INGO-GBDT-based observation and aiming system fault prediction method

Abstract

The invention relates to the technical field of artificial intelligence fault diagnosis, and discloses an observation and aiming system fault prediction method based on INGO-GBDT, and the method comprises the steps: carrying out the preprocessing of collected data through a VIKOR algorithm, carrying out the improvement of a northern eagle optimization algorithm NGO, introducing ICMIC chaotic mapping in an NGO population initialization stage, and introducing a Gaussian mutation strategy in an NGO prey recognition stage. A Levy flight strategy is introduced in the NGO chasing and escaping stage to optimize northern eagle position updating; an improved northern eagle optimization algorithm INGO is adopted to perform parameter optimization on a gradient boosting decision tree GBDT, an INGO-GBDT fault diagnosis model is constructed, and a training data set is trained; and inputting the test data set into the INGO-GBDT prediction model, and carrying out fault diagnosis on the observation and aiming system. According to the method, main parameters of the gradient boosting decision tree (GBDT) are optimized through the improved northern eagle optimization algorithm (INGO), the defect of blindness of parameter selection in the training process is overcome, the prediction precision of the classification prediction model is improved, and higher prediction precision and practicability are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis of sighting and aiming systems, and in particular to a fault prediction method of sighting and aiming systems based on INGO-GBDT. Background Art

[0002] The sighting and observation system is an important part of the fire control system. It is composed of gyro sensors, signal processing components, execution components, operating components and auxiliary components. The sighting and observation system is used in the fire control system to accurately control the recognition and tracking of targets to improve the efficiency and accuracy of aiming, shorten the target capture time, and enhance the hit rate of artillery while moving. The sighting and observation box plays a key role in battlefield operations. Frequent use also increases the incidence of failures. Fault prediction can improve battlefield combat mobility and firepower output and maximize combat capabilities.

[0003] At present, various artificial intelligence-based algorithms have been widely used in the field of fault diagnosis. The arbitrariness of kernel function selection and the limitation of large-scale training of support vector machines lead to the lack of accuracy of prediction results. The difficulty of expert system knowledge acquisition and the limitation of knowledge base on storage cannot guarantee the efficiency and correctness of prediction results. The fault tree analysis method has large calculation scale, cumbersome work, and excessive memory usage, which also leads to slower running speed in fault prediction.

[0004] Compared with other algorithms, the Gradient Boosting Decision Tree (GBDT) has the advantages of interpretability and accuracy, and perfectly solves the problem of complex internal mechanisms; the Gradient Boosting Decision Tree has strong generalization and expression capabilities, which also makes the fault prediction results more accurate. Through the method of ensemble learning, it has improved the limitations of other machine learning algorithms in the field of fault diagnosis to a certain extent, such as weak generalization ability. Summary of the invention

[0005] In view of the shortcomings of the fault analysis algorithm in the prior art, such as large calculation scale, high complexity, cumbersome work, and lack of accuracy in prediction results, the technical solution adopted by the present invention is: a method for predicting faults of an observation and aiming system based on INGO-GBDT, comprising the following steps: S1, collect data of the pin signal of the sighting system; S2. Preprocess the collected data through the multi-criteria compromise solution ranking method VIKOR, screen out the data set that can be used as model input, and divide the model input data set into a test data set and a training data set; S3. Improvements to the NGO optimization algorithm for northern goshawks, including the introduction of ICMIC chaotic mapping in the NGO population initialization stage to improve the uniformity of population initialization and calculate fitness; the introduction of Gaussian mutation strategy in the NGO prey recognition stage to enhance the algorithm's search efficiency; the introduction of Levy The flight strategy optimizes the position update of the northern goshawk and obtains the improved northern goshawk optimization algorithm INGO; S4, optimize the key parameters of the gradient boosting decision tree GBDT through the improved northern goshawk optimization algorithm INGO, build the INGO-GBDT fault diagnosis model, and train the fault diagnosis model through the training data set in S2; S5. Input the test data set in S2 into the trained INGO-GBDT model, perform fault diagnosis on the sighting system, and output the prediction results.

[0006] In step S2, the collected data is evaluated, sorted and normalized by the VIKOR algorithm as follows: S201, construct the initial decision matrix: Has The evaluation object of the parameter index is determined by and indicator set , construct a decision matrix ; S202, decision matrix Normalize the decision matrix using the standard 0-1 transformation Normalize and get , The specific calculation formula is: When the indicator When it is a benefit indicator, , (1-1) When the indicator When it is a cost indicator, , (1-2) In the formula is the normalized decision matrix, is the maximum value in the matrix, is the minimum value in the matrix; S203, according to the normalized matrix , determine the positive ideal solution and negative ideal solution : , (1-3) in, , (1-4) In the formula is the maximum value in the normalized matrix, is the minimum value of the normalized matrix; S204, calculate weight vector :Determine the index by constructing an optimization model Weight , the weight vector is calculated as follows: Using the positive ideal solution of the decision matrix and negative ideal solution Calculate the initial weight vector : , (1-5) For the initial vector After normalization, we can get: , (1-6) Then the indicator The weight is ; In the formula is the normalized solution of the positive ideal solution, is the normalized matrix; S205, calculate each evaluation object The group benefit value 、Individual regret value , compromise value : , (1-7) , (1-8) , (1-9) in, , is the maximum group benefit value; , is the minimum group benefit value; , is the maximum individual regret value; , is the minimum individual regret value; is the normalized weight of each indicator; is a positive ideal solution, is the negative ideal solution, v is the decision mechanism coefficient; For the evaluation object , and These three values ​​are sorted in descending order, resulting in three sorting methods: Sorting, Sorting and Sorting, the final sorting result is based on The value is determined; S206, determine the best solution: evaluate the object according to Sort the values ​​and get the solution ,choose The solution with the smallest value is regarded as the optimal solution; If the evaluation object in the program Value Satisfaction , then the compromise solution is obtained by verifying sorting conditions 1 and 2: Sorting condition 1: The smallest and second smallest solutions The value satisfies: , (1-10) Sorting condition 2: The smallest and second smallest solutions Value Sorting Satisfaction ,at the same time Sorting Satisfaction ; If sorting condition 1 and sorting condition 2 are met at the same time, then the solution The final compromise solution in the decision-making process; If only sorting condition 1 or only sorting condition 2 is met, the solution and solutions All are compromise solutions. As a final compromise solution; if sorting condition 1 and sorting condition 2 are not met at the same time, , get the maximum J, then the solution These are the final compromise solutions.

[0007] Step S3 includes the following steps: S301. In the NGO population initialization stage, a group of goshawk individuals are randomly generated, and random initial positions and speed parameters are assigned to these individuals respectively. The initial position is described as: , (2-1) Where: X is the population matrix of the northern goshawk, For the i The location of the northern goshawk, It refers to i The northern goshawk j The location of the dimension, N represents the population size of the northern goshawk, and s represents the dimension of the problem; The objective function vector of the population can be expressed as: , (2-2) In the formula, F is the objective function vector of the population; For the i The objective function value of the goshawk; For the N The objective function value of the goshawk; At this time, ICMIC chaotic mapping is introduced to improve the NGO algorithm and enhance the convergence of the algorithm. The improved initial position update formula is as follows: , (2-3) In the formula, For the i +1 northern goshawk position; a is a control parameter, and its value range is [0, +∞]; S302. In the NGO prey identification stage, each individual evaluates its own fitness based on its initial position and speed through a specific fitness function. In this process, individuals with higher fitness have a higher probability of being regarded as prey: , (2-4) , (2-5) Where: For the i The location of the prey of the northern goshawk; k is a random integer with a value range of [1, N ]; The updated prey identification phase i The Northern Goshawk j The location of the dimension; For the i The objective function value of the prey position of the northern goshawk; r is a random number with a value range of [0,1]; is a random integer with value 1 or 2; , (2-6) Where: Updated for the prey identification phase i The objective function value of the position of the northern goshawk only; At this time, the Gaussian mutation strategy is introduced to make the algorithm converge quickly and enhance the search efficiency. The position update of the goshawk after the introduction of Gaussian mutation is as follows: , (2-7) Where: For the i The new position of the northern goshawk in the vth dimension; The best position for the current northern goshawk; is the random position of the northern goshawk; and is the weight coefficient, and its value is 0.5; and All of them use the Gaussian distribution random number function with a mean of 0 and a variance of 1 as the Gaussian distribution function coefficient; S303. In the NGO chasing and escaping stage, after the goshawk captures the prey, the prey tries to escape. The mathematical expression of the chasing process is as follows: , (2-8) , (2-9) Where: Updated for the Chase and Escape phases i The Northern Goshawk j The position of the dimension; R is the radius of the attack range; t is the current number of iterations; T is the maximum number of iterations , (2-10) Where: Updated for the Chase and Escape phases i The objective function value of the position of the northern goshawk only; At this time, by introducing Levy The flight strategy optimizes the position update of the northern goshawk. The updated position expression is as follows: , (2-11) , (2-12) Where: is the step size distribution; h and g is the parameter that follows the normal distribution; To control the parameters, here The value is 1; To introduce Levy After the flight i The position of the northern goshawk in the vth dimension; For the i The northern goshawk is located in the vth dimension.

[0008] The improved northern goshawk optimization algorithm INGO in step S4 performs parameter optimization on the gradient boosting decision tree GBDT, specifically: using ICMIC chaotic mapping to generate the initial population position of the goshawk, selecting the first N individuals as the initial population of the goshawk, randomly selecting prey according to formula (2-7) in the prey recognition stage, and updating the position of the goshawk; in the chasing and escaping stage, updating the position of the goshawk chasing the prey in the second stage according to formula (2-10) and formula (2-12), and using Levy The advantage of the flying strategy is that it flies a long distance based on the original search step length, and the direction is constantly changed to determine whether the maximum number of iterations has been reached. If it is satisfied, the individual position and fitness value of the optimal group are output; if not, S3 is re-executed to update the individual position of the goshawk again.

[0009] Compared with the prior art, the present invention has the following beneficial technical effects and advantages: 1. Use the VIKOR algorithm to preprocess the data, which is more suitable for manifold learning capabilities, especially for high-dimensional data. This method can find the optimal linear approximation of the characteristic function of the Laplace-Beltrami operation on the manifold to obtain the optimal projection direction, which can better meet the needs of dimensionality reduction and reduce dependence on experience, thereby improving the reliability and effect of the algorithm; 2. The use of gradient boosting tree GBDT shows significant advantages in a variety of tasks. First, it has high prediction accuracy, especially suitable for complex data sets; second, GBDT can naturally handle missing values ​​without additional filling, reducing the complexity of data preprocessing; in addition, it provides feature importance evaluation, which helps to understand model decisions; finally, GBDT's flexibility is reflected in its support for a variety of loss functions, which enables it to adapt to different types of prediction tasks, greatly enhancing its scope of application; 3. The main parameters of the gradient boosting decision tree (GBDT) are optimized by using the improved northern goshawk optimization algorithm (INGO), which makes up for the blindness of parameter selection during the training process, improves the prediction accuracy of the classification prediction model, and has higher prediction accuracy and practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 This is a main framework diagram of a method for predicting faults in an observation and aiming system based on INGO-GBDT of the present invention; Figure 2 This is a flow chart of the VIKOR algorithm of the present invention. DETAILED DESCRIPTION

[0011] The present invention is described in detail below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited by the accompanying drawings.

[0012] The present invention provides a method for predicting faults of an observation and aiming system based on INGO-GBDT. Figure 1 As shown: The following steps are included: S1. Collect data of the pin signal of the sighting system.

[0013] S2. Preprocess the collected data through the multi-criteria compromise solution sorting method VIKOR to screen out the data set that can be used as model input; divide the model input data set into a test data set and a training data set; Specifically, in step S2, the collected data is evaluated, sorted and normalized by the VIKOR algorithm: S201, construct the initial decision matrix: Has The evaluation object of the parameter index is determined by and indicator set , construct a decision matrix .

[0014] S202, decision matrix Normalize the decision matrix using the standard 0-1 transformation Normalize and get , The specific calculation formula is: When the indicator When it is a benefit indicator, , (1-1) When the indicator When it is a cost indicator, , (1-2) In the formula is the normalized decision matrix, is the maximum value in the matrix, is the minimum value in the matrix.

[0015] S203, according to the normalized matrix , determine the positive ideal solution and negative ideal solution : , (1-3) in, , (1-4) In the formula is the maximum value in the normalized matrix, is the minimum value of the normalized matrix.

[0016] S204, calculate weight vector :Determine the index by constructing an optimization model Weight , the weight vector is calculated as follows: Using the positive ideal solution of the decision matrix and negative ideal solution Calculate the initial weight vector : , (1-5) For the initial vector After normalization, we can get: , (1-6) Then the indicator The weight is ; In the formula is the normalized solution of the positive ideal solution, is the normalized matrix, is a positive ideal solution, is a negative ideal solution.

[0017] S205, calculate each evaluation object The group benefit value 、Individual regret value , compromise value : , (1-7) , (1-8) , (1-9) in, , is the maximum group benefit value; , is the minimum group benefit value; , is the maximum individual regret value; , is the minimum individual regret value; is the normalized weight of each indicator; is a positive ideal solution, is the negative ideal solution, v is the decision mechanism coefficient; For the evaluation object , and These three values ​​are sorted in descending order, resulting in three sorting methods: Sorting, Sorting and Sorting, the final sorting result is based on The value is determined.

[0018] S206, determine the best solution: evaluate the object according to Sort the values ​​and get the solution ,choose The solution with the smallest value is regarded as the optimal solution; If the evaluation object in the program Value Satisfaction , then the compromise solution is obtained by verifying sorting conditions 1 and 2: Sorting condition 1: The smallest and second smallest solutions The value satisfies: , (1-10) Sorting condition 2: The smallest and second smallest solutions Value Sorting Satisfaction ,at the same time Sorting Satisfaction ; If sorting condition 1 and sorting condition 2 are met at the same time, then the solution The final compromise solution in the decision-making process; If only sorting condition 1 or only sorting condition 2 is met, the solution and solutions All are compromise solutions. As a final compromise solution; if sorting condition 1 and sorting condition 2 are not met at the same time, , get the maximum J, then the solution These are the final compromise solutions.

[0019] The Northern Goshawk optimization algorithm (NGO) was improved to solve the problems of poor convergence, low exploration efficiency and easy to fall into local optimality. The ICMIC chaotic mapping was introduced in the NGO population initialization stage to improve the uniformity of population initialization and calculate fitness; the Gaussian mutation strategy was introduced in the NGO prey recognition stage to enhance the search efficiency of the algorithm; and the NGO was introduced in the NGO chasing and escaping stages. Levy The flight strategy optimizes the position update of the northern goshawk.

[0020] Step S3 includes the following steps: S301. In the NGO population initialization stage, a group of goshawk individuals are randomly generated, and random initial positions and speed parameters are assigned to these individuals respectively. The initial position is described as: , (2-1) Where: X is the population matrix of the northern goshawk, For the i The location of the northern goshawk, It refers to i The northern goshawk j The location of the dimension, N represents the population size of the northern goshawk, and s represents the dimension of the problem; The objective function vector of the population can be expressed as: , (2-2) In the formula, F is the objective function vector of the population; For the i The objective function value of the goshawk; For the N The objective function value of the goshawk; The Northern Goshawk optimization algorithm still has the problem of uneven initialization in the initialization stage. At this time, the ICMIC chaotic mapping is introduced to improve the NGO algorithm and enhance the convergence of the algorithm. The improved initialization position update formula is as follows: , (2-3) In the formula, For the i +1 northern goshawk’s position; a is a control parameter, and its value range is [0,+∞].

[0021] S302. In the NGO prey identification stage, each individual evaluates its own fitness based on its initial position and speed through a specific fitness function. In this process, individuals with higher fitness have a higher probability of being regarded as prey: , (2-4) , (2-5) Where: For the i The location of the prey of the northern goshawk; k is a random integer with a value range of [1, N ]; The updated prey identification phase i The Northern Goshawk j The location of the dimension; For the i The objective function value of the prey position of the northern goshawk; r is a random number with a value range of [0,1]; is a random integer with value 1 or 2; , (2-6) Where: Updated for the prey identification phase i The objective function value of the position of the northern goshawk only; The use of flight parameters in the prey identification phase of the NGO algorithm will reduce the algorithm's search efficiency. At this time, the introduction of the Gaussian mutation strategy makes the algorithm converge quickly and enhances the search efficiency. The position update of the goshawk after the introduction of Gaussian mutation is as follows: , (2-7) Where: For the i The new position of the northern goshawk in the vth dimension; The best position for the current northern goshawk; is the random position of the northern goshawk; and is the weight coefficient, and its value is 0.5; and All of them use a Gaussian distribution random number function with a mean of 0 and a variance of 1 as the Gaussian distribution function coefficient.

[0022] S303. In the NGO chasing and escaping stage, after the goshawk captures the prey, the prey tries to escape. The mathematical expression of the chasing process is as follows: , (2-8) , (2-9) Where: Updated for the Chase and Escape phases i The Northern Goshawk j The position of the dimension; R is the radius of the attack range; t is the current number of iterations; T is the maximum number of iterations , (2-10) Where: Updated for the Chase and Escape phases i The objective function value of the position of the northern goshawk only; In the stage, the prey tries to escape. It can be seen from formula (2-9) that the iteration parameter R is limited, and the high-speed flight of the goshawk easily causes the process to fall into the local optimum. Levy The flight strategy optimizes the position update of the northern goshawk. The updated position expression is as follows: , (2-11) , (2-12) Where: is the step size distribution; h and g is the parameter that follows the normal distribution; To control the parameters, here The value is 1; To introduce Levy After the flight i The position of the northern goshawk in the vth dimension; For the i The northern goshawk is located in the vth dimension.

[0023] S4. Use the improved Northern Goshawk optimization algorithm INGO to optimize the parameters of the gradient boosting decision tree GBDT, build the INGO-GBDT fault diagnosis model, and train the training data set.

[0024] The specific steps are as follows: (1) Input the training data set, use the VIKOR algorithm to preprocess it and divide it into training sample set and test sample set; (2) Initialize the parameters such as the population size of goshawks and the maximum number of iterations in INGO; (3) Using ICMIC chaotic mapping to generate the initial population position, the first N individuals were selected as the initial population of northern goshawks; (4) In the prey identification stage, the prey is randomly selected according to formula (2-7) and the position of the goshawk is updated until the conditions are met; (5) In the chasing and escaping phases, the position of the goshawk chasing its prey in the second phase is updated according to equations (2-10) and (2-12), using Replace the original , that is, taking advantage of the Lévy flight, the hawk flies a long distance based on the original search step length, and the direction keeps changing. Determine whether the maximum number of iterations has been reached. If it is satisfied, output the individual position and fitness value of the optimal group; if not, re-execute S3 and update the individual position of the goshawk again (6) Save the optimal value so far; (7) Assign the optimal parameter combination to the GBDT model; (8) Use the training sample data to build the INGO-GBDT classification prediction model, and finally use the test sample data to verify the accuracy of the model.

[0025] More specifically: Let the input sample set , maximum number of iterations (number of base learners) M, loss function , output regression tree model .

[0026] Step 1: Initialization : (3-1); Assume the loss function is square loss: .

[0027] Step 2: Calculate the residual of the loss function in the current model The value of: (3-2); The residual obtained in a) As the new true value of the sample, we get the first regression tree, and the corresponding leaf node area is , j represents the number of leaf nodes.

[0028] right , j leaf node calculates the best fitting value; (3-3); yes Minimum square loss; Update the regression tree model : (3-4).

[0029] Step 3: Get the regression tree model; (3-5).

[0030] Step 1 is the initialization process, which calculates the minimum value of the loss function and takes it as the root node; Step 2 a) calculates the residual estimate through the negative gradient of the loss function in the current model; Step 2 b) fits the residual approximation to establish the mth regression tree; Step 2 c) calculates the loss function to reach the minimum value; Step 2 d) updates the regression tree model; The third step is to obtain the final regression tree model.

[0031] Improved algorithm optimization parameters: The parameters involved in GBDT algorithm modeling mainly include learning rate (learning_rate), which is used to control the step size of parameter update during learning. If the step size is too large, the learning process may diverge. Otherwise, the model will be iterated too many times, which will greatly increase the learning time. The maximum number of iterations (n_estimators) indicates the number of base learners and interacts with learning_rate. When learning_rate is small, the number of iterations needs to be increased to converge the training error. Subsampling (subsample) is used to control the proportion of data set samples involved in fitting. When it is set to less than 1, it can effectively reduce the variance of the overall model and prevent overfitting. The maximum depth of the decision tree (max_depth), the minimum number of samples required for internal node re-division (min_samples_split) and the minimum number of samples contained in the leaf node (min_samples_leaf) are all used to control the complexity of each tree. The specific value depends on the data distribution. If the value is too large, the model structure will be complicated and easily lead to overfitting. Otherwise, it will easily lead to underfitting. Step size factor, if the loss function is minimized too much, add a step size factor to avoid overfitting. The range of the step size factor is that under the same learning effect, if it is smaller, more iterations may be required, and if it is too large, the optimal point will be skipped.

[0032] S5. Input the test data set into the INGO-GBDT prediction model, perform fault diagnosis on the sighting system, and output the prediction results.

[0033] The main parameters of the gradient boosting decision tree (GBDT) are optimized by the improved Northern Goshawk Optimization algorithm (INGO), which makes up for the blindness of parameter selection in the training process and improves the prediction accuracy of the classification prediction model. It has higher prediction accuracy and practicality, and has a significant improvement compared with the NGO-GBDT prediction model, indicating that this model can effectively predict the faults of the pin signals of the power module of the sighting system.

[0034] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting faults in an observation and aiming system based on INGO-GBDT, characterized in that: The following steps are involved: S1, collect data of the pin signal of the sighting system; S2. Preprocess the collected data through the multi-criteria compromise solution ranking method VIKOR, screen out the data set that can be used as model input, and divide the model input data set into a test data set and a training data set; S3. Improvements to the NGO optimization algorithm for northern goshawks, including the introduction of ICMIC chaotic mapping in the NGO population initialization stage to improve the uniformity of population initialization and calculate fitness; the introduction of Gaussian mutation strategy in the NGO prey recognition stage to enhance the algorithm's search efficiency; the introduction of Levy The flight strategy optimizes the position update of the northern goshawk and obtains the improved northern goshawk optimization algorithm INGO; S4, optimize the key parameters of the gradient boosting decision tree GBDT through the improved northern goshawk optimization algorithm INGO, build the INGO-GBDT fault diagnosis model, and train the fault diagnosis model through the training data set in S2; S5. Input the test data set in S2 into the trained INGO-GBDT model, perform fault diagnosis on the sighting system, and output the prediction results.

2. The method for predicting faults of an observation and aiming system based on INGO-GBDT according to claim 1 is characterized in that: In step S2, the collected data is evaluated, sorted and normalized by the VIKOR algorithm as follows: S201, construct the initial decision matrix: Has The evaluation object of the parameter index is determined by and indicator set , construct a decision matrix ; S202, decision matrix Normalize the decision matrix using the standard 0-1 transformation Normalize and get , The specific calculation formula is: When the indicator When it is a benefit indicator, ，(1-1) When the indicator When it is a cost indicator, ，(1-2) In the formula is the normalized decision matrix, is the maximum value in the matrix, is the minimum value in the matrix; S203, according to the normalized matrix , determine the positive ideal solution and negative ideal solution : ，(1-3) in, , (1-4) In the formula is the maximum value in the normalized matrix, is the minimum value of the normalized matrix; S204, calculate weight vector :Determine the index by constructing an optimization model Weight , the weight vector is calculated as follows: Using the positive ideal solution of the decision matrix and negative ideal solution Calculate the initial weight vector : ，(1-5) For the initial vector After normalization, we can get: ，(1-6) Then the indicator The weight is ; In the formula is the normalized solution of the positive ideal solution, is the normalized matrix; S205, calculate each evaluation object The group benefit value 、Individual regret value , compromise value : ，(1-7) ，(1-8) ，(1-9) in, , is the maximum group benefit value; , is the minimum group benefit value; , is the maximum individual regret value; , is the minimum individual regret value; is the normalized weight of each indicator; is a positive ideal solution, is the negative ideal solution, v is the decision mechanism coefficient; For the evaluation object , and These three values ​​are sorted in descending order, resulting in three sorting methods: Sorting, Sorting and Sorting, the final sorting result is based on The value is determined; S206, determine the best solution: evaluate the object according to Sort the values ​​and get the solution ,choose The solution with the smallest value is regarded as the optimal solution; If the evaluation object in the program Value Satisfaction , then the compromise solution is obtained by verifying sorting conditions 1 and 2: Sorting condition 1: The smallest and second smallest solutions The value satisfies: , (1-10) Sorting condition 2: The smallest and second smallest solutions Value Sorting Satisfaction ,at the same time Sorting Satisfaction ; If sorting condition 1 and sorting condition 2 are met at the same time, then the solution The final compromise solution in the decision-making process; If only sorting condition 1 or only sorting condition 2 is met, the solution and solutions All are compromise solutions. As a final compromise solution; if sorting condition 1 and sorting condition 2 are not met at the same time, , get the maximum J, then the solution These are the final compromise solutions.

3. The method for predicting faults of an observation and aiming system based on INGO-GBDT according to claim 1, characterized in that: Step S3 includes the following steps: S301. In the NGO population initialization stage, a group of goshawk individuals are randomly generated, and random initial positions and speed parameters are assigned to these individuals respectively. The initial position is described as: ，(2-1) Where: X is the population matrix of the northern goshawk, For the i The location of the northern goshawk, It refers to i The northern goshawk j The location of the dimension, N represents the population size of the northern goshawk, and s represents the dimension of the problem; The objective function vector of the population can be expressed as: ，(2-2) In the formula, F is the objective function vector of the population; For the i The objective function value of the goshawk; For the N The objective function value of the goshawk; At this time, ICMIC chaotic mapping is introduced to improve the NGO algorithm and enhance the convergence of the algorithm. The improved initialization position update formula is as follows: ，(2-3) In the formula, For the i +1 northern goshawk position; a is a control parameter, and its value range is [0, +∞]; S302. In the NGO prey identification stage, each individual evaluates its own fitness based on its initial position and speed through a specific fitness function. In this process, individuals with higher fitness have a higher probability of being regarded as prey: ，(2-4) ，(2-5) Where: For the i The prey location of the northern goshawk; k is a random integer ranging from [1, N ]; The updated prey identification phase i The Northern Goshawk j The location of the dimension; For the i The objective function value of the prey position of the northern goshawk; r is a random number with a value range of [0,1]; is a random integer with value 1 or 2; ，(2-6) Where: Updated for the prey identification phase i The objective function value of the position of the northern goshawk only; At this time, the Gaussian mutation strategy is introduced to make the algorithm converge quickly and enhance the search efficiency. The position update of the goshawk after the introduction of Gaussian mutation is as follows: ，(2-7) Where: For the i The new position of the northern goshawk in the vth dimension; The best position for the current northern goshawk; is the random position of the northern goshawk; and is the weight coefficient, and its value is 0.5; and All of them use the Gaussian distribution random number function with a mean of 0 and a variance of 1 as the Gaussian distribution function coefficient; S303. In the NGO chasing and escaping stage, after the goshawk captures the prey, the prey tries to escape. The mathematical expression of the chasing process is as follows: ，(2-8) ，(2-9) Where: Updated for the Chase and Escape phases i The Northern Goshawk j The position of the dimension; R is the radius of the attack range; t is the current number of iterations; T is the maximum number of iterations , (2-10) Where: Updated for the Chase and Escape phases i The objective function value of the position of the northern goshawk only; At this time, by introducing Levy The flight strategy optimizes the position update of the northern goshawk. The updated position expression is as follows: ，(2-11) ，(2-12) Where: is the step size distribution; h and g is the parameter that follows the normal distribution; To control the parameters, here The value is 1; To introduce Levy After the flight i The position of the northern goshawk in the vth dimension; For the i The northern goshawk is located in the vth dimension.

4. The method for predicting faults of an observation and aiming system based on INGO-GBDT according to claim 3 is characterized in that: The improved northern goshawk optimization algorithm INGO in step S4 performs parameter optimization on the gradient boosting decision tree GBDT, specifically: using ICMIC chaotic mapping to generate the initial population position of the goshawk, selecting the first N individuals as the initial population of the goshawk, randomly selecting prey according to formula (2-7) in the prey recognition stage, and updating the position of the goshawk; in the chasing and escaping stage, updating the position of the goshawk chasing the prey in the second stage according to formula (2-10) and formula (2-12), and using Levy The advantage of the flying strategy is that it flies a long distance based on the original search step length, and the direction is constantly changed to determine whether the maximum number of iterations has been reached. If it is satisfied, the individual position and fitness value of the optimal group are output; if not, S3 is re-executed to update the individual position of the goshawk again.

Images