect the safety and stability of the overall system, increasing the uncertainty risk of the operation of power digital information systems. More importantly, existing heuristic optimization algorithms often respond slowly when faced with the ever-changing operating conditions and environment of power digital information systems, and lack the ability to quickly adapt to dynamic changes. This deficiency will cause the parameter configuration to gradually fail or even become inappropriate over time. Summary of the invention

[0004] The purpose of the present invention is to provide a method for regulating the operating parameters of an electric power digital information system based on a heuristic algorithm, so as to solve the problems that in the prior art, when facing a large and complex electric power digital information system network, the performance of the heuristic algorithm is often not as expected and the adaptability is weak; in the process of configuring the operating parameters, the heuristic algorithm usually consumes a lot of computing time and resources, and the processing efficiency is low, which affects the real-time and practicality of the optimization process; the sensitivity of the algorithm itself not only increases the complexity of the algorithm, but also affects the robustness of the algorithm; and the existing heuristic optimization algorithm often responds slowly when facing the constantly changing operating conditions and environment of the electric power digital information system, and lacks the ability to quickly adapt to dynamic changes. This deficiency will cause the parameter configuration to gradually fail or even become inappropriate over time.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for controlling operating parameters of a power digital information system based on a heuristic algorithm, the method comprising: S1: Construct a multi-objective heuristic algorithm model based on dynamic weight adjustment mechanism and improved adaptive search strategy to adapt to complex power digital information system; S2: Use multi-objective heuristic algorithms to process multiple optimization objectives in parallel; S3: Adjust algorithm parameters and search areas based on real-time status data of the power digital information system and predicted environmental changes; S4: A multi-level verification mechanism is used to verify the system stability after the operation parameters of the power digital information system are adjusted online and to automatically adjust the configuration parameters through feedback information to reduce manual intervention.

[0006] Preferably, the step S1 comprises: S11: Define the objective functions in the algorithm and their relative importance, and assign initial weights to each objective; S12: Automatically generate a dynamic weight table based on the change of weights; S13: Applying a predefined formula to adjust the weight ratio of the objective function in the dynamic weight table; S14: During the model initialization process, an initial weight table is set, and the weight value is dynamically adjusted according to the actual execution situation.

[0007] Preferably, the step S12 of automatically generating a dynamic weight table based on the weight change includes: Evaluate the satisfaction score of the system state for each objective function after each iteration; The weighted value is calculated by the unconventional exponential sliding average algorithm to update the current importance coefficient of each objective function, that is, , in, is the objective function weight after the next iteration update, is the smoothing factor, and its value range is , is the objective function weight in the current iteration, Indicates the newly measured score; Reset the dynamic weight distribution table based on the updated objective function importance to adapt to the changing needs; The redistributed weight ratio is used to carry out the objective function evaluation and optimization loop iteration process in the next stage.

[0008] Preferably, the non-conventional formula for calculating the weighted value by the exponential sliding average algorithm to update the current importance coefficient of each objective function also includes: Identify the target items that have the greatest impact on system performance ; When the importance score is less than the threshold And when the number of iterations exceeds the predetermined threshold, a reweighting event is triggered otherwise ; If triggered, , applying special correction factors ,make ,when Maintain normal operation ,in, Indicates the target item in the system that has the greatest impact on performance. represents the threshold value, A flag indicating a reweighting event, represents the special correction factor, represents the adjustment coefficient, Represents the current weight of the objective function; After the adjustment, the weights are used to readjust the multi-objective decision-making process of the entire system until the optimal solution is found.

[0009] Preferably, the if triggering the unconventional formula also includes a determination step of whether a stable state appears during the iteration process, which specifically includes the following: Monitor the rate of change of the current state relative to the previous state ; Based on the change rate being below a certain limit Indicates that it has entered a static area, so the stability flag is set ; When no substantial changes occur, , forced to reduce the acceleration parameter The value is set to zero to avoid oscillatory behavior; After the relative stagnation detection is completed, the system will re-enable the regular learning rate to continue searching for the optimal point until it is fully covered or one of the termination criteria is met; Wherein, the change rate is lower than a certain limit It shows that entering the static zone also includes a continuous monitoring mechanism for environmental adaptability, specifically: After a certain period of time , more recently The maximum difference between round calculation results Switch the prelude in recognition mode; Identify current environmental trends Increase, decrease or remain unchanged to adjust the local optimal path guidance rules; If a large deviation is detected in the environment, Adjusting the global governing equations Maintaining strategic consistency; Perform environment matching analysis.

[0010] Preferably, step S2 comprises: By using a multi-objective heuristic algorithm, the safe and economic operation of the power digital information system is decomposed into multiple independently executable goals; Implementing parallel optimization of these objectives on a parallel computing architecture reduces computation time while ensuring that each objective is solved efficiently; When the status of the power digital information system changes, re-evaluate the importance of each sub-goal and re-arrange the priority; Introduce additional safety constraints to verify in real time whether the minimum safety standards are met.

[0011] Preferably, the step S2 further comprises: The goal of safe operation is to minimize the probability of failure of the power digital information system, specifically: ; in, represents minimization, is the failure probability of the power digital information system, It is The safety index of the associated unit, Indicates the serial number of the associated unit, It represents the total number of all associated units in the power digital information system; In the economic operation goal, the operation cost is minimized, specifically: ; in, represents minimization, represents the operating cost of the power digital information system, Indicates the serial number corresponding to the fault type. represents the number of all fault types in the power digital information system, Indicates The cost corresponding to each failure type; Using the parallel computing architecture, the above objective function and Independent solution is performed and parallel optimization strategy is introduced. The parallel optimization process can be implemented through the master-slave parallel computing framework, including allocating the objective function to multiple computing nodes, each node independently processing part of the optimization task, each node using heuristic algorithm for local optimization, and exchanging results with the master node after each round of iteration. The master node summarizes the local optimization results, adjusts the weights through the global heuristic strategy, and updates the global optimal solution. For the local optimization of each node, the specific formula is: ; in, represents the solution vector for the next iteration, represents the solution vector of the current iteration, A sequence number indicating the number of iterations, represents the local optimal solution, represents the global optimal solution, represents the parameter that controls the local convergence speed, Represents the parameter that controls the global convergence speed; When the state of the power digital information system changes, the weight of the objective function is evaluated in real time, and the priority is reallocated based on the state change; At each iteration, it is verified whether the CPU occupancy and memory occupancy of the power digital information system are within the safe operating range. By performing security verification on the results of each round of iteration, if it is found that the minimum safety standard is not met, the optimization direction is adjusted or the weight is reallocated to ensure the safe operation of the power digital information system.

[0012] Preferably, step S3 comprises: Evaluate the rationality of current settings based on real-time operating data; Machine learning technology is used to predict state changes in the future and formulate parameter update strategies, including using time series prediction models to predict future state changes in the power digital information system. The specific formula is: ; in, Represents the predicted future time point The state of Indicates the time, represents the time interval of state prediction, represents a machine learning model, Indicates the current time point System status data, Respectively represent the system status data before the current time point, The number of time steps to be backtracked is used to dynamically adjust the parameters of the optimization algorithm according to the prediction results; Adjust the search area and expand the search range to introduce new potential solutions when it is detected that the initial parameters cannot produce optimized results.

[0013] Preferably, the multi-level verification mechanism in step S4 includes: Real-time monitoring of the core operating parameters of the power digital information system, including CPU occupancy and memory occupancy, using the threshold detection mechanism to set the safety range of the parameters. If the actual value of a certain operating parameter of the power digital information system is not within the preset safety range, an alarm will be triggered; Use performance indicators to evaluate the effect after adjustment. The specific formula is: ; in, represents the steady-state error, represents the target state value, Indicates at time The actual state value at Indicates the moment; The weighted comprehensive evaluation function of the multi-objective optimization results is used for comprehensive stability evaluation. The specific formula is: ; in, represents the value of the comprehensive stability evaluation function, Indicates the weight corresponding to the CPU usage. Indicates the weight corresponding to the memory usage, Indicates the CPU usage. Indicates the memory usage.

[0014] Preferably, the specific formula for automatically adjusting the configuration parameters through feedback information in step S4 is: ; in, Indicates the current time The parameter vector of Indicates the time, represents the learning rate, represents the system stability loss function, represents the parameter vector, Represents the loss function For parameter vector gradient.

[0015] It can be seen from the above technical solution that the present invention has the following beneficial effects: The method for controlling the operating parameters of the power digital information system based on the heuristic algorithm adapts to the complex power digital information system by constructing a multi-objective heuristic algorithm model based on a dynamic weight adjustment mechanism and an improved adaptive search strategy, uses a multi-objective heuristic algorithm to process multiple optimization objectives in parallel, adjusts the algorithm parameters and search areas in combination with the real-time status data of the power digital information system and the predicted environmental changes, and adopts a multi-level verification mechanism to verify the system stability after the operating parameters of the power digital information system are adjusted online, and automatically adjusts the configuration parameters through feedback information to reduce manual intervention, thereby improving the performance stability of the algorithm, overcoming the problem that the existing methods have weak adaptability in complex scenarios, improving the overall processing efficiency, making the optimization results more reliable and repeatable, and improving the stability and safety of the system operation. The response speed to environmental changes is enhanced, the application scope of the optimization method is expanded, and the problems of existing technologies, such as heuristic algorithms often perform less than expected and have weak adaptability when facing large and complex electric power digital information system networks; in the process of configuring operating parameters, heuristic algorithms usually consume a lot of computing time and resources, and have low processing efficiency, which affects the real-time and practicality of the optimization process; the sensitivity of the algorithm itself not only increases the complexity of the algorithm, but also affects the robustness of the algorithm; and existing heuristic optimization algorithms often respond slowly when facing the ever-changing operating conditions and environment of electric power digital information systems, and lack the ability to quickly adapt to dynamic changes. This deficiency will cause the parameter configuration to gradually become ineffective or even inappropriate over time. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The figure is a flow chart of the method for controlling the operating parameters of the power digital information system based on the heuristic algorithm. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] like Figure 1 As shown, a method for controlling operating parameters of a power digital information system based on a heuristic algorithm, the method comprising: S1: Construct a multi-objective heuristic algorithm model based on dynamic weight adjustment mechanism and improved adaptive search strategy to adapt to complex power digital information system; S2: Use multi-objective heuristic algorithms to process multiple optimization objectives in parallel; S3: Adjust algorithm parameters and search areas based on real-time status data of the power digital information system and predicted environmental changes; S4: A multi-level verification mechanism is used to verify the system stability after the operation parameters of the power digital information system are adjusted online and to automatically adjust the configuration parameters through feedback information to reduce manual intervention.

[0019] The present invention adopts a multi-objective heuristic algorithm, and constructs an optimization model adapted to a complex power digital information system through a dynamic weight adjustment mechanism and an improved adaptive search strategy. The model can process multiple optimization objectives in the power digital information system in parallel, such as economic operation and safety requirements. By combining real-time status data and predictions of future environmental changes, the algorithm parameters and search areas are dynamically adjusted to ensure the adaptability and flexibility of the algorithm. In the optimization process, a multi-level verification mechanism is used to monitor the adjustment effect of the operation parameters of the power digital information system in real time to ensure system stability, and the configuration parameters are automatically adjusted through feedback information to reduce dependence on manual intervention. Compared with the traditional power digital information system operation parameter configuration method, this method significantly improves the optimization efficiency and system stability. By dynamically adjusting weights and adaptive search strategies, the global optimal solution can be found more quickly, reducing the optimization time. At the same time, through the combination of real-time data and a multi-level verification mechanism, the operation risk of the power digital information system is effectively reduced, ensuring that the power digital information system can maintain a safe and stable operating state under various working conditions. In addition, the function of automatically adjusting configuration parameters reduces human operation errors and improves the automation level and reliability of the overall system.

[0020] In the above scheme, different heuristic algorithm models, such as particle swarm optimization algorithm, ant colony algorithm, etc., can be used to replace the original multi-objective heuristic algorithm model to adapt to the specific needs of different power digital information systems. The weight adjustment mechanism can be changed from dynamic adjustment to static adjustment based on historical data to reduce the computational complexity. The verification mechanism can be changed from multi-level verification to single-level real-time monitoring, which is suitable for scenarios with high requirements for system real-time performance. Other machine learning models, such as deep learning models, can be introduced into the parameter adjustment method for more complex environmental prediction and parameter tuning.

[0021] Step S1 includes: S11: Define the objective functions in the algorithm and their relative importance, and assign initial weights to each objective; S12: Automatically generate a dynamic weight table based on the change of weights; S13: Applying a predefined formula to adjust the weight ratio of the objective function in the dynamic weight table; S14: During the model initialization process, an initial weight table is set, and the weight value is dynamically adjusted according to the actual execution situation.

[0022] In step S1, the multi-objective heuristic algorithm model is optimized by weighting and dynamically adjusting the objective function. First, the initial weights of each objective function are defined to ensure that the optimization model can preliminarily consider the importance of different objectives. Subsequently, a dynamic weight table is generated by automatically detecting weight changes to reflect the importance changes of each objective at different optimization stages. The weight ratio is adjusted using a predefined formula to ensure that the model can adaptively adjust during operation. The setting and dynamic adjustment of the weight table are initialized to ensure that the algorithm maintains a reasonable weight distribution of the objective function at different operation stages, thereby improving the overall optimization effect. By defining the initial weights and the dynamic adjustment weight mechanism, the present invention can better balance the importance of multiple objective functions in the optimization process and avoid excessive influence of a certain objective function. This method ensures the diversity and stability of the optimization results and improves the flexibility and robustness of the parameter configuration optimization of the power digital information system. The use of the dynamic weight table enables the algorithm to have a more accurate grasp of the importance of the objective function, thereby improving the optimization efficiency and reducing the tuning time.

[0023] The weight allocation method can be changed from fixed weights to adaptive adjustment of initial weights using genetic algorithms or other optimization algorithms. The predefined formula can be replaced by a machine learning model based on historical data to more accurately predict weight adjustment trends. In the process of generating dynamic weight tables, different statistical or predictive methods, such as Monte Carlo simulation, can be used to adapt to different types of power digital information systems.

[0024] Step S12 automatically generates a dynamic weight table based on the weight changes, including: Evaluate the satisfaction score of the system state for each objective function after each iteration; The weighted value is calculated by the unconventional exponential sliding average algorithm to update the current importance coefficient of each objective function, that is, , in, is the objective function weight after the next iteration update, is the smoothing factor, and its value range is , is the objective function weight in the current iteration, Indicates the newly measured score; Reset the dynamic weight distribution table based on the updated objective function importance to adapt to the changing needs; The redistributed weight ratio is used to carry out the objective function evaluation and optimization loop iteration process in the next stage.

[0025] This implementation method evaluates the satisfaction degree of each objective function after each iteration, dynamically adjusts the weight of the objective function, and uses the exponential sliding average algorithm to calculate the updated importance coefficient according to the performance of each objective function in the current iteration. The adjustment of this importance coefficient can reflect the system's adaptation to each objective function in real time. The updated importance coefficient is used to reset the dynamic weight table to ensure that in the next round of optimization, the weight can adaptively reflect the importance change of the objective function. In this way, the optimization model can more flexibly adjust the weight of the objective function, thereby improving the overall optimization effect. The exponential sliding average algorithm is used to update the importance coefficient of the objective function in the present invention, so that the system can flexibly adjust the weight of the objective function in each iteration, thereby improving the adaptability and response speed of the optimization algorithm. Through the instant evaluation of the results of each iteration, the dynamic changes in the power digital information system can be effectively responded to, and the real-time and accuracy of the optimization process can be ensured. The automation of weight adjustment reduces the need for manual intervention, while improving the adaptability of the algorithm to a changing environment, and ultimately improving the safety and economy of the operation of the power digital information system.

[0026] When evaluating the degree of satisfaction of the objective function, other statistical methods, such as weighted average or weighted median, can be used to replace the exponential moving average algorithm. For the update of weighted values, a dynamic adjustment mechanism based on neural networks or other deep learning algorithms can be introduced to further improve the optimization efficiency. If the system stability requirements are high, constraints on the change of objective function weights can be added to prevent excessive fluctuations in weights from causing instability in the optimization process.

[0027] The exponential sliding average algorithm is used to calculate the weighted value to update the current importance coefficient of each objective function. Unconventional formulas also include: Identify the target items that have the greatest impact on system performance ; When the importance score is less than the threshold And when the number of iterations exceeds the predetermined threshold, a reweighting event is triggered otherwise ; If triggered, , applying special correction factors ,make ,when Maintain normal operation ,in, Indicates the target item in the system that has the greatest impact on performance. represents the threshold value, A flag indicating a reweighting event, represents the special correction factor, represents the adjustment coefficient, Represents the current weight of the objective function; After the adjustment, the weights are used to readjust the multi-objective decision-making process of the entire system until the optimal solution is found.

[0028] In this embodiment, the weight of the objective function is dynamically updated through an exponential sliding average algorithm, while considering the objective item with the greatest impact on system performance. When the importance score of a certain objective function is lower than the set threshold and the number of iterations exceeds the predetermined threshold, a re-weighting event will be triggered. The triggering of the re-weighting event will introduce a special correction coefficient to adjust the weight to ensure that the objective item with the greatest impact on the system's performance maintains normal operation. On this basis, the optimization algorithm readjusts the weight distribution of multi-objective decision-making to ensure rapid convergence of the optimal solution. Through this unconventional formula, the present invention can more accurately adjust the weight of the objective function to ensure that in critical situations (such as when a certain objective function has a significant impact on system performance), timely adjustments are made. This dynamic adjustment mechanism enhances the stability and robustness of the system, while also ensuring the rationality and convergence speed of the multi-objective optimization process. By introducing a special correction coefficient, it can effectively deal with potential optimization failure problems caused by too low weights of certain objective functions.

[0029] Other re-weighting mechanisms, such as methods based on Bayesian reasoning or reinforcement learning, can be used to further optimize the adjustment strategy of the objective function weight. The threshold setting can be flexibly adjusted according to different system requirements, or historical data can be used for adaptive learning optimization to adapt to the operating environment of different power digital information systems. The application of special correction coefficients can be changed from fixed values ​​to dynamic adjustments, and can change in real time according to feedback data from the system operating status, further improving the optimization efficiency.

[0030] If the unconventional formula in the trigger also includes a determination step of whether a stable state appears during the iteration process, the specific steps include the following: Monitor the rate of change of the current state relative to the previous state ; Based on the change rate being below a certain limit Indicates that it has entered a static area, so the stability flag is set ; When no substantial changes occur, , forced to reduce the acceleration parameter The value is set to zero to avoid oscillatory behavior; After the relative stagnation detection is completed, the system will re-enable the regular learning rate to continue searching for the optimal point until it is fully covered or one of the termination criteria is met; Wherein, the change rate is lower than a certain limit It shows that entering the static zone also includes a continuous monitoring mechanism for environmental adaptability, specifically: After a certain period of time , more recently The maximum difference between round calculation results Switch the prelude in recognition mode; Identify current environmental trends Increase, decrease or remain unchanged to adjust the local optimal path guidance rules; If a large deviation is detected in the environment, Adjusting the global governing equations Maintain policy consistency and perform environment matching analysis.

[0031] In this implementation, the rate of change between the current state and the previous state of the power digital information system is monitored to determine whether it has entered a stable state. When the rate of change is lower than a certain set threshold, the system determines that it has entered a static area and sets a stable flag to avoid system oscillation due to frequent adjustment of optimization parameters. If the system is detected to be in a stagnant state, the acceleration parameter value is forced to be reduced to zero to reduce the occurrence of oscillation behavior, and then the optimization is continued by re-enabling the regular learning rate. In this process, continuous monitoring of environmental adaptability helps the system identify the changing trend of the current mode (such as rising, falling or unchanged), and adjusts the guidance rules of the local optimal path accordingly. If it is found that the environment has a large If there is a deviation, the consistency of the strategy is maintained by adjusting the global control equation, thereby ensuring the effectiveness of the optimization process and the stability of the system operation. This implementation method effectively avoids the system oscillation phenomenon caused by frequent parameter adjustment during the optimization process by introducing a stable state determination step. By monitoring the rate of change of the system state and continuous monitoring of the environmental adaptability, the optimization algorithm can promptly identify the entry into the stable state and make corresponding adjustments, further enhancing the stability of the system and the robustness of the optimization process. At the same time, through the adjustment of the global control equation, the system can dynamically adapt to different operating environments to ensure that the optimization target is always efficiently executed in the complex power digital information system.

[0032] More complex state monitoring algorithms, such as Kalman filters or particle filters, can be used to more accurately identify stable states or system state changes. For the monitoring mechanism of environmental adaptability, machine learning technology can be combined for predictive analysis to more accurately adjust the global control equation. When oscillatory behavior is detected, other oscillation suppression methods can be introduced, such as adjusting the learning rate or using adaptive gain control methods, to further reduce the impact on system stability.

[0033] Step S2 includes utilizing a multi-objective heuristic algorithm to decompose the safe and economical operation of the electric power digital information system into multiple independently executable objectives; achieving parallel optimization of these objectives on a parallel computing architecture to reduce computing time while ensuring that each objective is effectively solved; re-evaluating the importance of each sub-objective and re-prioritizing it when the state of the electric power digital information system changes; and introducing additional safety constraints to verify in real time whether the minimum safety standards are met.

[0034] In this implementation, a multi-objective heuristic algorithm is used to decompose the safety and economy of the power digital information system in operation into multiple independent optimization objectives. These objectives can be optimized simultaneously on a parallel computing architecture. Through parallel processing, the calculation time is significantly reduced, while ensuring the effective solution of each objective. When the state of the power digital information system changes, the system will evaluate the importance of each sub-objective in real time and rearrange the priority to cope with the dynamically changing power digital information system operation environment. In addition, additional safety constraints are introduced in the optimization process to verify in real time whether the minimum safety standard is met to ensure that the power digital information system always maintains safe operation during the optimization process. By decomposing the power digital information system operation optimization problem into multiple independently executable sub-objectives and processing them on a parallel computing architecture, the optimization efficiency and response speed are greatly improved. The mechanism of real-time evaluation and adjustment of the priority of the sub-objectives ensures the flexibility and adaptability of the optimization process, and can quickly respond to changes in the state of the power digital information system. The introduction of safety constraints further enhances the safety of the operation of the power digital information system, so that the optimization process not only considers the economy, but also ensures the stability and reliability of the operation of the power digital information system.

[0035] The parallel optimization architecture can select different distributed computing platforms, such as cloud computing and edge computing, to adapt to power digital information systems of different scales and complexities. The priority evaluation method of sub-goals can be changed to a machine learning model driven by real-time data to further improve the ability to respond to state changes of the power digital information system. Safety constraints can be customized according to the different operating requirements of the power digital information system, such as increasing or decreasing certain safety indicators to meet the operating requirements in specific scenarios.

[0036] Step S2 further comprises: The goal of safe operation is to minimize the probability of failure of the power digital information system, specifically: ; in, represents minimization, is the failure probability of the power digital information system, It is The safety index of the associated unit, Indicates the serial number of the associated unit, It represents the total number of all associated units in the power digital information system; In the economic operation goal, the operation cost is minimized, specifically: ; in, represents minimization, represents the operating cost of the power digital information system, Indicates the serial number corresponding to the fault type. represents the number of all fault types in the power digital information system, Indicates The cost corresponding to each failure type; In this embodiment, for example, when the fault type is server fault, memory storage fault, or network fault, during the calculation process, the server fault mark number is 1, the memory storage fault mark number is 2, and the network fault mark number is 3. The total number of all fault types is 3. When the cost corresponding to the fault type is input, the cost is used to calculate the cost. To solve, this embodiment may also have other fault types; Using the parallel computing architecture, the above objective function and Independent solution is performed and parallel optimization strategy is introduced. The parallel optimization process can be implemented through the master-slave parallel computing framework, including allocating the objective function to multiple computing nodes, each node independently processing part of the optimization task, each node using heuristic algorithm for local optimization, and exchanging results with the master node after each round of iteration. The master node summarizes the local optimization results, adjusts the weights through the global heuristic strategy, and updates the global optimal solution. For the local optimization of each node, the specific formula is: ; in, represents the solution vector for the next iteration, represents the solution vector of the current iteration, A sequence number indicating the number of iterations, represents the local optimal solution, represents the global optimal solution, represents the parameter that controls the local convergence speed, Represents the parameter that controls the global convergence speed; When the state of the power digital information system changes, the weight of the objective function is evaluated in real time, and the priority is reallocated based on the state change; At each iteration, it is verified whether the CPU occupancy and memory occupancy of the power digital information system are within the safe operating range. By performing security verification on the results of each round of iteration, if it is found that the minimum safety standard is not met, the optimization direction is adjusted or the weight is reallocated to ensure the safe operation of the power digital information system.

[0037] In this implementation, the safety and economic operation goals of the power digital information system are decomposed into two independent optimization problems: minimizing the probability of failure of the power digital information system and minimizing the operating cost. These goals are solved independently through a parallel computing architecture, thereby significantly reducing the optimization calculation time. Through the master-slave parallel computing framework, each computing node independently processes local optimization tasks and summarizes local results to the master node. The master node adjusts the weight according to the global heuristic strategy and updates the global optimal solution. In each iteration, it is verified in real time whether the operating parameters of the power digital information system are within a safe range, such as CPU occupancy and memory occupancy. Once an abnormality is found, the system will dynamically adjust the optimization direction or reallocate the weight to ensure the safety of the power digital information system. Through the parallel optimization strategy, this implementation significantly improves the optimization efficiency and response speed, while ensuring the safety and economy of the operation of the power digital information system. Through the master-slave architecture, it can effectively handle complex optimization problems caused by changes in the state of the power digital information system, and maintain flexible responses to various goals. The safety verification and weight adjustment in each iteration ensure that the system is always in a safe operating state during the optimization process, reducing the operating risk of the power digital information system and improving the stability and reliability of the overall system.

[0038] The allocation and optimization process of the objective function can be implemented on other parallel computing frameworks, such as Hadoop and Spark, to adapt to different computing resources and task requirements. The update formulas of local and global optimal solutions can be replaced by dynamic adjustment mechanisms based on artificial intelligence, such as using deep reinforcement learning methods to optimize the adjustment of local and global weights. For the security verification of the CPU occupancy and memory occupancy of the power digital information system, more sophisticated monitoring mechanisms can be introduced, such as real-time monitoring algorithms based on interval estimation, to improve the recognition and response speed of abnormal conditions.

[0039] Step S3 includes: Evaluate the rationality of current settings based on real-time operating data; Machine learning technology is used to predict state changes in the future and formulate parameter update strategies, including using time series prediction models to predict future state changes in the power digital information system. The specific formula is: ; in, Represents the predicted future time point The state of Indicates the time, represents the time interval of state prediction, represents a machine learning model, Indicates the current time point System status data, Respectively represent the system status data before the current time point, The number of time steps to be backtracked is used to dynamically adjust the parameters of the optimization algorithm according to the prediction results; Adjust the search area and expand the search range to introduce new potential solutions when it is detected that the initial parameters cannot produce optimized results.

[0040] In this implementation, the rationality of the current optimization parameter settings is verified by evaluating the real-time operation data of the power digital information system, and the state changes in the future are predicted by combining machine learning technology to formulate parameter update strategies. The time series prediction model predicts the system state at future time points based on the current and historical system state data, and dynamically adjusts the parameters of the optimization algorithm based on the prediction results to ensure the foresight and accuracy of the optimization process. In addition, when it is detected that the initial parameters cannot produce satisfactory optimization results, the system will automatically adjust the search area, expand the search range, introduce new potential solutions, and improve the flexibility and convergence of the optimization process. By combining real-time data evaluation and machine learning prediction, the system can always maintain sensitivity to future state changes during the optimization process, effectively respond to the dynamic environment in complex power digital information systems, and the parameter adjustment strategy enables the optimization process to adapt to the new operating state in a timely manner, improving the overall optimization effect and system response speed. The dynamic adjustment function of the search area ensures that the algorithm can find the optimal solution in different situations, reducing the risk of suboptimal solutions caused by unreasonable parameter settings.

[0041] The time series prediction model can be replaced by a more complex prediction model, such as a long short-term memory network (LSTM) or a variational autoencoder (VAE) based on a neural network, to improve the prediction ability of nonlinear and complex state changes. In the parameter update strategy, an adaptive learning rate adjustment mechanism can be introduced to adjust the learning rate in real time according to system performance feedback to further improve the optimization efficiency. The expansion of the search area can use a combination of multiple heuristic algorithm strategies, such as simulated annealing or genetic algorithm, to further increase the diversity and robustness of the optimization results.

[0042] The multi-level verification mechanism in step S4 includes: Real-time monitoring of the core operating parameters of the power digital information system, including CPU occupancy and memory occupancy, using the threshold detection mechanism to set the safety range of the parameters. If the actual value of a certain operating parameter of the power digital information system is not within the preset safety range, an alarm will be triggered; Use performance indicators to evaluate the effect after adjustment. The specific formula is: ; in, represents the steady-state error, represents the target state value, Indicates at time The actual state value at Indicates the moment; The weighted comprehensive evaluation function of the multi-objective optimization results is used for comprehensive stability evaluation. The specific formula is: ; in, represents the value of the comprehensive stability evaluation function, Indicates the weight corresponding to the CPU usage. Indicates the weight corresponding to the memory usage, Indicates the CPU usage. Indicates the memory usage.

[0043] In this implementation, through a multi-level verification mechanism, the system performs real-time monitoring and security verification on the operating parameters of the power digital information system. The main operating parameters include CPU occupancy and memory occupancy. These parameters set a clear safety range. The system uses a threshold detection mechanism. When the actual value of the operating parameter exceeds the preset safety range, an alarm is triggered and corresponding measures are taken. In addition, a steady-state error calculation formula is used to evaluate the effect of parameter adjustment to quantify the effectiveness of the adjustment result. In the context of multi-objective optimization, a weighted evaluation function is used to comprehensively evaluate the CPU occupancy and memory occupancy to ensure that the optimization effect of the system in all aspects meets the expected standards. This implementation provides a complete multi-level verification mechanism, which effectively guarantees the safe operation of the power digital information system. Through real-time monitoring and threshold detection, the system can respond quickly to abnormal situations and reduce the risk of failure of the power digital information system. The use of steady-state error evaluation and comprehensive stability evaluation functions ensures the accuracy and reliability of the system adjustment effect, thereby improving the safety and stability of the operation of the power digital information system. This multi-level verification method helps to continuously monitor the core parameters of the system during the optimization process to ensure that the final optimization results meet safety standards and maximize economic benefits.

[0044] More advanced monitoring algorithms, such as anomaly detection models or machine learning classification models, can be introduced to further improve the accurate monitoring of system status. The safety range in the threshold detection mechanism can be dynamically adjusted to determine a more appropriate threshold based on statistical analysis of historical data to reduce false alarm and missed alarm rates. The weight values ​​in the weighted comprehensive evaluation function can be adaptively adjusted based on real-time feedback, or multi-criteria decision analysis methods (such as hierarchical analysis or TOPSIS) can be used to optimize weight distribution.

[0045] The specific formula for automatically adjusting the configuration parameters through feedback information in step S4 is: ; in, Indicates the current time The parameter vector of Indicates the time, represents the learning rate, represents the system stability loss function, represents the parameter vector, Represents the loss function For parameter vector gradient.

[0046] This implementation method automatically adjusts the configuration parameters through feedback information, and specifically uses the gradient descent method to optimize the parameters. In each iteration, the system calculates the stability loss function according to the current parameter configuration, and uses the gradient of the loss function to update the parameter value. The learning rate determines the step size of each update, ensuring that the parameter update will not be too fast to cause system instability, nor too slow to affect the optimization efficiency. Through continuous iterative updates, the system can gradually reduce the value of the stability loss function, thereby optimizing the parameter configuration of the power digital information system and achieving an ideal operating state. The feedback information is used to automatically adjust the configuration parameters, which greatly improves the adaptive ability of the system. Through the gradient descent method, the system can quickly respond and correct deviations, so that the operating parameters of the power digital information system are always kept in the optimal configuration state. This method reduces the need for manual intervention, improves the automation level and efficiency of the optimization process, and effectively avoids the performance degradation or instability caused by unreasonable parameter settings. This feedback-based automatic adjustment mechanism has a significant driving effect on the real-time optimization of the power digital information system, ensuring the safety and economy of the system.

[0047] Other optimization algorithms, such as the Adam optimization algorithm, can be used to improve the speed and stability of parameter updates. The learning rate can be changed from a fixed value to an adaptive adjustment strategy to automatically adjust the learning rate according to different system states to improve the efficiency and accuracy of parameter updates. Gradient calculations can be combined with automatic differentiation techniques or numerical methods to improve the accuracy and robustness of calculations, especially when the parameter space is large or the loss function is complex.

[0048] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for controlling operating parameters of a power digital information system based on a heuristic algorithm, characterized in that: The method comprises: S1: Construct a multi-objective heuristic algorithm model based on dynamic weight adjustment mechanism and improved adaptive search strategy to adapt to complex power digital information system; S2: Use multi-objective heuristic algorithms to process multiple optimization objectives in parallel; S3: Adjust algorithm parameters and search areas based on real-time status data of the power digital information system and predicted environmental changes; S4: A multi-level verification mechanism is used to verify the system stability after the operation parameters of the power digital information system are adjusted online and to automatically adjust the configuration parameters through feedback information to reduce manual intervention.

2. The method for controlling operating parameters of a power digital information system based on a heuristic algorithm according to claim 1, characterized in that: The step S1 comprises: S11: Define the objective functions in the algorithm and their relative importance, and assign initial weights to each objective; S12: Automatically generate a dynamic weight table based on the change of weights; S13: Applying a predefined formula to adjust the weight ratio of the objective function in the dynamic weight table; S14: During the model initialization process, an initial weight table is set, and the weight value is dynamically adjusted according to the actual execution situation.

3. The method for controlling operating parameters of a power digital information system based on a heuristic algorithm according to claim 2 is characterized in that: The step S12 automatically generates a dynamic weight table based on the weight change, including: Evaluate the satisfaction score of the system state for each objective function after each iteration; The weighted value is calculated by the unconventional exponential sliding average algorithm to update the current importance coefficient of each objective function, that is, , in, is the objective function weight after the next iteration update, is the smoothing factor, and its value range is , is the objective function weight in the current iteration, Indicates the newly measured score; Reset the dynamic weight distribution table based on the updated objective function importance to adapt to the changing needs; The redistributed weight ratio is used to carry out the objective function evaluation and optimization loop iteration process in the next stage.

4. The method for controlling operating parameters of a power digital information system based on a heuristic algorithm according to claim 3 is characterized in that: The unconventional formula for calculating the weighted value by the exponential sliding average algorithm to update the current importance coefficient of each objective function also includes: Identify the target items that have the greatest impact on system performance ; When the importance score is less than the threshold And when the number of iterations exceeds the predetermined threshold, a reweighting event is triggered otherwise ; If triggered, , applying special correction factors ,make ,when Maintain normal operation ,in, Indicates the target item in the system that has the greatest impact on performance. represents the threshold value, A flag indicating a reweighting event, represents the special correction factor, represents the adjustment coefficient, Represents the current weight of the objective function; After the adjustment, the weights are used to readjust the multi-objective decision-making process of the entire system until the optimal solution is found.

5. The method for controlling operating parameters of a power digital information system based on a heuristic algorithm according to claim 4 is characterized in that: The if triggering of the unconventional formula also includes a determination step of whether a stable state appears during the iteration process, which specifically includes the following: Monitor the rate of change of the current state relative to the previous state ; Based on the change rate being below a certain limit Indicates that it has entered a static area, so the stability flag is set ; When no substantial changes occur, , forced to reduce the acceleration parameter The value is set to zero to avoid oscillatory behavior; After the relative stagnation detection is completed, the system will re-enable the regular learning rate to continue searching for the optimal point until it is fully covered or one of the termination criteria is met; Wherein, the change rate is lower than a certain limit It shows that entering the static zone also includes a continuous monitoring mechanism for environmental adaptability, specifically: After a certain period of time , more recently The maximum difference between round calculation results Switch the prelude in recognition mode; Identify current environmental trends Increase, decrease or remain unchanged to adjust the local optimal path guidance rules; If a large deviation is detected in the environment, Adjusting the global governing equations Maintaining strategic consistency; Perform environment matching analysis.

6. The method for controlling operating parameters of a power digital information system based on a heuristic algorithm according to claim 1 is characterized in that: The step S2 comprises: By using a multi-objective heuristic algorithm, the safe and economic operation of the power digital information system is decomposed into multiple independently executable goals; Implementing parallel optimization of these objectives on a parallel computing architecture reduces computation time while ensuring that each objective is solved efficiently; When the status of the power digital information system changes, re-evaluate the importance of each sub-goal and re-arrange the priority; Introduce additional safety constraints to verify in real time whether the minimum safety standards are met.

7. The method for controlling operating parameters of a power digital information system based on a heuristic algorithm according to claim 6 is characterized in that: The step S2 further comprises: The goal of safe operation is to minimize the probability of failure of the power digital information system, specifically: ； in, represents minimization, is the failure probability of the power digital information system, It is The safety index of the associated unit, Indicates the serial number of the associated unit, It represents the total number of all associated units in the power digital information system; In the economic operation goal, the operation cost is minimized, specifically: ； in, represents minimization, represents the operating cost of the power digital information system, Indicates the serial number corresponding to the fault type. represents the number of all fault types in the power digital information system, Indicates The cost corresponding to each failure type; Using the parallel computing architecture, the above objective function and Independent solution is performed and parallel optimization strategy is introduced. The parallel optimization process can be implemented through the master-slave parallel computing framework, including allocating the objective function to multiple computing nodes, each node independently processing part of the optimization task, each node using heuristic algorithm for local optimization, and exchanging results with the master node after each round of iteration. The master node summarizes the local optimization results, adjusts the weights through the global heuristic strategy, and updates the global optimal solution. For the local optimization of each node, the specific formula is: ； in, represents the solution vector for the next iteration, represents the solution vector of the current iteration, A sequence number indicating the number of iterations, represents the local optimal solution, represents the global optimal solution, represents the parameter that controls the local convergence speed, represents the parameter that controls the global convergence speed; When the state of the power digital information system changes, the weight of the objective function is evaluated in real time, and the priority is reallocated based on the state change; At each iteration, it is verified whether the CPU occupancy and memory occupancy of the power digital information system are within the safe operating range. By performing security verification on the results of each round of iteration, if it is found that the minimum safety standard is not met, the optimization direction is adjusted or the weight is reallocated to ensure the safe operation of the power digital information system.

8. The method for controlling operating parameters of a power digital information system based on a heuristic algorithm according to claim 1 is characterized in that: The step S3 comprises: Evaluate the rationality of current settings based on real-time operating data; Machine learning technology is used to predict state changes in the future and formulate parameter update strategies, including using time series prediction models to predict future state changes in the power digital information system. The specific formula is: ； in, Represents the predicted future time point The state of Indicates the time, represents the time interval of state prediction, represents a machine learning model, Indicates the current time point System status data, Respectively represent the system status data before the current time point, The number of time steps to be backtracked is used to dynamically adjust the parameters of the optimization algorithm according to the prediction results; Adjust the search area and expand the search range to introduce new potential solutions when it is detected that the initial parameters cannot produce optimized results.

9. The method for controlling operating parameters of a power digital information system based on a heuristic algorithm according to claim 1, characterized in that: The multi-level verification mechanism in step S4 includes: Real-time monitoring of the core operating parameters of the power digital information system, including CPU occupancy and memory occupancy, using the threshold detection mechanism to set the safety range of the parameters. If the actual value of a certain operating parameter of the power digital information system is not within the preset safety range, an alarm will be triggered; Use performance indicators to evaluate the effect after adjustment. The specific formula is: ； in, represents the steady-state error, represents the target state value, Indicates at time The actual state value at Indicates the moment; The weighted comprehensive evaluation function of the multi-objective optimization results is used for comprehensive stability evaluation. The specific formula is: ； in, represents the value of the comprehensive stability evaluation function, Indicates the weight corresponding to the CPU usage. Indicates the weight corresponding to the memory usage, Indicates the CPU usage. Indicates the memory usage.

10. The method for controlling operating parameters of a power digital information system based on a heuristic algorithm according to claim 1, characterized in that: The specific formula for automatically adjusting the configuration parameters through feedback information in step S4 is: ； in, Indicates the current time The parameter vector of Indicates the time, represents the learning rate, represents the system stability loss function, represents the parameter vector, Represents the loss function For parameter vector gradient.

Images