A football team training optimization method based on chaos opposite learning

Abstract

The application discloses a kind of football team training optimization method based on chaos opposite learning, first design a novel chaotic mapping method, realize more random more chaotic population initialization, speed up the initial convergence speed of algorithm.Subsequently, successively use collective training strategy, joint set training strategy, grouping training strategy, individual additional training strategy to update population.Until meeting the maximum iteration number, output optimal solution.The method of the application compared with existing optimization method, not only can solve high-dimensional, non-convex optimization problem, but also has stronger global search capability and faster solution speed, can solve complex optimization problem to a certain extent.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent computing, and particularly relates to a football team training optimization method based on chaotic opposite learning. BACKGROUND

[0002] Optimization algorithms are fundamental tools in mathematics and computer science, aiming to identify the minimum or maximum value of a function subject to various constraints. These algorithms have wide applications in multiple engineering disciplines. For instance, in the military field, they are crucial for the optimization of unmanned aerial vehicle task allocation and path planning; in the field of intelligent systems, they serve as parameter optimizers, providing strong support for the collaborative control of multi-agent systems; in the financial industry, these algorithms can effectively predict the occurrence of financial crises; in electrical engineering, they significantly improve the operational efficiency of power systems; and in project management, optimization algorithms assist us in efficiently allocating tasks to team members within an organization.

[0003] Traditional optimization algorithms such as linear programming, quadratic programming, and convex optimization have wide applications in their respective fields. However, when faced with high-dimensional, nonlinear complex problems, these traditional algorithms often exhibit certain limitations in terms of precision and convergence speed. Against this background, meta-heuristic algorithms based on random search strategies have gradually attracted widespread attention from researchers and practitioners due to their wide applicability, efficiency, and flexibility. These meta-heuristic algorithms provide new ideas and methods for solving complex optimization problems in their unique ways. In recent years, inspired by human social behavior, evolutionary theory, biological population characteristics, and physical theories, scholars have conducted in-depth research and proposed various meta-heuristic algorithms, such as particle swarm optimization, grey wolf optimization, ant colony optimization, genetic algorithm, whale optimization algorithm, Harris hawk optimization, and football team training algorithm.

[0004] Meta-heuristics guide the search through some rules or empirical knowledge, thereby searching for optimization in the solution space within an acceptable time. It is worth noting that in meta-heuristic algorithms, the solution to the problem is usually referred to as a population, but this search strategy often has the following shortcomings:

[0005] 1. Prone to local optimum. The algorithm focuses only on finding the global optimal solution based on the current optimal solution found, which often falls into a local optimal solution. For example, simulated annealing algorithm, particle swarm optimization, and whale optimization algorithm all have similar problems.

[0006] 2. Slow convergence speed. Meta-heuristic algorithms often require a large number of iterations to find the optimal solution, not only slow convergence speed, but also occupy a lot of computing resources. For example, ant colony optimization, grey wolf optimization, and Harris hawk optimization all have similar problems. SUMMARY

[0007] The purpose of the present application is to overcome the deficiencies of the prior art, provide a football team training optimization method based on chaotic opposite learning, to solve the parameter optimization problem of the cooperative controller. The present application optimizes the randomly generated population, increases randomness and unpredictability, and can better simulate the behavior of athletes; and then updates the athlete state according to the collective training strategy, joint training strategy, grouping training strategy and individual additional training strategy, improves the global optimization ability and convergence speed of the algorithm.

[0008] The purpose of the present application is to overcome the deficiencies of the prior art, provide a football team training optimization method based on chaotic opposite learning, to solve the parameter optimization problem of the cooperative controller. The present application optimizes the randomly generated population, increases randomness and unpredictability, and can better simulate the behavior of athletes; and then updates the athlete state according to the collective training strategy, joint training strategy, grouping training strategy and individual additional training strategy, improves the global optimization ability and convergence speed of the algorithm.

[0009] S1, define a multi-dimensional optimization problem minf(x): for the cooperative control problem of satellite networking, a controller based on neural network is deployed on each satellite, and the goal of cooperative control is to make the attitude angles of all satellites consistent; the controller has two self-defined parameters to be optimized: controller gain and neural network update rate; the goal of the optimization problem is to select a pair of appropriate parameters to meet the following indicators: 1) minimize the total control input of the controller; 2) minimize the neural network estimation error; 3) minimize the standard deviation of the attitude angle convergence error; each controller needs to optimize these three indicators, so the optimization problem f(x) of satellite networking cooperative control is defined as follows:

[0010]

[0011] Wherein, N is the number of satellites, T is the total number of controller simulation iterations, K is the number of attitude angle convergence error vectors, in the embodiment, N=4, T=50001, K=2. alpha1, alpha2, alpha3 are the weights of the three optimization objectives, which need to satisfy alpha1+alpha2+alpha3=1, in the embodiment, alpha1=0.3, alpha2=0.3, alpha3=0.4; independent variable x={x 1 ,x 2 ,x 3} x 1 , x 2 , x 3 are the control output of the cooperative controller, the neural network estimation error, and the standard deviation of the attitude angle convergence error, respectively, wherein x 1 and x 2 are matrices with a size of N*T, and x 3 is a matrix with a size of K*T; is the sample data of the i-th satellite in x a for the j-th iteration, and a=1 or 2; is the sample data of the k-th attitude angle convergence for the j-th iteration, is the mean of x 3 ;

[0012] S2, in the football team recruitment stage, an infinite folding iterative logic-inclined tent mapping method with more chaos is used to iteratively generate pop athletes;

[0013] S3, collective training is carried out, the best athletes and the worst athletes are found out, and then the state of the athletes is updated according to the type of the athletes;

[0014] S4, joint training is carried out, opposite athletes are generated and absorbed, and all athletes are evaluated and ranked, and the top 50% of athletes are reserved to form a training team;

[0015] S5, group training is carried out: all athletes are grouped, and learning and communication are carried out within the group, and the state of the athletes is updated according to the four learning and communication effects;

[0016] S6, individual additional training: the best athletes in the football team are found out and are given individual additional training;

[0017] S7, the state and score of the best athletes are recorded and updated, if the maximum iteration number of the algorithm is not reached, return to S3, if the maximum iteration number of the algorithm is reached, output the state of the best athletes, which is the optimal solution obtained by the algorithm.

[0018] The application optimizes the randomly generated population by designing a novel chaotic mapping method, increases randomness and unpredictability, and can better simulate the behavior of athlete recruitment. Then the state of the athletes is updated according to the collective training strategy, joint training strategy, group training strategy and individual additional training strategy, which improves the global optimization ability and convergence speed of the algorithm. Compared with the prior art, the beneficial effects of the application mainly lie in the following three aspects:

[0019] 1. A novel football team training optimization method based on chaotic opposite learning is proposed, compared with genetic algorithm, grey wolf algorithm, Harris eagle algorithm and other methods, the application introduces a new and better football team training algorithm, which improves the optimization performance of the football team training algorithm;

[0020] 2. In the existing related research, most of the swarm intelligence optimization algorithms use uniform distribution to initialize the population, which does not conform to the behavior of the population in the natural system, the randomness of the global search of the algorithm is reduced, and the convergence speed is slow. In view of this problem, the application innovatively proposes an infinite folding iterative logic-inclined tent mapping method, which uses this chaotic mapping strategy when initializing the population, increases the randomness and unpredictability of the population, better simulates the behavior of football team recruitment, and makes the algorithm converge faster.

[0021] 3、The football team training optimization method based on chaotic opposite learning can be applied to parameter optimization problems of satellite networking cooperative controllers, and can be popularized to solve most optimization problems. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 A flowchart of the football team training optimization method based on chaotic opposite learning of the application;

[0023] Figure 2 A comparison experiment convergence curve diagram for the parameter optimization problem of the satellite networking cooperative controller.

[0024] Figure 3 A comparison experiment convergence curve diagram for the parameter optimization problem of the satellite networking cooperative controller. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical scheme and advantages of the application clearer, the application will be further described in detail below with reference to the drawings. It should be understood that the specific embodiments described herein are only used to explain the application and do not limit the application.

[0026] As shown in the figure, the football team training optimization method based on chaotic opposite learning of the application comprises the following steps: Figure 1

[0027] S1, define a multi-dimensional optimization problem minf(x): for the cooperative control problem of satellite networking, a controller based on a neural network is deployed on each satellite, and the goal of cooperative control is to make the attitude angles of all satellites consistent; the controller has two self-defined parameters to be optimized: controller gain and neural network update rate; the goal of the optimization problem is to select a pair of suitable parameters to meet the following indicators: 1) minimize the total control input of the controller; 2) minimize the neural network estimation error; 3) minimize the standard deviation of the attitude angle convergence error; each controller needs to optimize these three indicators, so the optimization problem f(x) of satellite networking cooperative control is defined as follows:

[0028]

[0029] Wherein, N is the number of satellites, T is the total number of controller simulation iterations, K is the number of attitude angle convergence error vectors, in the embodiment, N = 4, T = 50001, and K = 2. α1, α2, and α3 are the weights of the three optimization objectives, which need to satisfy α1+α2+α3=1, in the embodiment, α1=0.3, α2=0.3, and α3=0.4; the independent variable x = {x 1 ,x 2 ,x 3}, x 1 , x 2 ​, x 3 are the standard deviations of the control output of the cooperative controller, the estimation error of the neural network, and the attitude angle convergence error, respectively, wherein x 1 and x 2 are matrices of size N*T, and x 3 is a matrix of size K*T; is the sample data of the ith satellite in x a for the jth iteration, and a=1 or 2; is the sample data of the kth attitude angle convergence for the jth iteration, is the mean value of x 3 .

[0030] The present application solves the above optimization problem by using a football team training optimization based on chaotic opposite learning, and specifically includes the following steps:

[0031] S2, in the football team recruitment stage, an infinite folding iterative logistic-sloping tent mapping method with more chaos is used to iteratively generate pop athletes; the following sub-steps are included:

[0032] S21, a parameter pair composed of two to-be-optimized controller parameters is defined as an athlete; the number of athletes of the football team training optimization algorithm is defined as pop=30, the maximum number of iterations of the optimization algorithm is defined as IterMax=200; when the number of iterations is Iter, the solution in the jth dimension is

[0033] S22, the infinite folding iterative logistic-sloping tent mapping method is defined, and the iterative update formula of the t+1 athlete in one dimension is:

[0034]

[0035] wherein t∈[1,pop]; x t ∈[l b ,u b ] is a randomly generated athlete state, i.e., a random solution of the optimization problem (the state of the athlete in the present application is the specific value of the parameter pair); μ∈(0,4]、α∈[0,1] are control parameters;

[0036] S23, x1 is randomly generated, and then the infinite folding iterative logistic-sloping tent mapping method defined in S13 is used to iteratively generate pop athletes.

[0037] S3, collective training is performed, the best athlete and the worst athlete are found out, and then the state of the athlete is updated according to the athlete type; the following sub-steps are included:

[0038] S31, calculate the function value f(x) of the generated pop athletes, and find the best athlete in each dimension of the solution to the optimization problem and the worst athlete

[0039] S32, determine the type of each athlete, i.e., which of the four types of follower, discoverer, thinker, and fluctuator, using random interval sampling;

[0040] S33, if the athlete is a follower, he / she is more inclined to follow the steps of the best athlete, and the update calculation formula for his / her state is as follows:

[0041]

[0042] where rand∈[0,1] is a random number; is the solution of the optimization algorithm in the jth dimension at the ith iteration;

[0043] S34, if the athlete is a discoverer, he / she will avoid learning from the worst athlete while discovering the gap between himself / herself and the best athlete, and the update calculation formula for his / her state is as follows:

[0044]

[0045] where rand1∈[0,1] and rand2∈[0,1] are two different random numbers;

[0046] S35, if the athlete is a thinker, he / she will pay attention to the difference between the best athlete and the worst athlete, and the update calculation formula for his / her state is as follows:

[0047]

[0048] S36, if the athlete is a fluctuator, his / her behavior is not affected by others, and the update calculation formula for his / her state is as follows:

[0049]

[0050] where Iter is the current iteration number, and t(·) is a probability density function that meets the t distribution.

[0051] S4, perform joint set training, generate and absorb opposite athletes, evaluate and rank all athletes, and retain the top 50% of athletes to form a training team; the following sub-steps are included:

[0052] S41, based on the athletes updated in S3, generate opposite athletes using opposite learning rules:

[0053]

[0054] in is the state of the i-th opposing athlete; rand1∈[0,1] and rand2∈[0,1] are two different random numbers; k∈(0,∞) is the control parameter of the adversarial learning, and in this embodiment, k=12000;

[0055] S42. Incorporate opposing athletes into our team, calculate the status of all athletes, and retain the top 50% of athletes to form a training team.

[0056] S5. Conduct group training: Divide all athletes into groups, conduct learning and communication within the groups, and update the athlete status based on the four learning and communication results. This includes the following sub-steps:

[0057] S51. Group the training teams using the Gaussian mixture model expectation maximization method or the uniform random assignment method, dividing all athletes into four groups: forwards, centers, defenders, and goalkeepers.

[0058] S52. When training in groups, first identify the best athlete in each group;

[0059] S53. During group training, there are four situations: optimal learning, random learning, random communication, and random error. If athlete i can achieve the level of the best athlete in a certain dimension through learning, then it is optimal learning, and its status is The update calculation formula is:

[0060]

[0061] in represents the state of the best player in the group in the jth dimension, p study ∈[0,1] is the learning probability;

[0062] S54, if athlete i learns from a random athlete in the group, it is random learning, and its state The update calculation formula is:

[0063]

[0064] in represents the state of a random athlete k in the group in the jth dimension;

[0065] S55. If athlete i randomly communicates with athlete k on the jth dimension, it is a random communication. The states of the two athletes are and The update calculation formula is:

[0066]

[0067] where randn represents a random number obeying normal distribution, p comn ∈ [0, 1] is the alternating probability;

[0068] S56, if the athlete i accidentally learns the wrong dimension, it is a random error, and the state of the update calculation formula is:

[0069]

[0070] where is the state of the kth athlete in the lth dimension, p error ∈ [0, 1] is the error probability.

[0071] S6, individual additional training: find the best athlete in the football team and conduct individual additional training; there are the following sub-steps:

[0072] S61, evaluate all athletes after grouping training to find the best athlete after grouping training;

[0073] S62, update the state of the best athlete according to the following update formula:

[0074]

[0075] where randn and randc are random numbers obeying Gaussian distribution and Cauchy distribution respectively, is the state of the best athlete after individual additional training.

[0076] S7, record and update the state of the best athlete and its score, if the maximum number of iterations of the algorithm is not reached, return to S3; if the maximum number of iterations of the algorithm is reached, output the state of the best athlete, which is the optimal solution obtained by the algorithm.

[0077] In this embodiment, firstly, the CEC2022 test set is used to test the algorithm performance, and then simulation experiments are carried out for the controller parameter optimization problem proposed in the patent. In the test process, genetic algorithms (GA), grey wolf optimization (GWO), whale optimization algorithm (WOA), harris hawks optimization (HHO) and football team training algorithm (FTTA) are used for comparative simulation experiments. The CEC2022 test set has 12 benchmark functions, and the comparison result graph is as shown in Figure 2 Figure 3 The comparison experiment convergence curve diagram for the parameter optimization problem of the satellite networking cooperative controller. IFTTA is the convergence curve of the algorithm proposed in the application. The simulation results confirm that compared with the mainstream meta-heuristic optimization algorithm, the football team training optimization method based on chaotic opposite learning has stronger global search ability and convergence speed.

[0078] Those skilled in the art will realize that the embodiments described herein are to be used as examples only and are not to be construed as limiting the scope of the present application. Other various concrete modifications and combinations can be made to the embodiments disclosed in the present application according to the technical inspiration of the present application without departing from the essence of the present application, and these modifications and combinations are still within the scope of protection of the present application.​

Claims

1. A football team training optimization method based on chaotic adversarial learning, characterized in that: The following steps are involved: S1. Define the multidimensional optimization problem minf(x): For the collaborative control problem of satellite networking, a neural network-based controller is deployed on each satellite. The goal of collaborative control is to achieve consistency in the attitude angles of all satellites. The controller has two custom parameters to be optimized: controller gain and neural network update rate. The goal of the optimization problem is to select a pair of appropriate parameters that meet the following indicators: 1) minimize the total control input of the controller; 2) minimize the neural network estimation error; 3) minimize the standard deviation of the attitude angle convergence error. Each controller needs to optimize these three indicators, so the optimization problem f(x) for satellite network collaborative control is defined as follows: Where N is the number of satellites, T is the total number of controller simulation iterations, K is the number of attitude angle convergence error vectors, N = 4, T = 50001, K = 2; α1, α2, α3 are the weights of the three optimization objectives, which must satisfy α1+α2+α3=1, α1=0.3, α2=0.3, α3=0.4; independent variable x={x 1 ,x 2 ,x 3 }, x 1 、x 2 、x 3 are the control output of the cooperative controller, the neural network estimation error, and the standard deviation of the attitude angle convergence error, respectively. 1 and x 2 is a matrix of size N×T, x 3 is a matrix of size K×T; is x a The sample data of the jth iteration of the i-th satellite in , a=1 or 2; is the sample data of the jth iteration when the kth attitude angle converges, is x 3 The mean of S2. During the football team recruitment phase, a more chaotic infinite folding iterative logic, the tilted tent mapping method, is used to iteratively generate pop athletes. The sub-steps are as follows: S21. The parameter pair consisting of two controller parameters to be optimized is regarded as an athlete; the number of athletes in the football team training optimization algorithm is defined as pop, and the maximum number of iterations of the optimization algorithm is IterMax; when the number of iterations is Iter, the solution in the jth dimension is S22. Define the iterative logic of infinite folding - tilted tent mapping method. The iterative update formula of the t+1th athlete in one dimension is: Among them, t∈[1,pop]; x t ∈[l b ,u b ] is a randomly generated athlete state, i.e., a random solution to the optimization problem; μ∈(0,4],α∈[0,1] are control parameters; S23, randomly generate x1, and then iteratively generate pop athletes using the infinite folding iterative logic-tilted tent mapping method defined in S22; S3. Conduct group training to identify the best and worst athletes, and then update their status based on the athlete type; S4. Conduct joint training to generate and recruit opposing athletes, evaluate and rank all athletes, and retain the top 50% of athletes to form a training team. This includes the following sub-steps: S41. Based on the players iteratively updated in S3, an adversarial learning rule is used to generate an opposing player: in is the state of the i-th opponent; rand1∈[0,1] and rand2∈[0,1] are two different random numbers; k∈(0,∞) is the control parameter of the opponent learning, k=12000; S42. Absorb opposing athletes into our team, calculate the status of all athletes, and retain the top 50% of athletes to form a training team; S5. Conduct group training: Divide all athletes into groups, conduct learning and communication within the group, and update the athlete status based on the four learning and communication results; S6. Individual extra training: Find the best players in the football team and give them individual extra training; S7. Record and update the best athlete status and its score. If the maximum number of algorithm iterations has not been reached, return to S3; if the maximum number of algorithm iterations has been reached, output the best athlete status.

2. A football team training optimization method based on chaotic adversarial learning according to claim 1, characterized in that: In S3, there are the following sub-steps: S31. Calculate the function value f(x) of the generated pop athletes and find the best athlete for each dimension of the solution to the optimization problem. With the worst athletes S32. Use random interval sampling to determine the type of each athlete, i.e., follower, discoverer, thinker, or volatile type; S33. If the athlete is a follower, he or she is more inclined to follow the footsteps of the best athletes. The update calculation formula is: Where rand∈[0,1] is a random number; is the solution of the optimization algorithm in the jth dimension when the number of iterations is i; S34. If an athlete is a discoverer, then while discovering the gap between himself and the best athlete, he will also avoid learning from the worst athlete. The update calculation formula is: Where rand1∈[0,1] and rand2∈[0,1] are two different random numbers; S35. If athletes are thinkers, they will pay attention to the difference between the best and the worst athletes. The update calculation formula is: S36. If an athlete is a fluctuator, his behavior is not affected by others and his state The update calculation formula is: Where Iter is the current iteration number, and t(·) is the probability density function that conforms to the t distribution.

3. A football team training optimization method based on chaotic adversarial learning according to claim 1, characterized in that: In the S5, there are the following sub-steps: S51. Group the training teams using the Gaussian mixture model expectation maximization method or the uniform random assignment method, dividing all athletes into four groups: forwards, centers, defenders, and goalkeepers. S52. When training in groups, first identify the best athlete in each group; S53. During group training, there are four situations: optimal learning, random learning, random communication, and random error. If athlete i can achieve the level of the best athlete in a certain dimension through learning, then it is optimal learning, and its status is The update calculation formula is: in represents the state of the best player in the group in the jth dimension, p study ∈[0,1] is the learning probability; S54, if athlete i learns from a random athlete in the group, it is random learning, and its state The update calculation formula is: in represents the state of a random athlete k in the group in the jth dimension; S55. If athlete i randomly communicates with athlete k on the jth dimension, it is a random communication. The states of the two athletes are and The update calculation formula is: Where randn represents a random number that obeys the normal distribution, p comn ∈[0,1] is the probability of communication; S56. If athlete i accidentally learns the wrong dimension, it is a random error and its state The update calculation formula is: in is the random l-th dimension state of the k-th athlete, p error ∈[0,1] is the error probability.

4. A football team training optimization method based on chaotic adversarial learning according to claim 1, characterized in that: In S6, there are the following sub-steps: S61. Evaluate all athletes after group training and identify the best athletes after group training; S62. Update the status of the best athlete according to the following update formula: Among them, randn and randc are random numbers that conform to Gaussian distribution and Cauchy distribution respectively. This is the state of the best athletes after additional training alone.

Images