Fishery oxygenation control optimization method based on meta-heuristic algorithm

Abstract

The invention discloses a fishery oxygenation control optimization method based on a meta-heuristic algorithm, and belongs to the technical field of control optimization methods, and the method mainly comprises the following steps: S1, constructing a fishery oxygenation PID control system; s2, constructing an improved white sheep seat element heuristic algorithm module; s3, optimizing and setting parameters; and S4, parameter setting and effect optimization. According to the method, by introducing strategies such as interstellar distant symptoms, emotional outbreak and nebula aggregation, the global optimization and convergence capacity is remarkably enhanced, nearly-zero overshoot of dissolved oxygen adjustment is achieved through the PID controller set through the improved white sheep seat element heuristic algorithm, the response time can be effectively shortened, the steady-state error can approach zero, and the method is suitable for large-scale popularization and application. In this way, energy consumption and accumulative errors can be effectively reduced while rapid and accurate control is guaranteed, and the robustness of a fishery oxygenation PID control system can be remarkably improved.

Description

Technical Field

[0001] This invention belongs to the technical field of control optimization methods, and particularly relates to a fishery oxygenation control optimization method based on metaheuristic algorithms. Background Technology

[0002] Water aeration is an essential technical means to maintain good physiological function of aquatic organisms and the homeostasis of aquaculture ecosystem. Low oxygen stress not only inhibits the standard metabolic rate and induces oxidative stress, but also leads to the accumulation of reduced harmful substances by weakening the kinetics of nitrification. Meanwhile, supersaturation of dissolved oxygen in water can damage tissue function through the gas embolism effect and physiological damage mediated by reactive oxygen species. Therefore, more precise oxygenation control essentially seeks the global Pareto optimality among physiological homeostasis, metabolic efficiency, and biochemical cycle balance. However, existing oxygenation control schemes mainly rely on manual experience-based timing or fixed threshold triggering logic. This discrete control based on static judgment exhibits significant response lag and time delay characteristics when facing random disturbances such as atmospheric pressure fluctuations, sudden changes in irradiance, and differences in biological load. Furthermore, due to the lack of adaptability of the control strategy to time-varying environments, it often leads to violent fluctuations in dissolved oxygen levels, which can easily cause stress responses in aquaculture organisms. Therefore, it is necessary to introduce PID control into the oxygenation controller. In terms of PID controller parameter tuning, traditional engineering tuning methods or empirical trial-and-error methods have limited improvement in the accuracy of oxygenation regulation due to their strong subjectivity and low debugging efficiency. Metaheuristic optimization algorithms such as particle swarm optimization and genetic algorithms introduced during tuning have shown certain advantages in global search, but they still face bottlenecks such as convergence stagnation, susceptibility to local extrema, and poor dynamic adaptability. Summary of the Invention

[0003] To overcome the technical problems described in the background section, this invention provides a fishery oxygenation control optimization method based on a metaheuristic algorithm. By introducing strategies such as interstellar expedition, emotional burst, and nebula aggregation, the global optimization and convergence capabilities are significantly enhanced. The PID controller tuned using the improved Aries metaheuristic algorithm achieves near-zero overshoot in dissolved oxygen regulation and effectively shortens the response time. The steady-state error can approach zero. Thus, while ensuring fast and accurate control, it can effectively reduce energy consumption and cumulative error, and significantly improve the robustness of the fishery oxygenation PID control system.

[0004] The technical solution of this invention is: a fishery oxygenation control optimization method based on metaheuristic algorithms, comprising the following steps: S1. A fishery oxygenation PID control system is constructed based on the improved Aries meta-heuristic algorithm module. The fishery oxygenation PID control system includes a dissolved oxygen error calculation module, a dissolved oxygen PID controller module, an improved Aries meta-heuristic algorithm module, a dissolved oxygen regulation module, and a dissolved oxygen monitoring module. S2. Introducing an improved Aries metaheuristic algorithm: Constructing an improved Aries metaheuristic algorithm module. The improvement strategies for the improved Aries metaheuristic algorithm include: S21, the parameter adaptive time-varying strategy, dynamically adjusts the maturation impulse control step size and quantum expansion factor throughout the entire ephemeris evolution cycle, achieving a balanced optimization effect on the efficiency of global exploration and local development at different stages of the algorithm. S22, the interstellar expedition strategy, introduces the Lévy flight mechanism in the individual independent exploration stage, and uses its long-tail distribution characteristics to realize long-distance nonlinear jumps, which enhances the algorithm's ability to escape local optima traps. S23, the emotional burst search strategy, by combining Beta distribution and large perturbation coefficient during the individual independent exploration stage, achieves wide-area relocation of the solution space and improves the diversity preservation effect of the population. S24, Nebula Aggregation and Cooperation Strategy: In the Nebula Aggregation and Cooperation stage, quantum behavior and sage consensus mechanism are introduced to achieve a flexible and efficient convergence effect of the population towards the dominant region. S25, the mirror dimension reversal strategy, through stagnation detection and reverse learning mechanisms in the global update phase, achieves the effect of forcibly reorganizing inferior individuals when the algorithm stagnates, thereby restoring the vitality of the population; S3. The dissolved oxygen PID control parameters in the aeration PID control system of fisheries are tuned using an improved Aries metaheuristic algorithm, and the optimal dissolved oxygen PID control parameters are obtained through optimization. , , ; S4. The three optimal dissolved oxygen PID control parameters obtained by using the improved Aries metaheuristic algorithm are set as the parameters of the dissolved oxygen PID controller in the fishery oxygenation PID control system to optimize the dissolved oxygen regulation and control effect.

[0005] Furthermore, in the fishery aeration PID control system constructed in step S2, the dissolved oxygen monitoring module collects the actual dissolved oxygen feedback value and transmits it to the dissolved oxygen error calculation module. The dissolved oxygen error calculation module receives the preset target setpoint and compares the target setpoint with the actual feedback value to generate an error signal, which is then input to the dissolved oxygen PID controller module. Simultaneously, the improved Aries meta-heuristic algorithm module is used to process the internal parameters of the dissolved oxygen PID controller module. Dynamic optimization is performed to improve control accuracy. The dissolved oxygen PID controller module, after parameter optimization, calculates the control quantity based on the error signal and sends it to the dissolved oxygen regulation module to achieve closed-loop automatic regulation of dissolved oxygen concentration.

[0006] Furthermore, the parameter adaptive time-varying strategy in step S21 includes the following steps: Step S211: Set the initial maximum impulse step size And the range of variation of the quantum expansion factor; Step S212, in each iteration First, the maturation impulse control step size is calculated based on the current process. , , In the formula This indicates the step size for controlling the maturation impulse in the current iteration. This represents the preset initial maximum impulse step size. Indicates the current iteration number. This indicates the preset maximum number of iterations; Step S213: Simultaneously calculate the quantum expansion factor The calculation formula is: , In the formula Denotes the quantum expansion factor of the current iteration. This represents the preset upper limit of the quantum expansion factor. This represents the preset lower limit of the quantum expansion factor. Indicates the current iteration number. This indicates the preset maximum number of iterations.

[0007] Furthermore, the interstellar expedition strategy in step S22 includes the following steps: Step S221: Define the probability of triggering an interstellar expedition. Star Expedition Step Size Scaling Factor ,as well as Flight stability index ; Step S222: Generate random numbers during individual exploration. If the number is less than... If so, the policy will be triggered; Step S223: Using the Mantegna algorithm based on Generate Lévy flight step vector ; Step S224: Guide individuals toward the global optimum The direction is updated by jump, and the calculation formula is: , In the formula Indicates the updated position after the interstellar expedition. Indicates the individual's current location. Indicates based on preset stability index The generated Lévy flight stride vector, This represents the preset scaling factor for the interstellar expedition step size. This represents the upper bound of the search space. This represents the lower bound of the search space. This indicates pointing to the globally optimal position. The direction sign vector.

[0008] Furthermore, the emotional burst search strategy in step S23 includes the following steps: Step S231: Set the emotional arousal perturbation coefficient and the shape parameters of the Beta distribution This is used to construct perturbation operators with specific probability densities; Step S232: When the preset trigger probability is met... When the generation dimension is And a random vector that follows a Beta distribution , , Each dimension element is independently and identically distributed. ; Step S233: Perform a large-scale location update based on the Hadamard product. The calculation formula is as follows: , In the formula This indicates the updated position after an emotional outburst search. Indicates the individual's current location. Let represent a random vector that follows a standard normal distribution. Let represent a random vector that follows a Beta distribution. This represents the coefficient of emotional arousal disturbance. This represents the upper bound of the search space. This represents the lower bound of the search space, in the formula. This represents the Hadamard product of corresponding elements of a vector.

[0009] Furthermore, the nebula aggregation collaboration strategy in step S24 includes the following steps: Step S241: Calculate the consensus position of the ancient sages That is, the collective memory of all individuals' glory. The arithmetic mean is calculated using the following formula: , In the formula This indicates the position of consensus among the sages. Indicates population size, Indicates the first The location of each individual's glorious memory; Step S242: For each individual, calculate the distance vector from its current position to the consensus center of the ancients. ; Step S243: Introduce the quantum potential well model and utilize the quantum expansion factor. Generate aggregation location The calculation formula is: , In the formula Indicates the aggregation position after the quantum behavior update. This indicates the location of an individual's glorious memories. This represents a random direction symbol generated based on the standard normal distribution. Denotes the quantum expansion factor of the current iteration. This indicates the position of consensus among the sages. Indicates the individual's current location. Indicates the interval A random number that is uniformly distributed within the range; Step S244: Evaluate using a greedy strategy. An update is only performed if its fitness is better than the current state or the glory memory.

[0010] Furthermore, the mirror dimension reversal strategy in step S25 includes the following steps: Step S251: Set the stagnation detection window and fitness improvement threshold ; Step S252, if continuous The change in the global optimal fitness of each generation is less than The algorithm stalled. Step S253: Select the population with the lowest fitness ranking. Individuals as objects of reorganization; Step S254: Perform a mirror dimension reversal operation on the selected individual. The calculation formula is as follows: , In the formula This represents the reverse solution generated by reverse learning. This represents the lower bound of the search space. This represents the upper bound of the search space. This indicates the current position of the individual to be restarted; if the generated reverse solution is better, then the original individual and its glory memory will be replaced.

[0011] Furthermore, the specific steps in step S3 for tuning the dissolved oxygen PID control parameters of the fishery aeration PID control system using the improved Aries metaheuristic algorithm include: Step S31: Initialize the population dimension by setting the position of each individual star in the Aries constellation. Mapped to a set of PID control parameters And initialize the population position within the preset parameter value range; Step S32: Run the simulation of the fishery oxygenation control system, substitute each set of PID parameters into the controller, and calculate the objective function value based on the dissolved oxygen error signal output by the system as the individual's starlight brightness, i.e. fitness. Step S33: Execute the ephemeris evolution cycle, sequentially applying parameter adaptive time-varying, individual independent exploration, nebula aggregation and cooperation, and mirror dimension reversal strategies to the population. In each generation, update the individual glory memory by comparing fitness. and the global optimal position ; Step S34: Terminate the loop after reaching the preset maximum number of iterations, and output the globally optimal position. The three corresponding components are used as the optimal control parameters after tuning. .

[0012] Furthermore, the mathematical expression for the fitness function used in step S32 is: , In the formula This represents the fitness value; the smaller the value, the better. This represents the integral index of time multiplied by the absolute error. This indicates the overshoot penalty. Indicates the time penalty item for adjustment; This indicates the control of energy loss; For ITAE weights, For overshoot weights, To adjust the time weight, For energy weighting, This is the overshoot penalty scaling factor. To adjust the time penalty scaling factor.

[0013] Based on the above technical solution, the beneficial effects of this invention are as follows: 1. This invention significantly enhances the algorithm's global optimization capability and its ability to escape local optima. By introducing an interstellar expedition strategy, it utilizes the long-tail distribution characteristics of Lévy flight to achieve long-distance nonlinear jumps. Combined with an emotional burst search strategy and Beta distribution perturbation for wide-area relocation of the solution space, these two mechanisms jointly avoid the problem of the original algorithm easily getting trapped in local optima in the early stages. The fitness value of the improved algorithm at final convergence is significantly reduced, by about 37% compared to the original Aries metaheuristic algorithm, proving that the improved Aries metaheuristic algorithm can search for higher-quality global solutions. 2. This invention achieves a dynamic balance and efficient collaboration between global exploration and local development efficiency. Through a parameter adaptive time-varying strategy, the maturation impulse control step size and quantum expansion factor are dynamically adjusted, enabling the algorithm to automatically adjust its search behavior according to the iteration process. The improved Aries meta-heuristic algorithm can complete the switch from global detection to local fine search in about 10 generations, showing stronger evolutionary pressure. Combined with the nebula aggregation collaboration strategy that introduces the consensus mechanism of the sages, the collective intelligence is used to accelerate the flexible convergence of the population to the dominant region, significantly improving the convergence speed. 3. This invention effectively improves the ability to maintain population diversity and the robustness of the algorithm. The newly added mirror dimension reversal strategy, through the stagnation detection mechanism, forces individuals with poor fitness to undergo reverse learning and recombination when the algorithm is stuck. This mechanism provides the ability to restart, effectively restores the vitality of the population, prevents premature convergence due to loss of diversity, and ensures the stability of the algorithm when dealing with complex high-dimensional problems. 4. This invention significantly improves the dynamic response quality of the fishery oxygenation control system and reduces the control difficulty of large time delay systems. For fishery oxygenation controlled objects with large inertia and pure time delay characteristics, this invention uses an improved Aries metaheuristic algorithm to optimize the PID parameter combination with high differential gain, providing strong predictive compensation, realizing near-zero overshoot and smooth transition of system response, effectively eliminating the significant oscillation phenomenon in the original Aries metaheuristic algorithm, effectively shortening the settling time, and the steady-state error approaches zero, achieving fast and accurate control. 5. While ensuring rapid response, this invention effectively reduces the cumulative error and energy loss of the control system. It adopts a weighted comprehensive evaluation index including ITAE, overshoot, settling time and control energy as the fitness function to guide the algorithm to find a parameter combination that takes into account both dynamic performance and energy saving. It exhibits monotonically converging characteristics when approaching the target dissolved oxygen concentration, avoiding integral saturation or cyclic oscillation. It realizes high-precision and low-energy closed-loop automatic adjustment of dissolved oxygen concentration, which better meets the requirements of oxygenation control in aquaculture. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating the present invention.

[0015] Figure 2 This is a schematic diagram of the architecture of the fishery oxygenation PID control system of the present invention.

[0016] Figure 3 This is a comparison of the fitness convergence curves of the Aries metaheuristic algorithm of this invention and the improved Aries metaheuristic algorithm.

[0017] Figure 4This is a comparison chart of the step response curves of the Aries metaheuristic algorithm of this invention and the improved Aries metaheuristic algorithm.

[0018] Figure 5 This is a comparison chart of the PID control parameter optimization process curves of the Aries metaheuristic algorithm of this invention and the improved Aries metaheuristic algorithm. Detailed Implementation

[0019] Example 1: As Figures 1-5 As shown, this invention provides an optimization method for aquaculture oxygenation control based on a metaheuristic algorithm, comprising the following steps: S1. A fishery oxygenation PID control system is constructed based on the improved Aries meta-heuristic algorithm module. The fishery oxygenation PID control system includes a dissolved oxygen error calculation module, a dissolved oxygen PID controller module, an improved Aries meta-heuristic algorithm module, a dissolved oxygen regulation module, and a dissolved oxygen monitoring module. S2. Introducing an improved Aries metaheuristic algorithm: Constructing an improved Aries metaheuristic algorithm module. The improvement strategies for the improved Aries metaheuristic algorithm include: S21, the parameter adaptive time-varying strategy, dynamically adjusts the maturation impulse control step size and quantum expansion factor throughout the entire ephemeris evolution cycle, achieving a balanced optimization effect on the efficiency of global exploration and local development at different stages of the algorithm. S22, the interstellar expedition strategy, introduces the Lévy flight mechanism in the individual independent exploration stage, and uses its long-tail distribution characteristics to realize long-distance nonlinear jumps, which enhances the algorithm's ability to escape local optima traps. S23, the emotional burst search strategy, by combining Beta distribution and large perturbation coefficient during the individual independent exploration stage, achieves wide-area relocation of the solution space and improves the diversity preservation effect of the population. S24, Nebula Aggregation and Cooperation Strategy: In the Nebula Aggregation and Cooperation stage, quantum behavior and sage consensus mechanism are introduced to achieve a flexible and efficient convergence effect of the population towards the dominant region. S25, the mirror dimension reversal strategy, through stagnation detection and reverse learning mechanisms in the global update phase, achieves the effect of forcibly reorganizing inferior individuals when the algorithm stagnates, thereby restoring the vitality of the population; S3. The dissolved oxygen PID control parameters in the aeration PID control system of fisheries are tuned using an improved Aries metaheuristic algorithm, and the optimal dissolved oxygen PID control parameters are obtained through optimization. , , ; S4. The three optimal dissolved oxygen PID control parameters obtained by using the improved Aries metaheuristic algorithm are set as the parameters of the dissolved oxygen PID controller in the fishery oxygenation PID control system to optimize the dissolved oxygen regulation and control effect.

[0020] In summary, the dissolved oxygen monitoring module collects actual dissolved oxygen feedback values ​​and transmits them to the dissolved oxygen error calculation module. The dissolved oxygen error calculation module receives a preset target setpoint and compares it with the actual feedback value to generate an error signal, which is then input to the dissolved oxygen PID controller module. Simultaneously, an improved Aries meta-heuristic algorithm module is used to optimize the internal parameters of the dissolved oxygen PID controller module. The system performs optimization, and the dissolved oxygen PID controller module, after parameter optimization, calculates the control quantity based on the error signal and sends it to the dissolved oxygen regulation module to achieve closed-loop automatic regulation of dissolved oxygen concentration.

[0021] Accordingly, based on the thermodynamic characteristics of the fishery aeration system, the transfer function is: , In the formula Represents the transfer function of the controlled object. Indicates system gain. Represents the time constant. Indicates pure time delay. This represents the Laplace operator, after system identification. =3.5, =120.0, =15.0.

[0022] Accordingly, the mathematical expression for the output control quantity of the dissolved oxygen PID controller in the fishery aeration PID control system is: , In the formula This represents the output control quantity of the controller. Indicates proportional gain. This indicates the error between the target dissolved oxygen setpoint and the actual feedback value. Indicates integral gain. Represents the integral variable. Represents a time variable. This represents the differential gain.

[0023] Accordingly, in order to balance the dynamic performance and stability of the system, a weighted comprehensive evaluation index is adopted as the fitness function. The mathematical expression of the fitness function is: , In the formula This represents the fitness value; the smaller the value, the better. This represents the integral index of time multiplied by the absolute error. This indicates the overshoot penalty. Indicates the time penalty item for adjustment; This indicates the control of energy loss; This is the ITAE weight, with a value of 0.4; This is the overshoot weight, with a value of 0.2. To adjust the time weight, a value of 0.2 is set. This is the energy weight, with a value of 0.4; This is the overshoot penalty scaling factor, with a value of 100. The time penalty scaling factor is set to 10.

[0024] Specifically, step S3 involves tuning the dissolved oxygen PID control parameters of the fishery aeration PID control system using the improved Aries metaheuristic algorithm, which includes the following steps: Step 1: Star map initialization and parameter setting; Step 11: Define basic algorithm parameters: In this embodiment, the population size of the Aries constellation is specifically set. =50, maximum number of ephemeris evolution iterations =100, sets the dimension of the search space. These correspond to the proportional coefficients of the PID controller. Integral coefficient Differential coefficients And set a lower bound vector for the search space. =[0, 0, 0], the upper bound vector of the search space =[5.0, 2.0, 5.0]; Step 12: Initialize strategy control parameters: Set the maximum impulse step size Its value is 10% of the search space range, that is... Set the probability of triggering an interstellar expedition. =0.1 and the scaling factor for the interstellar expedition step size =0.01; setting Flight stability index =1.5; Set the shape parameter of the Beta distribution in the emotional arousal mechanism. =0.5 and =0.5, Emotional Excitation Disturbance Coefficient =0.5; sets the upper limit of the quantum expansion factor. =1.0 and lower limit =0.5; Set the stagnation detection window =20 and fitness improvement threshold ; Step 13, Initialize the Aries constellation state: in the search space Internal random generation The current position of each initial individual And calculate the starlight brightness corresponding to each individual position, i.e., the fitness value. ; Step 14: Establishing a Glorious Memory and Establishing a Leader: This involves defining the initial individual position. Directly recorded as the individual's glorious memory Record the corresponding fitness as Simultaneously, the individual with the highest starlight brightness is identified from all individuals, and its position is recorded as the initial global optimal position. ; Step 2: Ephemeris Evolution Cycle and Parameter Time Variation; At the current iteration number Less than or equal to the maximum number of iterations When the main loop is entered, parameter updates are performed first. Step 21: Update the step size for maturing impulse control. , , In the formula This indicates the step size for controlling the maturation impulse in the current iteration. This represents the initial maximum impulse step size. Indicates the current iteration number. Indicates the maximum number of iterations; Step 22: Update the quantum expansion factor , , In the formula This represents the quantum expansion factor for the current iteration, used to control the contraction behavior in the nebula aggregation mechanism. and Let represent the maximum and minimum values ​​of the factor, respectively, and let 1.0 and 0.5. Indicates the maximum number of iterations; Step 3: Individual independent exploration and memory updating; For each individual in the star cluster Perform the following exploration and update operations in sequence; Step 31: Perform an adaptive impulse exploration operation to generate a random vector that follows a standard normal distribution, and combine it with the current maturation impulse to control the step size. Perturb the individual's position to generate impulses to explore the location. , , In the formula Indicates candidate positions after impulsive exploration. Indicates the individual's current location. The dimension is A standard normally distributed random vector, Indicates the maturation impulse control step size for the current iteration; if Its fitness is better than Then update for ; Step 32: Execute the interstellar expedition operation to generate intervals. A random number is generated within the specified range. If the random number is less than the trigger probability... This guides individuals toward the global optimum. Flying and jumping to generate expedition locations , , In the formula Indicates the candidate locations after an interstellar expedition. Indicates based on set parameters Lévy flight stride vector generated by the Mantegna algorithm. This is the scaling factor for the interstellar expedition step size, with a value of 0.01. and These represent the upper and lower bounds of the search space, respectively. This indicates pointing to the globally optimal position. The direction sign vector; like Its fitness is better than Then update for ; Step 33: Perform the instinct-triggered search operation with probability. =0.1 Execute a global random reset, and with probability... Perform an emotional burst search based on a Beta distribution to generate exciting exploration locations. , , In the formula For dimension And a random vector that follows a Beta distribution, corresponding to and Each value is 0.5. This represents the emotional excitation disturbance coefficient, with a value of 0.5. For a standard normally distributed random vector, in the formula... The Hadamard product of corresponding elements of the vector; If the resulting new position fitness is better Then update ; Step 34: Update the Honor Memory operation and compare the current individual position. fitness Adaptability to the individual's honor memories ,like Superior Then update Glory Memory for ; Step 4: Nebula aggregation and collaboration, utilizing quantum behavior mechanisms to promote population convergence towards the dominant region; Step 41: Calculate the consensus position of the sages , , In the formula This represents the consensus position of the sages, which is the arithmetic mean of the glorious memories of all individuals. Indicates population size, Indicates the first The location of each individual's glorious memory; Step 42: Perform Nebula Aggregation Update. For each individual, based on their Glory Memory... The position of consensus with the sages Generate new aggregation locations , , In the formula Indicates the aggregation position after the quantum behavior update. This indicates the location of an individual's glorious memories. Indicates a random direction sign, taking either +1 or -1. Denotes the quantum expansion factor of the current iteration. Indicates the individual's current position To the position of consensus among the sages The distance vector, Representing an interval Uniformly distributed random numbers within; Step 43: Greedily accept updates and calculate the aggregation position. The fitness of the individual, if it is better than the individual's current fitness. Then update for If simultaneously superior to individual honor memory fitness Then, Honor Memory will be updated simultaneously. for ; Step 5: Mirror dimension reversal and global update; Step 51: Update the global optimum by traversing the glory memories of all current individuals. If there exists a position with a fitness better than the current global optimum... If the solution is found, then update. ; Step 52: Perform stall detection and reverse restart. If continuous... Inner generation The fitness improvement amount is less than the fitness improvement threshold. If the algorithm is in a stagnant state, then select the population with the lowest fitness ranking. =50% of the individuals perform mirror dimension reversal to generate the reverse solution , , In the formula This represents the candidate solutions generated by reverse learning. and These represent the lower and upper bounds of the search space, respectively. Indicates the current position of the individual to be restarted; like If the fitness of the individual is better than that of the original individual, then use Replace the original individual and its glorious memories; Step 6: Terminate judgment and output; Check the current iteration number Has the maximum number of iterations been reached? If it is not achieved, then Then return to step 2 to continue execution; if the target has been reached, terminate the algorithm and output the globally optimal position. As the optimal dissolved oxygen PID control parameter for the fishery aeration PID control system, the optimal dissolved oxygen PID control parameter is loaded into the PID controller module of the fishery aeration system, and the PID controller module then uses the real-time error signal as the control parameter. Calculate control quantity It also drives the aerator to operate, achieving more precise closed-loop control of dissolved oxygen concentration.

[0025] In the Matlab environment, the dissolved oxygen PID parameters of the oxygenation PID control system are tuned. The transfer function of the controlled object is:

[0026] The controlled object has a large inertial time constant of 120s and a pure time delay of 15s, which places high demands on the controller's advance compensation capability. The reference signal is set to jump from 4.0mg / L to 6.5mg / L.

[0027] The dissolved oxygen PID control parameters were optimized under the same conditions using the Aries metaheuristic algorithm and the improved Aries metaheuristic algorithm, and the results are shown in Table 1.

[0028] Table 1 Comparative Analysis Table

[0029] like Figure 2 As shown, the improved Aries metaheuristic algorithm finally converged to 160.5784, a reduction of approximately 37% compared to the original Aries metaheuristic algorithm's 255.7629. This indicates that the improved algorithm, while satisfying the system's dynamic constraints, more effectively reduced accumulated error and control energy. Furthermore, the improved Aries metaheuristic algorithm completed the switch from global probing to local fine-grained search around the 10th generation, demonstrating stronger evolutionary pressure and successfully avoiding the local optimum trap that the original algorithm fell into early on. Figure 3 As shown, due to the 15s delay of the controlled object, the response generated by the Aries metaheuristic algorithm exhibited significant oscillations, with the overshoot significantly exceeding the 5% constraint. In contrast, the improved Aries metaheuristic algorithm, through optimization of its high differential gain... =4.3042, providing stronger predictive compensation and achieving a smooth transition with near-zero overshoot. Meanwhile, the improved Aries metaheuristic algorithm exhibits monotonically converging characteristics when approaching the 6.5 reference value, avoiding the problems caused by... Integral saturation or cyclic oscillation caused by time delay coupling; such as Figure 4 As shown, all three parameters entered a stable period in the later stage of iteration, verifying the numerical stability of the improved Aries metaheuristic algorithm in handling constrained optimization problems. In summary, the improved Aries metaheuristic algorithm shows significant advantages in handling PID tuning tasks with complex time delays, not only reducing the objective function value at the mathematical level, but also improving the dynamic quality of the system at the control engineering level.

Claims

1. A method for optimizing aeration control in fisheries based on metaheuristic algorithms, characterized in that: Includes the following steps: S1. A fishery oxygenation PID control system is constructed based on the improved Aries meta-heuristic algorithm module. The fishery oxygenation PID control system includes a dissolved oxygen error calculation module, a dissolved oxygen PID controller module, an improved Aries meta-heuristic algorithm module, a dissolved oxygen regulation module, and a dissolved oxygen monitoring module. S2. Introducing an improved Aries metaheuristic algorithm: Constructing an improved Aries metaheuristic algorithm module. The improvement strategies for the improved Aries metaheuristic algorithm include: S21, the parameter adaptive time-varying strategy, dynamically adjusts the maturation impulse control step size and quantum expansion factor throughout the entire ephemeris evolution cycle, achieving a balanced optimization effect on the efficiency of global exploration and local development at different stages of the algorithm. S22, the interstellar expedition strategy, introduces the Lévy flight mechanism in the individual independent exploration stage, and uses its long-tail distribution characteristics to realize long-distance nonlinear jumps, which enhances the algorithm's ability to escape local optima traps. S23, the emotional burst search strategy, by combining Beta distribution and large perturbation coefficient during the individual independent exploration stage, achieves wide-area relocation of the solution space and improves the diversity preservation effect of the population. S24, Nebula Aggregation and Cooperation Strategy: In the Nebula Aggregation and Cooperation stage, quantum behavior and sage consensus mechanism are introduced to achieve a flexible and efficient convergence effect of the population towards the dominant region. S25, the mirror dimension reversal strategy, through stagnation detection and reverse learning mechanisms in the global update phase, achieves the effect of forcibly reorganizing inferior individuals when the algorithm stagnates, thereby restoring the vitality of the population; S3. The dissolved oxygen PID control parameters in the aeration PID control system of fisheries are tuned using an improved Aries metaheuristic algorithm, and the optimal dissolved oxygen PID control parameters are obtained through optimization. , , ; S4. The three optimal dissolved oxygen PID control parameters obtained by using the improved Aries metaheuristic algorithm are set as the parameters of the dissolved oxygen PID controller in the fishery oxygenation PID control system to optimize the dissolved oxygen regulation and control effect.

2. The fishery oxygenation control optimization method based on metaheuristic algorithm according to claim 1, characterized in that: In the fishery aeration PID control system constructed in step S2, the dissolved oxygen monitoring module collects the actual dissolved oxygen feedback value and transmits it to the dissolved oxygen error calculation module. The dissolved oxygen error calculation module receives the preset target setpoint and compares the target setpoint with the actual feedback value to generate an error signal, which is then input to the dissolved oxygen PID controller module. Simultaneously, the improved Aries metaheuristic algorithm module is used to optimize the internal parameters of the dissolved oxygen PID controller module. Dynamic optimization is performed to improve control accuracy. The dissolved oxygen PID controller module, after parameter optimization, calculates the control quantity based on the error signal and sends it to the dissolved oxygen regulation module to achieve closed-loop automatic regulation of dissolved oxygen concentration.

3. The fishery oxygenation control optimization method based on metaheuristic algorithm according to claim 2, characterized in that, The parameter adaptive time-varying strategy in step S21 includes the following steps: Step S211: Set the initial maximum impulse step size And the range of variation of the quantum expansion factor; Step S212, in each iteration First, the maturation impulse control step size is calculated based on the current process. , ， In the formula This indicates the step size for controlling the maturation impulse in the current iteration. This represents the preset initial maximum impulse step size. Indicates the current iteration number. Indicates the preset maximum number of iterations; Step S213: Simultaneously calculate the quantum expansion factor The calculation formula is: ， In the formula Denotes the quantum expansion factor of the current iteration. This represents the preset upper limit of the quantum expansion factor. This represents the preset lower limit of the quantum expansion factor. Indicates the current iteration number. This indicates the preset maximum number of iterations.

4. The fishery oxygenation control optimization method based on metaheuristic algorithm according to claim 3, characterized in that, The interstellar expedition strategy in step S22 includes the following steps: Step S221: Define the probability of triggering an interstellar expedition. Star Expedition Step Size Scaling Factor ,as well as Flight stability index ; Step S222: Generate random numbers during individual exploration. If the number is less than... If so, the policy will be triggered; Step S223: Using the Mantegna algorithm based on Generate Lévy flight step vector ; Step S224: Guide individuals toward the global optimum The direction is updated by jump, and the calculation formula is: ， In the formula Indicates the updated position after the interstellar expedition. Indicates the individual's current location. Indicates based on preset stability index The generated Lévy flight stride vector, This represents the preset scaling factor for the interstellar expedition step size. This represents the upper bound of the search space. This represents the lower bound of the search space. This indicates pointing to the globally optimal position. The direction sign vector.

5. The fishery oxygenation control optimization method based on metaheuristic algorithm according to claim 4, characterized in that, The emotional outburst search strategy in step S23 includes the following steps: Step S231: Set the emotional arousal perturbation coefficient and the shape parameters of the Beta distribution This is used to construct perturbation operators with specific probability densities; Step S232: When the preset trigger probability is met... When the generation dimension is And a random vector that follows a Beta distribution , ， Each dimension element is independently and identically distributed. ; Step S233: Perform a large-scale location update based on the Hadamard product. The calculation formula is as follows: ， In the formula This indicates the updated position after an emotional outburst search. Indicates the individual's current location. Let represent a random vector that follows a standard normal distribution. Let represent a random vector that follows a Beta distribution. This represents the coefficient of emotional arousal disturbance. This represents the upper bound of the search space. This represents the lower bound of the search space, in the formula. This represents the Hadamard product of corresponding elements of a vector.

6. The fishery oxygenation control optimization method based on metaheuristic algorithm according to claim 5, characterized in that, The nebula aggregation collaboration strategy in step S24 includes the following steps: Step S241: Calculate the consensus position of the ancient sages That is, the collective memory of all individuals' glory. The arithmetic mean is calculated using the following formula: ， In the formula This indicates the position of consensus among the sages. Indicates population size, Indicates the first The location of each individual's glorious memory; Step S242: For each individual, calculate the distance vector from its current position to the consensus center of the ancients. ; Step S243: Introduce the quantum potential well model and utilize the quantum expansion factor. Generate aggregation location The calculation formula is: ， In the formula This indicates the aggregation position after the quantum behavior is updated. This indicates the location of an individual's glorious memories. This represents a random direction symbol generated based on the standard normal distribution. Denotes the quantum expansion factor of the current iteration. This indicates the position of consensus among the sages. Indicates the individual's current location. Indicates the interval A random number that is uniformly distributed within the range; Step S244: Evaluate using a greedy strategy. An update is only performed if its fitness is better than the current state or the glory memory.

7. The fishery oxygenation control optimization method based on metaheuristic algorithm according to claim 6, characterized in that, The mirror dimension reversal strategy in step S25 includes the following steps: Step S251: Set the stagnation detection window and fitness improvement threshold ; Step S252, if continuous The change in the global optimal fitness of each generation is less than The algorithm stalled. Step S253: Select the population with the lowest fitness ranking. Individuals as objects of reorganization; Step S254: Perform a mirror dimension reversal operation on the selected individual. The calculation formula is as follows: ， In the formula This represents the reverse solution generated by reverse learning. This represents the lower bound of the search space. This represents the upper bound of the search space. This indicates the current position of the individual to be restarted; if the generated reverse solution is better, then the original individual and its glory memory will be replaced.

8. The fishery oxygenation control optimization method based on metaheuristic algorithm according to claim 7, characterized in that, The specific steps in step S3 for tuning the dissolved oxygen PID control parameters of the fishery aeration PID control system using the improved Aries metaheuristic algorithm include: Step S31: Initialize the population dimension by setting the position of each individual star in the Aries constellation. Mapped to a set of PID control parameters And initialize the population position within the preset parameter value range; Step S32: Run the simulation of the fishery oxygenation control system, substitute each set of PID parameters into the controller, and calculate the objective function value based on the dissolved oxygen error signal output by the system as the individual's starlight brightness, i.e. fitness. Step S33: Execute the ephemeris evolution cycle, sequentially applying parameter adaptive time-varying, individual independent exploration, nebula aggregation and cooperation, and mirror dimension reversal strategies to the population. In each generation, update the individual glory memory by comparing fitness. and the global optimal position ; Step S34: Terminate the loop after reaching the preset maximum number of iterations, and output the globally optimal position. The three corresponding components are used as the optimal control parameters after tuning. .

9. The fishery oxygenation control optimization method based on metaheuristic algorithm according to claim 8, characterized in that, The mathematical expression for the fitness function used in step S32 is: ， In the formula This represents the fitness value; the smaller the value, the better. This represents the integral index of time multiplied by the absolute error. This indicates the overshoot penalty. Indicates the time penalty item for adjustment; This indicates the control of energy loss; For ITAE weights, For overshoot weights, To adjust the time weight, For energy weighting, This is the overshoot penalty scaling factor. To adjust the time penalty scaling factor.

Images