High-density-resistant penaeus monodon breeding method and system based on breeding pressure simulation

Abstract

The invention discloses a breeding pressure simulation-based high-density-resistant penaeus monodon breeding method and system. The method comprises the following steps: determining pressure influence factors of penaeus monodon breeding through data retrieval, and extracting key environment variables by using principal component analysis; setting a plurality of groups of breeding pressure gradients, constructing a plurality of test groups by adopting an orthogonal test method, and collecting environment and phenotype data to evaluate the breeding effect and form a test data set. And constructing a breeding analysis model, and analyzing breeding evolution characteristics by constructing a correlation path and a topological structure chart of the breeding pressure gradient, the breeding effect and the monitoring data. And analyzing breeding demand information, inputting the breeding demand information into the model to obtain candidate breeding schemes, performing breeding feature extraction and cost evaluation on the candidate schemes, and finally recommending an optimal breeding scheme. The overall breeding benefit and production safety are improved, and the breeding period and complexity are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of Penaeus monodon breeding, and particularly to a method and system for breeding high-density tolerant Penaeus monodon based on simulated culture stress. Background Art

[0002] With the continuous development of aquaculture, high-density culture technology has become an important way to improve the yield per unit area. As a highly economically valuable aquaculture species, although Penaeus monodon can achieve increased yields under high-density culture conditions, it also faces a series of challenges. The high-density environment easily leads to water quality deterioration, insufficient dissolved oxygen, accumulation of ammonia nitrogen and other harmful substances, resulting in problems such as slow growth, increased mortality, and disease outbreaks of Penaeus monodon. In addition, under high-density culture conditions, Penaeus monodon has a strong stress response, and its immunity and disease resistance often fail to meet the culture requirements, seriously affecting the culture efficiency and production safety.

[0003] Traditional breeding methods mostly rely on screening under conventional culture conditions and fail to fully simulate the complex and variable stress factors in the high-density culture environment, resulting in unstable performance of the selected varieties in the actual high-density environment. In recent years, with the development of bioinformatics, data analysis, and artificial intelligence technologies, the use of culture stress simulation technology to precisely control and reproduce the stress factors in the high-density environment has gradually become a hot topic of concern in the academic and industrial circles. By simulating various stress factors (such as dissolved oxygen, ammonia nitrogen, water temperature, density, pH value, etc.) in the high-density culture environment, the growth, disease resistance, and adaptability of Penaeus monodon can be systematically evaluated in the laboratory or semi-closed environment, so as to screen out excellent varieties with high-density tolerance and stronger adaptability. Therefore, it is urgent to solve the problems of long breeding cycle, low efficiency, and poor adaptability of traditional breeding methods, and improve the overall culture efficiency and production safety. Summary of the Invention

[0004] The present invention overcomes the defects of the prior art and provides a method and system for breeding high-density tolerant Penaeus monodon based on simulated culture stress, and its important purpose is to improve the overall culture efficiency and production safety, and reduce the breeding cycle and complexity.

[0005] To achieve the above object, the first aspect of the present invention provides a method for breeding high-density tolerant Penaeus monodon based on simulated culture stress, including:

[0006] Based on data retrieval, obtain the stress influencing factors of Penaeus monodon culture, analyze the obtained stress influencing factors by principal component analysis, and define the key environmental variables of culture stress;

[0007] Set multiple groups of aquaculture pressure gradients according to the key environmental variables of aquaculture pressure, establish multiple experimental groups based on the orthogonal test method for aquaculture experiments, and evaluate the aquaculture effects through the collected aquaculture data to obtain an experimental data set;

[0008] Construct a breeding analysis model, establish the association path categories of aquaculture pressure gradient - aquaculture effect - aquaculture monitoring data according to the experimental data set, analyze the aquaculture evolution characteristics of each association path category, and construct a topological structure diagram to train the breeding analysis model;

[0009] Obtain breeding requirement information, extract features from the breeding requirement information to obtain expected aquaculture features, generate breeding requirement analysis information according to the expected aquaculture features and input it into the breeding analysis model for analysis to obtain candidate breeding plans;

[0010] Extract breeding features for each candidate breeding plan, evaluate the breeding costs of each candidate breeding plan based on the extracted breeding features, recommend breeding plans according to the evaluation results, and generate recommended breeding plans.

[0011] In the present invention, the pressure influencing factors for the culture of Penaeus monodon are obtained based on data retrieval, and the obtained pressure influencing factors are analyzed by the principal component analysis method to define the key environmental variables of aquaculture pressure, specifically including:

[0012] Based on data retrieval, obtain the relevant pressure influencing factors for the culture of Penaeus monodon, and obtain the influence degrees of each pressure influencing factor on the culture of Penaeus monodon through the expert analysis method to form a first data set;

[0013] Adopt the parallel coordinate method to map each pressure influencing factor in the first data set to the parallel coordinate system according to the influence degree of each pressure influencing factor to obtain a pressure influencing factor coordinate map;

[0014] Each data axis in the pressure influencing factor coordinate map corresponds to the dimensional range and type of pressure influencing factors of this characteristic. Screen the initial pressure influencing factors through the pressure influencing factor coordinate map, and eliminate the types of pressure influencing factors with small correlation to obtain an initial pressure influencing factor data set;

[0015] Introduce the principal component analysis method to perform dimensionality reduction processing on the initial pressure influencing factor data set, calculate the principal component scores corresponding to each pressure influencing factor, judge the calculated principal component scores with a preset threshold, and select the pressure influencing factors corresponding to the principal component scores greater than the preset threshold as the principal component factors;

[0016] Perform principal component direction projection using the said principal component factors to obtain a scatter plot of pressure influencing factors, select corresponding pressure influencing factors according to a preset selection range in the scatter plot of pressure influencing factors, and define key environmental variables of aquaculture pressure based on the selected pressure influencing factors.

[0017] In the present invention, multiple groups of aquaculture pressure gradients are set according to the key environmental variables of aquaculture pressure, multiple experimental groups are established based on the orthogonal test method for aquaculture experiments, and the aquaculture effect is evaluated through the collected aquaculture data, which specifically includes:

[0018] Obtain the key environmental variables of aquaculture pressure, preset variable level grades and set multiple groups of aquaculture pressure gradients, introduce the orthogonal test method, and construct an orthogonal test matrix based on the set aquaculture pressure gradients and aquaculture pressure environmental variables according to a preset orthogonal table;

[0019] Allocate each row in the orthogonal test matrix to an independent experimental group, and allocate the environmental parameter benchmarks based on the corresponding aquaculture pressure gradients, and conduct aquaculture pressure simulation experiments according to the preset aquaculture cycle;

[0020] Monitor the aquaculture conditions of each independent experimental group during aquaculture, collect aquaculture environment data during the aquaculture cycle using a sensor array arranged in the aquaculture pond, and detect and analyze the aquaculture behavior data of Penaeus monodon in each experimental group by using a camera array through machine vision technology to obtain an aquaculture monitoring data set;

[0021] Preset an aquaculture effect evaluation system, extract the aquaculture monitoring data corresponding to the aquaculture effect evaluation indicators of each experimental group from the aquaculture monitoring data set through the aquaculture effect evaluation system for aquaculture effect evaluation, and obtain aquaculture effect evaluation information;

[0022] Associate the aquaculture effect evaluation information with the corresponding aquaculture experimental group in the aquaculture monitoring data set to construct an experimental data set.

[0023] In the present invention, the breeding analysis model is constructed. According to the experimental data set, the association path categories of aquaculture pressure gradient - aquaculture effect - aquaculture monitoring data are constructed, the aquaculture evolution characteristics of each association path category are analyzed, and a topological structure diagram is constructed to train the breeding analysis model, which specifically includes:

[0024] Obtain the experimental data set, extract the aquaculture pressure gradient characteristics of each experimental group from the experimental data set, and use the K - means clustering algorithm to classify the aquaculture monitoring data of each experimental group based on the aquaculture pressure gradient characteristics to obtain classification information;

[0025] Extract the aquaculture effect evaluation characteristics corresponding to each aquaculture monitoring data in each aquaculture pressure gradient category according to the category division information, and perform secondary category division to construct the association path of aquaculture pressure gradient - aquaculture effect - aquaculture monitoring data, so as to obtain the second data set;

[0026] Extract the characteristic data of different association path categories based on the second data set and perform time series processing to form the time series of each association path category, and import the time series of each association path category as input into the LSTM network for aquaculture evolution characteristic analysis;

[0027] Capture the time series dependence relationship and coupling effect of each feature in the time series through the set multi-layer LSTM layer. For each input sequence, take the hidden state at the last moment of the LSTM network as the corresponding feature representation, and map the hidden state to the target output space through the fully connected layer to obtain the aquaculture evolution characteristic analysis information;

[0028] The aquaculture evolution characteristic analysis information includes aquaculture environment evolution characteristics and aquaculture phenotype evolution characteristics. Take the aquaculture environment evolution characteristics as the first node and the aquaculture phenotype evolution characteristics as the second node to construct a directed description relationship;

[0029] Connect the first node and the second node according to the directed description relationship to construct a topological structure diagram, construct a breeding analysis model based on the graph neural network, and input the topological structure diagram into the breeding analysis model for encoding learning;

[0030] Obtain the original feature map of each graph node in the topological structure diagram through the multi-head attention mechanism, calculate the attention weight by obtaining the channel descriptor through average pooling and weight the original feature map to obtain the attention feature map, update the state of the hidden layer and adjust and optimize the model parameters to obtain the desired breeding analysis model.

[0031] In the present invention, the obtaining of the breeding demand information, the feature extraction of the breeding demand information to obtain the desired aquaculture characteristics, and the generation of the breeding demand analysis information according to the desired aquaculture characteristics and inputting the breeding demand analysis information into the breeding analysis model for analysis to obtain the candidate breeding plan specifically include:

[0032] Obtain the breeding demand information, perform feature extraction on the breeding demand information to obtain the desired aquaculture area characteristics, desired aquaculture density characteristics and desired aquaculture trait characteristics of the current high-density tolerant Penaeus monodon, and obtain the breeding demand analysis information;

[0033] Input the breeding demand analysis information into the trained breeding analysis model to generate a target node, perform first-order neighborhood sampling according to the target node to obtain a number of neighbor nodes, and calculate the cosine similarity value between the target node and the neighbor nodes;

[0034] Sort each neighbor node according to the calculated cosine similarity value, select the neighbor node with the largest similarity as the target neighbor node, and obtain the breeding evolution characteristics corresponding to the target neighbor node to form the feature vector of the target neighbor node;

[0035] Fuse the feature vector of the target neighbor node with the target node to update the feature vector of the target node, and generate the next target node based on the updated feature vector of the target node, and perform iterative update until the preset number of times is reached to obtain the final target node;

[0036] Extract the feature vector of the final target node and use it as the final output feature to generate candidate breeding plans, and obtain a number of candidate breeding plans.

[0037] In the present invention, extracting breeding characteristics for each candidate breeding plan, evaluating the breeding cost of each candidate breeding plan based on the extracted breeding characteristics, and recommending a breeding plan according to the evaluation result to generate a recommended breeding plan, specifically including:

[0038] Use the random forest algorithm to construct a breeding plan analysis model, obtain the topological structure diagram, construct the adjacency matrix of different associated path categories based on the topological structure diagram, and train the breeding plan analysis model;

[0039] Obtain candidate breeding plans, input each candidate breeding plan into the trained breeding plan analysis model for analysis, analyze the breeding environment evolution characteristics and breeding phenotype characteristics of each candidate breeding plan, and obtain candidate breeding plan analysis information;

[0040] Extract the characteristics of each candidate breeding plan to obtain the breeding cycle characteristics and breeding required equipment characteristics of each candidate breeding plan, and generate the first characteristic information in combination with the candidate breeding plan analysis information;

[0041] Introduce the improved NSGA-II algorithm to evaluate the breeding cost of each candidate breeding plan, preset the objective function and establish the constraint conditions, generate the initial population based on the first characteristic information, and calculate the objective function value for each individual in the population;

[0042] Sort all individuals in the population according to the objective function value, divide them into several Pareto fronts, calculate the crowding distance for individuals within each non-dominated layer, use the tournament selection method to select elite individuals using the non-dominated sorting level and crowding distance, and perform crossover and mutation operations on the selected elite individuals;

[0043] Merge the parent generation and the offspring, re-perform non-dominated sorting and crowding distance calculation on all candidate plans, and select a preset number of elite individuals according to the Pareto front level and crowding distance to form the next generation population for iterative evolution;

[0044] When the change of the objective function of the population meets the convergence criterion or the preset number of iterations, the breeding cost evaluation results of each candidate breeding plan are output, and based on the breeding cost evaluation results, the final breeding plan is recommended to obtain the recommended breeding plan.

[0045] The second aspect of the present invention provides a high-density tolerant Penaeus monodon breeding system based on aquaculture stress simulation. The system includes: a memory and a processor. The memory contains a high-density tolerant Penaeus monodon breeding method program based on aquaculture stress simulation. When the high-density tolerant Penaeus monodon breeding method program is executed by the processor, the following steps are implemented:

[0046] Based on data retrieval, obtain the stress influencing factors of Penaeus monodon aquaculture, use the principal component analysis method to analyze the obtained stress influencing factors, and define the key environmental variables of aquaculture stress;

[0047] Set multiple groups of aquaculture stress gradients according to the key environmental variables of aquaculture stress, establish multiple experimental groups based on the orthogonal test method for aquaculture experiments, and evaluate the aquaculture effects through the collected aquaculture data to obtain an experimental data set;

[0048] Construct a breeding analysis model, construct the association path categories of aquaculture stress gradient - aquaculture effect - aquaculture monitoring data according to the experimental data set, analyze the aquaculture evolution characteristics of each association path category, and construct a topological structure diagram to train the breeding analysis model;

[0049] Obtain breeding requirement information, extract the characteristics of the breeding requirement information to obtain the expected aquaculture characteristics, generate breeding requirement analysis information according to the expected aquaculture characteristics, and input it into the breeding analysis model for analysis to obtain candidate breeding plans;

[0050] Extract the breeding characteristics of each candidate breeding plan, evaluate the breeding cost of each candidate breeding plan based on the extracted breeding characteristics, and recommend the breeding plan according to the evaluation results to generate a recommended breeding plan.

[0051] The present invention discloses a method and system for selecting high-density tolerant Penaeus monodon based on aquaculture stress simulation, including: determining the stress influencing factors of Penaeus monodon aquaculture through data retrieval, and extracting key environmental variables using principal component analysis. Setting multiple groups of aquaculture stress gradients, constructing multiple experimental groups using the orthogonal experimental method, collecting environmental and phenotypic data for aquaculture effect evaluation and forming an experimental data set. Constructing a breeding analysis model, analyzing the characteristics of aquaculture evolution by constructing the correlation path and topological structure diagram of aquaculture stress gradient, aquaculture effect and monitoring data. Analyzing the breeding requirement information and inputting it into the model to obtain a candidate breeding plan, and then extracting breeding characteristics and cost evaluation for the candidate plan, and finally recommending the optimal breeding plan. Improve the overall aquaculture efficiency and production safety, and reduce the breeding cycle and complexity. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments or exemplifications of the present invention, the following will briefly introduce the drawings required for use in the embodiments or exemplifications. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the drawings shown.

[0053] Figure 1 It is a flowchart of a method for selecting high-density tolerant Penaeus monodon based on aquaculture stress simulation provided by an embodiment of the present invention;

[0054] Figure 2 It is a flowchart for analyzing a breeding plan for high-density tolerant Penaeus monodon provided by an embodiment of the present invention;

[0055] Figure 3 It is a block diagram of a system for selecting high-density tolerant Penaeus monodon based on aquaculture stress simulation provided by an embodiment of the present invention;

[0056] The implementation, functional characteristics and advantages of the object of the present invention will be further described in conjunction with the embodiments and with reference to the drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] In order to be able to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described in detail below with reference to the drawings and specific embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0058] Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0059] Figure 1Flowchart of a method for selecting high-density tolerant Penaeus monodon based on simulated culture pressure provided by an embodiment of the present invention;

[0060] As Figure 1 shown, the present invention provides a flowchart of a method for selecting high-density tolerant Penaeus monodon based on simulated culture pressure, including:

[0061] S102, obtaining the stress influencing factors of Penaeus monodon culture based on data retrieval, analyzing the obtained stress influencing factors by principal component analysis, and defining the key environmental variables of culture stress;

[0062] S104, setting multiple groups of culture pressure gradients according to the key environmental variables of culture stress, establishing multiple experimental groups based on the orthogonal experimental method for culture experiments, and evaluating the culture effects through the collected culture data to obtain an experimental data set;

[0063] S106, constructing a selection analysis model, constructing the association path categories of culture pressure gradient - culture effect - culture monitoring data according to the experimental data set, analyzing the culture evolution characteristics of each association path category, and constructing a topological structure diagram to train the selection analysis model;

[0064] S108, obtaining the selection requirement information, extracting the characteristics of the selection requirement information to obtain the expected culture characteristics, generating selection requirement analysis information according to the expected culture characteristics and inputting it into the selection analysis model for analysis to obtain candidate selection schemes;

[0065] S110, extracting the selection characteristics of each candidate selection scheme, evaluating the selection costs of each candidate selection scheme based on the extracted selection characteristics, and recommending selection schemes according to the evaluation results to generate recommended selection schemes.

[0066] Further, in a preferred embodiment of the present invention, the obtaining the stress influencing factors of Penaeus monodon culture based on data retrieval, analyzing the obtained stress influencing factors by principal component analysis, and defining the key environmental variables of culture stress specifically includes:

[0067] Obtaining the relevant stress influencing factors of Penaeus monodon culture based on data retrieval, and obtaining the influence degree of each stress influencing factor on Penaeus monodon culture through expert analysis method to form a first data set;

[0068] Using the parallel coordinate method to map each stress influencing factor in the first data set to the parallel coordinate system according to the influence degree of each stress influencing factor to obtain a stress influencing factor coordinate diagram;

[0069] In the coordinate diagram of the pressure influence factors, each data axis corresponds to the dimensional range of the characteristics and the type of pressure influence factors. The initial pressure influence factors are screened through the coordinate diagram of the pressure influence factors, and the types of pressure influence factors with low correlation are eliminated to obtain the initial pressure influence factor data set;

[0070] The principal component analysis method is introduced to perform dimensionality reduction processing on the initial pressure influence factor data set, calculate the principal component scores corresponding to each pressure influence factor, and judge the calculated principal component scores with a preset threshold. The pressure influence factors corresponding to the principal component scores greater than the preset threshold are selected as the principal component factors;

[0071] Use the principal component factors for principal component direction projection to obtain a scatter diagram of pressure influence factors, and select the corresponding pressure influence factors according to the preset selection range in the scatter diagram of pressure influence factors, and define the key environmental variables of aquaculture pressure based on the selected pressure influence factors.

[0072] It should be noted that various pressure factors that may affect the Penaeus monodon farming process are obtained through data retrieval. At the same time, combined with the expert analysis method, the specific influence degree of each pressure influence factor on the Penaeus monodon farming is quantitatively evaluated to form a first data set including each factor and its influence degree. Then, using the parallel coordinate method, each pressure influence factor in this first data set is mapped into a parallel coordinate system according to its influence degree to generate a coordinate diagram of pressure influence factors, where each data axis represents the performance of this characteristic within a certain dimensional range and its corresponding pressure factor type. Thus, it is possible to intuitively screen out the pressure influence factors with relatively high initial correlation, eliminate those with lower correlation, and obtain the initial pressure influence factor data set. Subsequently, the principal component analysis method is introduced to perform dimensionality reduction processing on this initial data set. By calculating the principal component scores of each pressure influence factor and comparing these scores with a preset threshold, the pressure factors with scores higher than the threshold are selected as the main influence factors. Using the selected principal component factors, projection in the principal component direction is performed again to generate a scatter diagram of pressure influence factors, and the pressure factors that meet the requirements are further screened out in the scatter diagram according to the preset selection range, including water body space density, water body vibration frequency, dissolved oxygen gradient, ammonia nitrogen concentration, salinity fluctuation, etc. Finally, the key environmental variables of aquaculture pressure are defined, providing a scientific and effective environmental parameter basis for the subsequent high-density Penaeus monodon breeding.

[0073] Further, in a preferred embodiment of the present invention, multiple groups of aquaculture pressure gradients are set according to the key environmental variables of aquaculture pressure, multiple experimental groups are established based on the orthogonal test method for aquaculture experiments, and the aquaculture effect evaluation is carried out through the collected aquaculture data, specifically including:

[0074] Obtain the key environmental variables of aquaculture stress, preset the variable level grades, and set multiple groups of aquaculture stress gradients. Introduce the orthogonal experiment method to construct an orthogonal experiment matrix based on the preset orthogonal table according to the set aquaculture stress gradients and aquaculture stress environmental variables;

[0075] According to the orthogonal experiment matrix, allocate each row to an independent experimental group, and allocate the environmental parameter benchmarks based on the corresponding aquaculture stress gradients, and conduct aquaculture stress simulation experiments on the aquaculture cycle according to the preset;

[0076] During the aquaculture period, monitor the aquaculture status of each independent experimental group, collect the aquaculture environment data during the aquaculture cycle by using the sensor array arranged in the aquaculture pond, and use the camera array to detect and analyze the aquaculture behavior data of the Penaeus monodon in each experimental group through machine vision technology to obtain the aquaculture monitoring data set;

[0077] Preset an aquaculture effect evaluation system, and through the aquaculture effect evaluation system, extract the aquaculture monitoring data corresponding to the aquaculture effect evaluation indicators in each experimental group from the aquaculture monitoring data set to conduct aquaculture effect evaluation, and obtain aquaculture effect evaluation information;

[0078] Associate the aquaculture effect evaluation information with the corresponding aquaculture experimental group in the aquaculture monitoring data set to construct an experimental data set.

[0079] It should be noted that through systematic data retrieval and expert analysis, key environmental variables affecting the farming pressure of Penaeus monodon are identified and obtained, such as water temperature, dissolved oxygen, salinity, ammonia nitrogen concentration, farming density, light intensity, water flow rate, and bait supply. Different levels are set to form different farming pressure gradients in the experiment. The orthogonal experiment method is used to construct the experimental design, and the preset orthogonal table is used to reasonably arrange the variable combinations to ensure that the experiment covers as many factor interactions as possible, while reducing the number of experiments and improving the experimental efficiency. The finally formed orthogonal experiment matrix, each row represents a specific combination of environmental variables, that is, an independent experimental group. In the experimental implementation stage, each experimental group is allocated according to the combination of environmental variables set by the orthogonal experiment matrix, and the corresponding farming pressure gradient is set to ensure the consistency of the environmental parameter benchmark. During the farming cycle, through standardized environmental control means, such as a precise water quality management system and an automatic feeding system, it is ensured that the experimental conditions meet the set requirements, and farming pressure simulation experiments are carried out under different pressure gradients to observe the survival and growth status of Penaeus monodon under different environmental conditions. During the experiment, a sensor array arranged in the farming pond is used to collect real-time farming environment data, including key water quality parameters such as water temperature, pH value, and dissolved oxygen content, to form high-precision environmental monitoring data. At the same time, high-definition cameras and machine vision technology are used to analyze the behavioral characteristics of Penaeus monodon in each experimental group, such as feeding frequency, swimming pattern, aggregation behavior, and abnormal behavior, to ensure the comprehensiveness and accuracy of the data. After these data are preprocessed and stored, a complete farming monitoring data set is formed. After the experiment is completed, based on the preset farming effect evaluation system, in-depth analysis is carried out on the farming monitoring data set. By extracting core evaluation indicators such as growth rate, survival rate, bait utilization rate, and disease resistance ability of each experimental group during the farming cycle, the farming effect of each experimental group is evaluated. Finally, the farming effect evaluation information is associated and integrated with the corresponding farming monitoring data set to form an experimental data set. This data set not only contains the combinations of various farming environmental variables and their corresponding pressure gradients, but also covers all biological response data generated during the farming process and the final farming effect evaluation data, providing a scientific basis for subsequent high-density tolerance breeding analysis and laying a data foundation for the establishment of an intelligent farming decision-making system.

[0080] Furthermore, in a preferred embodiment of the present invention, for constructing the breeding analysis model, an association path category of farming pressure gradient - farming effect - farming monitoring data is constructed according to the experimental data set, the farming evolution characteristics of each association path category are analyzed, and a topological structure diagram is constructed to train the breeding analysis model, specifically including:

[0081] Obtain a test data set, extract the aquaculture pressure gradient characteristics of each test group from the test data set, and use the K-means clustering algorithm to classify the aquaculture monitoring data of each test group based on the aquaculture pressure gradient characteristics to obtain classification information;

[0082] Extract the aquaculture effect evaluation characteristics corresponding to each aquaculture monitoring data in each aquaculture pressure gradient category according to the classification information, and perform secondary classification to construct an association path of aquaculture pressure gradient - aquaculture effect - aquaculture monitoring data to obtain a second data set;

[0083] Extract the characteristic data of different association path categories based on the second data set and perform time series processing to form time series sequences of each association path category, and import the time series sequences of each association path category as inputs into the LSTM network for aquaculture evolution characteristic analysis;

[0084] Capture the time series dependence relationship and coupling effect of each feature in the time series sequence through the set multi-layer LSTM layer. For each input sequence, take the hidden state at the last moment of the LSTM network as the corresponding feature representation, and map the hidden state to the target output space through the fully connected layer to obtain aquaculture evolution characteristic analysis information;

[0085] The aquaculture evolution characteristic analysis information includes aquaculture environment evolution characteristics and aquaculture phenotype evolution characteristics. Take the aquaculture environment evolution characteristics as the first node and the aquaculture phenotype evolution characteristics as the second node to construct a directed description relationship;

[0086] Connect the first node and the second node according to the directed description relationship to construct a topological structure diagram, construct a breeding analysis model based on the graph neural network, and input the topological structure diagram into the breeding analysis model for encoding learning;

[0087] Obtain the original feature map of each graph node in the topological structure diagram through the multi-head attention mechanism, calculate the attention weight by obtaining the channel descriptor through average pooling and weight the original feature map to obtain the attention feature map, update the state of the hidden layer and adjust and optimize the model parameters to obtain the desired breeding analysis model.

[0088] It should be noted that, first of all, based on the experimental data set obtained from the aquaculture experiment, the aquaculture pressure gradient characteristics of each experimental group are extracted. These characteristics include the changes in environmental variables such as water temperature, dissolved oxygen, pH, aquaculture density, etc. The K-means clustering algorithm is used to classify the aquaculture monitoring data of the experimental groups, and then the category division information is obtained. Through the category division, the experimental groups under different aquaculture conditions can be initially classified. The characteristic data related to the aquaculture effect evaluation are extracted from each aquaculture pressure gradient category, and a secondary category division is carried out. An association path among the aquaculture pressure gradient, aquaculture effect, and aquaculture monitoring data is constructed to form a more refined data set, that is, the second data set. Through this data set, the changes in key indicators such as the growth performance, survival rate, and feed utilization efficiency of Penaeus monodon under different aquaculture pressure environments can be clarified. Subsequently, the characteristic data of each associated path category are arranged in chronological order to form a time series, and these time series are used as inputs and imported into the LSTM (Long Short-Term Memory) network for aquaculture evolution characteristic analysis. In the LSTM network, by setting multiple LSTM layers, the time-dependent relationships of each characteristic in the time series and the coupling effects between different variables are captured. During the calculation process of the LSTM network, for each input time series, the hidden state at the last moment is taken as the characteristic representation of the series, and the hidden state is mapped to the target output space through a fully connected layer, so as to obtain the aquaculture evolution characteristic analysis information. The aquaculture evolution characteristic analysis information mainly includes aquaculture environment evolution characteristics and aquaculture phenotype evolution characteristics. The aquaculture environment evolution characteristics mainly describe the change trends of environmental variables such as water quality, temperature, and aquaculture density during the aquaculture cycle, while the aquaculture phenotype evolution characteristics reflect the changes in the growth state, survival rate, behavior pattern, etc. of Penaeus monodon over time. The aquaculture environment evolution characteristics are defined as the first node, and the aquaculture phenotype evolution characteristics are defined as the second node. Based on the mutual influence relationship between the two, a directed description relationship is constructed to reflect how environmental variables affect the dynamic changes of aquaculture phenotypes. On this basis, according to the constructed directed description relationship, the first node and the second node are connected to construct a complete topological structure diagram. This topological structure diagram can intuitively describe the relationship between different environmental factors and aquaculture phenotype characteristics, providing a data basis for further breeding analysis. Finally, a breeding analysis model is established using a graph neural network (GNN), and the topological structure diagram is input into this model for encoding learning, and finally a trained breeding analysis model is obtained.

[0089] Furthermore, in a preferred embodiment of the present invention, the obtaining of the breeding requirement information, the extraction of characteristics from the breeding requirement information to obtain the expected aquaculture characteristics, and the generation of breeding requirement analysis information according to the expected aquaculture characteristics and inputting the breeding requirement analysis information into the breeding analysis model for analysis to obtain a candidate breeding plan specifically include:

[0090] Obtain the breeding requirement information, extract the features of the breeding requirement information to obtain the expected aquaculture area characteristics, expected aquaculture density characteristics, and expected aquaculture trait characteristics of the current high-density tolerant Penaeus monodon, and obtain the breeding requirement analysis information;

[0091] Input the breeding requirement analysis information into the trained breeding analysis model to generate a target node, perform first-order neighborhood sampling according to the target node to obtain a number of neighbor nodes, and calculate the cosine similarity value with the target checkpoint;

[0092] Sort each neighbor node according to the calculated cosine similarity value, select the neighbor node with the largest similarity as the target neighbor node, and obtain the aquaculture evolution characteristics corresponding to the target neighbor node to form the feature vector of the target neighbor node;

[0093] Fuse the feature vector of the target neighbor node with the target node to update the feature vector of the target node, and generate the next target node based on the updated feature vector of the target node, and perform iterative update until the preset number of times is reached to obtain the final target node;

[0094] Extract the feature vector of the final target node and use it as the final output feature to generate candidate breeding plans, and obtain a number of candidate breeding plans.

[0095] It should be noted that, first, breeding requirement information is obtained. For the breeding goal of high-density tolerant Penaeus monodon, key requirement characteristics are extracted, including expected aquaculture area characteristics, aquaculture density characteristics, and aquaculture trait characteristics. The expected aquaculture area characteristics involve environmental parameters such as water temperature, dissolved oxygen, and salinity. The expected aquaculture density characteristics mainly reflect the suitable stocking density and its impact on growth and survival rate. The expected aquaculture trait characteristics cover key indicators such as disease resistance, growth rate, and body shape. Subsequently, the breeding requirement analysis information is input into the trained breeding analysis model to generate a target node. The target node represents the position of the current breeding requirement in the topological structure and is used to further analyze the aquaculture evolution characteristics under similar aquaculture conditions. For this purpose, first-order neighborhood sampling is performed around the target node in the topological structure diagram, that is, several neighbor nodes directly connected to the target node are obtained, and the cosine similarity between each neighbor node and the target node is calculated. The cosine similarity is used to measure the similarity degree between the feature vectors of two nodes and can effectively characterize the relationship between different aquaculture environments and aquaculture characteristics. After the calculation is completed, all neighbor nodes are sorted according to the cosine similarity value, and the neighbor node with the highest similarity is selected as the target neighbor node. The feature vector of the target neighbor node represents the aquaculture evolution characteristics that are closest to the current breeding requirement in the historical aquaculture data. Therefore, by obtaining the aquaculture evolution characteristics of the target neighbor node, the optimal breeding plan under the current breeding requirement can be further inferred. Next, the feature vector of the target neighbor node is fused with the feature vector of the target node to update the feature vector of the target node to make it more in line with the actual aquaculture requirements. Based on the updated feature vector of the target node, a new target node is generated, and this process is repeated. The features of the target node are continuously optimized through iterative updates until the preset number of iterations is reached, so as to obtain the final target node. By analyzing the feature vector of the final target node, several candidate breeding plans that meet the requirements of high-density tolerant aquaculture are constructed.

[0096] Furthermore, in a preferred embodiment of the present invention, breeding characteristics are extracted from each candidate breeding plan, breeding cost evaluation is performed on each candidate breeding plan based on the extracted breeding characteristics, and breeding plan recommendation is performed according to the evaluation results to generate a recommended breeding plan, which specifically includes:

[0097] A breeding plan analysis model is constructed using the random forest algorithm to obtain a topological structure diagram. An adjacency matrix of different associated path categories is constructed based on the topological structure diagram, and the breeding plan analysis model is trained.

[0098] Candidate breeding plans are obtained, and each candidate breeding plan is respectively input into the trained breeding plan analysis model for analysis. The aquaculture environment evolution characteristics and aquaculture phenotype characteristics of each candidate breeding plan are analyzed to obtain candidate breeding plan analysis information.

[0099] Feature extraction is performed on each candidate breeding plan to obtain the breeding cycle features and breeding requirement equipment features of each candidate breeding plan, and the first feature information is generated by combining the analysis information of the candidate breeding plan;

[0100] The improved NSGA-II algorithm is introduced to evaluate the breeding cost of each candidate breeding plan. The objective function is preset and the constraint conditions are established. The initial population is generated based on the first feature information, and the objective function value of each individual in the population is calculated;

[0101] All individuals in the population are sorted non-dominated according to the objective function value, divided into several Pareto fronts. The crowding distance of the individuals in each non-dominated layer is calculated. The tournament selection method is used to select elite individuals using the non-dominated sorting level and the crowding distance, and the selected elite individuals are subjected to crossover and mutation operations;

[0102] The parent generation and the offspring are combined, and all candidate solutions are re-sorted non-dominated and the crowding distance is calculated again. The preset number of elite individuals are selected according to the Pareto front level and the crowding distance to form the next generation population for iterative evolution;

[0103] When the change of the objective function of the population meets the convergence criterion or the preset number of iterations, the breeding cost evaluation results of each candidate breeding plan are output, and the final breeding plan is recommended based on the breeding cost evaluation results to obtain the recommended breeding plan.

[0104] It should be noted that when using the random forest algorithm to construct a breeding plan analysis model, after training is completed, each candidate breeding plan is input into the breeding plan analysis model. The model analyzes the evolution characteristics of the aquaculture environment and the aquaculture phenotypic characteristics, extracts key parameters, and generates candidate breeding plan analysis information. Subsequently, further feature extraction is performed on each candidate breeding plan to obtain breeding cycle characteristics and breeding requirement equipment characteristics. The extracted feature information is integrated with the candidate breeding plan analysis information to generate first feature information, which serves as the input data for breeding cost evaluation. Next, an improved NSGA-II algorithm is introduced to evaluate the breeding costs of each candidate breeding plan. First, the objective function is preset and corresponding constraint conditions are established to ensure that the evaluation results can truly reflect the rationality of the breeding costs. The objective function consists of economic cost, risk cost, and benefit cost. Based on the first feature information, an initial population is generated, and the objective function value is calculated for each individual in the population. Then, all individuals are non-dominated sorted according to the objective function value and divided into different Pareto fronts. To further optimize individual selection, the crowding distance is calculated for each individual within each non-dominated layer, and the tournament selection method is used to screen elite individuals according to the non-dominated sorting level and the crowding distance. Crossover and mutation operations are performed on the selected elite individuals to increase the diversity and optimization ability of the population. Subsequently, the parent individuals and the offspring individuals are combined, and non-dominated sorting and crowding distance calculation are performed again. Based on the Pareto front level and the crowding distance, a preset number of elite individuals are selected to form the next generation population and iterative evolution is carried out. As the iteration progresses, the breeding plan converges continuously in terms of fitness and cost optimization. Finally, when the change in the objective function of the population meets the convergence criterion or reaches the preset number of iterations, the breeding cost evaluation results of each candidate breeding plan are output. Based on the breeding cost evaluation results, all candidate breeding plans are comprehensively compared, and the optimal breeding plan is selected as the final recommended plan. While meeting the aquaculture objectives, a better balance is achieved in terms of breeding cycle, equipment requirements, environmental adaptability, etc., providing a scientific basis and technical support for the efficient breeding of high-density tolerant Penaeus monodon.

[0105] Figure 2 This is a flowchart for analyzing a breeding plan for high-density tolerant Penaeus monodon provided by an embodiment of the present invention;

[0106] As Figure 2 shown, the present invention provides a flowchart for analyzing a breeding plan for high-density tolerant Penaeus monodon, including:

[0107] S202, obtaining candidate breeding plans, and respectively inputting each candidate breeding plan into the breeding plan analysis model after training for analysis, analyzing the evolution characteristics of the aquaculture environment and the aquaculture phenotypic characteristics of each candidate breeding plan, and obtaining candidate breeding plan analysis information;

[0108] S204. Extract features from each candidate breeding plan to obtain the breeding cycle features and breeding requirement equipment features of each candidate breeding plan, and generate first feature information by combining the analysis information of the candidate breeding plan;

[0109] S206. Preset an objective function and establish constraint conditions, and evaluate the breeding cost of each candidate breeding plan through an improved NSGA-II algorithm to obtain the breeding cost evaluation results of each candidate breeding plan;

[0110] S208. Recommend a final breeding plan based on the breeding cost evaluation results of each candidate breeding method to obtain a recommended breeding plan.

[0111] It should be noted that the breeding cost evaluation objectively compares different candidate breeding plans through scientific and reasonable quantification criteria to avoid biases caused by subjective judgments. In the breeding process of high-density tolerant Penaeus monodon, different plans may involve different breeding environments, equipment investments, breeding cycles, and growth performances. Through a systematic evaluation method, the differences in resource consumption, environmental adaptability, and breeding efficiency of each plan can be accurately measured, providing precise data support for breeding decisions. And it helps breeding enterprises and research institutions reduce breeding costs and improve resource utilization rates. Since the breeding process involves a large number of experiments and long-term monitoring, if the plan is blindly selected, it may lead to resource waste and extended experimental cycles. Through the breeding cost evaluation, more economical and feasible plans can be screened out at an early stage to avoid ineffective investment. In addition, the breeding cost evaluation can also improve the stability and predictability of target traits by optimizing the population genetic improvement path, thereby increasing the breeding yield and quality.

[0112] Figure 3 A high-density tolerant Penaeus monodon breeding system 3 based on simulated culture stress provided by an embodiment of the present invention, the system includes: a memory 31 and a processor 32. The memory 31 contains a program for the high-density tolerant Penaeus monodon breeding method based on simulated culture stress. When the program for the high-density tolerant Penaeus monodon breeding method based on simulated culture stress is executed by the processor 32, the following steps are implemented:

[0113] Obtain the stress influencing factors of Penaeus monodon breeding based on data retrieval, analyze the obtained stress influencing factors by using the principal component analysis method, and define the key environmental variables of the culture stress;

[0114] Set multiple groups of culture stress gradients according to the key environmental variables of the culture stress, establish multiple experimental groups based on the orthogonal test method for culture experiments, and evaluate the culture effects through the collected culture data to obtain an experimental data set;

[0115] Construct a breeding selection analysis model, construct an association path category of farming pressure gradient - farming effect - farming monitoring data based on the test data set, analyze the farming evolution characteristics of each association path category, and construct a topological structure diagram to train the breeding selection analysis model;

[0116] Obtain breeding selection requirement information, extract features from the breeding selection requirement information to obtain expected farming characteristics, generate breeding selection requirement analysis information according to the expected farming characteristics, and input it into the breeding selection analysis model for analysis to obtain candidate breeding selection plans;

[0117] Extract breeding selection characteristics for each candidate breeding selection plan, evaluate the breeding selection costs of each candidate breeding selection plan based on the extracted breeding selection characteristics, recommend breeding selection plans according to the evaluation results, and generate recommended breeding selection plans.

[0118] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed with each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be electrical, mechanical, or other forms.

[0119] The units described above as separate components may or may not be physically separated. The components shown as units may or may not be physical units; they can be located in one place or distributed to multiple network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0120] In addition, each functional unit in the embodiments of the present invention can be all integrated in a processing unit, or each unit can be separately used as a unit, or two or more units can be integrated in a unit; the above integrated units can be implemented in the form of hardware, or in the form of a combination of hardware and software functional units.

[0121] Those of ordinary skill in the art will understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including those of the above method embodiments. The foregoing storage medium includes various media that can store program codes, such as removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0122] Alternatively, if the above integrated units of the present invention are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention. The foregoing storage medium includes various media that can store program codes, such as removable storage devices, ROM, RAM, magnetic disks, or optical discs.

[0123] As described above, the above are only the specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for selecting high-density tolerant Penaeus monodon based on simulated farming pressure, characterized in that, Including: Based on data retrieval, obtain the stress influencing factors of Penaeus monodon farming, analyze the obtained stress influencing factors using the principal component analysis method, and define the key environmental variables of farming stress; Set multiple groups of farming stress gradients according to the key environmental variables of farming stress, establish multiple experimental groups based on the orthogonal experiment method for farming experiments, and evaluate the farming effects through the collected farming data to obtain an experimental data set; Construct a breeding analysis model, construct the association path categories of farming stress gradient - farming effect - farming monitoring data according to the experimental data set, analyze the farming evolution characteristics of each association path category, and construct a topological structure diagram to train the breeding analysis model; Obtain breeding requirement information, extract the characteristics of the breeding requirement information to obtain the expected farming characteristics, generate breeding requirement analysis information according to the expected farming characteristics and input it into the breeding analysis model for analysis to obtain candidate breeding plans; Extract the breeding characteristics of each candidate breeding plan, evaluate the breeding costs of each candidate breeding plan based on the extracted breeding characteristics, and recommend breeding plans according to the evaluation results to generate recommended breeding plans.

2. The method for selecting and breeding high-density tolerant Penaeus monodon based on simulated culture pressure according to claim 1, wherein The step of obtaining the stress influencing factors of Penaeus monodon farming based on data retrieval, analyzing the obtained stress influencing factors using the principal component analysis method, and defining the key environmental variables of farming stress specifically includes: Based on data retrieval, obtain the relevant stress influencing factors of Penaeus monodon farming, and obtain the influence degree of each stress influencing factor on Penaeus monodon farming through the expert analysis method to form a first data set; Use the parallel coordinate method to map each stress influencing factor in the first data set to the parallel coordinate system according to the influence degree of each stress influencing factor to obtain a stress influencing factor coordinate diagram; Each data axis in the stress influencing factor coordinate diagram corresponds to the dimension range and stress influencing factor type of the characteristic. Screen the initial stress influencing factors through the stress influencing factor coordinate diagram, and eliminate the stress influencing factor types with small correlation to obtain an initial stress influencing factor data set; Introduce the principal component analysis method to perform dimensionality reduction processing on the initial stress influencing factor data set, calculate the principal component scores corresponding to each stress influencing factor, judge the calculated principal component scores with a preset threshold, and select the stress influencing factors corresponding to the principal component scores greater than the preset threshold as the principal component factors; Use the principal component factors for principal component direction projection to obtain a stress influencing factor scatter diagram, select the corresponding stress influencing factors according to the preset selection range in the stress influencing factor scatter diagram, and define the key environmental variables of farming stress based on the selected stress influencing factors.

3. A method for selecting high-density tolerant Penaeus monodon based on simulated farming pressure according to claim 1, characterized in that, The step of setting multiple groups of farming stress gradients according to the key environmental variables of farming stress, establishing multiple experimental groups based on the orthogonal experiment method for farming experiments, and evaluating the farming effects through the collected farming data specifically includes: Obtain the key environmental variables of farming stress, preset the variable level grades and set multiple groups of farming stress gradients, introduce the orthogonal experiment method, and construct an orthogonal experiment matrix based on the set farming stress gradients and farming stress environmental variables according to the preset orthogonal table; Allocate each row according to the orthogonal test matrix as an independent test group, and allocate environmental parameter benchmarks based on the corresponding aquaculture pressure gradient, and conduct aquaculture pressure simulation tests on the aquaculture cycle according to the preset; During the aquaculture period, monitor the aquaculture status of each independent test group, collect aquaculture environment data during the aquaculture cycle using a sensor array arranged in the aquaculture pond, and use a camera array to detect and analyze the aquaculture behavior data of Penaeus monodon in each test group through machine vision technology to obtain an aquaculture monitoring data set; Preset an aquaculture effect evaluation system, and evaluate the aquaculture effect by extracting the aquaculture monitoring data corresponding to the aquaculture effect evaluation index of each test group from the aquaculture monitoring data set through the aquaculture effect evaluation system to obtain aquaculture effect evaluation information; Associate the aquaculture effect evaluation information with the corresponding aquaculture test group in the aquaculture monitoring data set to construct a test data set.

4. A method for selecting high-density tolerant Penaeus monodon based on aquaculture pressure simulation according to claim 1, characterized in that Construct the breeding analysis model. According to the test data set, construct the association path category of aquaculture pressure gradient-aquaculture effect-aquaculture monitoring data, analyze the aquaculture evolution characteristics of each association path category, and construct a topological structure diagram to train the breeding analysis model. Specifically, it includes: Obtain the test data set, extract the aquaculture pressure gradient characteristics of each test group from the test data set, and use the K-means clustering algorithm to classify the aquaculture monitoring data of each test group based on the aquaculture pressure gradient characteristics to obtain classification information; Extract the aquaculture effect evaluation characteristics corresponding to each aquaculture monitoring data in each aquaculture pressure gradient category according to the classification information, and conduct secondary classification to construct the association path of aquaculture pressure gradient-aquaculture effect-aquaculture monitoring data to obtain the second data set; Extract the feature data of different association path categories based on the second data set and perform temporal processing to form the temporal sequences of each association path category, and use the temporal sequences of each association path category as input and import them into the LSTM network for aquaculture evolution feature analysis; Capture the temporal dependence relationship and coupling effect of each feature in the temporal sequence through the set multi-layer LSTM layer. For each input sequence, take the hidden state at the last moment of the LSTM network as the corresponding feature representation, and map the hidden state to the target output space through the fully connected layer to obtain aquaculture evolution feature analysis information; The aquaculture evolution feature analysis information includes aquaculture environment evolution features and aquaculture phenotype evolution features. Use the aquaculture environment evolution features as the first node and the aquaculture phenotype evolution features as the second node to construct a directed description relationship; Connect the first node and the second node according to the directed description relationship to construct a topological structure diagram, construct a breeding analysis model based on the graph neural network, and input the topological structure diagram into the breeding analysis model for coding learning; Obtain the original feature map of each graph node in the topological structure diagram through the multi-head attention mechanism, calculate the attention weight by obtaining the channel descriptor through average pooling, weight the original feature map to obtain the attention feature map, update the state of the hidden layer, and adjust and optimize the model parameters to obtain the desired breeding analysis model.

5. The breeding method of high-density tolerant Penaeus monodon based on aquaculture stress simulation according to claim 1, characterized in that Obtaining breeding requirement information, extracting features from the breeding requirement information to obtain expected breeding characteristics, generating breeding requirement analysis information according to the expected breeding characteristics, and inputting the breeding requirement analysis information into the breeding analysis model for analysis to obtain candidate breeding plans, specifically including: Obtaining breeding requirement information, extracting features from the breeding requirement information to obtain the expected breeding area characteristics, expected breeding density characteristics, and expected breeding trait characteristics of the current high-density tolerant Penaeus monodon, and obtaining breeding requirement analysis information; Inputting the breeding requirement analysis information into the trained breeding analysis model to generate a target node, performing first-order neighborhood sampling according to the target node to obtain a number of neighbor nodes, and calculating the cosine similarity value between the target node and the neighbor nodes; Sorting each neighbor node according to the calculated cosine similarity value, selecting the neighbor node with the largest similarity as the target neighbor node, and obtaining the breeding evolution characteristics corresponding to the target neighbor node to form the feature vector of the target neighbor node; Fusing the feature vector of the target neighbor node with the target node to update the feature vector of the target node, and generating the next target node based on the updated feature vector of the target node, and performing iterative update until the preset number of times to obtain the final target node; Extracting the feature vector of the final target node and using it as the final output feature to generate candidate breeding plans, and obtaining a number of candidate breeding plans.

6. The selective breeding method for high-density tolerant Penaeus monodon based on aquaculture stress simulation according to claim 1, characterized in that, Performing breeding feature extraction on each candidate breeding plan, evaluating the breeding cost of each candidate breeding plan based on the extracted breeding features, and recommending breeding plans according to the evaluation results to generate recommended breeding plans, specifically including: Using the random forest algorithm to construct a breeding plan analysis model, obtaining a topological structure diagram, constructing an adjacency matrix of different associated path categories based on the topological structure diagram, and training the breeding plan analysis model; Obtaining candidate breeding plans, inputting each candidate breeding plan into the trained breeding plan analysis model for analysis, analyzing the breeding environment evolution characteristics and breeding phenotype characteristics of each candidate breeding plan, and obtaining candidate breeding plan analysis information; Performing feature extraction on each candidate breeding plan to obtain the breeding cycle characteristics and breeding requirement equipment characteristics of each candidate breeding plan, and generating first feature information in combination with the candidate breeding plan analysis information; Introducing the improved NSGA-II algorithm to evaluate the breeding cost of each candidate breeding plan, presetting an objective function and establishing constraint conditions, generating an initial population based on the first feature information, and calculating the objective function value for each individual in the population; Performing non-dominated sorting on all individuals in the population according to the objective function value, dividing them into several Pareto fronts, calculating the crowding distance for individuals in each non-dominated layer, using the tournament selection method to select elite individuals using the non-dominated sorting level and crowding distance, and performing crossover and mutation operations on the selected elite individuals; Combining the parent generation and the offspring, re-performing non-dominated sorting and crowding distance calculation on all candidate plans, and selecting a preset number of elite individuals according to the Pareto front level and crowding distance to form the next generation population for iterative evolution; When the change of the objective function of the population meets the convergence criterion or the preset number of iterations, the breeding cost evaluation results of each candidate breeding plan are output, and based on the breeding cost evaluation results, the final breeding plan is recommended to obtain the recommended breeding plan.

7. A high-density tolerant Penaeus monodon breeding system based on simulated farming pressure, characterized in that The system includes: a memory and a processor. The memory contains a program for the breeding method of high-density tolerant Penaeus monodon based on aquaculture stress simulation. When the program for the breeding method of high-density tolerant Penaeus monodon based on aquaculture stress simulation is executed by the processor, the following steps are implemented: Based on data retrieval, the stress influencing factors of Penaeus monodon aquaculture are obtained, and the obtained stress influencing factors are analyzed by the principal component analysis method to define the key environmental variables of aquaculture stress. According to the key environmental variables of aquaculture stress, multiple groups of aquaculture stress gradients are set, multiple experimental groups are established based on the orthogonal test method for aquaculture experiments, and the aquaculture effects are evaluated through the collected aquaculture data to obtain an experimental data set. A breeding analysis model is constructed, the association path categories of aquaculture stress gradient - aquaculture effect - aquaculture monitoring data are constructed according to the experimental data set, the aquaculture evolution characteristics of each association path category are analyzed, and a topological structure diagram is constructed to train the breeding analysis model. The breeding requirement information is obtained, the expected aquaculture characteristics are obtained by feature extraction of the breeding requirement information, and the breeding requirement analysis information is generated according to the expected aquaculture characteristics and input into the breeding analysis model for analysis to obtain candidate breeding plans. The breeding characteristics of each candidate breeding plan are extracted, the breeding cost of each candidate breeding plan is evaluated based on the extracted breeding characteristics, and the breeding plan is recommended according to the evaluation results to generate a recommended breeding plan.

Images