A Landing Gear Measurement Point Optimization Method and System Based on Binary Differential Evolution Algorithm

By using a landing gear measurement point optimization method based on binary differential evolution algorithm, a multi-signal flow graph model is built and a fault-testability matrix is ​​generated to optimize the measurement point layout of the landing gear system. This solves the problem of high testing complexity of the landing gear system and reduces testing efficiency and cost.

CN123116774APending Publication Date: 2026-11-17CHENGDU AIRCRAFT INDUSTRY GROUP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611089430.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-22
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In the existing technology, the optimization of the test point layout of the landing gear system is insufficient, resulting in the lack of evaluation of test completeness, large room for optimization of test methods, high test complexity, low test efficiency, and difficulty in applying existing algorithms to the landing gear system.

Method used

A landing gear test point optimization method based on binary differential evolution algorithm is adopted. By building a multi-signal flow graph model of landing gear, a fault-testability matrix is ​​generated. The binary-coded differential evolution algorithm is used for optimization, including mutation, crossover and selection operations, to generate the optimal test point combination.

Benefits of technology

It improves the testing efficiency of landing gear systems, ensures the completeness of testing, reduces testing costs and time, avoids local optima traps, and is highly adaptable and flexible.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123116774A_ABST
    Figure CN123116774A_ABST
Patent Text Reader

Abstract

This invention relates to the field of landing gear test point optimization technology, and particularly to a method and system for landing gear test point optimization based on a binary differential evolution algorithm. The method, from a testability evaluation perspective, uses a multi-signal flow graph model of the landing gear as the basis for testability evaluation. It analyzes the generated fault-testability matrix, transforms test information into chromosomes through encoding operations, and abstracts the test point optimization problem into a mathematical problem. A fitness function is designed using the system's testability index. By executing and repeating differential evolution operations, including mutation, crossover, and selection, the chromosome with the highest fitness value is finally output as the optimal combination of test points. This method and system serve the product design and test process design stages of landing gear systems, improving testing efficiency while ensuring test completeness.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of landing gear measurement point optimization technology, and in particular to a landing gear measurement point optimization method and system based on binary differential evolution algorithm. Background Technology

[0002] Testability refers to the ease with which a target system can be tested, and to what extent. As aircraft systems become increasingly complex, testing technology plays a growing role in aircraft manufacturing. The testing process and its complexity involve numerous aspects, and current production processes suffer from issues such as overtesting, undertesting, lack of test completeness assessment, and significant room for test method optimization. Both test completeness assessment and test method optimization involve test point layout. A reasonable test point layout not only aids in self-testability analysis during the product design phase but also plays a crucial role in the design of airborne system test processes. Timely optimization of test point layout can effectively improve test completeness, reduce the frequency of overtesting and undertesting, increase test reliability, and lower test cycles and costs.

[0003] In the field of measurement point optimization, various optimization algorithms have performed well in different industry contexts. A typical example is the particle swarm optimization technique for vehicle gearboxes, which considers the vibration signals of internal components and their transmission paths, and solves the measurement point configuration problem from the perspective of signal analysis. In the aerospace field, a typical example is the use of adaptive simulated annealing genetic algorithm to analyze the finite element model of the fuselage panel of a large aircraft, and to obtain the optimal arrangement method of panel detection points from the mathematical model. Another example is the effective independence method proposed by some scholars, which starts from the data required for parameter identification and solves the problem of structural sensor measurement point optimization for orbital spacecraft based on the information array.

[0004] The typical examples mentioned above demonstrate several main approaches to solving problems in the current field of measurement point optimization: 1) Optimize the number or location of sensors; 2) Analyze signal energy, approach it from a diagnostic perspective, and transform the problem of optimizing measurement points into the problem of finding the detection location with the maximum amount of information; 3) From a model perspective, establish mathematical models such as the finite element method for the object, conduct observations under different fault modes, and obtain the optimal measurement point scheme.

[0005] However, existing technologies rarely apply measurement point optimization techniques to landing gear systems, effectively using them as a test analysis method while ensuring test completeness. Summary of the Invention

[0006] To address the aforementioned technical issues, this invention proposes a landing gear measurement point optimization method and system based on a binary differential evolution algorithm. This method serves the product design and test process design stages of landing gear systems, improving test efficiency while ensuring test completeness.

[0007] This invention is achieved by adopting the following technical solution: The landing gear measurement point optimization method based on binary differential evolution algorithm includes the following steps: Step S1. Build a landing gear multi-signal flow graph model and generate a fault-testability matrix based on the model; wherein, the fault-testability matrix is ​​a binary knowledge matrix, with 1 and 0 respectively indicating whether there is a dependency relationship between the operating mode and the test point in the system; Step S2. Optimize the known discrete combinations of test points using a binary-coded differential evolution algorithm, specifically including the following steps: Step S 21 The test point combinations are represented using binary encoding; several chromosomes are randomly generated to obtain the initial population. Step S 22 Based on the failure-testability matrix and a pre-constructed fitness function, calculate the fitness value of each chromosome in the current generation; Step S 23 Differential evolution operations, including mutation, crossover, and selection, are performed to generate a new generation population. Mutation refers to: randomly selecting two different chromosomes from the current generation population, generating a covering code, and applying the covering code to the chromosome with the highest fitness value in the current generation to generate a mutated chromosome, thus obtaining a mutated population. Crossover refers to: selecting allotropes on chromosomes from the mutated population and the current generation population for crossover operations to generate experimental chromosomes, thus obtaining a progeny population. Selection refers to: selecting chromosomes from the current generation population and the progeny population to enter the next cycle. Step S3. Repeat step S. 22 To step S 23 The process continues until the convergence condition is met or the maximum number of iterations is reached, at which point the chromosome with the highest fitness value is output as the optimal test point combination.

[0008] The method for building a multi-signal flow graph model of the landing gear is as follows: divide the object structure according to the functional hierarchy of the landing gear system, determine the correspondence between system components, operating modes and test points, sort out the signal propagation path, realize hierarchical modeling, and complete the construction of the multi-signal flow graph model of the landing gear.

[0009] The combination of test points is represented using binary encoding, wherein the encoding on the chromosome is a string of length [length missing]. n binary sequence, n This represents the number of test points in the system. Code 1 indicates that a test point has been selected. This sequence represents a combination of test points.

[0010] The fitness function is: , In the formula, Indicates the first i One indicator, Indicates the first i The weight of each indicator, and , All n dimension.

[0011] The metrics include fault detection rate, fault isolation rate, and test point utilization rate, all derived from the fault-testability matrix.

[0012] The mutation specifically includes the following steps: Step 1. Randomly select two chromosomes in the current generation of the population and perform an XOR operation on them. If the numbers are the same, the result is 0; if the numbers are different, the result is 1. Use the string of 0s and 1s obtained by the XOR operation as the overlay code. Step 2. Apply the above covering code to the chromosome with the highest fitness value in the current generation to perform a covering operation, generating a mutated chromosome: if the covering code is... i If a position is 1, the number corresponding to the chromosome with the highest fitness value in the current generation is overwritten; if the original position is 0, it becomes 1; if the original position is 1, it becomes 0. Step 3. Repeat steps 1 and 2 above, for a total of [number missing]. N Next, a mutant population was obtained.

[0013] The crossover specifically refers to: generating a 0 / 1 mask of the same length as the chromosome for each chromosome in the current generation of the population; for the first 0 / 1 of the mask... i Each bit independently generates random numbers that follow a uniform distribution. R i , R i ∈[0,1]; based on preset crossover probability P c Generate experimental chromosomes: If R i ≤ P c When the mask's first i The position is 1, the experimental chromosome's... i The position is taken from the corresponding mutated chromosome; otherwise, the position of the chromosome is retained. i The above crossover operation is performed on each of the N chromosomes in the population, generating a total of N experimental chromosomes to obtain the offspring population.

[0014] It also includes dynamically adjusting the crossover probability. P c : , In the formula, , These are the crossover probability constants, For chromosome number The fitness value of the position. The average fitness value for each generation. This represents the maximum value of chromosome fitness for that generation.

[0015] The selection specifically refers to: merging the current generation population and the offspring population into a temporary population, calculating the fitness values ​​of all chromosomes in the temporary population, and selecting the value that is the highest among them. N Each chromosome enters the next round of iteration.

[0016] A landing gear measurement point optimization system based on binary differential evolution algorithm includes: The modeling and matrix generation module is used to build a landing gear multi-signal flow graph model and generate a fault-testability matrix based on the model; wherein, the fault-testability matrix is ​​a binary knowledge matrix, with 1 and 0 respectively indicating whether there is a dependency relationship between the operating mode and the test point in the system; The population initialization module is used to represent the combination of test points in binary encoding; and to randomly generate several chromosomes to obtain the initial population. The fitness evaluation module is used to calculate the fitness value of each chromosome in the current generation based on the failure-testability matrix and a pre-constructed fitness function. The differential evolution optimization module includes a mutation submodule, a crossover submodule, and a selection submodule. The mutation submodule randomly selects two different chromosomes, generates a covering code, and applies the covering code to the chromosome with the highest fitness value in the current generation to generate a mutated chromosome, thus obtaining a mutated population. The crossover submodule selects allotropes on chromosomes from the mutated population and the current generation population for crossover operations to generate experimental chromosomes, thus obtaining a progeny population. The selection submodule selects chromosomes from the current generation population and the progeny population to enter the next round of the cycle. The iteration control and output module is used to repeatedly call the fitness evaluation module and the differential evolution optimization module until the convergence condition is met or the maximum number of iterations is reached, and outputs the chromosome with the largest fitness value as the optimal test point combination.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The landing gear test point optimization technology based on binary differential evolution algorithm provided by this invention serves the product design and test process design stages of landing gear systems, achieving a coordinated match between test completeness, fault isolation rate, fault detection rate, and test cost cycle. Compared with similar algorithms, it can improve test efficiency and configuration flexibility while ensuring test completeness, and achieves better verification results on landing gear systems.

[0018] Specifically, this invention utilizes a binary differential evolution algorithm to perform test analysis and optimization of test points for a landing gear system test evaluation model. Compared to similar algorithms, it is less prone to getting trapped in local optima and has a high degree of automation. More specifically, this invention uses binary representation of chromosomes, analogizing solutions to chromosomes and performing three operations—mutation, crossover, and selection—on the sequence to generate new optimized solutions. Compared to existing technologies using improved particle swarm optimization algorithms based on simulated annealing and genetic algorithms, this is more intuitive, easier to operate and compute, and has better engineering adaptability. Furthermore, the accelerated convergence strategy used in this invention can largely avoid getting trapped in local optima.

[0019] 2. In existing technologies, measurement point layout schemes in other fields are generally geared towards engineering structural problems, establishing dynamic models of the structure based on relevant physical parameters. Applying these schemes to the landing gear system presents challenges such as modeling difficulties, computational complexity, and a lack of specificity. Due to the relatively fixed yet discontinuous nature of measurement points in the landing gear system, this invention, through fault analysis, focuses on optimizing measurement points based on structural response, thus enabling its application to fault measurement point systems in aircraft landing gear systems.

[0020] 3. This invention converts the relationships of a multi-signal flow graph model into a D-matrix, making optimization calculation and analysis more convenient.

[0021] 4. Based on the binary differential evolution algorithm used, this invention designs a mutation operation based on the covering code, which can solve the problem that the subtraction of binary encoded string vectors is meaningless.

[0022] 5. The metrics used in this invention include fault detection rate, fault isolation rate, and test point utilization rate, all derived from a fault-testability matrix. These metrics are specifically selected from a testability evaluation perspective, taking into account the unique characteristics of the landing gear system. Attached Figure Description

[0023] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments, wherein: Figure 1 This is a schematic diagram of the process of the present invention; Figure 2 This is a schematic diagram of the multi-signal flow graph model and D matrix in this invention; Figure 3 This is a schematic diagram of the encoding operation in this invention; Figure 4 This is a schematic diagram illustrating the changes in the D matrix before and after encoding in this invention; Figure 5 This is a schematic diagram of the mutation operation in this invention; Figure 6 This is a schematic diagram of the cross operation in this invention; Figure 7This is a schematic diagram showing the results of the measurement point optimization for the landing gear retraction system in this invention. Detailed Implementation

[0024] Example 1 As a basic embodiment of the present invention, the present invention includes a landing gear measurement point optimization method based on binary differential evolution algorithm, comprising the following steps: Step S1. Construct a landing gear multi-signal flow graph model and generate a fault-testability matrix based on the model. The fault-testability matrix is ​​a binary knowledge matrix, where 1 and 0 represent whether there is a dependency relationship between the system's operating conditions and the test points, respectively.

[0025] Step S2. Optimize the known discrete combinations of test points using a binary-coded differential evolution algorithm, specifically including the following steps: Step S 21 The test point combinations are represented using binary encoding; several chromosomes are randomly generated to obtain the initial population.

[0026] Step S 22 Based on the fault-testability matrix and a pre-constructed fitness function, the fitness value of each chromosome in the current generation is calculated.

[0027] Step S 23 Differential evolution operations, including mutation, crossover, and selection, are performed to generate a new generation population. Mutation refers to: randomly selecting two different chromosomes from the current generation population, generating a covering code, and applying the covering code to the chromosome with the highest fitness value in the current generation to generate a mutated chromosome, thus obtaining a mutated population. Crossover refers to: selecting alleles on chromosomes from the mutated population and the current generation population for crossover operations to generate experimental chromosomes, thus obtaining a progeny population. Selection refers to: selecting chromosomes from the current generation population and the progeny population to enter the next cycle.

[0028] Step S3. Repeat step S. 22 To step S 23 The process continues until the convergence condition is met or the maximum number of iterations is reached, at which point the chromosome with the highest fitness value is output as the optimal test point combination.

[0029] Example 2 As a preferred embodiment of the present invention, the present invention includes a landing gear measurement point optimization method based on binary differential evolution algorithm, comprising the following steps: Step S1. Construct a landing gear multi-signal flow graph model and generate a fault-testability matrix based on this model. The fault-testability matrix is ​​a binary knowledge matrix, where 1 and 0 represent the existence of dependencies between system operating modes and test points, respectively. The method for constructing the landing gear multi-signal flow graph model is as follows: The object structure is divided according to the functional hierarchy of the landing gear system; the correspondence between system components, operating modes, and test points is determined; signal propagation paths are identified; hierarchical modeling is achieved; and the construction of the landing gear multi-signal flow graph model is completed.

[0030] Step S2. Optimize the known discrete combinations of test points using a binary-coded differential evolution algorithm, specifically including the following steps: Step S 21 The combination of test points is represented using binary encoding. The encoding on the chromosome is a string of length [length missing]. n binary sequence, n This represents the number of test points in the system. Encoding 1 indicates that a test point has been selected; this sequence represents a combination of test points. Several chromosomes are randomly generated to obtain an initial population, which consists of several randomly generated binary solution vectors, each solution vector being called a chromosome.

[0031] Step S 22 Based on the fault-testability matrix and a pre-constructed fitness function, the fitness value of each chromosome in the current generation is calculated.

[0032] Step S 23 Differential evolution operations, including mutation, crossover, and selection, are performed to generate a new generation population. Mutation refers to: randomly selecting two different chromosomes from the current generation population, generating a covering code, and applying the covering code to the chromosome with the highest fitness value in the current generation to generate a mutated chromosome, thus obtaining a mutated population. Crossover refers to: selecting alleles on chromosomes from the mutated population and the current generation population for crossover operations to generate experimental chromosomes, thus obtaining a progeny population. Selection refers to: selecting chromosomes from the current generation population and the progeny population to enter the next cycle.

[0033] Step S3. Repeat step S. 22 To step S 23 The process continues until the convergence condition is met or the maximum number of iterations is reached, at which point the chromosome with the highest fitness value is output as the optimal test point combination.

[0034] Example 3 In another preferred embodiment of the present invention, the present invention includes a landing gear measurement point optimization method based on a binary differential evolution algorithm, comprising the following steps: Step S1. Construct a landing gear multi-signal flow graph model and generate a fault-testability matrix based on the model. The fault-testability matrix is ​​a binary knowledge matrix, where 1 and 0 represent whether there is a dependency relationship between the system's operating conditions and the test points, respectively.

[0035] Step S2. Optimize the known discrete combinations of test points using a binary-coded differential evolution algorithm, specifically including the following steps: Step S 21 The test point combinations are represented using binary encoding; several chromosomes are randomly generated to obtain the initial population.

[0036] Step S 22 Based on the fault-testability matrix and a pre-constructed fitness function, the fitness value of each chromosome in the current generation is calculated.

[0037] Step S 23 Perform differential evolution operations, including mutation, crossover, and selection, to generate a new generation of population. Mutation refers to: randomly selecting two different chromosomes from the current generation, generating a covering code, and applying the covering code to the chromosome with the highest fitness value in the current generation to generate a mutated chromosome, thus obtaining a mutated population.

[0038] The crossover refers to: selecting allotropes on chromosomes from the mutant population and the current generation population for crossover operations to generate experimental chromosomes and obtain offspring populations. Specifically, this includes: generating a 0 / 1 mask of the same length as the chromosome for each chromosome in the current generation population; for the 0th... i Each bit independently generates random numbers that follow a uniform distribution. R i , R i ∈[0,1]; based on preset crossover probability P c Generate experimental chromosomes: If R i ≤ P c When the mask's first i The position is 1, the experimental chromosome's... i The i-th position is taken from the corresponding mutated chromosome; otherwise, the i-th position of the chromosome is retained. The above crossover operation affects the population... N Each chromosome is executed separately, generating a total of N Experimental chromosomes were used to obtain offspring populations.

[0039] To improve convergence speed and avoid getting trapped in local optima, the crossover probability is also dynamically adjusted. P c : , In the formula, , These are the crossover probability constants, For chromosome number The fitness value of the position. The average fitness value for each generation. This represents the maximum value of chromosome fitness for that generation.

[0040] The selection refers to selecting chromosomes from the current generation and offspring populations to enter the next cycle.

[0041] Step S3. Repeat step S. 22 To step S 23 The process continues until the convergence condition is met or the maximum number of iterations is reached, at which point the chromosome with the highest fitness value is output as the optimal test point combination.

[0042] Example 4 In another preferred embodiment of the present invention, the present invention includes a landing gear test point optimization method based on a binary differential evolution algorithm. From the perspective of testability evaluation, the method uses a multi-signal flow graph model of the landing gear as the basis for testability evaluation, analyzes the generated fault-testability matrix (D matrix), transforms the test information into chromosomes through encoding operations, abstracts the test point optimization problem into a mathematical problem, designs a fitness function using the system's testability indicators, and completes the optimization process under the operation of various operators. (See attached specification.) Figure 1 The optimization method specifically includes the following steps: Step S1. Build a multi-signal flow graph model of the landing gear and generate a fault-testability matrix based on the model.

[0043] The method for building the landing gear multi-signal flow graph model is as follows: Based on the functional hierarchy of the landing gear system, the object structure is divided; the correspondence between system components, operating conditions (fault modes), and test points is determined; signal propagation paths are identified; hierarchical modeling is achieved; and the landing gear multi-signal flow graph model is completed. Based on the landing gear multi-signal flow graph model, a fault-testability matrix (D matrix) is extracted; static indicators such as fault detection rate and fault isolation rate, as well as dynamic indicators such as average isolation / maintenance cost and average fuzzy group size, are calculated; and a testability evaluation index system is established around the test points. The landing gear system's multi-signal flow graph model and D matrix are attached to the instruction manual. Figure 2 As shown.

[0044] The fault-testability matrix, or D matrix, is a binary knowledge matrix, where 1 and 0 represent whether there is a dependency relationship between the system's operating conditions and the test points, respectively. The mathematical description is as follows: , In matrix D, rows represent operating conditions (fault modes), and column elements represent test / measurement points. Representing the i The operating condition (fault) mode and the first j This is related to a test.

[0045] Step S2. The differential algorithm is a type of heuristic genetic algorithm. For the binary test knowledge mentioned above, during the optimization process of test points, each test point only has two states: "selected" and "not selected." Its solution space is not continuous, therefore, the traditional differential evolution algorithm using real-number encoding is not applicable. Therefore, this invention uses a binary-encoded differential evolution algorithm, optimizing the optimization process by changing the intermediate operators, thereby optimizing known discrete combinations of test points. Specifically, it includes the following steps: Step S 21 Initialize parameters, determining population size, maximum number of iterations, etc. Convert test information into chromosomes through encoding, i.e., represent the combinations of test points using binary encoding. The operation is as shown in the appendix of the manual. Figure 3 As shown. The problem solution is mapped to specific terms in the algorithm. A chromosome represents the sequence of test points, and the coding (gene) on the chromosome is a sequence of length... n binary sequence, n This represents the number of test points in the system. Encoding 1 indicates that a test point has been selected; this sequence represents a combination of test points. The changes in the D matrix before and after encoding are shown in the appendix to the instruction manual. Figure 4 As shown.

[0046] Several chromosomes are randomly generated to obtain the initial population.

[0047] Step S 22 Based on the fault-testability matrix and a pre-constructed fitness function, the fitness value of each chromosome in the current generation is calculated.

[0048] The optimization process requires a fitness function as the objective, which is the goal of the initial random solution to converge toward the optimal direction. For the landing gear system, from the perspective of testability evaluation, three indicators are selected to construct the fitness function: Fault Detection Rate (FDR), Fault Isolation Rate (FIR), and Test Rate (TR).

[0049] The fault detection rate metric refers to the proportion of detectable faults (operating conditions) out of all modes, and is derived from the D matrix. A row in the D matrix that is entirely zero indicates that the mode is undetectable. The fault isolation rate metric refers to the proportion of isolateable faults (operating conditions) out of all modes, and is derived from the D matrix. Two identical rows in the D matrix indicate that the two modes are not isolateable. The test point utilization rate metric refers to the proportion of test points used out of all test points under the current chromosome solution.

[0050] Under the constraints of the above indicators, the fitness function is designed as follows. , , The weights of each indicator are represented by 1, and the sum of the three is always 1. The actual weight allocation is freely chosen by the user according to the actual project needs.

[0051] .

[0052] The fitness function described is merely an example. In practical applications, users have a high degree of freedom in selecting metrics and allocating weights. Various test metrics, including coverage failure rate, average test cost, and system hierarchy, can be incorporated into the fitness function construction process. The general expression for the fitness function is: . in Indicates the first i One indicator, Indicates the first i The weight of each indicator, and , All n dimension.

[0053] Step S 23 Perform differential evolution operations, including mutation, crossover, and selection, to generate a new generation of population.

[0054] The mutation process refers to the generation of a mutated population based on the difference between two chromosomes and the best chromosome of the current generation (the chromosome with the highest fitness value). In traditional real-number encoded differential evolution algorithms, directly subtracting the two chromosomes is a standard vector subtraction operation. This invention uses a binary differential evolution algorithm and designs a mutation operation based on a covering code to solve the problem that subtracting binary encoded string vectors is meaningless. (Refer to the appendix of the specification.) Figure 5 Specifically, it includes the following steps: Step 1. Randomly select two chromosomes from the current generation of the population and perform an XOR operation on them. If the numbers are the same, the result is 0; if the numbers are different, the result is 1. The XOR result is the "difference" between the two chromosomes, which is a string of 0s and 1s. This string is used as the overlay code.

[0055] Step 2. Traverse the current population and select the best chromosome, i.e., the chromosome with the highest fitness value. Apply the covering code to the best chromosome to perform a covering operation, generating a new chromosome, i.e., a mutated chromosome: if the covering code is... i If a position is 1, the number corresponding to the chromosome with the highest fitness value in the current generation is overwritten; if the original position is 0, it becomes 1; if the original position is 1, it becomes 0.

[0056] Step 3. Repeat steps 1 and 2 above, for a total of [number missing]. N Next, a mutant population was obtained.

[0057] The crossover process refers to the selection of allotropes on chromosomes from both the variant population and the current generation population for crossover operations to generate experimental chromosomes and obtain offspring populations. (Refer to the appendix of the instruction manual.) Figure 6 The crossover process is as follows: For each chromosome in the current generation of the population, generate a 0 / 1 mask of the same length as the chromosome; for the 0th chromosome of the mask... i Each bit independently generates random numbers that follow a uniform distribution. R i , R i ∈[0,1]; based on preset crossover probability P c Generate experimental chromosomes: If R i ≤ P c When the mask's first i The position is 1, the experimental chromosome's... i The position is taken from the corresponding mutated chromosome; otherwise, the position of the chromosome is retained. i The above crossover operation affects the population; N Each chromosome is executed separately, generating a total of N Experimental chromosomes were used to obtain offspring populations.

[0058] In the later stages of traditional genetic algorithms, a dense concentration of fitness values ​​often occurs, significantly slowing down the convergence speed and making the algorithm prone to getting trapped in local optima. This invention proposes a strategy to accelerate convergence in response to this situation. During gene manipulation, when the fitness value of a crossover individual is high, the crossover probability is dynamically decreased to preserve better individuals; conversely, when the fitness value of a crossover individual is low, the crossover probability is dynamically increased to breed better individuals. When the fitness value of a mutated individual is high, the mutation probability is dynamically decreased to prevent the destruction of better individuals. (Mutation probability) P m With cross probability P c The design is as follows: , , In the formula, , The probability constant of variation, , The crossover probability constant; For chromosome number i The fitness value of the position. The average fitness value for each generation. This represents the maximum chromosome fitness value for that generation. This invention first calculates the average fitness value for each generation. If the fitness value of the current individual is greater than the average, the mutation probability constant is dynamically adjusted during the mutation operation. When performing crossover operations, the crossover probability constant is dynamically adjusted. This increases the probability of mutation and crossover, thus accelerating the convergence speed of the algorithm.

[0059] The selection refers to choosing a certain number of chromosomes from the current generation and offspring populations to enter the next round of the cycle. This invention employs a greedy algorithm for selecting individuals in the next generation population, that is, merging offspring and current generation chromosomes into a temporary population, calculating the fitness values ​​of the chromosomes in the temporary population, and selecting those with the highest fitness values. N Each chromosome enters the next round of iteration.

[0060] Step S3. Repeat step S. 22 To step S 23 The process continues until the convergence condition is met or the maximum number of iterations is reached, at which point the chromosome with the highest fitness value is output as the optimal test point combination.

[0061] Example 5 As a specific embodiment of the present invention, based on the measurement point optimization method described in any of Embodiments 1-4, a multi-signal flow graph model is established for the landing gear system, the D matrix is ​​extracted, and the weights of the three indicators in the fitness function are selected as α=0.4, β=0.3, and γ=0.3. A schematic diagram of the optimized measurement points for the landing gear retraction system is shown in the appendix to the specification. Figure 7 As shown in the diagram. Initially, the system included all the measurement points shown in the diagram. After optimization, the light-colored circles represent the selected measurement points, while the dark-colored circles represent the unselected measurement points. Firstly, the intuitive results from the system diagram model show that the number of measurement points used after optimization is reduced compared to before optimization, resulting in a corresponding decrease in cost. Secondly, this invention statistically analyzed the relevant indicators of the landing gear retraction system measurement points before and after optimization, as shown in Table 1 below: Table 1. Statistical table of relevant indicators before and after optimization of landing gear retraction system measuring points.

[0062] Table 1 above shows a comparison of relevant indicators of the system before and after optimization. The number of measurement points used in the system was reduced from 30 to 25; the average fuzzy group size decreased from 1.23 to 1.03; the average weighted cost of isolation maintenance decreased from 0.09 to 0.08; and the redundant test groups were cleared. Before optimization, the system's FDR and FIR were 83.33%, and after optimization, the corresponding values ​​were 96.7%. This improved the system testing efficiency and reduced the system testing cost while ensuring the completeness of system testing.

[0063] Example 6 As another preferred embodiment of the present invention, the present invention includes a landing gear measurement point optimization system based on a binary differential evolution algorithm, comprising: The modeling and matrix generation module is used to build a landing gear multi-signal flow graph model and generate a fault-testability matrix based on the model. The fault-testability matrix is ​​a binary knowledge matrix, with 1 and 0 representing whether there is a dependency relationship between the operating mode and the test point in the system.

[0064] The population initialization module is used to represent the combination of test points in binary encoding; and to randomly generate several chromosomes to obtain the initial population.

[0065] The fitness evaluation module is used to calculate the fitness value of each chromosome in the current generation based on the failure-testability matrix and a pre-built fitness function.

[0066] The differential evolution optimization module includes a mutation submodule, a crossover submodule, and a selection submodule. The mutation submodule is used to randomly select two different chromosomes, generate a covering code, and apply the covering code to the chromosome with the highest fitness value in the current generation to generate a mutated chromosome, thus obtaining a mutated population. The crossover submodule is used to select allotropes on chromosomes from the mutated population and the current generation population for crossover operations to generate experimental chromosomes, thus obtaining a progeny population. The selection submodule is used to select chromosomes from the current generation population and the progeny population to enter the next round of the cycle.

[0067] The iteration control and output module is used to repeatedly call the fitness evaluation module and the differential evolution optimization module until the convergence condition is met or the maximum number of iterations is reached, and outputs the chromosome with the largest fitness value as the optimal test point combination.

[0068] Specifically, each module in the system described in this embodiment is based on the same innovative concept as any of the methods in embodiments 1 to 4 above. When used, it adopts the same technical means as the above methods and can achieve the same technical effect, so it will not be described again here.

[0069] In summary, any other corresponding modifications made by those skilled in the art after reading this invention document, without requiring creative mental effort, based on the technical solutions and concepts of this invention, are all within the scope of protection of this invention.

Claims

1. A landing gear measurement point optimization method based on binary differential evolution algorithm, characterized in that: Includes the following steps: Step S1. Build a landing gear multi-signal flow graph model and generate a fault-testability matrix based on the model; wherein, the fault-testability matrix is ​​a binary knowledge matrix, with 1 and 0 respectively indicating whether there is a dependency relationship between the operating mode and the test point in the system; Step S2. Optimize the known discrete combinations of test points using a binary-coded differential evolution algorithm, specifically including the following steps: Step S 21 The test point combinations are represented using binary encoding; several chromosomes are randomly generated to obtain the initial population. Step S 22 Based on the failure-testability matrix and a pre-constructed fitness function, calculate the fitness value of each chromosome in the current generation; Step S 23 Differential evolution operations, including mutation, crossover, and selection, are performed to generate a new generation population. Mutation refers to: randomly selecting two different chromosomes from the current generation population, generating a covering code, and applying the covering code to the chromosome with the highest fitness value in the current generation to generate a mutated chromosome, thus obtaining a mutated population. Crossover refers to: selecting allotropes on chromosomes from the mutated population and the current generation population for crossover operations to generate experimental chromosomes, thus obtaining a progeny population. Selection refers to: selecting chromosomes from the current generation population and the progeny population to enter the next cycle. Step S3. Repeat step S. 22 To step S 23 The process continues until the convergence condition is met or the maximum number of iterations is reached, at which point the chromosome with the highest fitness value is output as the optimal test point combination.

2. The landing gear measurement point optimization method based on binary differential evolution algorithm according to claim 1, characterized in that: The method for building a multi-signal flow graph model of the landing gear is as follows: divide the object structure according to the functional hierarchy of the landing gear system, determine the correspondence between system components, operating modes and test points, sort out the signal propagation path, realize hierarchical modeling, and complete the construction of the multi-signal flow graph model of the landing gear.

3. The landing gear measurement point optimization method based on binary differential evolution algorithm according to claim 1, characterized in that: The combination of test points is represented using binary encoding, wherein the encoding on the chromosome is a string of length [length missing]. n binary sequence, n This represents the number of test points in the system. Code 1 indicates that a test point has been selected. This sequence represents a combination of test points.

4. The landing gear measurement point optimization method based on binary differential evolution algorithm according to claim 1, characterized in that: The fitness function is: , In the formula, Indicates the first i One indicator, Indicates the first i The weight of each indicator, and , All n dimension.

5. The landing gear measurement point optimization method based on binary differential evolution algorithm according to claim 4, characterized in that: The metrics include fault detection rate, fault isolation rate, and test point utilization rate, all derived from the fault-testability matrix.

6. The landing gear measurement point optimization method based on binary differential evolution algorithm according to claim 1, characterized in that: The mutation specifically includes the following steps: Step 1. Randomly select two chromosomes in the current generation of the population and perform an XOR operation on them. If the numbers are the same, the result is 0; if the numbers are different, the result is 1. Use the string of 0s and 1s obtained by the XOR operation as the overlay code. Step 2. Apply the above covering code to the chromosome with the highest fitness value in the current generation to perform a covering operation, generating a mutated chromosome: if the covering code is... i If a position is 1, the number corresponding to the chromosome with the highest fitness value in the current generation is overwritten; if the original position is 0, it becomes 1; if the original position is 1, it becomes 0. Step 3. Repeat steps 1 and 2 above, for a total of [number missing]. N Next, a mutant population was obtained.

7. The landing gear measurement point optimization method based on binary differential evolution algorithm according to claim 1, characterized in that: The crossover specifically refers to: generating a 0 / 1 mask of the same length as the chromosome for each chromosome in the current generation population; for the first 0 / 1 of the mask... i Each bit independently generates random numbers that follow a uniform distribution. R i , R i ∈[0,1]; based on preset crossover probability P c Generate experimental chromosomes: If R i ≤ P c When the mask's first i The position is 1, the experimental chromosome's... i The i-th position is taken from the corresponding mutated chromosome; otherwise, the i-th position of the chromosome is retained. The above crossover operation affects the population... N Each chromosome is executed separately, generating a total of N Experimental chromosomes were used to obtain offspring populations.

8. The landing gear measurement point optimization method based on binary differential evolution algorithm according to claim 7, characterized in that: It also includes dynamically adjusting the crossover probability. P c : , In the formula, , These are the crossover probability constants, For chromosome number i The fitness value of the position. The average fitness value for each generation. This represents the maximum value of chromosome fitness for that generation.

9. The landing gear measurement point optimization method based on binary differential evolution algorithm according to claim 1, characterized in that: The selection specifically refers to: merging the current generation population and the offspring population into a temporary population, calculating the fitness values ​​of all chromosomes in the temporary population, and selecting the value that is the highest among them. N Each chromosome enters the next round of iteration.

10. A landing gear measurement point optimization system based on binary differential evolution algorithm, characterized in that: include: The modeling and matrix generation module is used to build a landing gear multi-signal flow graph model and generate a fault-testability matrix based on the model; wherein, the fault-testability matrix is ​​a binary knowledge matrix, with 1 and 0 respectively indicating whether there is a dependency relationship between the operating mode and the test point in the system; The population initialization module is used to represent the combination of test points in binary encoding; and to randomly generate several chromosomes to obtain the initial population. The fitness evaluation module is used to calculate the fitness value of each chromosome in the current generation based on the failure-testability matrix and a pre-constructed fitness function. The differential evolution optimization module includes a mutation submodule, a crossover submodule, and a selection submodule. The mutation submodule randomly selects two different chromosomes, generates a covering code, and applies the covering code to the chromosome with the highest fitness value in the current generation to generate a mutated chromosome, thus obtaining a mutated population. The crossover submodule selects allotropes on chromosomes from the mutated population and the current generation population for crossover operations to generate experimental chromosomes, thus obtaining a progeny population. The selection submodule selects chromosomes from the current generation population and the progeny population to enter the next round of the cycle. The iteration control and output module is used to repeatedly call the fitness evaluation module and the differential evolution optimization module until the convergence condition is met or the maximum number of iterations is reached, and outputs the chromosome with the largest fitness value as the optimal test point combination.