PID (Proportion Integration Differentiation) parameter optimization method based on improved graffiti search algorithm

Abstract

The invention discloses a PID (Proportion Integration Differentiation) parameter optimization method based on an improved graffiti search algorithm, and relates to the technical field of PID. The invention aims to solve the problems of insufficient local search capability, slow convergence speed and poor system stability in the prior art, thereby realizing a better PID parameter optimization effect. Specifically, the PID parameter optimization method based on the improved graffiti search algorithm comprises the following steps: constructing a PID controller model; a step of replacing a fixed perception probability and a step length in the model with an adaptive perception probability and a Levy flight step length; a step of adding a crow misleading prevention factor into the model; and acquiring preset parameters in the PID controller model based on a particle swarm algorithm. The preset parameters comprise proportional (Kp), integral (Ki) and differential (Kd) parameters. Parameters of the Levy flight step length are dynamically adjusted through a Bayesian optimization algorithm. The method is suitable for being applied to PID parameter optimization work.

Description

Technical Field

[0001] The invention relates to the technical field of PID, and in particular to PID parameter optimization based on an improved crow search algorithm. Background Art

[0002] PID control technology is widely used in industrial control systems and has been widely recognized for its simplicity, reliability and adaptability. The PID controller achieves rapid response and stable control of the system by adjusting three parameters: proportional, integral and derivative. However, the adjustment effect of traditional PID control methods in nonlinear and complex dynamic systems is often not ideal. The main reason is that the selection of parameters is crucial to the performance of the system, and the process of manually adjusting PID parameters is time-consuming and difficult to achieve the best control effect.

[0003] In order to overcome this problem, various intelligent optimization algorithms have been introduced into the automatic adjustment of PID parameters, such as particle swarm optimization (PSO) and genetic algorithm (GA). These intelligent optimization algorithms can improve the effect of PID parameter optimization to a certain extent by simulating the behavior of certain organisms in nature and exploring the optimal solution. However, the existing optimization algorithms generally have the following technical problems:

[0004] 1. Weak local search capability: Many intelligent optimization algorithms are prone to falling into local optimality, especially when faced with complex search spaces, resulting in the parameter combination finally found not being the global optimal solution.

[0005] 2. Slow convergence: Due to the strong randomness of the algorithm, many optimization algorithms require a large number of iterations during the convergence process, resulting in low optimization efficiency and high time cost.

[0006] 3. Lack of stability: When faced with nonlinear systems and external interference, the search results of existing algorithms often lack robustness, making it difficult to ensure the stability of the control system under various conditions.

[0007] For example, the Crow Search Algorithm (CSA) is an optimization algorithm based on the intelligent behavior of crows, which performs global search by simulating the behavior of crows hiding food. However, the original Crow Search Algorithm has problems with fixed perception probability and step size, which leads to its weak local search ability on complex problems, and lacks the guidance of the optimal individual, making it difficult to achieve effective convergence. Summary of the invention

[0008] In view of the above technical problems, the present invention proposes an improved crow search algorithm, which enhances the global and local search capabilities of the algorithm by introducing adaptive perception probability, Lévy flight step, anti-misleading factor and particle swarm search strategy, and significantly improves the regulation effect of PID controller in complex dynamic systems. This improvement scheme aims to solve the problems of insufficient local search capability, slow convergence speed and poor system stability in the prior art, thereby achieving better PID parameter optimization effect. Specifically:

[0009] A PID parameter optimization method based on an improved crow search algorithm, the method comprising:

[0010] Steps to build a PID controller model;

[0011] The fixed sensing probability and step size in the model are replaced by the steps of adaptive sensing probability and Lévy flight step size;

[0012] In the model, a step of adding a crow anti-misleading factor;

[0013] The steps of collecting preset parameters in the PID controller model based on the particle swarm algorithm.

[0014] Furthermore, a preferred implementation is provided, in which preset parameters in the PID controller model are collected through the guidance of the optimal individual.

[0015] Furthermore, a preferred embodiment is provided, wherein the preset parameters include proportional (Kp), integral (Ki) and differential (Kd) parameters.

[0016] Furthermore, a preferred embodiment is provided, in which the controlled object is described by a transfer function in the PID controller model.

[0017] Furthermore, a preferred embodiment is provided, in which the ITAE system performance index is used as a fitness function in the PID controller model to evaluate the quality of each set of parameters.

[0018] Furthermore, a preferred embodiment is provided, in which the parameters of the Lévy flight step length are dynamically adjusted through a Bayesian optimization algorithm.

[0019] Based on the same inventive concept, the present invention also provides a PID parameter optimization device based on an improved crow search algorithm, the device comprising:

[0020] Modules for building PID controller models;

[0021] The fixed perception probability and step size in the model are replaced by a module with adaptive perception probability and Lévy flight step size;

[0022] In the model, a module of crow anti-misleading factor is added;

[0023] A module that collects preset parameters in the PID controller model based on the particle swarm algorithm.

[0024] Based on the same inventive concept, the present invention also provides a computer storage medium for storing a computer program, and when the computer reads the computer program, the computer executes the described method.

[0025] Based on the same inventive concept, the present invention also provides a computer, including a processor and a storage medium. When the processor reads a computer program stored in the storage medium, the computer executes the method described.

[0026] Based on the same inventive concept, the present invention also provides a computer program product, which is a computer program. When the computer program is executed, the method described above is implemented.

[0027] Compared with the prior art, the technical solution provided by the present invention is beneficial in that:

[0028] The introduction of adaptive perception probability and Lévy flight step has significantly enhanced the global search capability of the crow search algorithm. The adaptive perception probability can automatically adjust the perception range according to the progress of iteration, so that the algorithm has a larger exploration range in the early stage, and pays more attention to refined search in the later stage, reducing the possibility of falling into the local optimum. The Lévy flight step increases the coverage of individuals in the search space by introducing long-distance jumps, effectively improving the diversity of the population, especially in complex nonlinear systems, through a larger range of exploration, the probability of finding the global optimal solution is increased. Compared with the traditional fixed step method, these improvements perform well in balancing exploration and development capabilities, significantly shortening the convergence time and improving search efficiency.

[0029] The use of the crow anti-misleading factor significantly improves the stability and accuracy of local search. Traditional optimization algorithms are often susceptible to local optimal traps when dealing with multi-peak problems, causing the algorithm to converge to a suboptimal solution early. The anti-misleading factor adjusts the behavior of each individual so that when misled, the individual can effectively identify and correct it, avoiding the repetition of invalid searches. Especially in nonlinear dynamic systems, the addition of the anti-misleading factor makes the population more flexible in the face of misleading, improves the algorithm's search accuracy in local areas, and ensures the stability of the overall optimization process. This feature demonstrates stronger adaptability in complex industrial control scenarios.

[0030] The combination of particle swarm search strategy further improves the search accuracy near the optimal solution. By introducing the guiding role of the optimal individual on other individuals, the particle swarm search strategy ensures that the population can quickly concentrate and accelerate convergence when approaching the optimal solution, thereby obtaining a more accurate solution. Compared with using only the crow search algorithm, the guidance strategy of adding particle swarm has a significant effect on improving local development capabilities, reducing oscillation phenomena near the optimal solution, and improving convergence speed and accuracy. At the same time, this combined method effectively utilizes the advantages of the particle swarm algorithm while avoiding the premature convergence problem that may occur in the particle swarm algorithm, so that the entire optimization process maintains the ability of global exploration and performs well in local fine search. This plays a key role in maintaining excellent dynamic response and stability of the PID control system under various working conditions.

[0031] It is suitable for use in PID parameter optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 Algorithm flow chart for optimizing PID controller for improving crow search algorithm;

[0033] Figure 2 Schematic diagram of the process of optimizing PID parameters for improving the crow search algorithm;

[0034] Figure 3 This is the Simulink simulation model diagram of the PID controller;

[0035] Figure 4 It is a comparison chart of ITAE iteration curves of the original crow search algorithm and the improved crow search algorithm;

[0036] Figure 5 This is a comparison chart of the effects of optimizing PID controller using the original crow search algorithm and the improved crow search algorithm. DETAILED DESCRIPTION

[0037] In order to make the advantages and benefits of the technical solution provided by the present invention more clearly reflected, the technical solution provided by the present invention is now further described in detail with reference to the accompanying drawings, specifically:

[0038] Implementation method 1: This implementation method provides a PID parameter optimization method based on an improved crow search algorithm, the method comprising:

[0039] Steps to build a PID controller model;

[0040] The fixed sensing probability and step size in the model are replaced by the steps of adaptive sensing probability and Lévy flight step size;

[0041] In the model, a step of adding a crow anti-misleading factor;

[0042] The steps of collecting preset parameters in the PID controller model based on the particle swarm algorithm.

[0043] Specifically, this embodiment proposes a PID parameter optimization method based on an improved crow search algorithm, which mainly optimizes the proportional (Kp), integral (Ki) and differential (Kd) parameters of the PID controller by improving the existing crow search algorithm. The overall process includes building a PID controller model, improving the crow search algorithm, optimizing the PID controller parameters, and verifying the optimization effect through simulation. This solution can improve the control accuracy, convergence speed and robustness of the system.

[0044] include:

[0045] Build a PID controller model: Create a basic model of a PID controller to control the system.

[0046] Improved Crow Search Algorithm: The Crow Search Algorithm is improved both globally and locally to improve its search efficiency and robustness.

[0047] PID parameter optimization: Use the improved crow search algorithm to optimize the Kp, Ki, and Kd parameters of the PID controller.

[0048] Simulation verification: The improved PID control system is simulated through MATLAB and Simulink to verify the optimization effect.

[0049] The specific steps are:

[0050] Building a PID Controller Model

[0051] First, build a PID controller model to simulate the basic architecture of system control. The PID controller model contains the following modules:

[0052] Parameter input module: input three parameters Kp, Ki, and Kd.

[0053] Deviation calculation module: Calculates the difference between the target value and the real-time value.

[0054] PID controller module: adjusts the system output according to the set PID parameters.

[0055] Controlled object module: describes the controlled object through the system transfer function.

[0056] This model is the basis of the entire optimization process and its output will serve as input in subsequent steps.

[0057] Improved crow search algorithm:

[0058] In this step, the global and local search capabilities of the crow search algorithm are improved:

[0059] Improvement of global mobility strategy: Adaptive perception probability and Lévy flight step size are used instead of fixed perception probability and step size, so that the population has stronger global search ability in the early stage and improves the ability of local refined search in the later stage.

[0060] Improvement of local search strategy: Introducing the crow anti-misleading factor and particle swarm search strategy. The anti-misleading factor enables the crow to have a higher probability of identifying misleading and adjusting its search behavior when it is misled; the particle swarm search strategy guides the optimal individual to enhance the search accuracy of the population near the optimal solution.

[0061] These improvements enable the crow search algorithm to strike a balance between diversity and stability, ensuring more efficient parameter optimization. The output of the improved algorithm is used to guide the search process for PID parameter optimization.

[0062] PID parameter optimization:

[0063] The three key parameters (Kp, Ki, Kd) of the PID controller are optimized using the improved crow search algorithm:

[0064] Parameter initialization: set population size, search dimension, boundary value, maximum number of iterations, etc.

[0065] Initial position setting: Randomly select N groups of parameters within the parameter range of the PID control system as the initial positions of the individuals in the population.

[0066] Fitness function construction: Use ITAE system performance indicators as the fitness function to evaluate the pros and cons of each set of parameters.

[0067] Iterative optimization: By calculating the perception probability, anti-misleading probability and individual fitness value, the position of each individual in the population is updated, and iteration is continued until the optimal solution is reached or the maximum number of iterations is reached.

[0068] Through the above process, the optimized PID control parameters are finally obtained. These optimized parameters are used to improve the response speed and stability of the system.

[0069] There is also simulation verification:

[0070] Finally, MATLAB and Simulink are used to simulate and test the optimized PID control system:

[0071] Verify the control effect: Compare the performance differences between the original crow search algorithm and the improved algorithm in PID parameter optimization, mainly through ITAE index curve, overshoot, system stabilization time and other parameters for evaluation.

[0072] Implementation method 2: This implementation method further limits the PID parameter optimization method based on the improved crow search algorithm provided in implementation method 1. The preset parameters in the PID controller model are collected through the guidance of the optimal individual.

[0073] Implementation method three: This implementation method is a further limitation of the PID parameter optimization method based on the improved crow search algorithm provided in implementation method one, and the preset parameters include proportional (Kp), integral (Ki) and differential (Kd) parameters.

[0074] Implementation mode 4: This implementation mode further limits the PID parameter optimization method based on the improved crow search algorithm provided in implementation mode 1. In the PID controller model, the controlled object is described by a transfer function.

[0075] Implementation method five: This implementation method is a further limitation of the PID parameter optimization method based on the improved crow search algorithm provided in implementation method one. In the PID controller model, the ITAE system performance index is used as the fitness function to evaluate the advantages and disadvantages of each group of parameters.

[0076] Implementation method six: This implementation method is a further limitation of the PID parameter optimization method based on the improved crow search algorithm provided in implementation method one, and the parameters of the Lévy flight step length are dynamically adjusted through a Bayesian optimization algorithm.

[0077] Embodiment 7: This embodiment provides a PID parameter optimization device based on an improved crow search algorithm, the device comprising:

[0078] Modules for building PID controller models;

[0079] The fixed perception probability and step size in the model are replaced by a module with adaptive perception probability and Lévy flight step size;

[0080] In the model, a module of crow anti-misleading factor is added;

[0081] A module that collects preset parameters in the PID controller model based on the particle swarm algorithm.

[0082] Embodiment 8: This embodiment provides a computer storage medium for storing a computer program. When the computer reads the computer program, the computer executes the method provided in embodiment 1.

[0083] Embodiment 9: This embodiment provides a computer, including a processor and a storage medium. When the processor reads a computer program stored in the storage medium, the computer executes the method provided in embodiment 1.

[0084] Embodiment 10: This embodiment provides a computer program product, which is a computer program. When the computer program is executed, the method provided in embodiment 1 is implemented.

[0085] Implementation Method XI: Combination Figure 1-5 This embodiment is described in detail by using specific examples to further describe the technical solution provided above. Specifically:

[0086] include:

[0087] Step 1: Build a PID controller model.

[0088] Step 2: Use adaptive perception probability and Lévy flight step size to replace the original fixed perception probability and fixed step size to improve the global movement strategy of the crow search algorithm; introduce the crow anti-misleading factor and particle swarm search strategy to improve the local search strategy of the crow search algorithm.

[0089] Step 3: Use the improved crow search algorithm to optimize the parameters of the PID controller. After a certain number of iterations, the three parameters Kp, Ki, and Kd of the optimized PID controller are obtained.

[0090] Step 4: Use MATLAB and Simulink to simulate the PID control system.

[0091] The specific steps are:

[0092] In step one, the constructed PID controller model includes a parameter input module, a target value and a real-time value deviation calculation module, a deviation value output module, a PID controller module, and a controlled object module.

[0093] In the step 1, the transfer function formula of the selected controlled object is:

[0094]

[0095] Where G(s) is the transfer function and s is the function variable.

[0096] In step 2, adaptive perception probability AP is used to replace the original fixed perception probability, so that the individuals in the population have a lower perception probability in the early and middle stages of the algorithm, and the probability is exponentially improved in the later stages of the algorithm, aiming to increase the diversity of search strategies while maintaining the global search capability of the original algorithm. The formula for adaptive perception probability AP is as follows:

[0097]

[0098] Where T represents the current number of iterations; Tmax represents the maximum number of iterations; AP TRepresents the perception probability of individuals in the population at the Tth iteration.

[0099] In step 2, the Lévy flight step length ls is used to replace the original fixed step length, which increases the difference in the individual moving distances in the crow population and improves the global search capability in the early stage of the algorithm. The improved formula is as follows:

[0100]

[0101] In the formula Represents the new position of individual i in the search space; Indicates the current position of individual i; represents the best position stored in the memory of individual k (i.e., the location where crow k hides the food); ls represents the Lévy flight step length, as shown below:

[0102]

[0103] Where α represents the step size scaling factor, which can be adjusted according to needs; β represents the power law exponent, which is used to adjust the distribution characteristics; u and v obey the normal distribution, σ v =1,σ u The formula is as follows:

[0104]

[0105] In step 2, a crow anti-misleading factor AM is introduced to give the crows an anti-misleading ability that gradually improves with iterations. When the crows are misled to a random location by the tracking object, they are more likely to realize that they are being tricked, making the search behavior of the population more diverse and intelligent; in the early stage of the search, they are more inclined to global search, and in the later stage of the search, they are more inclined to local search, in order to obtain higher quality results. The formula of the anti-misleading factor is as follows:

[0106]

[0107] Where T represents the current number of iterations; Tmax represents the maximum number of iterations; AM T It represents the anti-misleading probability of individuals in the population at the Tth iteration.

[0108] In step 2, a particle swarm search strategy is introduced to improve the local search strategy of the crow search algorithm, and the population is guided by the optimal individual to conduct a more detailed search near the optimal individual. The formula of the search strategy is as follows:

[0109]

[0110] Where ls represents the Lévy flight step length; C1 and C2 represent weight coefficients; represents the position of the global optimal individual; Represents the position of the optimal individual in the population.

[0111] In step three, the improved crow search algorithm is used to optimize the PID controller parameters, and the three parameters Kp, Ki, and Kd of the optimized PID controller are iteratively obtained. The specific steps are as follows:

[0112] S1. Initialize algorithm parameters: Set the population size N, dimension dim, maximum boundary ub, minimum boundary lb, and maximum number of iterations Tmax of the improved crow search algorithm.

[0113] S2. The parameter value range of the PID control system is used as the population search space of the crow search algorithm. A group of PID control system parameters are randomly selected within the value range as the initial position of the population individuals. A total of N groups of parameters are selected as the initial position of the population.

[0114] S3. Construct a fitness function based on the ITAE system performance indicators. The calculation formula is as follows:

[0115]

[0116] Where t represents time; e(t) represents the difference between the current value and the expected value.

[0117] S4. Calculate the current fitness value of each individual in the crow population And store the fitness value and position of each individual in memory, expressed as Select the minimum fitness value at the current number of iterations as the optimal fitness value Select the individual with the smallest fitness value at the current iteration as the global optimal individual

[0118] S5. Calculate the perception probability AP of the current iteration number T and the probability of anti-misleading AM T .

[0119] S6. Using the process of crows searching for food to update the position of each individual in the population, three strategies are mainly used, namely, strategy 1: tracking behavior; strategy 2: misled behavior; strategy 3: anti-misleading behavior. The specific descriptions of the three strategies are as follows:

[0120] Strategy 1: Tracking behavior. Crow i will randomly track a crow k. When the current perception probability AP of crow k is T When the value is less than a random number Rand between 0 and 1, crow i will hide the food according to the location of crow k. To update the position, the update formula is as follows:

[0121]

[0122] Where v follows a standard normal distribution, σ u The formula is as follows:

[0123]

[0124] Strategy 2: Misled behavior. When crow i randomly follows crow k, if crow k’s perception probability AP T Greater than Rand1 and the anti-misleading probability AM of crow i T If it is less than Rand2, crow i will be misled by crow k and randomly generate a new position in the search space. The position update formula is as follows:

[0125]

[0126] Where ub represents the maximum boundary of the search space; lb represents the minimum boundary of the search space.

[0127] Strategy 3: Anti-misleading behavior. When crow i randomly follows crow k, if crow k’s perception probability AP T Greater than Rand1 and the anti-misleading probability AM of crow i T If Rand2 is greater than crow i, crow i will not be misled by crow k, and crow i will follow the position of the global optimal individual. and the position of the optimal individual in the population To update the position, the update formula is as follows:

[0128]

[0129] S7. Calculate the current fitness value of each individual in the crow population after updating its position like Greater than Then Updated to Updated to Otherwise the fitness value and position stored in memory remain unchanged.

[0130] S8. Select the individual with the smallest fitness value in the population at the current number of iterations as the optimal individual in the population And its fitness value With the optimal fitness value For comparison, if Less than Then Updated to Updated to Otherwise, the optimal fitness value and the position of the global optimal individual remain unchanged.

[0131] S9. Determine whether the maximum number of iterations has been reached. If so, stop the optimization and output the optimal parameters; otherwise, return to S4 to continue the optimization.

[0132] S10, assigning the searched optimal position to Kp, Ki, and Kd respectively, and transferring the optimized parameters to the PID controller to complete the optimization of the PID controller parameters.

[0133] In this example, the parameter settings involved include N=20, dim=3, ub=50, lb=0, Tmax=100, α=0.1, β=1, C1=2, C2=1. Using the above parameters, according to Figures 1 to 3 The design process finally Figure 4 and Figure 5 The simulation results are as follows. Figure 4 and Figure 5 It can be seen that the improved crow search algorithm has stronger search ability and higher search efficiency than the original algorithm, and can converge to a better solution at a faster convergence speed. The PID controller optimized by the improved algorithm has a significantly smaller overshoot than the original algorithm, which enables the system to enter a stable state faster. This fully demonstrates the superiority of the improved crow search algorithm.

[0134] Based on the original crow search algorithm, this implementation method uses adaptive perception probability and Lévy flight step length to replace the original fixed perception probability and fixed step length to improve the global movement strategy of the crow search algorithm; introduces crow anti-misleading factor and particle swarm search strategy to improve the local search strategy of the crow search algorithm. The problems of weak local search ability, low search efficiency, slow convergence speed and lack of stability of the crow search algorithm are solved. The search strategy for different search periods is adaptively adjusted to enhance the stability and search performance of the algorithm. At the same time, the improved crow search algorithm is used to optimize the PID controller, which improves the adaptability and robustness of the PID control system.

[0135] The technical solution provided by the present invention is further described in detail above through several specific implementation modes in order to highlight the advantages and benefits of the technical solution provided by the present invention. However, the several specific implementation modes described above are not intended to be used as limitations on the present invention. Any reasonable modification and improvement of the present invention, combination of implementation modes and equivalent substitution within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A PID parameter optimization method based on an improved crow search algorithm, characterized in that: Methods include: Steps to build a PID controller model; The fixed sensing probability and step size in the model are replaced by the steps of adaptive sensing probability and Lévy flight step size; In the model, a step of adding a crow anti-misleading factor; The steps of collecting preset parameters in the PID controller model based on the particle swarm algorithm.

2. The PID parameter optimization method based on the improved crow search algorithm according to claim 1 is characterized in that: Through the guidance of the optimal individual, the preset parameters in the PID controller model are collected.

3. The PID parameter optimization method based on the improved crow search algorithm according to claim 1 is characterized in that: The preset parameters include proportional (Kp), integral (Ki) and differential (Kd) parameters.

4. The PID parameter optimization method based on the improved crow search algorithm according to claim 1 is characterized in that: In the PID controller model, the controlled object is described by a transfer function.

5. The PID parameter optimization method based on the improved crow search algorithm according to claim 1 is characterized in that: In the PID controller model, the ITAE system performance index is used as the fitness function to evaluate the quality of each set of parameters.

6. The PID parameter optimization method based on the improved crow search algorithm according to claim 1 is characterized in that: The parameters of the Lévy flight step are dynamically adjusted through a Bayesian optimization algorithm.

7. A PID parameter optimization device based on an improved crow search algorithm, characterized in that: The device includes: Modules for building PID controller models; The fixed perception probability and step size in the model are replaced by a module with adaptive perception probability and Lévy flight step size; In the model, a module of crow anti-misleading factor is added; A module that collects preset parameters in the PID controller model based on the particle swarm algorithm.

8. A computer storage medium for storing a computing program, characterized in that: When the computer reads the computer program, the computer executes the method of claim 1.

9. A computer, comprising a processor and a storage medium, characterized in that: When the processor reads the computer program stored in the storage medium, the computer executes the method of claim 1 .

10. A computer program product, being a computer program, characterized in that When the computer program is executed, the method of claim 1 is implemented.

Images