CN111414927A - Method for evaluating seawater quality - Google Patents
Method for evaluating seawater quality Download PDFInfo
- Publication number
- CN111414927A CN111414927A CN201910009673.6A CN201910009673A CN111414927A CN 111414927 A CN111414927 A CN 111414927A CN 201910009673 A CN201910009673 A CN 201910009673A CN 111414927 A CN111414927 A CN 111414927A
- Authority
- CN
- China
- Prior art keywords
- individual
- neural network
- selecting
- class
- mode
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/18—Water
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Chemical & Material Sciences (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Immunology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Food Science & Technology (AREA)
- Medicinal Chemistry (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a method for evaluating seawater quality, which comprises the steps of obtaining an initial optimal value of a brainstorming, matching an optimal value of a BP neural network and verifying an optimal result. The initial value of the brain storm optimization is obtained, and the brain storm algorithm in the swarm intelligence optimization algorithm is adopted, so that the problem that the calculation of the weight value and the threshold value of the BP neural network is complex and difficult is solved. And (3) optimizing value matching and optimizing result verification of the BP neural network, constructing a seawater quality evaluation model by adopting the optimized BP neural network, training the model by using historical case data, and finishing the training when the prediction error is within an allowable range or the network training reaches the maximum iteration times. And (4) analyzing a prediction result, evaluating the seawater quality by using a trained model, wherein the root mean square error of the evaluation result is smaller than that of the traditional BP neural network before optimization, and the evaluation precision is higher. The invention can provide a new solution for seawater quality evaluation.
Description
Technical Field
The invention relates to the technical field of seawater quality monitoring, in particular to a method for evaluating seawater quality.
Background
With the continuous development of marine economy, the marine environment is under increasing pressure. The basis of seawater quality evaluation is to comprehensively evaluate the water quality level according to the value of the seawater quality index so as to determine the main pollution problem of the evaluation water area and provide a basis for environmental management and decision of coastal water areas.
A plurality of evaluation methods are researched and developed by domestic and foreign environment workers, such as a single index method, a fuzzy comprehensive evaluation method, a fuzzy clustering method, a gray clustering method and the like, the marine water body is a comprehensive system with a large number of nonlinear and non-stable problems, the seawater pollution is influenced by a plurality of factors, the pollutants have a complex nonlinear relation, and the methods have the problems of limited evaluation modes, rough evaluation results and the like and can not meet the requirements of the current seawater quality evaluation.
Disclosure of Invention
The invention aims to provide a seawater quality evaluation method, which has the characteristic of accurate evaluation result and solves the problems of rough evaluation result and the like of the traditional seawater quality evaluation method.
In order to achieve the purpose, the invention provides the following technical scheme: a method for evaluating seawater quality comprises two stages of brainstorming optimization and BP neural network calculation, wherein the brainstorming optimization stage comprises the steps of acquiring a brainstorming optimization initial value and iteratively updating in an optimization process, and the BP neural network calculation stage comprises the steps of optimizing value matching and optimizing value result verification.
The method for evaluating the seawater quality comprises the following steps of a first storm optimization stage of the seawater quality evaluation method, wherein the initial value is obtained by the following steps:
for the initial value of the brainstorm method, according to the central idea of the brainstorm optimization, the more individuals are, the higher the possibility of obtaining the optimal solution is, the complexity of the algorithm is increased due to the increase of the individuals, through the comprehensive simulation test of the algorithm, the size of the individuals is set to be 100, the maximum iteration time is 500 times, and through the test, the four probabilities of Pa, Pb, Pc and Pd are respectively set to be 0.8, 0.2, 0.4 and 0.5.
A brainstorming optimization stage step, wherein the iterative updating in the optimization process comprises the following steps:
in the feasible solution space, generating n feasible solution individuals of the potential problem, determining a fitness function and calculating the fitness of the n feasible solution individuals;
clustering n individuals into m classes by using a k-means clustering algorithm, wherein the selected probability of each class is in direct proportion to the number of the individuals in the class; sorting the fitness values of the individuals in each class, and regarding the individual with the best fitness value as a class center of the class;
randomly selecting a class, and adding random disturbance to the class to generate a new individual; b. randomly selecting a class, randomly selecting an individual in the selected class, and adding random disturbance to generate a new individual; c. randomly selecting 2 classes, firstly fusing class centers of the two classes, and generating a new individual by random disturbance; d. randomly selecting 2 classes, firstly randomly selecting an individual from each of the two classes for fusion, and then generating a new individual by random disturbance;
and comparing the fitness value of the newly generated individual with the original individual, and replacing the original individual if the new individual is better. And updating each individual one by one, stopping iteration if an iteration stopping condition is met, and returning to the third step until iteration is stopped.
The BP neural network calculation stage step of the seawater quality method, wherein the optimization value matching has the following steps:
establishing a BP neural network, designing a hidden layer of the BP neural network to comprise 5 neurons, and designing an output layer to comprise 1 neuron. The learning rate and impulse coefficient of the BP network are Ir = 0.001, mc = 0.05 after continuous trial, and logsig function is used as a transfer function of the hidden layer.
The BP neural network calculation stage step of the seawater quality evaluation method, wherein the optimization value result verification comprises the following steps:
and (3) adopting the optimal solution obtained by optimizing the brainstorming as a weight value and a threshold value of the BP neural network, and carrying out neural network simulation by using data.
Drawings
Fig. 1 is a workflow diagram of the brainstorming optimization according to the present invention.
Fig. 2 is a working flow chart of BP neural network computation according to the present invention.
Detailed description of the preferred embodiments
The seawater quality evaluation method comprises two working stages: the brain storm optimization process and the neural network calculation process comprise the following steps:
A. brain storm optimization process
The brain storm optimization parameters comprise an initial individual number n, a population type number m, a maximum iteration number iteration, a randomly disturbed gradient adjustment parameter K and an initial solution dimension D, and values of the four probabilities of Pa, Pb, Pc and Pd are determined and used for selecting different individual updating modes to control and realize the individual iterative updating. a. Randomly selecting a class, and adding random disturbance to the center of the class to generate a new individual; b. randomly selecting a class, randomly selecting an individual in the selected class, and adding random disturbance to generate a new individual; c. randomly selecting 2 classes, firstly fusing class centers of the two classes, and generating a new individual by random disturbance; d. randomly selecting 2 classes, randomly selecting an individual from each of the two classes for fusion, and generating a new individual by random disturbance. Selecting one of the four new ways to generate a new individual, comparing the fitness value of the new individual with the original individual, and replacing the original individual if the new individual is better. And continuously and iteratively updating the individuals until the optimal condition is reached or the termination condition is reached, and stopping updating.
B. Neural network computing process
Establishing a BP neural network, selecting inorganic nitrogen, active phosphate, COD and petroleum as 4 input neurons, designing a hidden layer of the BP neural network to comprise 5 neurons and an output layer to comprise 1 neuron according to experiments and calculation. The learning rate and impulse coefficient of the BP network are Ir = 0.001, mc = 0.05 after continuous trial, and logsig function is used as a transfer function of the hidden layer. And initializing the fitness value of each individual by utilizing the training set and the fitness function of the BP neural network. And taking the obtained optimal solution as a weight value and a threshold value of the BP neural network, and performing neural network simulation by using data.
Claims (5)
1. A method for evaluating seawater quality is characterized by comprising the following steps:
s1, obtaining the optimal weight and threshold of the BP neural network in continuous iteration updating by using a brainstorming optimization algorithm;
and S2, taking the seawater quality evaluation factor as a reference factor and a BP neural network input neuron, and calculating by using the optimized weight and threshold value to evaluate the water quality grade.
2. The seawater quality evaluation method according to claim 1, wherein the step S1 specifically comprises:
assigning a group of initial values to parameters of a brainstorm optimization algorithm, wherein the parameters of the brainstorm optimization algorithm comprise an initial individual number n, a population type number m, a maximum iteration number iteration, a randomly disturbed gradient adjusting parameter K and an initial solution dimension D, determining a fitness calculation function and calculating the fitness value of each individual;
using the mean square error MSE as a way to calculate the fitness value: in the model, the smaller the fitness value, the better the individual;
determining values of Pa, Pb, Pc and Pd probabilities, and selecting different individual updating modes to control and realize individual iterative updating;
initializing the fitness value of each individual by utilizing a training set and a fitness function of the BP neural network;
selecting one of the four new ways to generate a new individual, comparing the fitness value of the new individual with the original individual, and replacing the original individual if the new individual is better.
3. The seawater quality evaluation method according to claim 1, wherein the step S2 specifically comprises:
selecting a single hidden layer feedforward network, namely a 3-layer BP neural network, selecting inorganic nitrogen, active phosphate, COD and petroleum as 4 input neurons, designing a hidden layer of the BP neural network to comprise 5 neurons according to experiments and calculation, and designing an output layer to comprise 1 neuron;
the learning rate and impulse coefficient of the BP network are Ir = 0.001, mc = 0.05 after continuous trial, and a logsig function is used as a transfer function of the hidden layer;
and taking the obtained optimal solution of the step S1 as a weight value and a threshold value of the BP neural network, and performing neural network simulation by using data.
4. The method of optimizing a brainstorming according to claim 2, wherein said method further comprises:
randomly selecting a class, and adding random disturbance to the class to generate a new individual; b. randomly selecting a class, randomly selecting an individual in the selected class, and adding random disturbance to generate a new individual; c. randomly selecting two classes, firstly fusing class centers of the two classes, and generating a new individual by random disturbance; d. two classes are randomly selected, firstly, an individual is randomly selected from each of the two classes to be fused, and then random disturbance is added to generate a new individual.
5. The method of optimizing a brainstorming according to claim 2, wherein said method further comprises:
generating three different random numbers rand1, rand2, rand3 and rand4 in the range of 0 to 1; randomly selecting a class, and if rand1< Pa, randomly generating an individual to replace the class center of the selected class; pb is the probability of selecting two updating individual modes of a and b or c and d of the individual updating mode of the brainstorm optimization algorithm, and if rand2< Pb, the two updating modes of a and b are selected for individual updating; pc is the probability of selecting one of a mode and a mode b to update the individual, Pm is the probability of selecting the mth class, if rand3< Pm and rand3< Pc, the individual is updated according to the mode a, otherwise, the individual is updated according to the mode b; if rand2> = Pb, the probability that Pd is one of c and d to update the individual mode, if rand4< Pd exists, the individual mode is updated according to the c mode, otherwise, the individual mode is updated according to the d mode.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910009673.6A CN111414927A (en) | 2019-01-05 | 2019-01-05 | Method for evaluating seawater quality |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910009673.6A CN111414927A (en) | 2019-01-05 | 2019-01-05 | Method for evaluating seawater quality |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111414927A true CN111414927A (en) | 2020-07-14 |
Family
ID=71492599
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910009673.6A Pending CN111414927A (en) | 2019-01-05 | 2019-01-05 | Method for evaluating seawater quality |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111414927A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111950942A (en) * | 2020-10-19 | 2020-11-17 | 平安国际智慧城市科技股份有限公司 | Model-based water pollution risk assessment method and device and computer equipment |
CN113343583A (en) * | 2021-06-29 | 2021-09-03 | 河北工程大学 | Water quality evaluation method based on continuous time neurodynamic network |
-
2019
- 2019-01-05 CN CN201910009673.6A patent/CN111414927A/en active Pending
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111950942A (en) * | 2020-10-19 | 2020-11-17 | 平安国际智慧城市科技股份有限公司 | Model-based water pollution risk assessment method and device and computer equipment |
CN111950942B (en) * | 2020-10-19 | 2021-01-19 | 平安国际智慧城市科技股份有限公司 | Model-based water pollution risk assessment method and device and computer equipment |
CN113343583A (en) * | 2021-06-29 | 2021-09-03 | 河北工程大学 | Water quality evaluation method based on continuous time neurodynamic network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103581188B (en) | A kind of network security situation prediction method and system | |
Yan et al. | Adaptive neuro fuzzy inference system for classification of water quality status | |
Idri et al. | Can neural networks be easily interpreted in software cost estimation? | |
CN108764540B (en) | Water supply network pressure prediction method based on parallel LSTM series DNN | |
Rotshtein et al. | Fuzzy evidence in identification, forecasting and diagnosis | |
Shih et al. | Grey number prediction using the grey modification model with progression technique | |
CN112906982A (en) | GNN-LSTM combination-based network flow prediction method | |
CN106453293A (en) | Network security situation prediction method based on improved BPNN (back propagation neural network) | |
CN103105246A (en) | Greenhouse environment forecasting feedback method of back propagation (BP) neural network based on improvement of genetic algorithm | |
Tang et al. | An optimal model based on multifactors for container throughput forecasting | |
Ning et al. | GA-BP air quality evaluation method based on fuzzy theory. | |
Abdalla et al. | Optimizing the multilayer feed-forward artificial neural networks architecture and training parameters using genetic algorithm | |
CN112529685A (en) | Loan user credit rating method and system based on BAS-FNN | |
CN111414927A (en) | Method for evaluating seawater quality | |
CN116663419A (en) | Sensorless equipment fault prediction method based on optimized Elman neural network | |
Edinson et al. | Performance analysis of fcm based anfis and elman neural network in software effort estimation. | |
CN114021432A (en) | Stress corrosion cracking crack propagation rate prediction method and system | |
Umoh et al. | A fuzzfuzzy-neural intelligent trading model for stock price prediction | |
Ma et al. | Highway construction investment risk evaluation using BP neural network model | |
El-Dahshan et al. | Artificial neural network and genetic algorithm hybrid technique for nucleus–nucleus collisions | |
Tarsauliya et al. | Financial time series volatility forecast using evolutionary hybrid artificial neural network | |
Sujamol et al. | A genetically optimized method for weight updating in fuzzy cognitive maps | |
CN113837443A (en) | Transformer substation line load prediction method based on depth BilSTM | |
Skakalina | Investigation of intelligent technologies for formation forecasting models | |
Mourhir et al. | Empirical comparison of fuzzy cognitive maps and dynamic rule-based fuzzy cognitive maps |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200714 |
|
WD01 | Invention patent application deemed withdrawn after publication |