A fish breeding suitability assessment method and system based on an inference rule base

Abstract

The application provides a fish breeding suitability evaluation method and system based on an inference rule base, relates to the technical field of ecological environment protection, and comprises the following steps: mining the distribution of hydrodynamic parameters and a fish breeding suitability evaluation index to obtain a frequency distribution function of each hydrodynamic parameter and a frequency distribution function of the fish breeding suitability evaluation index; establishing a mapping relationship from the combination of each state level of each hydrodynamic parameter to the state level of the fish breeding suitability evaluation index with the highest occurrence frequency under the combination, so as to construct an inference rule base; and evaluating the fish breeding suitability of a fish spawning ground in a target fish breeding period of a research river section according to the inference rule base. The application constructs an interpretable mapping of a fish breeding suitability evaluation index and flow velocity and water depth, and realizes rapid, accurate and objective evaluation of the fish breeding suitability of a target river section under a given flow condition by coupling the mapping with two-dimensional hydrodynamic calculation.

Description

Technical Field

[0001] This invention relates to the fields of water conservancy engineering and ecological environment protection technology, specifically to a method and system for assessing fish reproductive suitability based on a reasoning rule base. Background Technology

[0002] Assessing fish reproductive suitability is a crucial step in ecological flow and ecological regulation research. Existing methods often rely on empirical thresholds or single membership functions, which are highly subjective and struggle to simultaneously characterize the uncertainties caused by randomness and fuzziness. Furthermore, under different flow conditions, the mapping between hydrodynamic parameters such as flow velocity (v) and water depth (d) calculated from a two-dimensional hydrodynamic model and habitat response generally exhibits nonlinear and non-stationary characteristics, making direct regression or fixed thresholds prone to assessment errors. Therefore, there is an urgent need for a technical solution that, without compromising interpretability, collects actual data and establishes inference rules between hydrodynamic parameters and fish reproductive suitability factors to improve the robustness and portability of spatial discrimination of fish reproductive suitability factors. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a method and system for assessing fish reproductive suitability based on a reasoning rule base, which can effectively solve the aforementioned problems.

[0004] The technical solution adopted in this invention is as follows:

[0005] This invention provides a method for assessing fish reproductive suitability based on a reasoning rule base, comprising:

[0006] Step S1: Determine the target fish species to be evaluated, the breeding season of the target fish species, and the spawning grounds of the fish species in the study section; determine at least one hydrodynamic parameter affecting fish reproductive behavior and at least one fish reproductive behavior indicator reflecting fish reproductive behavior;

[0007] Step S2: Divide the fish spawning grounds of the study river section into several monitoring sections; select several monitoring points for each monitoring section; collect data during the multi-year breeding season of the target fish, and obtain the daily representative vector of each monitoring point during the breeding season of the target fish: including the daily representative value of each hydrodynamic parameter and the daily representative value of the fish breeding suitability assessment index calculated based on the fish breeding behavior index;

[0008] Step S3: Select typical samples from the daily representative vectors of each monitoring point at each monitoring section to construct a sample set. ={ };in, For sample set The number of samples in the middle; each sample = , , ; For the sample hydrodynamic parameters eigenvalues, The number of hydrodynamic parameters, For the sample Fish reproductive suitability assessment index eigenvalues;

[0009] Step S4, for the sample set The distribution of each hydrodynamic parameter and the fish reproductive suitability assessment index was mined to obtain the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index.

[0010] Step S5: Using the Gaussian distribution decomposition algorithm, the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index are decomposed to obtain several state levels of each hydrodynamic parameter and several state levels of the fish reproductive suitability assessment index; each state level has four state level characteristics, namely: expectation, standard deviation, amplitude coefficient and fluctuation coefficient.

[0011] Step S6: Establish a mapping relationship between the combination of each state level of each hydrodynamic parameter and the state level of the fish reproductive suitability assessment index that appears most frequently under that combination, thereby constructing a reasoning rule base;

[0012] Step S7: Evaluate the suitability of the fish spawning grounds in the study section for fish reproduction during the breeding season of the target fish based on the reasoning rule base.

[0013] Furthermore, the hydrodynamic parameters include one or more of the following: flow velocity, water depth, flow field gradient modulus, water temperature, and substrate type; the fish reproductive behavior indicators include one or more of the following: environmental DNA concentration, instantaneous movement vector of individual fish, and the ratio of the area of ​​healthy fish eggs to the area of ​​the monitoring point.

[0014] Furthermore, step S2 specifically involves:

[0015] Step S2.1: Data collection during the multi-year breeding season of the target fish species.

[0016] Several monitoring sections were set up in the fish spawning grounds of the studied river section; several vertical monitoring lines were planned along the transverse direction of the river at each monitoring section; several monitoring points were set up along the vertical direction of each monitoring line, and the monitoring points in each vertical direction covered different water depth levels.

[0017] Set daily during the target fish's breeding season. During a fixed monitoring period, hydrodynamic parameter data and fish reproductive behavior index data were collected at each monitoring point during each monitoring period, and the fish reproductive behavior index data were converted into a fish reproductive suitability assessment index.

[0018] Step S2.2: Process the collected data to obtain the daily representative vector for each monitoring point:

[0019] Each monitoring period is assigned a corresponding weight coefficient based on its importance.

[0020] For each hydrodynamic parameter, the parameters obtained daily at each monitoring point are... The hydrodynamic parameter data from different monitoring periods are weighted and summed to obtain the daily representative value of each hydrodynamic parameter at that monitoring point on that day.

[0021] Daily data obtained at each monitoring point The fish reproductive suitability assessment index at different monitoring periods is weighted and summed to obtain the daily representative value of the fish reproductive suitability assessment index at that monitoring point on that day.

[0022] This yields a daily representative vector for each monitoring point, including the daily representative value of each hydrodynamic parameter and the daily representative value of the fish reproductive suitability assessment index.

[0023] Furthermore, step S3 specifically includes:

[0024] Step S3.1: For each monitoring section, construct the sample subset corresponding to that monitoring section:

[0025] Step S3.1.1: For each monitoring section, construct a daily representative vector set consisting of the daily representative vectors obtained by all monitoring points during the breeding season of the target fish.

[0026] Step S3.1.2: Calculate the daily representative value of the fish reproductive suitability assessment index for all monitoring points of each monitoring section during the breeding season of the target fish. Construct the numerical range of the daily representative value of the fish reproductive suitability assessment index based on the minimum and maximum values ​​obtained. Divide this numerical range into several equally spaced suitability intervals.

[0027] Step S3.1.3: Set the baseline sampling quantity ;

[0028] Step S3.1.4: Count the number of daily representative vectors contained in each suitability interval; for intervals with more than [number missing] daily representative vectors... To determine the suitability range, a daily representative vector optimization method based on K-Means clustering was used for data extraction. A number of daily representative vectors are extracted and retained within the suitability interval, while other daily representative vectors within the suitability interval are deleted. The data extraction method is as follows: all daily representative vectors within the suitability interval are used as clustering input, and clustering is performed as follows: For each sub-cluster, calculate the centroid of each sub-cluster and select the representative vector that is closest to the Euclidean distance of the centroid of each sub-cluster, and keep it in the suitability interval;

[0029] The number of daily representative vectors does not exceed The suitability interval is determined by directly retaining all daily representative vectors within that suitability interval;

[0030] In steps S3.1.5 and S3.1.4, the daily representative vector combination retained for each suitability interval is the sample subset corresponding to the monitoring section.

[0031] Step S3.2: Combine the sample subsets corresponding to all monitoring sections to obtain the sample set. ;in, For sample set The number of samples in the sample; each sample Let a day represent a vector, denoted as = ; For the sample hydrodynamic parameters The daily representative value is called the hydrodynamic parameter. eigenvalues; For the sample Fish reproductive suitability assessment index The daily representative value is called the fish reproductive suitability assessment index. eigenvalues.

[0032] Furthermore, step S5 specifically involves:

[0033] Step S5.1: Represent each hydrodynamic parameter and the fish reproductive suitability assessment index as a unified element. ;element The frequency distribution function is expressed as ;

[0034] Step S5.2, apply a Gaussian distribution to the elements. Frequency distribution function The decomposition yields N Gaussian distribution functions, each with four characteristics: expectation, standard deviation, amplitude coefficient, and volatility coefficient.

[0035] Step S5.3: Cluster the N Gaussian distribution functions into M clusters; for each cluster, merge all Gaussian distribution functions within the cluster, and the merged Gaussian distribution function is called the Gaussian distribution representative function; thus, a total of M Gaussian distribution representative functions are obtained, each representing an element. A state level, where each Gaussian distribution represents a characteristic of the function, i.e., an element. The characteristics of the corresponding state level.

[0036] Furthermore, a Gaussian distribution is used for the elements. Frequency distribution function The decomposition yields N Gaussian distribution functions, specifically:

[0037] Step S5.2.1, calculate the sample set ={ } elements Standard deviation According to the element Standard deviation Determine the neighborhood length L: 𝐿 = 0.5 ⋅ 𝑚𝑖𝑛(𝑆𝐷, IQR / 1.34) ⋅ 𝑛 −1 / 5 , For elements Interquartile range in sample set S;

[0038] Step S5.2.2, let the variable =1;

[0039] Step S5.2.3, let the frequency distribution function =Frequency distribution function ;

[0040] Step S5.2.4, based on the frequency distribution function Determine the Gaussian distribution function Expectations Standard deviation Amplitude coefficient and volatility coefficient :

[0041] Take frequency distribution function The x-coordinate corresponding to the peak point of the curve is used as the Gaussian distribution function. Expectations ;

[0042] Take frequency distribution function The ordinate corresponding to the peak point of the curve is used as the Gaussian distribution function. amplitude coefficient ;

[0043] In frequency distribution function In the process of collecting sample sets ={ } is located in the expected The neighborhood length L range elements The value is represented as: , for Quantity, The values ​​of the neighborhood samples are expressed using the formula. Obtain the Gaussian distribution function Standard deviation ;

[0044] Using formula Obtain the Gaussian distribution function Fluctuation coefficient ;

[0045] Step S5.2.5, using the formula Generate Gaussian distribution function The function expression;

[0046] Step S5.2.6: Generate the frequency distribution function ;

[0047] Step S5.2.7, determine the frequency distribution function peak Does it meet the following conditions: , ε is the set stop threshold; β is the control coefficient;

[0048] If satisfied, proceed to step S5.2.8; otherwise, set the variable... = +1, return to step S5.2.4;

[0049] Step S5.2.8 outputs each Gaussian distribution function and its expected value, standard deviation, amplitude coefficient, and volatility coefficient.

[0050] Furthermore, step S5.3 specifically includes:

[0051] Step S5.3.1: Using the expected value of each Gaussian distribution function as the clustering feature parameter, a clustering algorithm is used to cluster the N Gaussian distribution functions into M clusters;

[0052] Step S5.3.2: For each cluster, merge all Gaussian distribution functions within the cluster. The merged Gaussian distribution function is called the Gaussian distribution representative function. The merging method is as follows:

[0053] ① Suppose there are C Gaussian distribution functions in the cluster. Calculate the Gaussian distribution function for each Gaussian distribution function. Effective contribution , :

[0054] In Gaussian distribution function In the middle, determine according to its expectations The sampling interval centered on a preset sampling length. ; These are the sampling start position and sampling end position, respectively; within the sampling interval Within, samples were obtained at equal intervals. Each sampling point , If it follows a Gaussian distribution function ordinate value If the value of the sampling point is greater than the ordinate value of its other C-1 Gaussian distribution functions, then let the sampling point... Effective contribution for Otherwise, let the sampling points Effective contribution =0;

[0055] Using formula ,calculate The effective contribution of each sampling point is used to obtain the Gaussian distribution function. Effective contribution ;

[0056] ②Based on the expected value, standard deviation, amplitude coefficient, volatility coefficient, and effective contribution of each Gaussian distribution function within the cluster, calculate the expected value, standard deviation, amplitude coefficient, and volatility coefficient of the representative Gaussian distribution function of the cluster:

[0057] The expected value of the Gaussian distribution representative function is obtained by weighting the expected value of each Gaussian distribution function within the cluster with the effective contribution of each Gaussian distribution function within the cluster.

[0058] The fluctuation coefficient of the Gaussian distribution representative function is obtained by weighting the effective contribution of each Gaussian distribution function within the cluster and averaging the fluctuation coefficients of each Gaussian distribution function within the cluster.

[0059] Calculate the effective contribution / amplitude coefficient of each Gaussian distribution function within the cluster, and then sum the effective contribution / amplitude coefficients of each Gaussian distribution function to obtain the standard deviation of the representative Gaussian distribution function;

[0060] Calculate the effective contribution / standard deviation of each Gaussian distribution function within the cluster, and then sum the effective contribution / standard deviations of each Gaussian distribution function to obtain the amplitude coefficient of the representative Gaussian distribution function;

[0061] ③ Generate the functional expression of the Gaussian distribution representative function based on the expected value, standard deviation, amplitude coefficient, and fluctuation coefficient of the Gaussian distribution representative function;

[0062] ④ Correct the Gaussian distribution representative function obtained in step ③ to obtain the corrected Gaussian distribution representative function. The correction method is as follows:

[0063] If compared with the Gaussian distribution functions within the cluster, the Gaussian distribution representative function is closest to the element If the boundary of the domain is defined, the curve of the Gaussian distribution representative function inside the domain is retained, and the Gaussian distribution function value between its expected value and the boundary of the domain is corrected to 1, thus transforming it into a Gaussian distribution function with a one-sided plateau characteristic.

[0064] Furthermore, step S7 specifically includes:

[0065] Step S7.1: Obtain each hydrodynamic parameter of each cell grid of the fish spawning grounds in the study section during the target fish breeding season for the simulated or predicted time period. eigenvalues ;

[0066] Step S7.2: Determine each hydrodynamic parameter State level and certainty :

[0067] Assume hydrodynamic parameters There are A types of status levels. Iterate through each state level , Hydrodynamic parameters were calculated. eigenvalues Corresponding to each state level Certainty Each degree of certainty The state level corresponding to the maximum value is the determined hydrodynamic parameter. State level; each degree of certainty The maximum value is the determined hydrodynamic parameter Certainty Certainty The formula is expressed as follows:

[0068] ;

[0069] in: Status level Expectations; To be based on state level Standard deviation As expected, in terms of state level The square of the fluctuation coefficient Random numbers generated for variance;

[0070] Step S7.3, based on the determined hydrodynamic parameters Based on the state level and the reasoning rule base, a fish reproductive suitability assessment index is derived. Status level This leads to the state level. Expectations Standard deviation Amplitude coefficient and volatility coefficient ;

[0071] Step S7.4: Obtain the fish reproductive suitability assessment index according to the following formula. Evaluation value :

[0072]

[0073] in: For As expected, with volatility coefficient square Random numbers generated for variance; Hydrodynamic parameters eigenvalues Fish reproductive suitability assessment index The directional parameter of action, It is either 1 or -1, and its calculation formula is:

[0074]

[0075] in:

[0076] For the determined hydrodynamic parameters The expected state level;

[0077] Hydrodynamic parameters The optimal value is determined as follows: In the reasoning rule base, all reasoning rules corresponding to the highest state level of the fish reproductive suitability assessment index are found; among all the found reasoning rules, the hydrodynamic parameters are determined. The various state levels, and the hydrodynamic parameters The mean of the expected values ​​of each state level is taken as the hydrodynamic parameter. The optimal value;

[0078] Step S7.5: Repeat steps S7.2-S7.4 multiple times, and take the evaluation value obtained each time. The average value is used as the final fish breeding suitability assessment index.

[0079] This invention also provides a fish reproductive suitability assessment system based on a reasoning rule base, comprising:

[0080] The determination module is used to determine the target fish species to be evaluated, the breeding season of the target fish species, and the spawning grounds of the fish species in the study section; and to determine at least one hydrodynamic parameter affecting fish reproductive behavior and at least one fish reproductive behavior index reflecting fish reproductive behavior.

[0081] The data acquisition module is used to divide the fish spawning grounds of the study river section into several monitoring sections; select several monitoring points for each monitoring section; collect data during the multi-year breeding season of the target fish, and obtain the daily representative vector of each monitoring point during the breeding season of the target fish: including the daily representative value of each hydrodynamic parameter and the daily representative value of the fish breeding suitability assessment index calculated based on the fish breeding behavior index;

[0082] The sample set construction module is used to select typical samples from the daily representative vectors of each monitoring point at each monitoring section to construct the sample set. ={ };in, For sample set The number of samples in the middle; each sample = , , ; For the sample hydrodynamic parameters eigenvalues, The number of hydrodynamic parameters, For the sample Fish reproductive suitability assessment index eigenvalues;

[0083] The frequency distribution function construction module is used for the sample set. The distribution of each hydrodynamic parameter and the fish reproductive suitability assessment index was mined to obtain the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index.

[0084] The frequency distribution function decomposition module is used to decompose the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index using the Gaussian distribution decomposition algorithm, to obtain several state levels of each hydrodynamic parameter and several state levels of the fish reproductive suitability assessment index; each state level has four state level characteristics, namely: expectation, standard deviation, amplitude coefficient and fluctuation coefficient.

[0085] The reasoning rule base construction module is used to establish a mapping relationship between the combination of various state levels of each hydrodynamic parameter and the state level of the fish reproductive suitability assessment index that appears most frequently under that combination, thereby constructing the reasoning rule base;

[0086] The fish reproduction suitability assessment module is used to assess the fish reproduction suitability of the fish spawning grounds in the study section during the breeding season of the target fish, based on the inference rule base.

[0087] The fish reproductive suitability assessment method and system based on a reasoning rule base provided by this invention has the following advantages:

[0088] This invention, based on the autonomous extraction of rules and grade boundaries from Gaussian distribution in sample data, constructs an interpretable mapping between a fish reproductive suitability assessment index and flow velocity and water depth. By coupling this mapping with two-dimensional hydrodynamic calculations, it achieves a rapid, accurate, and objective evaluation of fish reproductive suitability in a target river segment under given flow conditions. This invention reduces the subjectivity and evaluation errors of traditional fish reproductive suitability assessment techniques, demonstrating significant practicality and broad application prospects. Attached Figure Description

[0089] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0090] Figure 1 A flowchart of the fish reproductive suitability assessment method based on inference rule base provided by the present invention;

[0091] Figure 2 The flow rate in the embodiments of the present invention , water depth and fish reproductive suitability assessment index The result after Gaussian distribution decomposition;

[0092] Figure 3 The flow rate in the embodiments of the present invention , water depth and fish reproductive suitability assessment index Status level result diagram;

[0093] Figure 4 This is a graph showing the results of the fish reproductive suitability assessment index in the target river section spawning grounds under a given flow condition, according to an embodiment of the present invention. Detailed Implementation

[0094] To make the technical problems solved, the technical solutions, and the beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the invention.

[0095] This invention develops a method for assessing fish reproductive suitability based on inference rules. By establishing an inference rule base, a mapping relationship between the fish reproductive suitability assessment index and hydrodynamic parameters is created in a sample-driven manner. The inference rules are coupled with two-dimensional hydrodynamic calculations, enabling rapid assessment of fish reproductive suitability while maintaining physical constraints. This results in a fish reproductive suitability assessment approach that combines objectivity, accuracy, and computational efficiency.

[0096] See Figure 1 This invention provides a method for assessing fish reproductive suitability based on a reasoning rule base, comprising:

[0097] Step S1: Determine the target fish species to be evaluated, the breeding season of the target fish species, and the spawning grounds of the fish species in the study section; determine at least one hydrodynamic parameter affecting fish reproductive behavior and at least one fish reproductive behavior indicator reflecting fish reproductive behavior;

[0098] In this step, the hydrodynamic parameters include one or more of the following: flow velocity, water depth, flow field gradient modulus, water temperature, and substrate type; the fish reproductive behavior indicators include one or more of the following: environmental DNA concentration, instantaneous movement vector of individual fish, and the ratio of the area of ​​healthy fish eggs to the area of ​​the monitoring point.

[0099] Step S2: Divide the fish spawning grounds of the study river section into several monitoring sections; select several monitoring points for each monitoring section; collect data during the multi-year breeding season of the target fish, and obtain the daily representative vector of each monitoring point during the breeding season of the target fish: including the daily representative value of each hydrodynamic parameter and the daily representative value of the fish breeding suitability assessment index calculated based on the fish breeding behavior index;

[0100] This step is specifically as follows:

[0101] Step S2.1: Data collection during the multi-year breeding season of the target fish species.

[0102] Several monitoring sections were set up in the fish spawning grounds of the studied river section; several vertical monitoring lines were planned along the transverse direction of the river at each monitoring section; several monitoring points were set up along the vertical direction of each monitoring line, and the monitoring points in each vertical direction covered different water depth levels.

[0103] Set daily during the target fish's breeding season. During a fixed monitoring period, hydrodynamic parameter data and fish reproductive behavior index data were collected at each monitoring point during each monitoring period, and the fish reproductive behavior index data were converted into a fish reproductive suitability assessment index.

[0104] Step S2.2: Process the collected data to obtain the daily representative vector for each monitoring point:

[0105] Each monitoring period is assigned a corresponding weight coefficient based on its importance.

[0106] For each hydrodynamic parameter, the parameters obtained daily at each monitoring point are... The hydrodynamic parameter data from different monitoring periods are weighted and summed to obtain the daily representative value of each hydrodynamic parameter at that monitoring point on that day.

[0107] Daily data obtained at each monitoring point The fish reproductive suitability assessment index at different monitoring periods is weighted and summed to obtain the daily representative value of the fish reproductive suitability assessment index at that monitoring point on that day.

[0108] This yields a daily representative vector for each monitoring point, including the daily representative value of each hydrodynamic parameter and the daily representative value of the fish reproductive suitability assessment index.

[0109] Step S3: Select typical samples from the daily representative vectors of each monitoring point at each monitoring section to construct a sample set. ={ };in, For sample set The number of samples in the middle; each sample = , , ; For the sample hydrodynamic parameters eigenvalues, The number of hydrodynamic parameters, For the sample Fish reproductive suitability assessment index eigenvalues;

[0110] This step is specifically as follows:

[0111] Step S3.1: For each monitoring section, construct the sample subset corresponding to that monitoring section:

[0112] Step S3.1.1: For each monitoring section, construct a daily representative vector set consisting of the daily representative vectors obtained by all monitoring points during the breeding season of the target fish.

[0113] Step S3.1.2: Calculate the daily representative value of the fish reproductive suitability assessment index for all monitoring points of each monitoring section during the breeding season of the target fish. Construct the numerical range of the daily representative value of the fish reproductive suitability assessment index based on the minimum and maximum values ​​obtained. Divide this numerical range into several equally spaced suitability intervals.

[0114] Step S3.1.3: Set the baseline sampling quantity ;

[0115] Step S3.1.4: Count the number of daily representative vectors contained in each suitability interval; for intervals with more than [number missing] daily representative vectors... To determine the suitability range, a daily representative vector optimization method based on K-Means clustering was used for data extraction. A number of daily representative vectors are extracted and retained within the suitability interval, while other daily representative vectors within the suitability interval are deleted. The data extraction method is as follows: all daily representative vectors within the suitability interval are used as clustering input, and clustering is performed as follows: For each sub-cluster, calculate the centroid of each sub-cluster and select the representative vector that is closest to the Euclidean distance of the centroid of each sub-cluster, and keep it in the suitability interval;

[0116] The number of daily representative vectors does not exceed The suitability interval is determined by directly retaining all daily representative vectors within that suitability interval;

[0117] In steps S3.1.5 and S3.1.4, the daily representative vector combination retained for each suitability interval is the sample subset corresponding to the monitoring section.

[0118] Step S3.2: Combine the sample subsets corresponding to all monitoring sections to obtain the sample set. ;in, For sample set The number of samples in the sample; each sample Let a day represent a vector, denoted as = ; For the sample hydrodynamic parameters The daily representative value is called the hydrodynamic parameter. eigenvalues; For the sample Fish reproductive suitability assessment index The daily representative value is called the fish reproductive suitability assessment index. eigenvalues.

[0119] Step S4, for the sample set The distribution of each hydrodynamic parameter and the fish reproductive suitability assessment index was mined to obtain the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index.

[0120] Step S5: Using the Gaussian distribution decomposition algorithm, the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index are decomposed to obtain several state levels of each hydrodynamic parameter and several state levels of the fish reproductive suitability assessment index; each state level has four state level characteristics, namely: expectation, standard deviation, amplitude coefficient and fluctuation coefficient.

[0121] This step is specifically as follows:

[0122] Step S5.1: Represent each hydrodynamic parameter and the fish reproductive suitability assessment index as a unified element. ;element The frequency distribution function is expressed as ;

[0123] Step S5.2, apply a Gaussian distribution to the elements. Frequency distribution function The decomposition yields N Gaussian distribution functions, each with four characteristics: expectation, standard deviation, amplitude coefficient, and volatility coefficient.

[0124] Step S5.2.1, calculate the sample set ={ } elements Standard deviation According to the element Standard deviation Determine the neighborhood length L: 𝐿 = 0.5 ⋅ 𝑚𝑖𝑛(𝑆𝐷, IQR / 1.34) ⋅ 𝑛 −1 / 5 , For elements Interquartile range in sample set S;

[0125] Step S5.2.2, let the variable =1;

[0126] Step S5.2.3, let the frequency distribution function =Frequency distribution function ;

[0127] Step S5.2.4, based on the frequency distribution function Determine the Gaussian distribution function Expectations Standard deviation Amplitude coefficient and volatility coefficient :

[0128] Take frequency distribution function The x-coordinate corresponding to the peak point of the curve is used as the Gaussian distribution function. Expectations ;

[0129] Take frequency distribution function The ordinate corresponding to the peak point of the curve is used as the Gaussian distribution function. amplitude coefficient ;

[0130] In frequency distribution function In the process of collecting sample sets ={ } is located in the expected The neighborhood length L range elements The value is represented as: , for Quantity, The values ​​of the neighborhood samples are expressed using the formula. Obtain the Gaussian distribution function Standard deviation ;

[0131] Using formula Obtain the Gaussian distribution function Fluctuation coefficient ;

[0132] In this step, the Gaussian distribution function Standard deviation It is used to quantify the average deviation of each sample point in a Gaussian distribution from the mean; volatility coefficient This is used to reflect the uncertainty that arises when conducting fish reproductive suitability assessments;

[0133] Step S5.2.5, using the formula Generate Gaussian distribution function The function expression;

[0134] Step S5.2.6: Generate the frequency distribution function ;

[0135] Step S5.2.7, determine the frequency distribution function peak Does it meet the following conditions: , ε is the set stop threshold; β is the control coefficient;

[0136] If satisfied, proceed to step S5.2.8; otherwise, set the variable... = +1, return to step S5.2.4;

[0137] Step S5.2.8 outputs each Gaussian distribution function and its expected value, standard deviation, amplitude coefficient, and volatility coefficient.

[0138] Step S5.3: Cluster the N Gaussian distribution functions into M clusters; for each cluster, merge all Gaussian distribution functions within the cluster, and the merged Gaussian distribution function is called the Gaussian distribution representative function; thus, a total of M Gaussian distribution representative functions are obtained, each representing an element. A state level, where each Gaussian distribution represents a characteristic of the function, i.e., an element. The characteristics of the corresponding state level.

[0139] Step S5.3.1: Using the expected value of each Gaussian distribution function as the clustering feature parameter, a clustering algorithm is used to cluster the N Gaussian distribution functions into M clusters;

[0140] Step S5.3.2: For each cluster, merge all Gaussian distribution functions within the cluster. The merged Gaussian distribution function is called the Gaussian distribution representative function. The merging method is as follows:

[0141] ① Suppose there are C Gaussian distribution functions in the cluster. Calculate the Gaussian distribution function for each Gaussian distribution function. Effective contribution , :

[0142] In Gaussian distribution function In the middle, determine according to its expectations The sampling interval centered on a preset sampling length. ; These are the sampling start position and sampling end position, respectively; within the sampling interval Within, samples were obtained at equal intervals. Each sampling point , If it follows a Gaussian distribution function ordinate value If the value of the sampling point is greater than the ordinate value of its other C-1 Gaussian distribution functions, then let the sampling point... Effective contribution for Otherwise, let the sampling points Effective contribution =0;

[0143] Using formula ,calculate The effective contribution of each sampling point is used to obtain the Gaussian distribution function. Effective contribution ;

[0144] ②Based on the expected value, standard deviation, amplitude coefficient, volatility coefficient, and effective contribution of each Gaussian distribution function within the cluster, calculate the expected value, standard deviation, amplitude coefficient, and volatility coefficient of the representative Gaussian distribution function of the cluster:

[0145] The expected value of the Gaussian distribution representative function is obtained by weighting the expected value of each Gaussian distribution function within the cluster with the effective contribution of each Gaussian distribution function within the cluster.

[0146] The fluctuation coefficient of the Gaussian distribution representative function is obtained by weighting the effective contribution of each Gaussian distribution function within the cluster and averaging the fluctuation coefficients of each Gaussian distribution function within the cluster.

[0147] Calculate the effective contribution / amplitude coefficient of each Gaussian distribution function within the cluster, and then sum the effective contribution / amplitude coefficients of each Gaussian distribution function to obtain the standard deviation of the representative Gaussian distribution function;

[0148] Calculate the effective contribution / standard deviation of each Gaussian distribution function within the cluster, and then sum the effective contribution / standard deviations of each Gaussian distribution function to obtain the amplitude coefficient of the representative Gaussian distribution function;

[0149] ③ Generate the functional expression of the Gaussian distribution representative function based on the expected value, standard deviation, amplitude coefficient, and fluctuation coefficient of the Gaussian distribution representative function;

[0150] ④ Correct the Gaussian distribution representative function obtained in step ③ to obtain the corrected Gaussian distribution representative function. The correction method is as follows:

[0151] If compared with the Gaussian distribution functions within the cluster, the Gaussian distribution representative function is closest to the element If the boundary of the domain is defined, the curve of the Gaussian distribution representative function inside the domain is retained, and the Gaussian distribution function value between its expected value and the boundary of the domain is corrected to 1, thus transforming it into a Gaussian distribution function with a one-sided plateau characteristic.

[0152] Step S6: Establish a mapping relationship between the combination of each state level of each hydrodynamic parameter and the state level of the fish reproductive suitability assessment index that appears most frequently under that combination, thereby constructing a reasoning rule base;

[0153] Step S7: Evaluate the suitability of the fish spawning grounds in the study section for fish reproduction during the breeding season of the target fish based on the reasoning rule base.

[0154] This step is specifically as follows:

[0155] Step S7.1: Obtain each hydrodynamic parameter of each cell grid of the fish spawning grounds in the study section during the target fish breeding season for the simulated or predicted time period. eigenvalues ;

[0156] Step S7.2: Determine each hydrodynamic parameter State level and certainty :

[0157] Assume hydrodynamic parameters There are A types of status levels. Iterate through each state level , Hydrodynamic parameters were calculated. eigenvalues Corresponding to each state level Certainty Each degree of certainty The state level corresponding to the maximum value is the determined hydrodynamic parameter. State level; each degree of certainty The maximum value is the determined hydrodynamic parameter Certainty Certainty The formula is expressed as follows:

[0158] ;

[0159] in: Status level Expectations; To be based on state level Standard deviation As expected, in terms of state level The square of the fluctuation coefficient Random numbers generated for variance;

[0160] Step S7.3, based on the determined hydrodynamic parameters Based on the state level and the reasoning rule base, a fish reproductive suitability assessment index is derived. Status level This leads to the state level. Expectations Standard deviation Amplitude coefficient and volatility coefficient ;

[0161] Step S7.4: Obtain the fish reproductive suitability assessment index according to the following formula. Evaluation value :

[0162]

[0163] in: For As expected, with volatility coefficient square Random numbers generated for variance; Hydrodynamic parameters eigenvalues Fish reproductive suitability assessment index The directional parameter of action, It is either 1 or -1, and its calculation formula is:

[0164]

[0165] in:

[0166] For the determined hydrodynamic parameters The expected state level;

[0167] Hydrodynamic parameters The optimal value is determined as follows: In the reasoning rule base, all reasoning rules corresponding to the highest state level of the fish reproductive suitability assessment index are found; among all the found reasoning rules, the hydrodynamic parameters are determined. The various state levels, and the hydrodynamic parameters The mean of the expected values ​​of each state level is taken as the hydrodynamic parameter. The optimal value;

[0168] Step S7.5: Due to the randomness of the calculation, repeat steps S7.2-S7.4 multiple times and take the evaluation value obtained each time. The average value is used as the final fish breeding suitability assessment index.

[0169] The reasoning rule base constructed using the method of this invention can be used to support the formulation of ecological dispatching operation plans for hydropower stations during the fish breeding season. By inputting different flow dispatching plans, the corresponding fish breeding suitability assessment results can be obtained quickly.

[0170] This invention also provides a fish reproductive suitability assessment system based on a reasoning rule base, comprising:

[0171] The determination module is used to determine the target fish species to be evaluated, the breeding season of the target fish species, and the spawning grounds of the fish species in the study section; and to determine at least one hydrodynamic parameter affecting fish reproductive behavior and at least one fish reproductive behavior index reflecting fish reproductive behavior.

[0172] The data acquisition module is used to divide the fish spawning grounds of the study river section into several monitoring sections; select several monitoring points for each monitoring section; collect data during the multi-year breeding season of the target fish, and obtain the daily representative vector of each monitoring point during the breeding season of the target fish: including the daily representative value of each hydrodynamic parameter and the daily representative value of the fish breeding suitability assessment index calculated based on the fish breeding behavior index;

[0173] The sample set construction module is used to select typical samples from the daily representative vectors of each monitoring point at each monitoring section to construct the sample set. ={ };in, For sample set The number of samples in the middle; each sample = , , ; For the sample hydrodynamic parameters eigenvalues, The number of hydrodynamic parameters, For the sample Fish reproductive suitability assessment index eigenvalues;

[0174] The frequency distribution function construction module is used for the sample set. The distribution of each hydrodynamic parameter and the fish reproductive suitability assessment index was mined to obtain the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index.

[0175] The frequency distribution function decomposition module is used to decompose the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index using the Gaussian distribution decomposition algorithm, to obtain several state levels of each hydrodynamic parameter and several state levels of the fish reproductive suitability assessment index; each state level has four state level characteristics, namely: expectation, standard deviation, amplitude coefficient and fluctuation coefficient.

[0176] The reasoning rule base construction module is used to establish a mapping relationship between the combination of various state levels of each hydrodynamic parameter and the state level of the fish reproductive suitability assessment index that appears most frequently under that combination, thereby constructing the reasoning rule base;

[0177] The fish reproduction suitability assessment module is used to assess the fish reproduction suitability of the fish spawning grounds in the study section during the breeding season of the target fish, based on the inference rule base.

[0178] This invention, based on the autonomous extraction of rules and grade boundaries from Gaussian distribution in sample data, constructs an interpretable mapping between a fish reproductive suitability assessment index and flow velocity and water depth. By coupling this mapping with two-dimensional hydrodynamic calculations, it achieves a rapid, accurate, and objective evaluation of fish reproductive suitability in a target river segment under given flow conditions. This invention reduces the subjectivity and evaluation errors of traditional fish reproductive suitability assessment techniques, demonstrating significant practicality and broad application prospects.

[0179] The following is an example:

[0180] This embodiment uses hydrodynamic parameters as the flow velocity. With water depth Taking environmental DNA concentration as an example, we will introduce the reproductive behavior indicators of fish:

[0181] 1. Spawning grounds and indicator fish species in the target river section

[0182] This embodiment uses a spawning ground of a certain river as an example for illustration. The target spawning ground has a river section topographic data of about 1.7 km in length. The convex bank of the river channel forms gravel and pebble shoals, with deep pools and shallow shoals interspersed, and the water flow pattern is relatively diverse. The bottom material is mostly pebbles and gravel. The species composition of fish collected in the spawning ground is richer than that in other river sections, and the resource quantity is relatively abundant.

[0183] Among the fish caught during the survey of the spawning grounds in the target river section, a certain genus of schizothorax was absolutely dominant, while other species were occasionally observed. Therefore, this species of schizothorax was selected as the target fish species for the spawning grounds in this target river section.

[0184] 2. Collection of fish reproductive suitability samples

[0185] Because a reservoir is planned to be built upstream of the target river spawning grounds, in order to investigate the impact of the upstream reservoir construction on the reproductive suitability of fish in the target river spawning grounds, it is necessary to collect environmental DNA concentrations of the target fish under different hydrodynamic parameters during the breeding season of the target river spawning grounds. The environmental DNA concentrations will be normalized to [0, 100], and the fish reproductive suitability assessment index will be obtained through a fish reproductive suitability assessment index calculation model. Among these, hydrodynamic parameters include flow velocity. With water depth The collected flow rate The range is 0-5.274 m / s, and the water depth is... The range is 0.146-0.866m. Using collected fish reproductive suitability samples, a method based on an inference rule base is employed to establish the correspondence between the target fish's hydrodynamic parameters and the fish reproductive suitability assessment index. This allows for the rapid determination of the fish reproductive suitability assessment index for the target river segment by providing the hydrodynamic parameters.

[0186] 3. Generate parameter status levels

[0187] By collecting samples of fish reproductive suitability, the flow rate was analyzed. , water depth and fish reproductive suitability assessment index Data processing is performed. A frequency distribution function for each element is constructed. , Using Gaussian distribution for elements Frequency distribution function Decompose; where elements Represents flow rate , water depth and fish reproductive suitability assessment index .

[0188] First, let k=1, and determine the frequency distribution function. Through frequency distribution function Determine the Gaussian distribution function Features, including expectations Standard deviation Amplitude coefficient and volatility coefficient :

[0189] frequency distribution function The x-axis corresponding to the peak value is used as the expectation. The vertical axis represents the amplitude coefficient. Determining the expected value based on the sample standard deviation and interquartile range. neighborhood length Collect neighborhood The standard deviation is obtained by calculating the average deviation of the sample from the mean. The fluctuation coefficient is obtained by calculating the variance of the sample deviation. Using formulas Generate Gaussian distribution function The function expression generates the frequency distribution function. Repeat the iteration until the frequency distribution function is obtained. The peak value is less than the set stop threshold. In this embodiment, the set stop threshold is... β is the control coefficient.

[0190] Through the above steps, the flow rate was respectively... , water depth and fish reproductive suitability assessment index After Gaussian distribution decomposition, the flow velocity yielded 11 Gaussian distribution functions, the water depth yielded 21 Gaussian distribution functions, and the fish reproductive suitability assessment index yielded 21 Gaussian distribution functions. For example... Figure 2 As shown, the flow rate is... , water depth and fish reproductive suitability assessment index The result after Gaussian distribution decomposition.

[0191] Furthermore, the Gaussian distribution function of each element x needs to be clustered to identify the state level of each element x. In this embodiment, six state levels are chosen as the final number for each of the three elements, and a clustering method is used to divide the Gaussian distribution function of each element x into six clusters.

[0192] For each element x, there are 6 clusters. All Gaussian distribution functions within each cluster are merged. The merged Gaussian distribution function is called the Gaussian distribution representative function. The merging method is as follows:

[0193] First, calculate the Gaussian distribution function for each cluster. Effective contribution ;

[0194] Then, using the effective contribution of each Gaussian distribution function within the cluster as the weight, the expected value of each Gaussian distribution function within the cluster is weighted and averaged to obtain the expected value of the representative Gaussian distribution function.

[0195] The fluctuation coefficient of the Gaussian distribution representative function is obtained by weighting the effective contribution of each Gaussian distribution function within the cluster and averaging the fluctuation coefficients of each Gaussian distribution function within the cluster.

[0196] Calculate the effective contribution / amplitude coefficient of each Gaussian distribution function within the cluster, and then sum the effective contribution / amplitude coefficients of each Gaussian distribution function to obtain the standard deviation of the representative Gaussian distribution function;

[0197] Calculate the effective contribution / standard deviation of each Gaussian distribution function within the cluster, and then sum the effective contribution / standard deviations of each Gaussian distribution function to obtain the amplitude coefficient of the representative Gaussian distribution function;

[0198] Finally, based on the expectation, standard deviation, amplitude coefficient, and volatility coefficient of the Gaussian distribution representative function, a functional expression for the Gaussian distribution representative function is generated; if compared with each Gaussian distribution function within the cluster, the Gaussian distribution representative function is closest to the element... If the boundary of the domain is defined, the curve of the Gaussian distribution representative function inside the domain is retained, and the Gaussian distribution function value between its expected value and the boundary of the domain is corrected to 1, thus transforming it into a Gaussian distribution function with a one-sided plateau characteristic, thereby realizing the correction of the Gaussian distribution representative function.

[0199] Through the above steps, for each element x, six Gaussian distribution representative functions are obtained, each representing an element. A state level, where each Gaussian distribution represents a characteristic of the function, i.e., an element. The corresponding state level has characteristics including expectation, standard deviation, amplitude coefficient, and volatility coefficient.

[0200] like Figure 3 As shown, the flow rate is... , water depth and fish reproductive suitability assessment index The resulting state level diagram yielded six state levels for flow velocity, six for water depth, and six for the reproductive suitability assessment index. Tables 1-3 show the state levels for flow velocity, water depth, and reproductive suitability assessment index, respectively. , water depth and fish reproductive suitability assessment index Characteristics of each state level:

[0201] Table 1: Flow velocity State level characteristics

[0202] Status level number expect Standard deviation Volatility coefficient Amplitude coefficient 1 0.316 0.173 0.157 0.300 2 0.694 0.199 0.157 0.496 3 1.232 0.266 0.161 0.544 4 1.832 0.165 0.164 0.229 5 2.253 0.173 0.158 0.157 6 2.621 0.282 0.172 0.040

[0203] Table 2: Water Depth State level characteristics

[0204] Status level number expect Standard deviation Volatility coefficient Amplitude coefficient 1 0.718 0.560 0.265 0.092 2 1.844 0.587 0.241 0.171 3 3.711 0.666 0.251 0.163 4 5.141 0.472 0.269 0.095 5 6.446 0.594 0.263 0.041 6 7.655 0.588 0.242 0.035

[0205] Table 3: Reproduction Index State level characteristics

[0206] Status level number expect Standard deviation Volatility coefficient Amplitude coefficient 1 3.340 4.702 1.459 0.020 2 18.321 6.766 7.765 0.005 3 33.918 5.821 7.700 0.008 4 50.625 6.801 7.867 0.006 5 69.175 7.668 7.609 0.004 6 91.369 8.235 5.809 0.006

[0207] 4. Establish a reasoning rule base

[0208] This embodiment constructs a dual-input, single-output inference rule. Based on historical data, it statistically analyzes the most frequently occurring fish reproductive suitability assessment index status levels under different combinations of flow velocity and water depth status levels, generating an inference rule base. In this embodiment, there are a total of 36 rules.

[0209] The total number of rules is 36, the first... Rules For "if the water depth status level is And the flow velocity state level is The corresponding fish reproductive suitability assessment index status level is: The specific reasoning rules are shown in the table below.

[0210] Table 4: Reasoning Rule Base

[0211] Serial Number Flow rate status level Water depth status level Fish reproductive suitability assessment index status level Serial Number Flow rate status level Water depth status level Fish reproductive suitability assessment index status level 1 1 1 1 19 4 1 6 2 1 2 4 20 4 2 6 3 1 3 1 21 4 3 4 4 1 4 1 22 4 4 1 5 1 5 1 23 4 5 1 6 1 6 1 24 4 6 1 7 2 1 5 25 5 1 6 8 2 2 5 26 5 2 6 9 2 3 3 27 5 3 2 10 2 4 1 28 5 4 1 11 2 5 1 29 5 5 1 12 2 6 1 30 5 6 1 13 3 1 6 31 6 1 4 14 3 2 6 32 6 2 4 15 3 3 4 33 6 3 2 16 3 4 1 34 6 4 1 17 3 5 1 35 6 5 1 18 3 6 1 36 6 6 1

[0212] 5. Assess the reproductive suitability of fish based on the reasoning rule base.

[0213] Underwater topographic data of the spawning grounds in the target river section were collected. A two-dimensional hydrodynamic model was used to divide the target river section into several cell grids. According to the reservoir scheduling design results, after the upstream reservoir is built, the upstream inflow will be approximately Q = 760 m³ / s at the beginning of the target fish's breeding season in a normal year. This embodiment calculates the characteristic values ​​of the hydrodynamic parameters of each cell grid in the target river section under the condition of a flow rate of Q = 760 m³ / s. (flow rate) and water depth Based on the established reasoning rule base, the fish reproductive suitability assessment index is calculated for each cell grid.

[0214] For each given hydrodynamic parameter characteristic value There are a total of 6 status levels. Iterate through each state level , Hydrodynamic parameters were calculated. eigenvalues Corresponding to each state level Certainty Each degree of certainty The state level corresponding to the maximum value is the determined hydrodynamic parameter. State level; each degree of certainty The maximum value is the determined hydrodynamic parameter Certainty Certainty The formula is expressed as follows:

[0215]

[0216] Where A is the total number of status levels, A=6; Status level Expectations Using standard deviation For the expected value, squared by the volatility coefficient Random numbers generated for variance.

[0217] The state level with the highest certainty is determined as the state level of the current parameters. Based on the determined state levels of flow velocity and water depth, the corresponding state level of the reproductive suitability assessment index is found in the inference rule base. To obtain their expectations Standard deviation and volatility coefficient .

[0218] Calculate the assessment value of the fish reproductive suitability assessment index. :

[0219]

[0220] Where: m is a hydrodynamic parameter The number of possibilities, in this embodiment, m=2; For As expected, with volatility coefficient square Random numbers generated for variance; Hydrodynamic parameters eigenvalues Fish reproductive suitability assessment index The directional parameter of action, It is either 1 or -1, and its calculation formula is:

[0221]

[0222] in:

[0223] For the determined hydrodynamic parameters The expected state level;

[0224] Hydrodynamic parameters The optimal value is determined as follows: In the reasoning rule base, all reasoning rules corresponding to the highest state level of the fish reproductive suitability assessment index are found; among all the found reasoning rules, the hydrodynamic parameters are determined. The various state levels, and the hydrodynamic parameters The mean of the expected values ​​of each state level is taken as the hydrodynamic parameter. The optimal value; in this embodiment, =1.77m / s, =1.28m.

[0225] Due to the calculation process The generation of [something] is random and subject to fluctuation coefficients. To ensure the stability and reliability of the results, the above calculation steps were repeated 500 times for each grid, and the average value was taken as the final fish reproductive suitability assessment index for that grid.

[0226] like Figure 4 The figure shown is a graph illustrating the results of the fish reproductive suitability assessment index in the spawning grounds of the target river section under this flow condition.

[0227] Therefore, this invention autonomously extracts rules and grade boundaries from sample data based on Gaussian distribution, constructs an interpretable mapping between the fish reproductive suitability assessment index and flow velocity and water depth, and couples it with two-dimensional hydrodynamic calculation to achieve a rapid, accurate, and objective evaluation of the fish reproductive suitability of a target river section under a given flow condition.

[0228] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for assessing fish reproductive suitability based on a reasoning rule base, characterized in that, include: Step S1: Determine the target fish species to be evaluated, the breeding season of the target fish species, and the spawning grounds of the fish species in the study section of the river. To determine at least one hydrodynamic parameter that influences fish reproductive behavior and at least one fish reproductive behavior index that reflects fish reproductive behavior; Step S2: Divide the fish spawning grounds of the study river section into several monitoring sections; select several monitoring points for each monitoring section; collect data during the multi-year breeding season of the target fish, and obtain the daily representative vector of each monitoring point during the breeding season of the target fish: including the daily representative value of each hydrodynamic parameter and the daily representative value of the fish breeding suitability assessment index calculated based on the fish breeding behavior index; Step S3: Select typical samples from the daily representative vectors of each monitoring point at each monitoring section to construct a sample set. ={ };in, For sample set The number of samples in the middle; each sample = , , ; For the sample hydrodynamic parameters eigenvalues, The number of hydrodynamic parameters, For the sample Fish reproductive suitability assessment index eigenvalues; Step S4, for the sample set The distribution of each hydrodynamic parameter and the fish reproductive suitability assessment index was mined to obtain the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index. Step S5: Using the Gaussian distribution decomposition algorithm, the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index are decomposed to obtain several state levels of each hydrodynamic parameter and several state levels of the fish reproductive suitability assessment index; each state level has four state level characteristics, namely: expectation, standard deviation, amplitude coefficient and fluctuation coefficient. Step S6: Establish a mapping relationship between the combination of each state level of each hydrodynamic parameter and the state level of the fish reproductive suitability assessment index that appears most frequently under that combination, thereby constructing a reasoning rule base; Step S7: Evaluate the suitability of the fish spawning grounds in the study section for fish reproduction during the breeding season of the target fish based on the reasoning rule base; Step S5 is as follows: Step S5.1: Represent each hydrodynamic parameter and the fish reproductive suitability assessment index as a unified element. ;element The frequency distribution function is expressed as ; Step S5.2, apply a Gaussian distribution to the elements. Frequency distribution function The decomposition yields N Gaussian distribution functions, each possessing four characteristics: expectation, standard deviation, amplitude coefficient, and volatility coefficient. Step S5.2.1, calculate the sample set ={ } elements Standard deviation According to the element Standard deviation Determine the neighborhood length L: 𝐿 = 0.5 ⋅ 𝑚𝑖𝑛(𝑆𝐷, IQR / 1.34) ⋅ 𝑛 −1 / 5 , For elements Interquartile range in sample set S; Step S5.2.2, let the variable =1; Step S5.2.3, let the frequency distribution function =Frequency distribution function ; Step S5.2.4, based on the frequency distribution function Determine the Gaussian distribution function Expectations Standard deviation Amplitude coefficient and volatility coefficient : Take frequency distribution function The x-coordinate corresponding to the peak point of the curve is used as the Gaussian distribution function. Expectations ; Take frequency distribution function The ordinate corresponding to the peak point of the curve is used as the Gaussian distribution function. amplitude coefficient ; In frequency distribution function In the process of collecting sample sets ={ } is located in the expected The neighborhood length L range elements The value is represented as: , for Quantity, The values ​​of the neighborhood samples are expressed using the formula. Obtain the Gaussian distribution function Standard deviation ; Using formula Obtain the Gaussian distribution function Fluctuation coefficient ; Step S5.2.5, using the formula Generate Gaussian distribution function The function expression; Step S5.2.6: Generate the frequency distribution function ; Step S5.2.7, determine the frequency distribution function peak Does it meet the following conditions: , ε is the set stop threshold; β is the control coefficient; If satisfied, proceed to step S5.2.8; otherwise, set the variable... = +1, return to step S5.2.4; Step S5.2.8: Output each Gaussian distribution function obtained, along with its expected value, standard deviation, amplitude coefficient, and volatility coefficient. Step S5.3: Cluster the N Gaussian distribution functions into M clusters; for each cluster, merge all Gaussian distribution functions within the cluster, and the merged Gaussian distribution function is called the Gaussian distribution representative function; thus, a total of M Gaussian distribution representative functions are obtained, each representing an element. A state level, where each Gaussian distribution represents a characteristic of the function, i.e., an element. The characteristics of the corresponding state level.

2. The method for assessing fish reproductive suitability based on a reasoning rule base according to claim 1, characterized in that, The hydrodynamic parameters include one or more of the following: flow velocity, water depth, flow field gradient modulus, water temperature, and substrate type; the fish reproductive behavior indicators include one or more of the following: environmental DNA concentration, instantaneous movement vector of individual fish, and the ratio of the area of ​​healthy fish eggs to the area of ​​the monitoring point.

3. The method for assessing fish reproductive suitability based on a reasoning rule base according to claim 1, characterized in that, Step S2 is as follows: Step S2.1: Data collection during the multi-year breeding season of the target fish species. Several monitoring sections were set up in the fish spawning grounds of the studied river section; several vertical monitoring lines were planned along the transverse direction of the river at each monitoring section; several monitoring points were set up along the vertical direction of each monitoring line, and the monitoring points in each vertical direction covered different water depth levels. Set daily during the target fish's breeding season. During a fixed monitoring period, hydrodynamic parameter data and fish reproductive behavior index data were collected at each monitoring point during each monitoring period, and the fish reproductive behavior index data were converted into a fish reproductive suitability assessment index. Step S2.2: Process the collected data to obtain the daily representative vector for each monitoring point: Each monitoring period is assigned a corresponding weight coefficient based on its importance. For each hydrodynamic parameter, the parameters obtained daily at each monitoring point are... The hydrodynamic parameter data from different monitoring periods are weighted and summed to obtain the daily representative value of each hydrodynamic parameter at that monitoring point on that day. Daily data obtained at each monitoring point The fish reproductive suitability assessment index at different monitoring periods is weighted and summed to obtain the daily representative value of the fish reproductive suitability assessment index at that monitoring point on that day. This yields a daily representative vector for each monitoring point, including the daily representative value of each hydrodynamic parameter and the daily representative value of the fish reproductive suitability assessment index.

4. The method for assessing fish reproductive suitability based on a reasoning rule base according to claim 3, characterized in that, Step S3 is as follows: Step S3.1: For each monitoring section, construct the sample subset corresponding to that monitoring section: Step S3.1.1: For each monitoring section, construct a daily representative vector set consisting of daily representative vectors obtained from all monitoring points during the breeding season of the target fish. Step S3.1.2: Calculate the daily representative value of the fish reproductive suitability assessment index for all monitoring points of each monitoring section during the breeding season of the target fish. Construct the numerical range of the daily representative value of the fish reproductive suitability assessment index based on the minimum and maximum values ​​obtained. Divide this numerical range into several equally spaced suitability intervals. Step S3.1.3: Set the reference sampling quantity ; Step S3.1.4: Count the number of daily representative vectors contained in each suitability interval; for intervals with more than [number missing] daily representative vectors... To determine the suitability range, a daily representative vector optimization method based on K-Means clustering was used for data extraction. A number of daily representative vectors are extracted and retained within the suitability interval, while other daily representative vectors within the suitability interval are deleted. The data extraction method is as follows: all daily representative vectors within the suitability interval are used as clustering input, and clustering is performed as follows: For each sub-cluster, calculate the centroid of each sub-cluster and select the representative vector that is closest to the Euclidean distance of the centroid of each sub-cluster, and keep it in the suitability interval; The number of daily representative vectors does not exceed The suitability interval is determined by directly retaining all daily representative vectors within that suitability interval; In steps S3.1.5 and S3.1.4, the daily representative vector combination retained for each suitability interval is the sample subset corresponding to the monitoring section. Step S3.2: Combine the sample subsets corresponding to all monitoring sections to obtain the sample set. ;in, For sample set The number of samples in the sample; each sample Let be a daily representative vector, denoted as = ; For the sample hydrodynamic parameters The daily representative value is called the hydrodynamic parameter. eigenvalues; For the sample Fish reproductive suitability assessment index The daily representative value is called the fish reproductive suitability assessment index. eigenvalues.

5. The method for assessing fish reproductive suitability based on a reasoning rule base according to claim 1, characterized in that, Step S5.3 specifically includes: Step S5.3.1: Using the expected value of each Gaussian distribution function as the clustering feature parameter, a clustering algorithm is used to cluster the N Gaussian distribution functions into M clusters; Step S5.3.2: For each cluster, merge all Gaussian distribution functions within the cluster. The merged Gaussian distribution function is called the Gaussian distribution representative function. The merging method is as follows: ① Suppose there are C Gaussian distribution functions in the cluster. Calculate the Gaussian distribution function for each Gaussian distribution function. Effective contribution , : In Gaussian distribution function In the middle, determine according to its expectations The sampling interval centered on a preset sampling length. ; These are the sampling start position and sampling end position, respectively; within the sampling interval Within, samples were obtained at equal intervals. Each sampling point , If it follows a Gaussian distribution function ordinate value If the value of the sampling point is greater than the ordinate value of its other C-1 Gaussian distribution functions, then let the sampling point... Effective contribution for Otherwise, let the sampling points Effective contribution =0; Using formula ,calculate The effective contribution of each sampling point is used to obtain the Gaussian distribution function. Effective contribution ; ②Based on the expected value, standard deviation, amplitude coefficient, volatility coefficient, and effective contribution of each Gaussian distribution function within the cluster, calculate the expected value, standard deviation, amplitude coefficient, and volatility coefficient of the representative Gaussian distribution function of the cluster: The expected value of the Gaussian distribution representative function is obtained by weighting the expected value of each Gaussian distribution function within the cluster with the effective contribution of each Gaussian distribution function within the cluster. The fluctuation coefficient of the Gaussian distribution representative function is obtained by weighting the effective contribution of each Gaussian distribution function within the cluster and averaging the fluctuation coefficients of each Gaussian distribution function within the cluster. Calculate the effective contribution / amplitude coefficient of each Gaussian distribution function within the cluster, and then sum the effective contribution / amplitude coefficients of each Gaussian distribution function to obtain the standard deviation of the representative Gaussian distribution function; Calculate the effective contribution / standard deviation of each Gaussian distribution function within the cluster, and then sum the effective contribution / standard deviations of each Gaussian distribution function to obtain the amplitude coefficient of the representative Gaussian distribution function; ③ Generate the functional expression of the Gaussian distribution representative function based on the expected value, standard deviation, amplitude coefficient, and fluctuation coefficient of the Gaussian distribution representative function; ④ Correct the Gaussian distribution representative function obtained in step ③ to obtain the corrected Gaussian distribution representative function. The correction method is as follows: If compared with the Gaussian distribution functions within the cluster, the Gaussian distribution representative function is closest to the element If the boundary of the domain is defined, the curve of the Gaussian distribution representative function inside the domain is retained, and the Gaussian distribution function value between its expected value and the boundary of the domain is corrected to 1, thus transforming it into a Gaussian distribution function with a one-sided plateau characteristic.

6. The method for assessing fish reproductive suitability based on a reasoning rule base according to claim 1, characterized in that, Step S7 is as follows: Step S7.1: Obtain each hydrodynamic parameter of each cell grid of the fish spawning grounds in the study section during the target fish breeding season for the simulated or predicted time period. eigenvalues ; Step S7.2: Determine each hydrodynamic parameter State level and certainty : Assume hydrodynamic parameters There are A types of status levels. Iterate through each state level , Hydrodynamic parameters were calculated. eigenvalues Corresponding to each state level Certainty Each degree of certainty The state level corresponding to the maximum value is the determined hydrodynamic parameter. State level; each degree of certainty The maximum value is the determined hydrodynamic parameter Certainty ; Certainty The formula is expressed as follows: ； in: Status level Expectations; To be based on state level Standard deviation As expected, in terms of state level The square of the fluctuation coefficient Random numbers generated for variance; Step S7.3, based on the determined hydrodynamic parameters Based on the state level and the reasoning rule base, a fish reproductive suitability assessment index is derived. Status level This leads to the state level. Expectations Standard deviation Amplitude coefficient and volatility coefficient ; Step S7.4: Obtain the fish reproductive suitability assessment index according to the following formula. Evaluation value : ； in: For As expected, with volatility coefficient square Random numbers generated for variance; Hydrodynamic parameters eigenvalues Fish reproductive suitability assessment index The directional parameter of action, It is either 1 or -1, and its calculation formula is: ； in: For the determined hydrodynamic parameters The expected state level; Hydrodynamic parameters The optimal value is determined as follows: In the reasoning rule base, all reasoning rules corresponding to the highest state level of the fish reproductive suitability assessment index are found; among all the found reasoning rules, the hydrodynamic parameters are determined. The various state levels, and the hydrodynamic parameters The mean of the expected values ​​of each state level is taken as the hydrodynamic parameter. The optimal value; Step S7.5: Repeat steps S7.2-S7.4 multiple times, and take the evaluation value obtained each time. The average value is used as the final fish breeding suitability assessment index.

7. A fish reproductive suitability assessment system based on a reasoning rule base, characterized in that, A method for assessing fish reproductive suitability based on a reasoning rule base, as described in any one of claims 1 to 6, comprises: The determination module is used to determine the target fish species to be evaluated, the breeding season of the target fish species, and the spawning grounds of the fish species in the study section; and to determine at least one hydrodynamic parameter affecting fish reproductive behavior and at least one fish reproductive behavior index reflecting fish reproductive behavior. The data acquisition module is used to divide the fish spawning grounds of the study river section into several monitoring sections; select several monitoring points for each monitoring section; collect data during the multi-year breeding season of the target fish, and obtain the daily representative vector of each monitoring point during the breeding season of the target fish: including the daily representative value of each hydrodynamic parameter and the daily representative value of the fish breeding suitability assessment index calculated based on the fish breeding behavior index; The sample set construction module is used to select typical samples from the daily representative vectors of each monitoring point at each monitoring section to construct the sample set. ={ };in, For sample set The number of samples in the middle; each sample = , , ; For the sample hydrodynamic parameters eigenvalues, The number of hydrodynamic parameters, For the sample Fish reproductive suitability assessment index eigenvalues; The frequency distribution function construction module is used for the sample set. The distribution of each hydrodynamic parameter and the fish reproductive suitability assessment index was mined to obtain the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index. The frequency distribution function decomposition module is used to decompose the frequency distribution function of each hydrodynamic parameter and the frequency distribution function of the fish reproductive suitability assessment index using the Gaussian distribution decomposition algorithm, to obtain several state levels of each hydrodynamic parameter and several state levels of the fish reproductive suitability assessment index; each state level has four state level characteristics, namely: expectation, standard deviation, amplitude coefficient and fluctuation coefficient. The reasoning rule base construction module is used to establish a mapping relationship between the combination of various state levels of each hydrodynamic parameter and the state level of the fish reproductive suitability assessment index that appears most frequently under that combination, thereby constructing the reasoning rule base; The fish reproduction suitability assessment module is used to assess the fish reproduction suitability of the fish spawning grounds in the study section during the breeding season of the target fish, based on the inference rule base.

Images