Self-adaptive adjusting method of fishery oxygen adjusting control system

Abstract

The invention discloses a self-adaptive adjustment method for a fishery oxygen adjustment control system, and belongs to the technical field of PID control optimization, and the method specifically comprises the steps: 1, constructing the fishery oxygen adjustment control system; step 2, improving a football team training optimization algorithm, wherein a specific improvement strategy comprises: S1, using a dynamic multi-dimensional weight updating strategy to improve an updating strategy of followers in a collective training stage; s2, using a distance-based optimal learning strategy to improve an optimal learning strategy in group training; step 3, optimizing PID control in the fishery oxygen regulation control system by using an improved football team training optimization algorithm; 4, inputting a group of optimal parameters obtained through optimization in the step 3 into a PID controller in the fishery oxygen regulation control system; the PID controller in the fishery oxygen regulation control system is optimized through the improved football team training optimization algorithm, and the dynamic response speed and the steady-state precision of the whole control system can be remarkably improved.

Description

Technical Field

[0001] The invention belongs to the technical field of PID control optimization, and in particular relates to a self-adaptive adjustment method for a fishery oxygen regulation control system. Background Art

[0002] With the continuous development of modern fishery farming, the control of water quality in breeding ponds has become a key factor in ensuring the healthy growth of farmed organisms, especially the regulation of oxygen concentration in water. Establishing an effective oxygen control system to dynamically adjust the oxygen concentration in the pond has become an important task in breeding pond management. By accurately controlling the operating status of the aeration equipment and ensuring that the oxygen concentration in the water body remains near the target value, it helps to optimize the breeding environment and improve breeding efficiency.

[0003] Traditional PID controllers (proportional, integral, and differential controllers) are widely used in many industrial control systems. Their basic principle is to adjust the control signal according to the current value, cumulative value, and rate of change of the error, thereby achieving stable regulation of the system. In the oxygen regulation of fishery ponds, the PID controller can adjust the oxygenation equipment in real time according to the error of oxygen concentration, so as to achieve stability at the target concentration. However, traditional PID controllers have certain limitations. In the case of environmental changes or large disturbances, fixed parameters are difficult to adapt and the control effect may be affected. Secondly, the parameter adjustment process of the PID controller requires manual setting, and the performance is unstable under different working conditions. In order to solve these shortcomings, in recent years, intelligent optimization algorithms have gradually been introduced into the parameter optimization of PID controllers. The PID parameters are adaptively adjusted in real time according to environmental feedback through the algorithm, thereby improving the system's adaptability to dynamic changes.

[0004] The football team training optimization algorithm is a meta-heuristic optimization algorithm inspired by the football team training process. It simulates the three stages of team training: collective training, group training, and individual training. By simulating the phased collaboration mechanism of football team training, it shows significant advantages in global optimization and engineering applications. Its core lies in phased dynamic balance exploration and development, and improves model flexibility through unconstrained weight strategy, providing an efficient solution to complex optimization problems. Summary of the invention

[0005] The purpose of the invention is to improve the standard football team training optimization algorithm, improve the optimization speed and optimization accuracy of the algorithm during the optimization process, and increase the algorithm's adaptability and stability, so that it is more suitable for the application scenario of fishery oxygen control, and optimize the PID controller parameters in the fishery oxygen control system by using the improved football team training optimization algorithm, improve the overall control performance of the system through algorithm optimization, reduce the complexity of manual parameter adjustment, and improve the stability of oxygen concentration regulation, thereby improving the environment and breeding efficiency of fishery breeding.

[0006] In order to achieve the above object, the present invention adopts the following technical solution.

[0007] A self-adaptive adjustment method for a fishery oxygen regulation control system, the specific steps are as follows.

[0008] Step 1: Construct a fishery oxygen control system and convert the fishery oxygen control problem into a mathematical model to be optimized. The fishery oxygen control system includes: error calculation module, PID controller, improved football team training optimization algorithm module, real-time detection module and oxygen control module.

[0009] Step 2: Improve the football team training optimization algorithm. The specific improvements include: S1. A dynamic multi-dimensional weight update strategy is used to improve the update strategy of followers in the collective training stage of the football team training optimization algorithm. A dynamic multi-dimensional weight update strategy is to select the top 6 individuals in the search population with the highest fitness value. The weight of the first position increases nonlinearly with the increase of the number of algorithm iterations. The remaining 5 individual position vectors are decomposed according to different dimensions, and three new individuals are randomly formed according to different dimensions. The weights of these three new individuals are dynamically distributed by the remaining weights, and the sum of all weights is 1. S2. Use a distance-based optimal learning strategy to improve the optimal learning strategy in the group training of the football team training optimization algorithm. A distance-based optimal learning strategy is to guide the current individual by using the position of the optimal individual in the population, and adjust the position of the current individual through the distance between the optimal individual and the current individual, thereby achieving position update.

[0010] Step 3: Use the improved football team training optimization algorithm to optimize the PID control in the fishery oxygen control system, and obtain the optimal set of PID control parameters Kp, Ki, and Kd through iterative optimization of the algorithm.

[0011] Step 4: Input the optimal set of Kp, Ki, and Kd parameters obtained by optimization in step 3 into the PID controller in the fishery oxygen control system to optimize the control effect of the entire fishery oxygen control system.

[0012] Preferably, in step 1, the execution process of each module in the constructed fishery oxygen control system is as follows: by inputting the given target oxygen concentration and the actual oxygen concentration obtained by the real-time detection module into the error calculation module, the real-time error e(t) is calculated, and then the error is input into the PID controller, and the PID controller is optimized using the improved football team training optimization algorithm module to obtain the optimal parameters. The optimized PID controller calculates and outputs the control amount u(t) according to these optimal parameters, and the control amount is then transmitted to the oxygen control module, thereby adjusting the oxygen supply to achieve the required oxygen concentration, wherein the mathematical model of the oxygen control module is shown in formula (1): (1); In formula (1), O(t) represents the oxygen concentration in water, Kc represents the oxygen consumption coefficient, which is related to the biomass and temperature in water, Nc represents the biomass coefficient in water, O_min represents the minimum oxygen concentration in water, which is related to water quality, and u(t) represents the control quantity output by the PID controller. The specific formula is shown in formula (2): (2); In formula (2), Kp represents proportional gain, Ki represents integral gain, Kd represents differential gain, and e(t) represents real-time error.

[0013] Preferably, in S1, a dynamic multi-dimensional weight update strategy is used to improve the update strategy of followers in the collective training phase of the football team training optimization algorithm. The improved specific mathematical model is shown in formula (3): (3); In formula (3), Indicates the updated individual position, F(iter) indicates the current individual position, The calculation formula of the dynamic weight factor is shown in formula (4). represents the position of the best individual in the population, It means that the individual positions ranked 2-6 in the population by fitness value are split into three new individual positions randomly formed by different dimensions. In the optimization of PID controller, the dimension value dim=3. represents the weight of the i-th individual among the three new individuals. The weight values ​​of the three new individuals are (1- ) for random assignment; (4); In formula (4), iter represents the current number of iterations, max_iter represents the maximum number of iterations, and the value of the dynamic weight changes nonlinearly from 0.25 to 0.85 as the number of iterations increases.

[0014] Preferably, a dynamic multi-dimensional weight update strategy takes the optimal individual as the main guide, and as the number of iterations increases, the weight changes dynamically, and the weight changes nonlinearly from low to high, so that the algorithm can better gradually shift from global search to local search. This strategy takes into account the positions of multiple excellent individuals, and improves the algorithm's exploration ability by selecting combined individuals in multiple dimensions, which helps the algorithm to quickly jump out when it falls into a local optimal solution, thereby improving the algorithm's optimization accuracy.

[0015] Preferably, in S2, an optimal learning strategy based on distance is used to improve the optimal learning strategy in the group training of the football team training optimization algorithm. The improved specific mathematical formula is shown in formula (5): (5); In formula (5), Indicates updating the position of the individual, F(iter) indicates the position of the current individual, rand indicates a random number between [0, 1], and the dist() function indicates calculating the Euclidean distance between the positions of two individuals. represents the maximum distance of the entire search space, α represents the local search weight factor, Represents the position of the optimal individual in the population, α=log(1+iter / max_iter).

[0016] Preferably, a distance-based optimal learning strategy is dynamically adjusted by combining the distance between the current individual and the global optimal individual. Individuals with a long distance rely on the guidance of the global optimal individual for extensive search, while individuals with a short distance rely more on local search and focus on fine search. This strategy effectively balances the global and local optimization of the algorithm, so that the algorithm can have excellent optimization performance at all stages.

[0017] Preferably, in step three, the PID control in the fishery oxygen regulation control system is optimized using an improved football team training optimization algorithm, and the specific steps are: Step 1, initialize the parameters of the improved football team training optimization algorithm, including the population size N, the problem dimension dim of the algorithm optimization, the maximum number of iterations max_iter, the upper limit ub and lower limit lb of the algorithm's search space, and generate the initialized population through the initial parameters of the algorithm; Step 2, map the improved football team training optimization algorithm with the PID controller in the fishery oxygen control system, and establish a connection between the individual position vector F(iter) in the algorithm population and the three parameters Kp, Ki, and Kd of the PID controller. The problem dimension dim of the algorithm is 3, so the individual position vector corresponds to the values ​​of the three dimensions, which correspond to Kp, Ki, and Kd respectively. The position in the population is updated through the iterative update of the algorithm, and the three control parameters of the PID controller are continuously adjusted, and finally a set of optimal control parameters of the PID controller is obtained; Step 3: Set the fitness value function of the improved football team training optimization algorithm, calculate the fitness value of the individuals in the initial population through the set fitness value function, sort them according to the size of the fitness value, and select the individual with the best fitness value as , the specific fitness value function is shown in formula (6): (6); In formula (6), J represents the fitness value, e(t) represents the real-time error value, L represents the running time of the system, M represents the overshoot of the system response, that is, the peak value exceeding the target value, and λ represents the weight of the overshoot, which is 0.1; Step 4: Update the positions of individuals in the population through the mathematical model of the improved football team training optimization algorithm. The improved football team training optimization algorithm is divided into three stages, specifically: Phase 1: Collective training phase. In the collective training phase, individuals in the population are divided into four types: followers, discoverers, thinkers, and oscillators. When the algorithm enters the collective training phase, the individuals are first classified and the corresponding mathematical models are used to update their positions. The update strategy of followers is shown in formula (3), and the update strategies of the other three types are shown in formula (7): (7); In formula (7), Indicates the updated individual position, F(iter) indicates the current individual position, Indicates the position of the best individual in the population. rand, rand1, and rand2 all represent random numbers between [0,1]. It represents the position of the worst individual in the population, iter represents the current iteration number, and t(iter) represents a random number that follows t distribution; In the second stage, the group training stage, the MGEM clustering method is used to group individuals. There are three ways to train members in each group, including optimal learning, random learning and random communication. The update strategy of optimal learning is shown in formula (5), and the update strategy of random learning and random communication is shown in formula (8): (8); In formula (8), Indicates the position of the updated individual, F_rand(iter) indicates the position of the randomly selected individual in the current stage, and randn indicates a number randomly generated from a normal distribution; Phase 3: Individual training phase. After the group training phase, the best individual in the population will be selected for training. The specific update strategy is shown in formula (9): (9); In formula (9), represents the optimal position of the update, Indicates the position of the best individual in the population, iter indicates the current number of iterations, Gauss indicates Gaussian distribution, and Gauchy indicates Cauchy distribution; Each iteration compares the fitness value of the updated position with the fitness value before the update to decide whether to keep the updated position; Step 5: Determine whether the current number of iterations has reached the maximum number of iterations. If so, complete the optimization and output the optimal solution. Otherwise, continue to execute step 4 for optimization.

[0018] By adopting the above technical scheme, the beneficial effects of the present invention are as follows: the performance of the football team training optimization algorithm is significantly improved through two improvement strategies, the dynamic multi-dimensional weight update strategy better realizes the smooth transition from the global search of the algorithm to the local search, and at the same time, the multi-dimensional reorganization is used to enhance the ability of the algorithm to jump out of the local optimal solution and improve the optimization accuracy, and the distance-based optimal learning strategy guides the direction by dynamically adjusting the distance between the individual and the global optimal, strengthens the adaptability of the algorithm at different stages, and through the synergy of the two strategies, takes into account the exploration and development efficiency, significantly improves the convergence and solution quality of the algorithm, and optimizes the PID controller in the fishery oxygen control system through the improved football team training optimization algorithm, which can significantly improve the dynamic response speed and steady-state accuracy of the entire control system, effectively adapt to the fluctuation of environmental parameters, reduce overshoot and oscillation phenomena, and enhance the anti-interference ability of the system, so as to achieve more accurate oxygen control, take into account the safety and economy of aquaculture, and achieve efficient and reliable adaptive control. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flow chart of an adaptive adjustment method of a fishery oxygen control system; Figure 2 This is a model diagram of the fishery oxygen regulation control system; Figure 3 A comparison chart of the fitness value changes during the optimization process of the improved football team training optimization algorithm and the standard football team training optimization algorithm; Figure 4Comparison chart of the effect of optimizing PID controller for improved football team training optimization algorithm and standard football team training optimization algorithm. DETAILED DESCRIPTION

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0021] The present invention provides a technical solution: a method for adaptively adjusting a fishery oxygen control system, which specifically comprises the following steps: Figure 1 shown.

[0022] Step 1: Construct a fishery oxygen control system and convert the fishery oxygen control problem into a mathematical model to be optimized. The fishery oxygen control system includes: error calculation module, PID controller, improved football team training optimization algorithm module, real-time detection module, oxygen control module, such as Figure 2 shown.

[0023] Furthermore, in step 1, the execution process of each module in the constructed fishery oxygen control system is as follows: by inputting the given target oxygen concentration and the actual oxygen concentration obtained by the real-time detection module into the error calculation module, the real-time error e(t) is calculated, and then the error is input into the PID controller, and the PID controller is optimized using the improved football team training optimization algorithm module to obtain the optimal parameters. The optimized PID controller calculates and outputs the control quantity u(t) according to these optimal parameters, and the control quantity is then transmitted to the oxygen control module, thereby adjusting the oxygen supply to achieve the required oxygen concentration, wherein the mathematical model of the oxygen control module is shown in formula (1): (1); In formula (1), O(t) represents the oxygen concentration in water, Kc represents the oxygen consumption coefficient, which is related to the biomass and temperature in water, Nc represents the biomass coefficient in water, O_min represents the minimum oxygen concentration in water, which is related to water quality, and u(t) represents the control quantity output by the PID controller. The specific formula is shown in formula (2): (2); In formula (2), Kp represents proportional gain, Ki represents integral gain, Kd represents differential gain, and e(t) represents real-time error.

[0024] Step 2: Improve the football team training optimization algorithm. The specific improvements include: S1. A dynamic multi-dimensional weight update strategy is used to improve the update strategy of followers in the collective training stage of the football team training optimization algorithm. A dynamic multi-dimensional weight update strategy is to select the top 6 individuals in the search population with the highest fitness value. The weight of the first position increases nonlinearly with the increase of the number of algorithm iterations. The remaining 5 individual position vectors are decomposed according to different dimensions, and three new individuals are randomly formed according to different dimensions. The weights of these three new individuals are dynamically distributed by the remaining weights, and the sum of all weights is 1. S2. Use a distance-based optimal learning strategy to improve the optimal learning strategy in the group training of the football team training optimization algorithm. A distance-based optimal learning strategy is to guide the current individual by using the position of the optimal individual in the population, and adjust the position of the current individual through the distance between the optimal individual and the current individual, thereby achieving position update.

[0025] Furthermore, in S1, a dynamic multi-dimensional weight update strategy is used to improve the update strategy of followers in the collective training phase of the football team training optimization algorithm. The improved specific mathematical model is shown in formula (3): (3); In formula (3), Indicates the updated individual position, F(iter) indicates the current individual position, The calculation formula of the dynamic weight factor is shown in formula (4). represents the position of the best individual in the population, It means that the individual positions ranked 2-6 in the population by fitness value are split into three new individual positions randomly formed by different dimensions. In the optimization of PID controller, the dimension value dim=3. represents the weight of the i-th individual among the three new individuals. The weight values ​​of the three new individuals are (1- ) for random assignment; (4); In formula (4), iter represents the current number of iterations, max_iter represents the maximum number of iterations, and the value of the dynamic weight changes nonlinearly from 0.25 to 0.85 as the number of iterations increases.

[0026] Furthermore, in S2, an optimal learning strategy based on distance is used to improve the optimal learning strategy in the group training of the football team training optimization algorithm. The specific mathematical formula after improvement is shown in formula (5): (5); In formula (5), Indicates updating the position of the individual, F(iter) indicates the position of the current individual, rand indicates a random number between [0, 1], and the dist() function indicates calculating the Euclidean distance between the positions of two individuals. represents the maximum distance of the entire search space, α represents the local search weight factor, Represents the position of the optimal individual in the population, α=log(1+iter / max_iter).

[0027] Step 3: Use the improved football team training optimization algorithm to optimize the PID control in the fishery oxygen control system, and obtain the optimal set of PID control parameters Kp, Ki, and Kd through iterative optimization of the algorithm.

[0028] Furthermore, in step three, the PID control in the fishery oxygen control system is optimized using the improved football team training optimization algorithm, and the specific steps are as follows: Step 1, initialize the parameters of the improved football team training optimization algorithm, including the population size N, the problem dimension dim of the algorithm optimization, the maximum number of iterations max_iter, the upper limit ub and lower limit lb of the algorithm's search space, and generate the initialized population through the initial parameters of the algorithm; Step 2, map the improved football team training optimization algorithm with the PID controller in the fishery oxygen control system, and establish a connection between the individual position vector F(iter) in the algorithm population and the three parameters Kp, Ki, and Kd of the PID controller. The problem dimension dim of the algorithm is 3, so the individual position vector corresponds to the values ​​of the three dimensions, which correspond to Kp, Ki, and Kd respectively. The position in the population is updated through the iterative update of the algorithm, and the three control parameters of the PID controller are continuously adjusted, and finally a set of optimal control parameters of the PID controller is obtained; Step 3: Set the fitness value function of the improved football team training optimization algorithm, calculate the fitness value of the individuals in the initial population through the set fitness value function, sort them according to the size of the fitness value, and select the individual with the best fitness value as , the specific fitness value function is shown in formula (6): (6); In formula (6), J represents the fitness value, e(t) represents the real-time error value, L represents the running time of the system, M represents the overshoot of the system response, that is, the peak value exceeding the target value, and λ represents the weight of the overshoot, which is 0.1; Step 4: Update the positions of individuals in the population through the mathematical model of the improved football team training optimization algorithm. The improved football team training optimization algorithm is divided into three stages, specifically: Phase 1: Collective training phase. In the collective training phase, individuals in the population are divided into four types: followers, discoverers, thinkers, and oscillators. When the algorithm enters the collective training phase, the individuals are first classified and the corresponding mathematical models are used to update their positions. The update strategy of followers is shown in formula (3), and the update strategies of the other three types are shown in formula (7): (7); In formula (7), Indicates the updated individual position, F(iter) indicates the current individual position, Indicates the position of the best individual in the population. rand, rand1, and rand2 all represent random numbers between [0,1]. It represents the position of the worst individual in the population, iter represents the current iteration number, and t(iter) represents a random number that follows t distribution; In the second stage, the group training stage, the MGEM clustering method is used to group individuals. There are three ways to train members in each group, including optimal learning, random learning and random communication. The update strategy of optimal learning is shown in formula (5), and the update strategy of random learning and random communication is shown in formula (8): (8); In formula (8), Indicates the position of the updated individual, F_rand(iter) indicates the position of the randomly selected individual in the current stage, and randn indicates a number randomly generated from a normal distribution; Phase 3: Individual training phase. After the group training phase, the best individual in the population will be selected for training. The specific update strategy is shown in formula (9): (9); In formula (9), represents the optimal position of the update, Indicates the position of the best individual in the population, iter indicates the current number of iterations, Gauss indicates Gaussian distribution, and Gauchy indicates Cauchy distribution; Each iteration compares the fitness value of the updated position with the fitness value before the update to decide whether to keep the updated position; Step 5: Determine whether the current number of iterations has reached the maximum number of iterations. If so, complete the optimization and output the optimal solution. Otherwise, continue to execute step 4 for optimization.

[0029] Step 4: Input the optimal set of Kp, Ki, and Kd parameters obtained by optimization in step 3 into the PID controller in the fishery oxygen control system to optimize the control effect of the entire fishery aquaculture pond oxygen control system.

[0030] Furthermore, in order to verify the superiority of the improved football team training optimization algorithm for the fishery oxygen control system after optimization, simulation experiments were carried out using MATLAB and Simulink simulation tools. The mathematical model of the standard football team training optimization algorithm was improved through MATLAB, and the simulation model of the fishery oxygen control system was constructed through Simulink. The initial parameters of the algorithm were initialized as follows: the maximum number of iterations was 30, the problem dimension was 3, the population size was 100, and the search space range was [50, 0.001]. The mathematical model of the oxygen control module was Laplace transformed, and specific parameter values ​​were given. The formula is: , s is a complex variable, run the experimental code and model, the experimental results are as follows Figure 3-4 shown.

[0031] Furthermore, Figure 3 This is a comparison chart of the fitness value changes during the optimization process of the improved football team training optimization algorithm and the standard football team training optimization algorithm. From the figure, it can be seen that after the improved algorithm quickly reaches the vicinity of the optimal solution during the optimization process, it will continue to optimize downward instead of falling into the local optimal solution. The fitness value of the solution found is lower, and the optimization accuracy is better. Figure 4 This is a comparison chart of the effects of optimizing the PID controller using the improved football team training optimization algorithm and the standard football team training optimization algorithm. The target value is set to 1 unit. The response curve of the PID controller optimized using the improved algorithm is more stable and reaches a stable state in a shorter time, indicating that the optimized system is more robust.

Claims

1. A method for adaptively adjusting a fishery oxygen control system, characterized in that: The specific steps are as follows: Step 1: construct a fishery oxygen control system, convert the fishery oxygen control problem into a mathematical model to be optimized, and the fishery oxygen control system includes: an error calculation module, a PID controller, an improved football team training optimization algorithm module, a real-time detection module, and an oxygen control module; Step 2: Improve the football team training optimization algorithm. The specific improvements include: S1. Use a dynamic multi-dimensional weight update strategy to improve the update strategy of followers in the collective training phase of the football team training optimization algorithm; S2, using a distance-based optimal learning strategy to improve the optimal learning strategy in group training of the football team training optimization algorithm; Step 3: Optimize the PID control in the fishery oxygen control system using the improved football team training optimization algorithm, and obtain the optimal set of PID control parameters Kp, Ki, and Kd through iterative optimization of the algorithm; Step 4: Input the optimal set of Kp, Ki, and Kd parameters obtained by optimization in step 3 into the PID controller in the fishery oxygen control system to optimize the control effect of the entire fishery oxygen control system.

2. The method for adaptively adjusting a fishery oxygen control system according to claim 1, characterized in that: In the step 1, the execution process of each module in the constructed fishery oxygen control system is as follows: by inputting the given target oxygen concentration and the actual oxygen concentration obtained by the real-time detection module into the error calculation module, the real-time error e(t) is calculated, and then the error is input into the PID controller, and the PID controller is optimized using the improved football team training optimization algorithm module to obtain the optimal parameters. The optimized PID controller calculates and outputs the control amount u(t) according to these optimal parameters, and the control amount is then transmitted to the oxygen control module, thereby adjusting the oxygen supply to achieve the required oxygen concentration.

3. The adaptive adjustment method of a fishery oxygen control system according to claim 2, characterized in that: In S1, a dynamic multi-dimensional weight update strategy is used to improve the update strategy of followers in the collective training stage of the football team training optimization algorithm. A dynamic multi-dimensional weight update strategy is to select the top 6 individuals in the search population with the fitness value, and the weight of the first position increases nonlinearly with the increase of the number of algorithm iterations. The remaining 5 individual position vectors are decomposed according to different dimensions, and three new individuals are randomly formed according to different dimensions. The weights of these three new individuals are dynamically allocated by the remaining weights, and the sum of all weights is 1. The improved specific mathematical model is shown in formula (3): (3)； In formula (3), Indicates the updated individual position, F(iter) indicates the current individual position, The calculation formula of the dynamic weight factor is shown in formula (4). represents the position of the best individual in the population, It means that the individual positions ranked 2-6 in the population by fitness value are split into three new individual positions randomly formed by different dimensions. In the optimization of PID controller, the dimension value dim=3. represents the weight of the i-th individual among the three new individuals. The weight values ​​of the three new individuals are (1- ) for random assignment; (4)； In formula (4), iter represents the current number of iterations, max_iter represents the maximum number of iterations, and the value of the dynamic weight changes nonlinearly from 0.25 to 0.85 as the number of iterations increases.

4. The method for adaptively adjusting a fishery oxygen control system according to claim 3, characterized in that: In S2, an optimal learning strategy based on distance is used to improve the optimal learning strategy in the group training of the football team training optimization algorithm. An optimal learning strategy based on distance is to guide the current individual by using the position of the optimal individual in the population, and adjust the position of the current individual by the distance between the optimal individual and the current individual, so as to achieve the update of the position. The improved specific mathematical formula is shown in formula (5): (5)； In formula (5), Indicates updating the position of the individual, F(iter) indicates the position of the current individual, rand indicates a random number between [0, 1], and the dist() function indicates calculating the Euclidean distance between the positions of two individuals. represents the maximum distance of the entire search space, α represents the local search weight factor, Represents the position of the optimal individual in the population, α=log(1+iter / max_iter).

5. The method for adaptively adjusting a fishery oxygen control system according to claim 4, characterized in that: In step 3, the PID control in the fishery oxygen regulation control system is optimized using the improved football team training optimization algorithm, and the specific steps are as follows: Step 1, initialize the parameters of the improved football team training optimization algorithm, including the population size N, the problem dimension dim of the algorithm optimization, the maximum number of iterations max_iter, the upper limit ub and lower limit lb of the algorithm's search space, and generate the initialized population through the initial parameters of the algorithm; Step 2, map the improved football team training optimization algorithm with the PID controller in the fishery oxygen control system, and establish a connection between the individual position vector F(iter) in the algorithm population and the three parameters Kp, Ki, and Kd of the PID controller. The problem dimension dim of the algorithm is 3, so the individual position vector corresponds to the values ​​of the three dimensions, which correspond to Kp, Ki, and Kd respectively. The position in the population is updated through the iterative update of the algorithm, and the three control parameters of the PID controller are continuously adjusted, and finally a set of optimal control parameters of the PID controller is obtained; Step 3: Set the fitness value function of the improved football team training optimization algorithm, calculate the fitness value of the individuals in the initial population through the set fitness value function, sort them according to the size of the fitness value, and select the individual with the best fitness value as , the specific fitness value function is shown in formula (6): (6)； In formula (6), J represents the fitness value, e(t) represents the real-time error value, L represents the running time of the system, M represents the overshoot of the system response, that is, the peak value exceeding the target value, and λ represents the weight of the overshoot, which is 0.1; Step 4: Update the positions of individuals in the population through the mathematical model of the improved football team training optimization algorithm. The improved football team training optimization algorithm is divided into three stages, specifically: Phase 1: Collective training phase. In the collective training phase, individuals in the population are divided into four types: followers, discoverers, thinkers, and volatiles. When the algorithm enters the collective training phase, individuals are first classified and their positions are updated using the corresponding mathematical models. In the second stage, group training, the MGEM clustering method is used to divide individuals into groups. Members in each group are trained in three ways, including optimal learning, random learning, and random communication. Phase 3: Individual training phase: After the group training phase, the best individuals in the population will be selected for training; Each iteration compares the fitness value of the updated position with the fitness value before the update to decide whether to keep the updated position; Step 5: Determine whether the current number of iterations has reached the maximum number of iterations. If so, complete the optimization and output the optimal solution. Otherwise, continue to execute step 4 for optimization.

Images