Optimization method of horizontal deployment position of submerged buoys for island and reef detection based on genetic algorithm

Abstract

The present invention discloses a method for optimizing the horizontal placement of buoys for detecting island and reef areas based on a genetic algorithm. This method utilizes online open-source data to extract topographic data for the island and reef area. This topographic data is then divided into an N×N grid. The resulting sound field data is then converted from polar coordinates to a Cartesian coordinate system based on angle and distance, and the entire data is stored in a rectangular sound field environment. Subsequently, a genetic algorithm is used to select, overlay, and optimize the buoys. After overlaying the buoy combinations, the proportion of the detectable area, excluding the area occupied by the terrain, is calculated. Through sound field simulation and optimization of the horizontal placement of the buoys, the present invention can obtain the optimal configuration solution with a small number of input parameters, significantly reducing the need for manual calculations and experiments and improving the convenience and efficiency of operation.

Description

Technical Field

[0001] The present invention belongs to the technical field of optimization methods, and in particular relates to a method for optimizing the horizontal placement of a buoy for detecting island and reef areas based on a genetic algorithm. Background Art

[0002] With the development of marine resources and the increasing attention paid to the protection of marine rights and interests, marine exploration plays an indispensable role in seabed topography mapping and underwater target search. In complex island and reef areas, acoustic field detection faces huge challenges due to the existence of seabed topography such as slopes, mountains and trenches. In particular, the complex topography of the island and reef area greatly affects the propagation process of sound waves, including phenomena such as scattering, reflection, shielding and attenuation of sound waves. These factors make it difficult to effectively perform acoustic field detection using a single buoy. However, the effective deployment of multiple buoys is a technical challenge, and the position of each buoy needs to be optimized to maximize the acoustic field detection effect. Although the traditional brute force optimization method can find the optimal solution in theory, it will encounter huge challenges in computing resources in practical applications. The specific analysis is as follows:

[0003] (1) The impact of complex terrain on acoustic field detection. In existing technologies, the selection of buoy locations is usually based on experience or simplified environmental models, and the complexity of the actual marine environment is rarely taken into account. Especially in areas surrounded by islands and reefs, unreasonable buoy layout may lead to an increase in detection blind spots, and the effective detection range and accuracy cannot be guaranteed. For example, when the buoy is located above the shallow sea slope, the sound wave propagation loss received by the deep-sea sound channel axis is reduced; in a seabed environment with a certain slope, the sound line will be reflected by the inclined seabed interface, and its propagation direction will be horizontally deflected; the sound wave propagation loss after being blocked by the seamount increases significantly, and the reflection on the seamount slope surface enhances the sound wave energy; the sound waves emitted from the top of the slope propagate down the slope to the deep-sea sound channel axis and achieve long-distance propagation. The above laws all emphasize the important influence of seabed topography on underwater sound propagation. The undulating changes in these terrains, such as seamounts or trenches, have become the key to the study of acoustic field laws.

[0004] (2) Underwater buoy deployment position optimization method based on greedy algorithm. Since the computational complexity of directly searching for the global optimal solution from all possible combinations is as high as more than 10 million, the efficiency of solving the global optimal solution is too low. To solve this problem, a greedy algorithm is used to first perform local optimization to obtain the local optimal solution, and then use it as the heuristic input for global optimization, which greatly improves the computational efficiency. However, existing methods usually fail to fully consider the complementary effects of multiple buoys and the maximization of sound field coverage under different deployment schemes, lack adaptability and efficiency to specific terrain, which often leads to unsatisfactory results in practical applications. In addition, even if the greedy algorithm is used for optimization, the screening process still takes about five hours, and the efficiency still needs to be improved. Therefore, there is an urgent need for an efficient technology that can quickly and accurately select the horizontal position of the buoy.

[0005] Genetic algorithms are highly efficient search algorithms that, by simulating the processes of natural selection and genetics, can effectively find near-optimal solutions in complex search spaces, thereby reducing the required computing resources and accelerating the problem-solving process. This approach not only addresses the challenges posed by the complex terrain of island reef areas but also provides strong technical support for the further application of ocean acoustic field detection technology. Summary of the Invention

[0006] To overcome the shortcomings of the existing technology, the present invention provides a genetic algorithm-based method for optimizing the horizontal placement of buoys for detecting island and reef areas. This method utilizes online open-source data to extract topographic data of the island and reef area. This topographic data is then divided into an N×N grid. The resulting sound field data is then converted from polar coordinates to a Cartesian coordinate system based on angle and distance, and the entire data is stored in a rectangular sound field environment. A genetic algorithm is then used to select, overlay, and optimize the buoys. After overlaying the buoy combinations, the proportion of the detectable area, excluding the terrain, is calculated. Through sound field simulation and optimization of the horizontal placement of buoys, the present invention can obtain the optimal configuration with a small number of input parameters, significantly reducing the need for manual calculations and experiments and improving operational convenience and efficiency.

[0007] The technical solutions adopted by the present invention to solve the technical problems are as follows:

[0008] Step 1: Select the detection area, set the propagation loss threshold, receiving depth, receiving distance, and sound source frequency;

[0009] Step 2: Define the grid size as N×N, where N≥20; the x-axis and y-axis ranges from negative to positive values; each interval of the x-axis and y-axis is divided into N-1 equal parts, allowing equal detection opportunities within each grid cell; using the generated x-axis and y-axis coordinate points, form a two-dimensional coordinate grid; each coordinate point represents the center point of a cell in the grid, which is used to arrange potential buoys;

[0010] Step 3: Extract the horizontal position corresponding to each point in the grid from the terrain file, and find the corresponding terrain depth according to the horizontal position; the buoy depth is set to 100m higher than the terrain, and the environmental file under the terrain condition is bellhop3d to obtain the sound field data of the buoy at N×N different horizontal positions; for the superposition of N×N sound field data, it is necessary to unify the sound field data from polar coordinates to Cartesian coordinates as shown in formulas (1) and (2):

[0011] x=r·cos(th) (1)

[0012] y=r·sin(th) (2)

[0013] Where r is the acceptance distance, th is the acceptance angle, x is the horizontal coordinate in the Cartesian coordinate system, and y is the vertical coordinate in the Cartesian coordinate system;

[0014] Step 4: The nearest neighbor interpolation method is used to map the calculated sound field data onto the specified grid;

[0015] By finding the sound field points closest to the grid points, and directly assigning the sound pressure values ​​of these points to the grid points; in the process of mapping the sound field data to the grid, the Euclidean distance is used to determine the distance between each grid point and the recording point in the sound field as shown in formula (3):

[0016]

[0017] Where x1 and y1 represent the horizontal and vertical coordinates of the first point; x2 and y2 represent the horizontal and vertical coordinates of the second point; d represents the Euclidean distance between the two points;

[0018] By calculating the minimum distance, the nearest sound field point is selected to represent the sound pressure value of the grid point; the sound field part beyond the rectangular area is removed; finally, a rectangular three-dimensional sound field storing N×N sound field data is obtained;

[0019] Step 5: By simulating the distribution of the acoustic field at the set grid and depth and calculating the area that meets the threshold conditions, a genetic algorithm is used to find the optimal buoy placement position to maximize the acoustic field detection range, as follows:

[0020] Step 5-1: Initialize a set of random seed arrays; configure the genetic algorithm based on the number of potential targets, population size, number of iterations, crossover ratio, and mutation rate;

[0021] Step 5-2: In the main loop of the genetic algorithm, a population is initialized for each random seed, where each individual represents a potential buoy combination; the fitness of each generation is calculated based on the sound field coverage of each potential buoy combination; the sound field data of each potential buoy are superimposed into a three-dimensional array, and then the accumulated sound field energy is converted into decibel values ​​using formula (4); based on these decibel values, the size of the effective coverage area that does not exceed the set maximum threshold is calculated, and the coverage area is expressed as the number of grid cells, which provides an indicator to quantify the sound field performance:

[0022]

[0023] Among them, L p represents the sound pressure level in decibels; p represents the actual measured sound pressure in Pascals; P ref is the reference sound pressure;

[0024] Preferably, the reference sound pressure Pref Take it as 20 micropascals.

[0025] Preferably, the calculation method of the sound field coverage is as follows:

[0026] Considering the impact of the horizontal placement of buoys on the detection range, acoustic field data are obtained from the simulation, including the sound pressure level data of each buoy at different depths and positions. The sound field data in polar coordinates are converted into sound field data in a Cartesian coordinate system. The sound pressure level data are converted into decibel values ​​for easy analysis and comparison. The coverage area in the sound field that meets the specified decibel threshold, which defines the minimum acceptable hearing level, is calculated. The effectiveness of different buoy combination configuration methods is evaluated to guide the horizontal optimization layout of buoys.

[0027] Preferably, the method for configuring the potential marker combination is as follows:

[0028] Use genetic algorithms to automatically select the optimal horizontal position and depth of the buoy to maximize the sound field coverage in a specific area;

[0029] Initialize the population, where each individual represents a buoy configuration scheme, including the number, position and depth of the buoys;

[0030] Calculate the fitness of each configuration scheme based on the coverage of the sound field and the distribution of sound pressure levels;

[0031] The roulette wheel method is used to select excellent individuals from the population for reproduction, and crossover and mutation operations are used to generate a new population. The above steps are iteratively performed until the termination condition is met, and the optimal potential buoy configuration scheme is selected.

[0032] Preferably, in step 4, the nearest neighbor interpolation method is used to effectively map the sound field data onto a predefined grid, and the specific steps are as follows:

[0033] (1) Data correspondence: First, for each point in the predefined grid, the distance between the point and the recorded point in the sound field is calculated using the Euclidean distance formula (3);

[0034] (2) Closest point selection: Identify the closest sound field point to each grid point by comparing the calculated distances;

[0035] (3) Sound pressure value allocation: the sound pressure value of the nearest sound field point is directly allocated to the corresponding grid point;

[0036] (4) Data truncation processing: The sound field data outside the predefined rectangular area is removed to ensure the accuracy and practicality of the sound field model;

[0037] (5) Rectangular sound field storage: The final result is a rectangular three-dimensional sound field array containing N×N sound field data points. The data of each point comes directly from the actual sound field point closest to the grid point.

[0038] The beneficial effects of the present invention are as follows:

[0039] (1) The present invention uses a genetic algorithm to optimize the buoy deployment configuration, which can effectively increase the coverage of the sound field and ensure that the sound field is uniform and continuous in the target area, thereby improving the detection efficiency and accuracy of the acoustic equipment. The simulation time is significantly shorter than the brute force solution that calculates all possible buoy combinations, which is more efficient and time-saving.

[0040] (2) The method of the present invention can adjust the layout of the submerged buoy according to different marine environmental characteristics (such as water depth, bottom quality, and topography), making the acoustic field layout scheme highly adaptable to the environment and able to maintain good performance in a changing marine environment. This invention is not only applicable to island and reef areas, but can also be extended to other fields requiring acoustic field optimization, such as port monitoring and underwater pipeline inspection, and has broad application prospects.

[0041] (3) The present invention can obtain the optimal configuration scheme by simulating the sound field and optimizing the horizontal position of the buoy with a small amount of input parameters, which greatly reduces the need for manual calculations and experiments and improves the convenience and efficiency of operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 The simulated terrain area of ​​the present invention (the area with a distance from -100 km to 100 km from the islands and reefs of the Xisha Islands in the South China Sea of ​​my country);

[0043] Figure 2 This is a flow chart of the genetic algorithm structure of the present invention;

[0044] Figure 3 The simulation results of the present invention when the propagation loss threshold is 80dB, (a) the genetic algorithm is used to select the optimal buoy placement position for three buoys, (b) the 500-meter cross-sectional sound field diagram;

[0045] Figure 4 The simulation results of the present invention when the propagation loss threshold is 90dB, (a) the genetic algorithm is used to select the optimal buoy placement position for three buoys, (b) the 500-meter cross-sectional sound field diagram;

[0046] Figure 5 The simulation results of the present invention when the propagation loss threshold is 95dB, (a) the genetic algorithm is used to select the optimal buoy placement position for three buoys, (b) the 500-meter cross-sectional sound field diagram;

[0047] Figure 6These are the convergence diagrams of four seeds when the propagation loss threshold is 95 dB according to the present invention, where (a) the seed is 50, (b) the seed is 100, (c) the seed is 150, and (d) the seed is 200. DETAILED DESCRIPTION

[0048] The present invention will be further described below with reference to the accompanying drawings and examples.

[0049] The present invention proposes a method for optimizing the horizontal deployment position of buoys for detecting island and reef areas based on a genetic algorithm. The method can be applied to superimpose multiple sound field files at different horizontal positions into the same rectangular sound field environment by establishing a mapping relationship from polar coordinates to Cartesian coordinates and a nearest neighbor interpolation method.

[0050] The main process is divided into the following three steps:

[0051] Step 1: Select the Xisha Islands and reefs in the South China Sea (such as Figure 1 For example, set the propagation loss threshold to 90dB, the receiving depth to the part above 1000m, the receiving distance to 100km, and the sound source frequency to 400Hz.

[0052] Step 2: First, define the size of the entire grid as N×N (N≥20), with the range of the x-axis and y-axis covering from negative to positive values, ensuring the symmetry of the detection area so that all possible areas can be covered. Each interval of the x-axis and y-axis is divided equally. Divide the interval of each axis into N-1 equal parts (a total of N points, including the start and end points). This equally spaced division method simplifies the grid generation process and ensures a uniform distribution of the grid, allowing equal detection opportunities within each grid cell. The generated x-axis and y-axis coordinate points are used in combination to form a two-dimensional coordinate grid. Each coordinate point represents the center point of a cell in the grid and is used to arrange potential markers.

[0053] Step 3: Extract the horizontal position corresponding to each point in the grid from the terrain file and find the corresponding terrain depth based on this horizontal position. The buoy depth is set to 100m above the terrain. Perform a bellhop3d on this to obtain the acoustic field data of the buoy at N×N different horizontal positions. To superimpose the N×N acoustic field data, it is necessary to unify the acoustic field data from polar coordinates to a Cartesian coordinate system as shown in Formulas (1) and (2).

[0054] x=r·cos(th) (5)

[0055] y=r·sin(th) (6)

[0056] Step 4: The nearest neighbor interpolation method is used to map the calculated sound field data onto the specified grid. This method simplifies the complex interpolation calculation by finding the sound field points closest to the grid points and directly assigning the sound pressure values ​​of these points to the grid points. In the process of mapping the sound field data to the grid, the Euclidean distance is used to determine the distance between each grid point and the recording point in the sound field, as shown in Formula 3. By calculating the minimum distance, the nearest sound field point is selected to represent the sound pressure value of the grid point. In addition, the part of the sound field that exceeds the rectangular area is removed. Finally, a rectangular three-dimensional sound field storing N×N sound field data is obtained.

[0057]

[0058] Step 5: Genetic Algorithm is used to find the optimal buoy placement to maximize the range of acoustic field detection. This is done by simulating the distribution of the acoustic field at a certain grid and depth and calculating the area that meets the threshold condition.

[0059] The genetic algorithm process is as follows Figure 2 ;

[0060] First, to ensure algorithm reproducibility, a random seed array is initialized. Next, the genetic algorithm is configured based on the key parameters listed in Table 1, including the number of latent markers, population size, number of iterations, crossover ratio, and mutation rate. A large population and high crossover ratio are selected to maintain genetic diversity, while a high mutation rate increases the algorithm's randomness to prevent the solution from becoming trapped in a local optimum. Using multiple initial seeds helps improve global search efficiency.

[0061] In the main loop of the genetic algorithm, a population is initialized for each random seed, with each individual representing a potential beacon combination. Fitness is calculated for each generation based on the sound field coverage of each potential beacon combination. This process involves stacking the sound field data for each potential beacon into a three-dimensional array and then converting the accumulated sound field energy into decibel values ​​using Equation 4. Based on these decibel values, the effective coverage area is calculated, expressed as the number of grid cells, providing a metric for quantifying sound field performance.

[0062]

[0063] Among them, L p Represents the sound pressure level in decibels (dB); p represents the actual measured sound pressure in Pascals (Pa); reference sound pressure, P ref It is usually taken as 20 microPascals (20μPa), which is the minimum sound pressure that the human ear can perceive.

[0064] The calculation method for evaluating acoustic coverage involves taking into account the impact of the horizontal placement of buoys on the detection range. Acoustic field data, including sound pressure level data for each buoy at different depths and positions, is acquired from simulations. The polar coordinate sound field data is converted to Cartesian coordinates. The sound pressure level data is converted to decibel values ​​for easier analysis and comparison. The coverage area within the sound field that meets a specific decibel threshold, which defines the minimum acceptable hearing level, is calculated. This method is used to evaluate the effectiveness of different buoy configurations, thereby guiding the optimal horizontal placement of buoys.

[0065] A method for buoy configuration with sound field optimization includes the following steps: A genetic algorithm is used to automatically select the optimal horizontal position and depth of buoys to maximize sound field coverage within a specific area. A population is initialized, where each individual represents a buoy configuration solution, including the number, location, and depth of buoys. The fitness of each configuration solution is calculated based on the sound field coverage and sound pressure level distribution. A roulette wheel method is used to select outstanding individuals from the population for reproduction, and crossover and mutation operations are used to generate a new population to explore the solution space and increase population diversity. These steps are iterated until a termination condition is met (sufficient fitness), and the optimal buoy configuration solution is selected.

[0066] The nearest neighbor interpolation method is used to effectively map the sound field data onto a predefined grid. This process follows the following steps: (1) Data correspondence: First, for each point in the predefined grid, the Euclidean distance formula (Formula 3) is used to calculate the distance between the point and the recorded point in the sound field. (2) Nearest point selection: By comparing the calculated distances, the sound field point closest to each grid point is identified. This step omits the complex mathematical interpolation calculation and directly selects the nearest data point. (3) Sound pressure value assignment: The sound pressure value of the nearest sound field point is directly assigned to the corresponding grid point. This ensures that the spatial distribution of the sound field can be reflected concisely and intuitively on the grid model. (4) Data truncation processing: For sound field data outside the predefined rectangular area, it is eliminated to ensure the accuracy and practicality of the sound field model. (5) Rectangular sound field storage: Ultimately, this method produces a rectangular three-dimensional sound field array containing N×N sound field data points, and the data of each point is directly derived from the actual sound field point closest to the grid point. By adopting the nearest neighbor interpolation method, complex sound field data can be quickly and effectively converted into a grid model, while ensuring the accuracy of the data and the simplicity of processing, meeting the efficiency and practicality requirements of the sound field detection system design.

[0067] Example:

[0068] Figure 1 The terrain data for acoustic simulation is given, with a horizontal distance of 200km. Based on this island and reef terrain data, the acoustic field simulation steps are carried out.

[0069] Figure 3 The simulation results when N is 20 and the propagation loss threshold is 80dB are given to screen the three-buoy combination with the largest acceptable range within the sea depth of 0 to 1000m. Figure 3 (a) is the optimal placement position of three buoys selected by genetic algorithm. Figure 3 (b) is the sound field propagation loss diagram of the 500-meter cross section after the optimal buoy combination is superimposed. The optimal buoy combination deployment position (horizontal position x, y and depth z) is obtained:

[0070] Optimal buoy combination placement points: [142, 134, 230]

[0071] Buoy 1: x = -89.000000 km, y = -26.000000 km, z = 1848 m

[0072] Buoy 2: x = 36.000000 km, y = -36.000000 km, z = 1864 m

[0073] Buoy 3: x = -5.000000 km, y = 15.000000 km, z = 1773 m

[0074] Figure 4 The simulation results when N is 20 and the propagation loss threshold is 90dB are given to screen the three-buoy combination with the largest acceptable range within the sea depth of 0 to 1000m. Figure 4 (a) is the optimal placement position of three buoys selected by genetic algorithm. Figure 4 (b) is the sound field propagation loss diagram of the 500-meter cross section after the optimal buoy combination is superimposed. The optimal buoy combination deployment position (horizontal position x, y and depth z) is obtained:

[0075] Optimal buoy combination placement points: [238, 60, 106]

[0076] Buoy 1: x = 75.000000 km, y = 15.000000 km, z = 528 m

[0077] Buoy 2: x = 100.000000 km, y = -78.000000 km, z = 1602 m

[0078] Buoy 3: x = -47.000000 km, y = -47.000000 km, z = 1000 m

[0079] Figure 5 The simulation results are given when N is 20 and the propagation loss threshold is 95dB. Figure 5 (a) The genetic algorithm selects the optimal placement position of the three buoys and selects the three buoy combinations with the largest acceptable range within the sea depth of 0 to 1000m. Figure 5 (b) is the sound field propagation loss diagram of the 500-meter cross section after the optimal buoy combination is superimposed. The optimal buoy combination deployment position (horizontal position x, y and depth z) is obtained:

[0080] Optimal buoy combination placement points: [89, 140, 285]

[0081] Buoy 1: x = -15.000000km, y = -57.000000km, z = 1000m

[0082] Buoy 2: x = 100.000000 km, y = -36.000000 km, z = 927 m

[0083] Buoy 3: x = -57.000000 km, y = 47.000000 km, z = 1000 m

[0084] Figure 6 This is the convergence diagram of 4 seeds when the propagation loss threshold is 95dB. Figure 6 (a) The seed is 50, Figure 6 (b) The seed is 100, Figure 6 (c) The seed is 150, Figure 6 (d) The seed is 200.

[0085] Table 1 shows the parameter settings of the genetic algorithm. When the solution space is large, a larger population and a higher ratio of parents participating in crossover operations are selected to maintain genetic diversity. A higher mutation rate can improve randomness, and multiple initial seeds are added to better avoid the answers being concentrated in the local optimal area.

[0086] Table 2 shows the proportion of effective detection range when the propagation loss threshold is 80dB, 90dB and 95dB. As the propagation loss threshold increases, the proportion of effective detection range increases significantly. When the propagation loss threshold is 95dB, the convergence diagram of the genetic algorithm is as follows: Figure 6 shown.

[0087] Table 1 Genetic algorithm parameter settings

[0088]

[0089] Table 2 Effective signal reception ratio under different propagation loss thresholds

[0090]

[0091] The method proposed in the present invention uses about half an hour to find the optimal horizontal deployment position of the three-buoy combination in the example of the islands and reefs of the Xisha Islands in the South my country Sea, and can provide the propagation loss threshold with the best detection effect. This invention is not only applicable to island and reef areas, but can also be extended to other areas that require sound field optimization. By changing the acceptance depth range, the optimal deployment position of the three-buoy combination can be found for detecting different areas, and it has a wide range of applications.

Claims

1. A method for optimizing the horizontal placement of submerged buoys for island and reef area detection based on genetic algorithms, characterized in that: The steps include: Step 1: Select the detection area, set the propagation loss threshold, receiving depth, receiving distance, and sound source frequency; Step 2: Define the grid size as N×N, where N≥20; the x-axis and y-axis ranges from negative to positive values; each interval of the x-axis and y-axis is divided into N-1 equal parts, allowing equal detection opportunities within each grid cell; using the generated x-axis and y-axis coordinate points, form a two-dimensional coordinate grid; each coordinate point represents the center point of a cell in the grid, which is used to arrange potential buoys; Step 3: Extract the horizontal position corresponding to each point in the grid from the terrain file, and find the corresponding terrain depth according to the horizontal position; the buoy depth is set to 100m higher than the terrain, and the environmental file under the terrain condition is bellhop3d to obtain the sound field data of the buoy at N×N different horizontal positions; for the superposition of N×N sound field data, it is necessary to unify the sound field data from polar coordinates to Cartesian coordinates as shown in formulas (1) and (2): x=r·cos(th) (1) y=r·sin(th) (2) Where r is the acceptance distance, th is the acceptance angle, x is the horizontal coordinate in the Cartesian coordinate system, and y is the vertical coordinate in the Cartesian coordinate system; Step 4: The nearest neighbor interpolation method is used to map the calculated sound field data onto the specified grid; By finding the sound field points closest to the grid points, and directly assigning the sound pressure values ​​of these points to the grid points; in the process of mapping the sound field data to the grid, the Euclidean distance is used to determine the distance between each grid point and the recording point in the sound field as shown in formula (3): Where x1 and y1 represent the horizontal and vertical coordinates of the first point; x2 and y2 represent the horizontal and vertical coordinates of the second point; d represents the Euclidean distance between the two points; By calculating the minimum distance, the nearest sound field point is selected to represent the sound pressure value of the grid point; the sound field part beyond the rectangular area is removed; finally, a rectangular three-dimensional sound field storing N×N sound field data is obtained; Step 5: By simulating the distribution of the acoustic field at the set grid and depth and calculating the area that meets the threshold conditions, a genetic algorithm is used to find the optimal buoy placement position to maximize the acoustic field detection range, as follows: Step 5-1: Initialize a set of random seed arrays; configure the genetic algorithm based on the number of potential targets, population size, number of iterations, crossover ratio, and mutation rate; Step 5-2: In the main loop of the genetic algorithm, a population is initialized for each random seed, where each individual represents a potential buoy combination; the fitness of each generation is calculated based on the sound field coverage of each potential buoy combination; the sound field data of each potential buoy are superimposed into a three-dimensional array, and then the accumulated sound field energy is converted into decibel values ​​using formula (4); based on these decibel values, the size of the effective coverage area that does not exceed the set maximum threshold is calculated, and the coverage area is expressed as the number of grid cells, which provides an indicator to quantify the sound field performance: Among them, L p represents the sound pressure level in decibels; p represents the actual measured sound pressure in Pascals; P ref is the reference sound pressure.

2. The method for optimizing the horizontal placement of submerged buoys for island and reef area detection based on genetic algorithm according to claim 1, characterized in that: The reference sound pressure P ref Take it as 20 micropascals.

3. The method for optimizing the horizontal placement of submerged buoys for island and reef area detection based on genetic algorithm according to claim 1 is characterized in that: The calculation method of the sound field coverage is as follows: Considering the impact of the horizontal placement of buoys on the detection range, acoustic field data are obtained from the simulation, including the sound pressure level data of each buoy at different depths and positions. The sound field data in polar coordinates are converted into sound field data in a Cartesian coordinate system. The sound pressure level data are converted into decibel values ​​for easy analysis and comparison. The coverage area in the sound field that meets the specified decibel threshold, which defines the minimum acceptable hearing level, is calculated. The effectiveness of different buoy combination configuration methods is evaluated to guide the horizontal optimization layout of buoys.

4. The method for optimizing the horizontal placement of a submerged buoy for detecting islands and reefs based on a genetic algorithm according to claim 3 is characterized in that: The method for configuring the potential marker combination is as follows: Use genetic algorithms to automatically select the optimal horizontal position and depth of the buoy to maximize the sound field coverage in a specific area; Initialize the population, where each individual represents a buoy configuration scheme, including the number, position and depth of the buoys; Calculate the fitness of each configuration scheme based on the coverage of the sound field and the distribution of sound pressure levels; The roulette wheel method is used to select excellent individuals from the population for reproduction, and crossover and mutation operations are used to generate a new population. The above steps are iteratively performed until the termination condition is met, and the optimal potential buoy configuration scheme is selected.

5. The method for optimizing the horizontal placement of a submerged buoy for detecting islands and reefs based on a genetic algorithm according to claim 1, characterized in that: In step 4, the nearest neighbor interpolation method is used to effectively map the sound field data onto a predefined grid. The specific steps are as follows: (1) Data correspondence: First, for each point in the predefined grid, the distance between the point and the recorded point in the sound field is calculated using the Euclidean distance formula (3); (2) Closest point selection: Identify the closest sound field point to each grid point by comparing the calculated distances; (3) Sound pressure value allocation: the sound pressure value of the nearest sound field point is directly allocated to the corresponding grid point; (4) Data truncation processing: The sound field data outside the predefined rectangular area is removed to ensure the accuracy and practicality of the sound field model; (5) Rectangular sound field storage: The final result is a rectangular three-dimensional sound field array containing N×N sound field data points. The data of each point comes directly from the actual sound field point closest to the grid point.

Images