A time-scale-based complexity analysis method and system for marine harmful algal blooms

Abstract

The present invention proposes a complexity analysis method and system for harmful algal blooms in the ocean based on time scale, which relates to the technical field of data processing for simulation of harmful algal blooms in the ocean, and targets the following problems: the existing simulation predictions have insufficient descriptions of algae behavior under nonlinear conditions, which makes it impossible to accurately analyze the complex characteristics of harmful algal blooms in the ocean, thereby affecting the stability and accuracy of simulation and prediction. The method more accurately describes the complexity and dynamics of algae growth in the marine environment by constructing a time-scale fractional-order algal bloom model; the constructed fractional-order set is analyzed for complexity using methods such as symmetry index and box-counting dimension, and the occurrence time, range and intensity of harmful algal blooms are simulated based on the analysis results of the model. The present invention solves the problem of insufficient descriptions of algae behavior under nonlinear conditions, realizes accurate complex characteristic analysis and effective prediction of harmful algal blooms in the ocean, and improves the accuracy of simulation and prediction.

Description

Technical Field

[0001] The present invention belongs to the technical field of data processing for simulation of harmful algal blooms in the ocean, and in particular to a method and system for analyzing complexity of harmful algal blooms in the ocean based on a time scale. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Marine Harmful Algal Blooms, also known as water blooms, are a type of disaster that causes damage to aquatic ecosystems due to the rapid proliferation of microscopic algae in water bodies. Marine Harmful Algal Blooms (HABs) are a global marine environmental problem that has a serious impact on marine ecosystems and human activities. In recent years, marine harmful algal blooms have continued to occur frequently in coastal waters, which has had a serious impact on the marine ecological environment along the coast of my country and caused huge economic losses. Therefore, there is an urgent need to effectively predict marine harmful algal blooms, and effective prediction requires accurate analysis of the complex characteristics of marine harmful algal blooms.

[0004] The data processing corresponding to the traditional marine harmful algal bloom simulation method, such as statistical simulation method, has problems such as difficulty in solving and insufficient simulation accuracy. In recent years, with the development of computer technology, simulation methods based on machine learning have begun to be applied to marine harmful algal bloom simulation, such as patent CN116050645A, a method and system for predicting the concentration of marine harmful algae based on deep learning, which uses machine learning methods to simulate and predict marine harmful algal blooms. However, it is only a simulation prediction of the concentration of marine harmful algae, and there is still a problem of insufficient description of algae behavior under nonlinear conditions, which makes it impossible to accurately analyze the complex characteristics of marine harmful algal blooms, thereby affecting the stability and accuracy of simulation and prediction. Summary of the invention

[0005] In order to overcome the deficiencies of the above-mentioned prior art, the present invention provides a method and system for analyzing the complexity of harmful algal blooms in the ocean based on a time scale. By constructing a fractional-order algal bloom model based on a time scale, a fractional-order (q, h)-Julia set family of the model is generated, and the complexity analysis of the fractional-order (q, h)-Julia set family is performed using methods such as symmetry index and box-counting dimension. This solves the problem of inability to accurately describe the nonlinear dynamic behavior of harmful algal blooms in the ocean, achieves accurate analysis and prediction of the complex characteristics of harmful algal blooms in the ocean, and can improve the stability and accuracy of simulation and prediction to a certain extent.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0007] The first aspect of the present invention provides a method for analyzing the complexity of marine harmful algal blooms based on time scale, comprising:

[0008] Acquiring marine harmful algal bloom monitoring data, wherein the marine harmful algal bloom monitoring data at least includes the biomass of phytoplankton and the concentration of nutrients that cause marine harmful algal blooms;

[0009] Preprocessing the marine harmful algal bloom monitoring data to obtain preprocessed data;

[0010] The preprocessed data were input into the time scale fractional algal bloom model to conduct complexity analysis of marine harmful algal blooms;

[0011] Among them, the complexity analysis of marine harmful algal blooms is as follows:

[0012] The preprocessed data are input into the time scale fractional algal bloom model to determine the model parameters;

[0013] According to the model parameters, determine the current fractional order set;

[0014] Calculate the indicator parameters according to the current fractional order set;

[0015] The index parameters are compared with the threshold range of index parameters of the time scale fractional algal bloom model to obtain the results of fractional set complexity analysis;

[0016] Based on the results of fractional-order set complexity analysis, simulation and early warning of harmful algal blooms in the ocean are carried out.

[0017] As an implementation method, the construction process of the time scale fractional algal bloom model is as follows:

[0018] Based on the nutrient-phytoplankton population evolution model, the generation system of the traditional fractional order set is obtained;

[0019] Based on the generation system and time scale of traditional fractional-order sets, a time-scale fractional-order algal bloom model is constructed.

[0020] As an implementation mode, the marine harmful algal bloom monitoring data is preprocessed to obtain preprocessed data, specifically: data cleaning is performed on the marine harmful algal bloom monitoring data to remove abnormal values ​​in the data;

[0021] The cleaned data is standardized to obtain a time series.

[0022] As an implementation method, after the time scale fractional algal bloom model is constructed, a fractional cluster is generated, and the complexity analysis of marine harmful algal blooms is performed based on the fractional cluster to determine the threshold range of indicator parameters. The specific process is as follows:

[0023] The event analysis method is used to determine the perturbed fractional order clusters for the time scale fractional algal bloom model.

[0024] The complexity analysis of marine harmful algal blooms is carried out based on the symmetry index and box-counting dimension for fractional set families.

[0025] Determine the indicator parameter threshold range based on the complexity analysis results.

[0026] As an implementation method, the time scale fractional algal bloom model is subjected to event analysis to determine the perturbed fractional order cluster, specifically:

[0027] Performing data-valued processing on the fractional-order difference operator to obtain a data-valued fractional-order difference operator;

[0028] The numerically quantified fractional-order difference operator is mapped in the model to obtain a fractional-order recursive mapping;

[0029] By performing event tree analysis on the properties of the fractional-order recursive mapping, we obtain the perturbed fractional-order family of the model.

[0030] As an implementation method, the preprocessed data is input into a time-scale fractional-order algal bloom model to determine the model parameters. Specifically, the model parameters are determined through a multi-objective neural network algorithm.

[0031] As an implementation method, the model parameters are determined by a multi-objective neural network algorithm, and the specific process is as follows:

[0032] Build a multi-layer neural network;

[0033] Based on multi-layer neural network, use Huber function to construct loss function;

[0034] A set of model parameters is obtained by solving the minimum value of the loss function.

[0035] A second aspect of the present invention provides a system for analyzing the complexity of harmful algal blooms in the ocean based on a time scale, comprising:

[0036] A data acquisition module, used to acquire marine harmful algal bloom monitoring data, wherein the marine harmful algal bloom monitoring data at least includes the biomass of phytoplankton and the concentration of nutrients that cause marine harmful algal blooms;

[0037] A data processing module, used for preprocessing the marine harmful algal bloom monitoring data to obtain preprocessed data;

[0038] The marine harmful algal bloom complexity analysis and early warning module is used to input the preprocessed data into the time scale fractional algal bloom model to conduct complexity analysis of marine harmful algal blooms;

[0039] Among them, the complexity analysis of marine harmful algal blooms is as follows:

[0040] The preprocessed data are input into the time scale fractional algal bloom model to determine the model parameters;

[0041] According to the model parameters, determine the current fractional order set;

[0042] Calculate the indicator parameters according to the current fractional order set;

[0043] The index parameters are compared with the threshold range of index parameters of the time scale fractional algal bloom model to obtain the results of fractional set complexity analysis;

[0044] Based on the results of fractional-order set complexity analysis, simulation and early warning of harmful algal blooms in the ocean are carried out.

[0045] A third aspect of the present invention provides a computer device, comprising a memory, a processor, and a program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method described in the first aspect of the present invention are implemented.

[0046] The fourth aspect of the present invention aims to provide a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in the method described in the first aspect of the present invention.

[0047] One or more of the above technical solutions have the following beneficial effects:

[0048] In this embodiment, a time scale fractional algal bloom model is constructed, and the time scale generated by the model is used to generate Analyzing the fractal dynamics of the fractional-order (q,h)-Julia set on the surface can more accurately describe the nonlinear dynamic behavior of marine harmful algal blooms, and can better describe the complexity and dynamics of the growth of marine harmful algal blooms in the marine environment, thereby realizing accurate modeling and prediction of marine harmful algal blooms.

[0049] In this embodiment, the complexity analysis of the fractional-order (q, h)-Julia set using methods such as symmetry index and box-counting dimension can reveal the complexity characteristics of marine harmful algal blooms and provide a more accurate basis for simulation and prediction.

[0050] In this embodiment, based on the time scale The complexity analysis method of marine harmful algal blooms characterized by the fractional-order (q,h)-Julia set on the surface has the ability to resist disturbances and can improve the stability and accuracy of simulation and prediction to a certain extent.

[0051] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0053] Figure 1 This is a process flow chart of the time-scale-based marine harmful algal bloom complexity analysis method of the first embodiment;

[0054] Figure 2 is the perturbed fractional-order (q, h)-Julia set generated in the first embodiment;

[0055] Among them, (a) is , , The fractional (q,h)-Julia set of , (b) is , , The fractional order (q,h)-Julia set of , , The fractional (q,h)-Julia set of , (d) is , , The fractional order (q,h)-Julia of (e) is , , The fractional (q,h)-Julia set of , (f) is , , Fractional (q,h)-Julia sets of ;

[0056] Figure 3 is the symmetric index generated in the first embodiment of the present invention;

[0057] Among them, (a) is the evolution of the symmetry index with respect to the algebraic parameter, (b) is the evolution of the symmetry index with respect to the geometric parameter, and (c) is the evolution of the symmetry index with respect to the memory parameter;

[0058] Figure 4 is the box counting dimension calculated in the first embodiment;

[0059] Among them, (a) is the evolution of dimension with respect to algebraic parameters, (b) is the evolution of dimension with respect to geometric parameters, and (c) is the evolution of dimension with respect to memory parameters. DETAILED DESCRIPTION

[0060] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0061] It should be noted that the terms used herein are for describing specific embodiments only and are not intended to be limiting of exemplary embodiments according to the present invention.

[0062] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0063] Explanation of the name:

[0064] Time scale is any non-empty closed subset of real numbers, belonging to the category of mathematical analysis. It was proposed by Stefan Hilger in 1988. It is a theory that unifies and expands different theories such as differential equations, difference equations, and q-difference equations into "dynamic equations". The time scale involved in this invention , which is a special time scale that combines both geometric and algebraic scales.

[0065] A fractional order system is a dynamic system described by a non-integer order (discrete) calculus equation, which has the characteristics of memory, nonlinearity and complexity. Common fractional order operators include Riemann-Liouville type, Caputo type and Grümwald-Letnikov type. The Riemann-Liouville type fractional order (q, h)-system involved in the present invention is a special type of discrete fractional order system.

[0066] The classical Julia set is composed of quadratic mappings One of the generated fractal sets was proposed by French mathematician Gaston Julia in 1918. In 1979, French-American mathematician Benoit Mandelbrot studied the Julia set with the help of computers, classified all possible shapes, and proposed another famous fractal set, the Mandelbro set. The fractional order (q, h)-Julia set involved in the present invention is a generalization of the classic Julia set.

[0067] The Escape-time Algorithm (ETA) is one of the earliest computer graphics coloring algorithms. The algorithm classifies and colors points based on the number of iterations required for the system state trajectory sequence to tend to infinity.

[0068] Box-counting Dimension, also known as Minkowski Dimension, is a concept of fractal dimension that measures how the complexity of fractal details changes with the scale at which the fractal is observed. for dimensional vector space Any non-empty bounded subset of , To cover and the diameter does not exceed The minimum number of sets. The box-counting dimension of is defined as The larger the fractal dimension of a fractal set, the more self-similar the fractal is and the higher its complexity.

[0069] The overall idea proposed in the present invention is: by constructing a time-scale fractional algal bloom model, a complexity analysis is performed on the marine harmful algal bloom fractional (q, h)-Julia set family, and according to the complexity analysis results, the threshold range of indicator parameters (symmetry index and box counting dimension) is determined, and then the marine harmful algal bloom monitoring data to be analyzed is input into the time-scale fractional algal bloom model to obtain the indicator parameters, and the indicator parameters are compared with the indicator parameter threshold range to determine whether they are within the indicator parameter threshold range, and an early warning prompt is issued.

[0070] Embodiment 1

[0071] This embodiment discloses a method for analyzing the complexity of marine harmful algal blooms based on time scale.

[0072] In order to more clearly illustrate this embodiment, the implementation process of a method for analyzing the complexity of marine harmful algal blooms based on time scale can be specifically described as follows:

[0073] A time-scale-based complexity analysis method for marine harmful algal blooms, including:

[0074] S1. Acquire marine harmful algal bloom monitoring data, wherein the marine harmful algal bloom monitoring data at least includes the biomass of phytoplankton and the concentration of nutrients that cause marine harmful algal blooms;

[0075] S2. preprocessing the marine harmful algal bloom monitoring data to obtain preprocessed data;

[0076] S3, input the preprocessed data into the time scale fractional algal bloom model to conduct complexity analysis of marine harmful algal blooms;

[0077] Among them, the complexity analysis of marine harmful algal blooms is as follows:

[0078] The preprocessed data are input into the time scale fractional algal bloom model to determine the model parameters;

[0079] According to the model parameters, determine the current fractional order set;

[0080] Calculate the indicator parameters according to the current fractional order set;

[0081] Compare the indicator parameters with the indicator parameter threshold range to obtain the fractional order set complexity analysis results;

[0082] Based on the results of fractional-order set complexity analysis, simulation and early warning of harmful algal blooms in the ocean are carried out.

[0083] like Figure 1 As shown, in step S1, marine harmful algal bloom monitoring data is obtained.

[0084] In this embodiment, the marine harmful algal bloom monitoring data includes at least the biomass of phytoplankton and the concentration of nutrients that cause the marine harmful algal bloom.

[0085] Utilize sensors carried by ocean buoys and satellite remote sensing technology to monitor and collect data on the ocean's physical and chemical environment that are closely related to harmful algal blooms, including but not limited to: monitoring the biomass of phytoplankton that causes harmful algal blooms, and collecting concentration data on nutrients (such as nitrogen, phosphorus, etc.). Excessive nutrient input is one of the main causes of harmful algal blooms.

[0086] Through the above steps, data closely related to harmful algal blooms in the ocean were collected, ensuring the validity of input data for subsequent model construction.

[0087] like Figure 1 As shown, in step S2, the marine harmful algal bloom monitoring data is preprocessed to obtain preprocessed data.

[0088] In this embodiment, data cleaning is first performed to remove outliers in the data, and then the data is standardized.

[0089] Specifically, through data cleaning, we can identify and deal with logical errors and inconsistencies in the data, such as null values ​​and redundant data. We can identify and eliminate outliers in the data through range testing, peak testing, and the Rhineda criterion.

[0090] Use unified standards for formatting and standardization to ensure consistency and accuracy of subsequent analysis.

[0091] After preprocessing, the time series of phytoplankton biomass (density) and nutrient level (concentration) are obtained, and the formula is:

[0092] (1)

[0093] in, for The data value at the moment, and Respectively Phytoplankton biomass and nutrient levels at each moment.

[0094] After the above steps, the time series data of marine harmful algal blooms were obtained, providing a data basis for the model.

[0095] like Figure 1 As shown, in step S3, the preprocessed data is input into the time scale fractional algal bloom model to perform complexity analysis of harmful algal blooms.

[0096] In this embodiment, S31, construct a time scale fractional algal bloom model.

[0097] S311. Based on the nutrient-phytoplankton population evolution model, the generation system of the traditional fractional order set is obtained.

[0098] In this embodiment, the classic algal bloom discrete dynamics model is adopted, and the formula is:

[0099] (2)

[0100] Model formula (2) includes only two variables: nutrient level and phytoplankton biomass , and for and The difference is expressed as and Suppose a small amount of nutrient salt changes at a slow but constant rate. The system takes into account the enrichment effect of nutrients themselves and their consumption by phytoplankton, which absorb nutrients to grow and are removed from the water column by dying and sinking.

[0101] make , , and substitute it into the formula (2) according to the classic algal bloom discrete dynamics model to obtain the generation system of the traditional Julia set, the formula is:

[0102] (3)

[0103] in, For the The coupled state variables of phytoplankton biomass and nutrient levels at the first iteration are It is the system reduced parameter, which represents the comprehensive environmental effect that characterizes the system.

[0104] Considering complex variables The purpose is to and The two state variables are coupling.

[0105] S312. Based on the generation system and time scale of the traditional fractional-order set, a time-scale fractional-order algal bloom model is constructed.

[0106] Introducing the time scale and fractional order parameters into formula (3), the time scale fractional order algal bloom model is constructed, and the formula is:

[0107] (4)

[0108] in, is the state variable of marine harmful algal blooms, for The Riemann-Liouville-type fractional-order (q,h)-difference operator is defined as

[0109] , (5)

[0110] parameter To characterize the memory of the system, the parameter Expressed as the geometric scale of the system, the parameter Expressed as an algebraic scale that characterizes the system, for The fractional order back-jump operator, for The fractional forward jump operator, ,and That is not less than The smallest positive integer.

[0111] After the above steps, a time-scale fractional-order algal bloom model was constructed. The introduction of two key elements, time scale and fractional-order parameters, endowed the model with higher flexibility and adaptability. It not only considered the time dependence of algal growth, but also incorporated fractional-order operator theory to effectively capture the nonlinear dynamic characteristics of algal growth.

[0112] S32. Conduct complexity analysis of harmful algal blooms in the ocean based on fractional-order clusters and determine the threshold range of indicator parameters.

[0113] S321. The event analysis method is used to determine the perturbed fractional-order clusters for the time-scale fractional algal bloom model.

[0114] (1) Perform data value processing on the fractional-order difference operator to obtain the data-valued fractional-order difference operator.

[0115] In this embodiment, according to the definition and properties of formula (5), formula (5) is numerically processed and the formula is:

[0116] (6)

[0117] in, for forward jump operator, and .

[0118] (2) The digitized fractional-order difference operator is mapped in the model to obtain a fractional-order recursive mapping.

[0119] Specifically, substitute formula (6) into formula (4) and get the fractional-order recursive mapping as follows:

[0120] (7)

[0121] in, Indicates The coupled state variables of phytoplankton biomass and nutrient levels at the first iteration are Indicates The scale shift during the step iteration, is the kernel function of the fractional-order (q, h)-mapping, characterizing the mapping memory, which is a unique property of the fractional-order (q, h)-mapping; ,and .

[0122] (3) Perform event tree analysis on the properties of the fractional-order recursive mapping and obtain the perturbed fractional-order family of the model.

[0123] By performing event tree analysis on the properties of the fractional-order recursive mapping, the perturbed fractional-order recursive mapping of the model is obtained, and from this, the fractional-order (q,h)-Julia set family is calculated.

[0124] Since there are still uncertainties and random factors in the real environment, noise following uniform distribution is introduced into the mapping formula (7) to obtain a perturbed fractional-order (q, h)-mapping, which is:

[0125] (8)

[0126] in, , , They are , and The correction after introducing additive noise is , , , To obey The noise is uniformly distributed on , .

[0127] In the above steps, the noise contained in the fractional-order (q, h)-mapping (8) constitutes the internal uncertainty of the system and reflects the inaccuracy of modeling.

[0128] In this embodiment, the fractional-order (q, h)-Julia family of model formula (4) in a noisy environment can be obtained through ETA event tree analysis, which is denoted as .

[0129] Specifically: 1) Initialize complex variables .

[0130] The coordinates of each pixel on the screen with a grid range of [-1.5, 2.5] × [-2.0, 2.0] and a grid resolution of 1500 correspond to complex variables. The two components of the complex variable One-to-one correspondence with grid points.

[0131] 2) The complex variable Substitute into formula (7) and iterate. In each iteration, update The value of .

[0132] 3) After each iteration, check the complex variables Is the modulus (i.e., the distance from the origin) greater than a preset sufficiently large value, i.e., the escape radius (usually 10^10)? If the modulus is greater than the escape radius, then the point is considered has "escaped" and the number of iterations is recorded.

[0133] 4) Color the pixels according to the number of iterations.

[0134] The number of iterations can be mapped to color values, generating colored Julia set images. Typically, escaped pixels are assigned a color based on the number of iterations before escaping, while points belonging to the Julia set are colored maroon.

[0135] 5) Traverse all pixels on the screen, repeat the above process, and finally generate the image of the Julia set.

[0136] After the above steps, the visualization of the fractional-order (q,h)-Julia set family is realized. Figure 2 is a perturbed fractional (q,h)-Julia family In the figure, the cluster is represented by the maroon part, and the escape velocity of the initial point is characterized by the color.

[0137] S322. Complexity analysis of marine harmful algal blooms based on the symmetry index and box-counting dimension for fractional-order sets.

[0138] Symmetry and dimension are two important indicators to measure the complexity of fractal sets, especially Julia sets, which can quantify the complex geometric properties with self-similar properties. At present, the box-counting dimension can be used as an approximation of the fractal dimension, which can easily estimate the dimension of the Julia set; however, there is no unified measurement tool for the characterization of the symmetry of the Julia set. Therefore, the symmetry index of the fractional order (q,h)-Julia set family is first defined.

[0139] (1) Based on the fractional-order family, calculate the symmetry index of the fractional-order (q,h)-Julia family on the horizontal axis.

[0140] In this embodiment, the disturbance will affect the cluster The structure includes symmetry, and symmetry reflects the spatial complexity of the family of sets, so it is necessary to give a symmetry index to quantitatively analyze the symmetry of the family of fractional-order (q,h)-Julia sets.

[0141] Specifically, let the background space be , the two parts divided by the horizontal axis can be expressed as

[0142] (9)

[0143] Among them, point is the initial point of mapping (8).

[0144] For any two initial points symmetric about the horizontal axis and , according to the definition of the fractional (q,h)-Julia family, the symmetry between two points about the horizontal axis is defined as:

[0145] (10)

[0146] Symmetric index of fractional (q,h)-Julia sets about the horizontal axis , the calculation formula is:

[0147] (11)

[0148] in, Represents a family The cardinality of .

[0149] After the above steps, we can get the symmetric index of the fractional (q,h)-Julia set family , providing a data basis for subsequent complexity analysis.

[0150] (2) According to the box counting method, the dimension of the family of fractional-order (q,h)-Julia sets is obtained.

[0151] In this embodiment, the dimension of the fractional (q, h)-Julia set family is obtained by the box counting dimension formula: , the formula is:

[0152] (12)

[0153] in, represents different grid sizes, represents the family of fractional (q,h)-Julia sets Minimum number of meshes required.

[0154] After the above steps, we can get the dimension of the fractional (q,h)-Julia set family , providing a data basis for subsequent complexity analysis.

[0155] (3) Based on the calculated symmetry index and dimension , judge and analyze memory parameters , geometric parameters and algebraic parameters Symmetry index for indicator parameters and dimension impact.

[0156] Using the control variable method, we examine the relationship between the memory parameters of the fractional (q,h)-Julia set family in a noisy environment. , geometric parameters and algebraic parameters When testing one of the parameters, the other two parameters are fixed at 0.5 and the process is repeated. The symmetry index and dimension The evolution of Figure 3 and Figure 4 The red dotted line indicates and The average value in the range [0,2].

[0157] Specifically, 1) for memory parameters When performing analysis, , Fixed to 0.5;

[0158] when When it approaches 0, the dimension and symmetry index All experienced a sudden increase phase, namely The increase in the impact of the dimension and symmetry index ;

[0159] On the contrary, for the dimension and symmetry index The impact is relatively small.

[0160] 2) Geometric parameters When performing analysis, , Fixed to 0.5;

[0161] when Within [0,0.95], and The average values ​​were maintained at 0.9997 and 1.8676 respectively, which were almost unaffected by The impact of

[0162] when When it increases to [0.95,1.07], and The value of drops sharply, indicating that the complexity of the fractional (q,h)-Julia set family increases significantly;

[0163] when When it increases to [1.07,2], and All become 0, and the set family degenerates into an empty set; close to (corresponding to the case of a continuous system), the family The dynamic behavior of the system shows extremely high complexity and its robustness is relatively weak.

[0164] 3) Algebraic parameters When performing analysis, , Fixed to 0.5;

[0165] when In [0,0.65], and There was no significant effect, with increase, and Basically negatively correlated, It exhibits chaotic behavior, but the overall amplitude is very small, and the maximum peak-to-valley difference is only .

[0166] After the above steps, the memory parameters can be obtained , geometric parameters , algebraic parameters Dimension and symmetry index The impact of The greatest impact.

[0167] (4) According to the memory parameters , geometric parameters , algebraic parameters Symmetric index to indicator parameters and dimension impact analysis and complexity analysis of marine harmful algal blooms.

[0168] 1) According to Determine when harmful algal blooms will occur in the ocean.

[0169] when When it is close to 1, the group The dynamic behavior of the ocean environment shows great complexity and fragility, that is, the instability of the ocean environment increases under certain conditions (such as when the delicate balance of factors such as temperature, salinity, and light is broken).

[0170] 2) According to the dimension Determine the extent of harmful algal blooms in the ocean.

[0171] Although right The impact is not obvious, but As a group A dimensionality measurement can reflect the range and complexity of its spatial distribution to a certain extent. Increase leads to clustering As complexity increases, if The sharp decline in the number of algal blooms is associated with the spatial expansion of harmful algal blooms in the ocean, so it can be monitored to predict the extent of algal blooms.

[0172] 3) According to the symmetry index Determine the intensity of harmful algal blooms in the ocean.

[0173] The chaotic behavior and sudden changes in algal growth reflect the dramatic fluctuations in marine environments, which are associated with the intensity of marine harmful algal blooms. The overall amplitude is small, but its sudden increase or decrease indicates a sharp change in algal biomass.

[0174] S323. Determine the indicator parameter threshold range based on the complexity analysis result.

[0175] 1) When Near 1:00 a.m., harmful algal blooms may occur in the ocean. Close to 1 is a warning signal of the time when harmful algal blooms may occur in the ocean. The threshold range is , when the model predicts When the value enters this range, a time warning is triggered, indicating that harmful algal blooms may occur.

[0176] 2) Through monitoring to predict the extent of algal blooms, Threshold , when the model predicts When the value falls below this threshold, a range warning is triggered, indicating that the scope of harmful algal blooms in the ocean may be expanding.

[0177] 3) Despite The overall amplitude is small, but its sudden increase or decrease may indicate a sharp change in algal biomass. , that is, the change exceeds a small threshold, when the model predicts When the value changes significantly, a severe warning is triggered, indicating that the intensity of harmful algal blooms may be increasing.

[0178] After the above steps, the complexity analysis of marine harmful algal blooms was carried out according to geometric parameters, symmetry index and dimension. According to the results of complexity analysis, the threshold range of indicator parameters was obtained, the time, scope and intensity of marine harmful algal blooms were determined, and the accurate complexity analysis of marine harmful algal blooms was achieved.

[0179] S33. Input the preprocessed data into the time scale fractional algal bloom model to determine the model parameters.

[0180] Through the above operations, the complexity analysis of various parameters of the time scale fractional algal bloom model was obtained. Through the multi-objective neural network algorithm, according to the time series formula (1) of phytoplankton biomass (density) and nutrient level (concentration), the parameters , , and Perform parameter identification and determine the model parameters. The details are as follows:

[0181] (1) Design a multi-layer neural network.

[0182] Design a multi-layer neural network ,in is the output of the neural network, is a vector parameter containing weights and biases. is an unknown vector parameter of model (6).

[0183] (2) The Huber function is used to construct the loss function.

[0184] Huber function, the formula is:

[0185] (13)

[0186] in, Represents the true value With the predicted value The difference between.

[0187] The loss function is constructed according to formula (13):

[0188] (14)

[0189] in, It is Status The estimated value of It is a tuning parameter.

[0190] (3) A set of model parameters is obtained by solving the minimum value of the loss function.

[0191] Solve the optimization problem (14) and find the estimated value of the vector parameter that minimizes the loss function of problem (14). The formula is:

[0192] (15)

[0193] in, and Represents vector parameters and vector parameters The estimated value of .

[0194] According to this method, a set of fractional algal bloom model parameters that meet the current data on the actual time scale are obtained. .

[0195] S34, determining the current fractional order set according to the model parameters;

[0196] Substituting the identified parameters into the fractional-order (q, h)-mapping (8), the current fractional-order (q, h)-Julia set can be obtained by the ETA algorithm, which is denoted as .

[0197] S35. Calculate the index parameters according to the current fractional order set;

[0198] The obtained fractional (q,h)-Julia set Substituting the previously defined symmetry index and dimension , that is, the calculated fractional-order (q,h)-Julia set The symmetry index and dimension .

[0199] S36. Compare the indicator parameters with the indicator parameter threshold range to simulate and warn of harmful algal blooms in the ocean.

[0200] (1) Compare the indicator parameter with the indicator parameter threshold range to obtain the fractional order (q,h)-Julia set Harmful operations complexity analysis results.

[0201] According to the geometric parameters Is it in the interval , it can be predicted whether harmful algal blooms will occur in the ocean. Falling in the range , then harmful algal blooms may occur in the ocean. On this basis, according to the dimension of the fractional (q,h)-Julia set and symmetry index , predict the extent of algal blooms and fluctuations in algal biomass, , the extent of harmful algal blooms in the oceans has expanded; , the fluctuation intensity of harmful algal blooms in the ocean has increased.

[0202] (2) Based on the results of the complexity analysis of harmful operations of fractional sets, simulate and warn of harmful algal blooms in the ocean.

[0203] when In the interval , issuing early warnings that harmful algal blooms may occur in the ocean.

[0204] Then, when , issued early warning alerts, and the scope of harmful algal blooms in the ocean expanded.

[0205] when , issued early warning alerts, and the fluctuation intensity of harmful algal blooms in the ocean increased.

[0206] Through the above steps, while realizing the complexity analysis of harmful algal blooms in the ocean, it also realizes the effective prediction of harmful algal blooms in the ocean, issues early warning for harmful algal blooms in the ocean, and takes corresponding measures in advance, which can reduce the damage caused by the outbreak of harmful algal blooms in the ocean.

[0207] Embodiment 2

[0208] The purpose of this embodiment is to provide a system for analyzing the complexity of harmful algal blooms in the ocean based on time scale, including:

[0209] A data acquisition module, used to acquire marine harmful algal bloom monitoring data, wherein the marine harmful algal bloom monitoring data at least includes the biomass of phytoplankton and the concentration of nutrients that cause marine harmful algal blooms;

[0210] A data processing module, used for preprocessing the marine harmful algal bloom monitoring data to obtain preprocessed data;

[0211] The marine harmful algal bloom complexity analysis and early warning module is used to input the preprocessed data into the time scale fractional algal bloom model to conduct complexity analysis of marine harmful algal blooms;

[0212] Among them, the complexity analysis of marine harmful algal blooms is as follows:

[0213] The preprocessed data are input into the time scale fractional algal bloom model to determine the model parameters;

[0214] According to the model parameters, determine the current fractional order set;

[0215] Calculate the indicator parameters according to the current fractional order set;

[0216] The index parameters are compared with the threshold range of index parameters of the time scale fractional algal bloom model to obtain the results of fractional set complexity analysis;

[0217] Based on the results of fractional-order set complexity analysis, simulation and early warning of harmful algal blooms in the ocean are carried out.

[0218] Based on providing a time-scale-based marine harmful algal bloom complexity analysis system, the method steps in Example 1 are implemented.

[0219] Embodiment 3

[0220] The purpose of this embodiment is to provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the program.

[0221] Embodiment 4

[0222] The purpose of this embodiment is to provide a computer-readable storage medium.

[0223] A computer-readable storage medium stores a computer program, which executes the steps of the above method when executed by a processor.

[0224] Embodiment 5

[0225] The purpose of this embodiment is to provide a computer program product containing instructions, which, when running on a computer, enables the computer to execute the methods and functions involved in any of the above embodiments.

[0226] The steps involved in the apparatus of the above embodiment correspond to the method embodiment 1, and the specific implementation method can refer to the relevant description part of embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood to include any medium that can store, encode or carry an instruction set for execution by a processor and enable the processor to execute any method in the present invention.

[0227] Those skilled in the art should understand that the modules or steps of the present invention described above can be implemented by a general-purpose computer device, or alternatively, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. The present invention is not limited to any specific combination of hardware and software.

[0228] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A time-scale-based analysis method for the complexity of marine harmful algal blooms, characterized in that: include: Acquiring marine harmful algal bloom monitoring data, wherein the marine harmful algal bloom monitoring data at least includes the biomass of phytoplankton and the concentration of nutrients that cause marine harmful algal blooms; Preprocessing the marine harmful algal bloom monitoring data to obtain preprocessed data; The preprocessed data were input into the time scale fractional algal bloom model to conduct complexity analysis of marine harmful algal blooms; Among them, the complexity analysis of marine harmful algal blooms is as follows: The preprocessed data are input into the time scale fractional algal bloom model to determine the model parameters; According to the model parameters, determine the current fractional order set; Calculate the indicator parameters according to the current fractional order set; The index parameters are compared with the threshold range of index parameters of the time scale fractional algal bloom model to obtain the results of fractional set complexity analysis; Based on the results of fractional set complexity analysis, simulate and warn of harmful algal blooms in the ocean; The construction process of the time scale fractional algal bloom model is as follows: Based on the nutrient-phytoplankton population evolution model, the generation system of the traditional fractional order set is obtained; Based on the generation system and time scale of traditional fractional order sets, a time scale fractional algal bloom model is constructed; The time scale and fractional order parameters are introduced to construct the time scale fractional order algal bloom model. The formula is: ； in, is the state variable of marine harmful algal blooms, for The Riemann-Liouville-type fractional-order (q,h)-difference operator is defined as: ， ； parameter To characterize the memory of the system, the parameter Expressed as the geometric scale of the system, the parameter Expressed as an algebraic scale that characterizes the system, for The fractional order back-jump operator, for The fractional forward jump operator, ,and That is not less than The smallest positive integer.

2. A method for analyzing complexity of marine harmful algal blooms based on time scale as claimed in claim 1, characterized in that: Preprocessing the marine harmful algal bloom monitoring data to obtain preprocessed data, specifically: cleaning the marine harmful algal bloom monitoring data to remove abnormal values ​​in the data; The cleaned data is standardized to obtain time series data.

3. A method for analyzing complexity of marine harmful algal blooms based on time scale as claimed in claim 1, characterized in that: After the time scale fractional algal bloom model is constructed, a fractional cluster is generated, and the complexity analysis of marine harmful algal blooms is carried out based on the fractional cluster to determine the threshold range of indicator parameters. The specific process is as follows: The event analysis method is used to determine the perturbed fractional order clusters for the time scale fractional algal bloom model. The complexity analysis of marine harmful algal blooms is carried out based on the symmetry index and box-counting dimension for fractional set families. Determine the indicator parameter threshold range based on the complexity analysis results.

4. A method for analyzing complexity of marine harmful algal blooms based on time scale as claimed in claim 1, characterized in that: The event analysis method is used to determine the perturbed fractional-order clusters for the time-scale fractional algal bloom model, specifically: Performing data-valued processing on the fractional-order difference operator to obtain a data-valued fractional-order difference operator; The numerically quantified fractional-order difference operator is mapped in the model to obtain a fractional-order recursive mapping; By performing event tree analysis on the properties of the fractional-order recursive mapping, we obtain the perturbed fractional-order family of the model.

5. The method for analyzing complexity of marine harmful algal blooms based on time scale according to claim 1, characterized in that: The preprocessed data is input into the time scale fractional algal bloom model to determine the model parameters. Specifically, the model parameters are determined through a multi-objective neural network algorithm.

6. A method for analyzing complexity of marine harmful algal blooms based on time scale as claimed in claim 5, characterized in that: The model parameters are determined through the multi-objective neural network algorithm. The specific process is as follows: Build a multi-layer neural network; Based on multi-layer neural network, use Huber function to construct loss function; A set of model parameters is obtained by solving the minimum value of the loss function.

7. A time-scale-based marine harmful algal bloom complexity analysis system, characterized in that: Implementing a time-scale-based marine harmful algal bloom complexity analysis method as described in any one of claims 1 to 6, comprising: A data acquisition module, used to acquire marine harmful algal bloom monitoring data, wherein the marine harmful algal bloom monitoring data at least includes the biomass of phytoplankton and the concentration of nutrients that cause marine harmful algal blooms; A data processing module, used for preprocessing the marine harmful algal bloom monitoring data to obtain preprocessed data; The marine harmful algal bloom complexity analysis and early warning module is used to input the preprocessed data into the time scale fractional algal bloom model to conduct complexity analysis of marine harmful algal blooms; Among them, the complexity analysis of marine harmful algal blooms is as follows: The preprocessed data are input into the time scale fractional algal bloom model to determine the model parameters; According to the model parameters, determine the current fractional order set; Calculate the indicator parameters according to the current fractional order set; The index parameters are compared with the threshold range of index parameters of the time scale fractional algal bloom model to obtain the results of fractional set complexity analysis; Based on the results of fractional-order set complexity analysis, simulation and early warning of harmful algal blooms in the ocean are carried out.

8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method described in any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 6 are performed.

Images