CN114202063A - Fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization - Google Patents

Fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization Download PDF

Info

Publication number
CN114202063A
CN114202063A CN202111639593.2A CN202111639593A CN114202063A CN 114202063 A CN114202063 A CN 114202063A CN 202111639593 A CN202111639593 A CN 202111639593A CN 114202063 A CN114202063 A CN 114202063A
Authority
CN
China
Prior art keywords
neural network
fuzzy neural
fuzzy
network model
genetic algorithm
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.)
Withdrawn
Application number
CN202111639593.2A
Other languages
Chinese (zh)
Inventor
王丽娜
刘锦杰
李雪
戴文彬
张延政
王斌锐
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Jiliang University
Original Assignee
China Jiliang University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by China Jiliang University filed Critical China Jiliang University
Priority to CN202111639593.2A priority Critical patent/CN114202063A/en
Publication of CN114202063A publication Critical patent/CN114202063A/en
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/043—Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
    • 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
    • G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/02—Computing arrangements based on specific mathematical models using fuzzy logic

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Biophysics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Automation & Control Theory (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Fuzzy Systems (AREA)
  • Physiology (AREA)
  • Genetics & Genomics (AREA)
  • Algebra (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization, which belongs to the technical field of intelligent agriculture and comprises the steps of taking initial parameters optimized by a genetic algorithm as initial parameters of fuzzy neural network iteration, returning errors obtained by prediction after the fuzzy neural network is trained as target values to the genetic algorithm for next generation inheritance, alternately cooperating the two algorithms to search for optimal parameters of the network training, finding out initial input parameters for establishing an optimal prediction model in limited iteration times and inheritance, and inputting meteorological station temperature, relative humidity and accumulated radiation data in an area in the fuzzy neural network model based on the genetic algorithm optimization to obtain the predicted greenhouse temperature. The method has better prediction accuracy, can effectively predict the temperature of the greenhouse, and provides a new idea for realizing temperature control of the greenhouse environment.

Description

Fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization
Technical Field
The invention relates to the technical field of intelligent agriculture, in particular to a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization.
Background
The greenhouse environment control system is a complex dynamic system with characteristics of multi-coupling, nonlinearity and large hysteresis performance, and is a comprehensive control technology mainly applied to computers. The purpose of the control is finally to obtain an ideal environment which is beneficial to the growth of crops. The technology can realize an industrial scale production mode in the aspect of resource saving, has the advantages of high quality, high efficiency, low consumption and the like, and the currently adopted methods for intelligent control of the greenhouse mainly comprise fuzzy control, neural network, hybrid control and the like.
The fuzzy control does not need to establish an accurate mathematical model for a researched object, has short transition time, small overshoot and high response speed, is superior to PID control in the aspects of regulation speed and robustness, and can only realize rough control. The fuzzy control rule is optimized by adopting a genetic algorithm, and the control of the genetic algorithm is added into the fuzzy control rule, so that the optimization rate is improved on the basis of avoiding the premature phenomenon in the optimization process. The fuzzy control has fuzzy logic reasoning capability, and the strong learning capability of the neural network can avoid the defect of the fuzzy control and can better adapt to the nonlinear and time-varying characteristics of the greenhouse system.
The fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization combines the advantages of the genetic algorithm and the fuzzy neural network algorithm, effectively solves the problem of random selection of initial parameters of the fuzzy neural network, and accordingly improves the training efficiency of the network and the prediction accuracy of a training model.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization, and the fusion algorithm has a good prediction effect and provides a new thought for the research of the greenhouse temperature prediction method.
The embodiment of the invention provides a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization, which comprises the following steps:
acquiring the temperature, the relative humidity and the accumulated radiation data of a meteorological station in an area in a past period of time;
selecting a fuzzy neural network model structure;
determining parameters of the fuzzy neural network model according to the structure of the fuzzy neural network model;
encoding parameters of the fuzzy neural network model to form chromosomes of a genetic algorithm, and initializing population individuals of the genetic algorithm;
randomly initializing a plurality of initial parameters in the fuzzy neural network model;
encoding a plurality of initial parameters of the fuzzy neural network model;
decoding a plurality of initial parameters and endowing the initial parameters to a fuzzy neural network model;
carrying out repeated iterative training on the fuzzy neural network model by using a training set, and testing the fuzzy neural network model by using a testing set to obtain a testing error;
calculating the fitness value of the genetic algorithm according to the test error;
carrying out genetic selection, crossing and variation according to the fitness to obtain optimized initial parameters;
taking the optimized initial parameters as initial parameters of the fuzzy neural network model;
after the network parameters are alternately optimized for multiple times through the fuzzy neural network model and the genetic algorithm, the optimal network parameters are taken as final input parameters of the fuzzy neural network model, and the fuzzy neural network model based on the genetic algorithm optimization is constructed according to the final input parameters;
and inputting the weather station temperature, the relative humidity and the accumulated radiation data of a past period of time in the region in a fuzzy neural network model optimized based on a genetic algorithm, and outputting the predicted greenhouse temperature at the future moment.
Further, the fuzzy neural network model adopts a T-S fuzzy system, and the T-S fuzzy system comprises:
calculating each input variable x according to fuzzy rulejDegree of membership of:
Figure BDA0003442326270000021
in the formula (I), the compound is shown in the specification,
Figure BDA0003442326270000022
are respectively the ith fuzzy set AiThe center value and the width of (c); k is the number of input parameters; n is the number of fuzzy subsets, and the input quantity x is [ x1, x2, …, xk]Mu represents the language variable membership coefficient;
fuzzy calculation is carried out on each membership degree, and a fuzzy operator is adopted as a connecting operator:
Figure BDA0003442326270000023
wherein u represents a linguistic variable membership coefficient,
Figure BDA0003442326270000031
is xiAt the jth degree of membership of linguistic variables, k denotes the number of input parameters, wiRepresenting the fitness of the ith rule;
calculating the output value y of the ith node of the fuzzy model according to the fuzzy calculation resulti：
Figure BDA0003442326270000032
Wherein the content of the first and second substances,
Figure BDA0003442326270000033
input value x representing the kth nodekThe weight of (c).
Further, the fuzzy neural network model parameters determined according to the fuzzy neural network model structure comprise:
learning target, learning rate, sample coefficient p0, sample coefficient p1, sample coefficient p2, sample coefficient p3, and membership function center c and width b.
Further, randomly initializing a plurality of starting parameters in the fuzzy neural network model, including:
the fuzzy membership function center c, the width b, the sample coefficient p0, the sample coefficient p1, the sample coefficient p2 and the sample coefficient p3 are initialized randomly, the learning target is set to be 0.001, and the learning rate is 0.05.
Further, the test error of the fuzzy neural network model is calculated by the following formula:
error calculation
Figure BDA0003442326270000034
In the formula, ydIs the desired output of the network; y iscIs the actual output of the network; e is the error of the desired output from the actual output.
Further, still include:
systematic modification of the fuzzy neural network model, wherein the calculation formula comprises:
Figure BDA0003442326270000035
Figure BDA0003442326270000036
in the formula (I), the compound is shown in the specification,
Figure BDA0003442326270000037
is a neural network coefficient; alpha is the network learning rate; x is the number ofjInputting parameters for the network; w is aiIs the product of the membership degree of the input parameters and k is the number of the input parameters.
Further, still include:
and (3) correcting parameters of the fuzzy neural network model, wherein the calculation formula comprises:
Figure BDA0003442326270000038
Figure BDA0003442326270000041
in the formula (I), the compound is shown in the specification,
Figure BDA0003442326270000042
the central value and the width of the membership function are respectively, and beta represents an error correction coefficient.
Further, the method comprises the steps of carrying out repeated iterative training on the fuzzy neural network model by utilizing a training set, and testing the fuzzy neural network model by utilizing a testing set, wherein the steps comprise:
training a fuzzy neural network model, wherein training data comprises meteorological station temperature, relative humidity and accumulated radiation data of the region in the last decade, and the average temperature in the last decade in the region is used as an output value of the training, and the daily average accumulated radiation quantity corresponds to a parameter P1, daily average air relative humidity, a parameter P2, daily average earth surface temperature and a parameter P3 which are used as input values of the training;
and testing the fuzzy neural network model, wherein the test data comprises the daily average temperature, the average surface temperature, the average relative humidity and the accumulated radiation quantity of the meteorological station in a certain year in the region, the temperature data in the test data is used as a test output expected value, and the rest of the test data is used as a test input value.
Further, the network parameters are optimized through a fuzzy neural network model and a genetic algorithm in a repeated alternating mode, and the optimal network parameters are obtained, wherein the optimal network parameters comprise:
network iteration is carried out for 200 times after each inheritance, and the optimal network parameters are obtained after 20 times of inheritance.
The embodiment of the invention provides a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization, and compared with the prior art, the fuzzy neural network greenhouse temperature prediction method has the following beneficial effects:
the method takes initial parameters optimized by a genetic algorithm as initial parameters of a fuzzy neural network, takes errors predicted by the network as target values to carry out next generation heredity on the genetic algorithm, and combines the advantages of the two algorithms to obtain better prediction effect. Simulation experiments show that: the method has better prediction accuracy and higher precision.
Drawings
FIG. 1 is a flowchart of a fuzzy neural network method based on genetic algorithm optimization for a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization according to an embodiment of the present invention;
FIG. 2 is a diagram of a fuzzy neural network training result of a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization according to an embodiment of the present invention;
FIG. 3 is a diagram of a test result of a fuzzy neural network based on a fuzzy neural network greenhouse temperature prediction method optimized by a genetic algorithm according to an embodiment of the present invention;
FIG. 4 is a training result diagram of a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization according to an embodiment of the present invention;
FIG. 5 is a test result diagram of a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization according to an embodiment of the present invention;
FIG. 6 is a flow chart of a fuzzy neural network model algorithm of a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1 to 6, an embodiment of the present invention provides a fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization, including:
s1: acquiring the temperature, the relative humidity and the accumulated radiation data of a meteorological station in an area in a past period of time;
s2: selecting a fuzzy neural network model structure;
in particular, the fuzzy neural network in the application adopts a T-S fuzzy system with strong self-adaptive capacity. The Takagi-Sugeno (T-S) model is a new fuzzy inference model proposed by Takagi and Sugeno in 1985, and can generate more complex nonlinear functions by using a small number of fuzzy rules, thereby effectively reducing the number of fuzzy rules when processing a multivariable system.
The definition rule is as follows:
the fuzzy inference in the case of a rule Ri is as follows:
Ri:If x1 is
Figure BDA0003442326270000051
x2 is
Figure BDA0003442326270000052
…，xk is
Figure BDA0003442326270000053
then
Figure BDA0003442326270000054
wherein the content of the first and second substances,
Figure BDA0003442326270000055
a fuzzy set which is a fuzzy system;
Figure BDA0003442326270000056
is a fuzzy system parameter; yi is the output according to the fuzzy rule, the input part (if part) is fuzzy, the output part (then part) is deterministic, the fuzzy inference means that the output is a linear combination of the inputs.
Suppose for an input quantity x ═ x1, x2,…，xk]First, each input variable x is calculated according to a fuzzy rulejDegree of membership of:
Figure BDA0003442326270000061
in the formula (I), the compound is shown in the specification,
Figure BDA0003442326270000062
respectively the center and the width of the membership function; k is the number of input parameters; n is the number of fuzzy subsets.
Fuzzy calculation is carried out on each membership degree, and a fuzzy operator is adopted as a connecting operator:
Figure BDA0003442326270000063
calculating the output value y of the fuzzy model according to the fuzzy calculation resulti：
Figure BDA0003442326270000064
The learning method of the network is as follows:
(1) error calculation
Figure BDA0003442326270000065
In the formula, ydIs the desired output of the network; y iscIs the actual output of the network; e is the error of the desired output from the actual output.
(2) System correction
Figure BDA0003442326270000066
Figure BDA0003442326270000067
In the formula (I), the compound is shown in the specification,
Figure BDA0003442326270000068
is a neural network coefficient; alpha is the network learning rate; x is the number ofjInputting parameters for the network; w is aiIs the product of the membership of the input parameters.
(3) Parameter correction
Figure BDA0003442326270000069
Figure BDA00034423262700000610
In the formula (I), the compound is shown in the specification,
Figure BDA00034423262700000611
respectively the center and width of the membership function.
S3: determining parameters of the fuzzy neural network model according to the structure of the fuzzy neural network model;
s4: encoding parameters of the fuzzy neural network model to form chromosomes of a genetic algorithm, and initializing population individuals of the genetic algorithm;
specifically, according to the number of nodes determined by the fuzzy neural network algorithm and the characteristics of the genetic algorithm, a learning target, a learning rate, a sample coefficient (p 0-p 3), a membership function center c and a width b in the fuzzy neural network algorithm are selected as 8 genes of an individual to be screened generation by generation.
S5: randomly initializing a plurality of initial parameters in the fuzzy neural network model;
specifically, the fuzzy membership function center c, the width b and the coefficients p0 to p3 are randomly initialized, the learning target is set to 0.001, and the learning rate is set to 0.05.
S6: encoding a plurality of initial parameters of the fuzzy neural network model;
s7: decoding a plurality of initial parameters and endowing the initial parameters to a fuzzy neural network model;
s8: carrying out repeated iterative training on the fuzzy neural network model by using a training set, and testing the fuzzy neural network model by using a testing set to obtain a testing error;
specifically, proper training and prediction data are selected, the adopted data are meteorological station temperature, relative humidity and accumulated radiation data of the Hangzhou city (2010-2019) in the last decade, a model of a fuzzy neural network is established, part of the selected data is used as input, part of the selected data is used as output, an accurate fuzzy neural network model is trained, the prediction accuracy of the model is verified, the training data are meteorological station temperature, relative humidity and accumulated radiation data of the Hangzhou city (2010-2019) in the last decade, and the data used for prediction testing are meteorological station daily average temperature, average earth surface temperature, average relative humidity and accumulated radiation quantity of the Hangzhou city in the 2016 year. The data are recorded once every 5-6 minutes, so that the change condition of the greenhouse temperature in one day is described by 240-288 data, the average temperature in nearly ten years in Hangzhou city is used as a training output value, the average daily cumulative radiant quantity (corresponding to a parameter P1), the average daily air relative humidity (corresponding to a parameter P2) and the average daily surface temperature (corresponding to a parameter P3) are used as training input values, the temperature data in the test data are used as a test output expected value, the rest of the test data are used as test input values, and the data sequence is disturbed to eliminate the rule in the time sequence before training so as to obtain a better prediction model. The structure of the established fuzzy neural network is 3-7-1, and on the basis, the network is repeatedly trained for 200 times.
S9: calculating the fitness value of the genetic algorithm according to the test error;
s10: carrying out genetic selection, crossing and variation according to the fitness to obtain optimized initial parameters;
s11: taking the optimized initial parameters as initial parameters of the fuzzy neural network model;
s12: after the network parameters are alternately optimized for multiple times through the fuzzy neural network model and the genetic algorithm, the optimal network parameters are taken as final input parameters of the fuzzy neural network model, and the fuzzy neural network model based on the genetic algorithm optimization is constructed according to the final input parameters;
specifically, a fuzzy neural network is optimized by adopting a genetic algorithm, the genetic algorithm encodes problem parameters into chromosomes, and then the information of the chromosomes in the population is exchanged by performing operations such as selection, crossing and variation in an iterative mode, so that the chromosomes meeting the optimization target are generated finally. The genetic algorithm has the capability of solving the problem of complex nonlinear optimization, and the optimization capability of the genetic algorithm can effectively solve the random selection of initial parameters of the fuzzy neural network, so that the training efficiency of the network and the prediction accuracy of a training model are improved.
And (3) performing network iteration for 200 times after each inheritance, and finally obtaining a final prediction model by taking the optimal parameters after 20 times of inheritance as initial input parameters for fuzzy neural network training.
S13: and inputting the weather station temperature, the relative humidity and the accumulated radiation data of a past period of time in the region in a fuzzy neural network model optimized based on a genetic algorithm, and outputting the predicted greenhouse temperature at the future moment.
Example 1:
the results of simulations of the fuzzy neural network algorithm optimized based on the genetic algorithm and the fuzzy neural network algorithm were evaluated using a prediction decision coefficient R2 (the square of the correlation coefficient R). According to the established model, MatlabR2019b software is adopted for programming simulation, and a fuzzy neural network training and testing result graph based on genetic algorithm optimization are obtained, and are shown in figures 3-5. Fig. 2 and fig. 3 are fuzzy neural network training and testing results, where an error between a training predicted value and an expected value fluctuates in a small range around a value of 0, a model trained by a sample has a high prediction accuracy, and an error value in the testing result fluctuates in a large range around the value of 0, the error is significant, and a decision coefficient R2 is 0.9545, which has a certain accuracy but is not accurate enough; after the genetic algorithm optimization, under the condition of the same iteration times, the fuzzy neural network training result based on the genetic algorithm optimization is shown in fig. 4, the error curve is more stable and is very close to the value of 0, the test result in fig. 5 shows that the test error has small amplitude fluctuation only above and below the value of 0, and the decision coefficient R20.9724, compared to fuzzy neural netsIn addition, the prediction accuracy is greatly improved.
Although the embodiments of the present invention have been disclosed in the foregoing for illustrative purposes, those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying drawings.

Claims (9)

1. A fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization is characterized by comprising the following steps:
acquiring the temperature, the relative humidity and the accumulated radiation data of a meteorological station in an area in a past period of time;
selecting a fuzzy neural network model structure;
determining parameters of the fuzzy neural network model according to the structure of the fuzzy neural network model;
encoding parameters of the fuzzy neural network model to form chromosomes of a genetic algorithm, and initializing population individuals of the genetic algorithm;
randomly initializing a plurality of initial parameters in the fuzzy neural network model;
encoding a plurality of initial parameters of the fuzzy neural network model;
decoding a plurality of initial parameters and endowing the initial parameters to a fuzzy neural network model;
carrying out repeated iterative training on the fuzzy neural network model by using a training set, and testing the fuzzy neural network model by using a testing set to obtain a testing error;
calculating the fitness value of the genetic algorithm according to the test error;
carrying out genetic selection, crossing and variation according to the fitness to obtain optimized initial parameters;
taking the optimized initial parameters as initial parameters of the fuzzy neural network model;
after the network parameters are alternately optimized for multiple times through the fuzzy neural network model and the genetic algorithm, the optimal network parameters are taken as final input parameters of the fuzzy neural network model, and the fuzzy neural network model based on the genetic algorithm optimization is constructed according to the final input parameters;
and inputting the weather station temperature, the relative humidity and the accumulated radiation data of a past period of time in the region in a fuzzy neural network model optimized based on a genetic algorithm, and outputting the predicted greenhouse temperature at the future moment.
2. The fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization as claimed in claim 1, wherein the fuzzy neural network model adopts a T-S fuzzy system, and the T-S fuzzy system comprises:
calculating each input variable x according to fuzzy rulejDegree of membership of:
Figure FDA0003442326260000021
in the formula (I), the compound is shown in the specification,
Figure FDA0003442326260000022
are respectively the ith fuzzy set AiThe center value and the width of (c); k is the number of input parameters; n is the number of fuzzy subsets, and the input quantity x is [ x1, x2, …, xk]Mu represents the language variable membership coefficient;
fuzzy calculation is carried out on each membership degree, and a fuzzy operator is adopted as a connecting operator:
Figure FDA0003442326260000023
wherein u represents a linguistic variable membership coefficient,
Figure FDA0003442326260000024
is xiAt the jth degree of membership of linguistic variables, k denotes the number of input parameters, wiRepresenting the fitness of the ith rule;
calculating the output value y of the ith node of the fuzzy model according to the fuzzy calculation resulti：
Figure FDA0003442326260000025
Wherein the content of the first and second substances,
Figure FDA0003442326260000027
input value x representing the kth nodekThe weight of (c).
3. The fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization as claimed in claim 1, wherein the fuzzy neural network model parameters determined according to the fuzzy neural network model structure comprise:
learning target, learning rate, sample coefficient p0, sample coefficient p1, sample coefficient p2, sample coefficient p3, and membership function center c and width b.
4. The fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization as claimed in claim 3, wherein said randomly initializing a plurality of initial parameters in the fuzzy neural network model comprises:
the fuzzy membership function center c, the width b, the sample coefficient p0, the sample coefficient p1, the sample coefficient p2 and the sample coefficient p3 are initialized randomly, the learning target is set to be 0.001, and the learning rate is 0.05.
5. The fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization as claimed in claim 1, wherein the test error of the fuzzy neural network model is calculated by a formula comprising:
error calculation
Figure FDA0003442326260000026
In the formula, ydIs the desired output of the network; y iscIs a networkActual output; e is the error of the desired output from the actual output.
6. The fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization as claimed in claim 1, further comprising:
systematic modification of the fuzzy neural network model, wherein the calculation formula comprises:
Figure FDA0003442326260000031
Figure FDA0003442326260000032
in the formula (I), the compound is shown in the specification,
Figure FDA0003442326260000036
is a neural network coefficient; alpha is the network learning rate; x is the number ofjInputting parameters for the network; w is aiIs the product of the membership degree of the input parameters and k is the number of the input parameters.
7. The fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization as claimed in claim 1, further comprising:
and (3) correcting parameters of the fuzzy neural network model, wherein the calculation formula comprises:
Figure FDA0003442326260000033
Figure FDA0003442326260000034
in the formula (I), the compound is shown in the specification,
Figure FDA0003442326260000035
the central value and the width of the membership function are respectively, and beta represents an error correction coefficient.
8. The method for predicting the greenhouse temperature of the fuzzy neural network based on genetic algorithm optimization as claimed in claim 1, wherein the training set is used for training the fuzzy neural network model for a plurality of iterations, and the testing set is used for testing the fuzzy neural network model, and the method comprises the following steps:
training a fuzzy neural network model, wherein training data comprises meteorological station temperature, relative humidity and accumulated radiation data of the region in the last decade, and the average temperature in the last decade in the region is used as an output value of the training, and the daily average accumulated radiation quantity corresponds to a parameter P1, daily average air relative humidity, a parameter P2, daily average earth surface temperature and a parameter P3 which are used as input values of the training;
and testing the fuzzy neural network model, wherein the test data comprises the daily average temperature, the average surface temperature, the average relative humidity and the accumulated radiation quantity of the meteorological station in a certain year in the region, the temperature data in the test data is used as a test output expected value, and the rest of the test data is used as a test input value.
9. The method for predicting the greenhouse temperature based on the fuzzy neural network optimized by the genetic algorithm as claimed in claim 1, wherein the step of optimizing the network parameters through a fuzzy neural network model and a plurality of times of alternation of the genetic algorithm to obtain the optimal network parameters comprises the following steps:
network iteration is carried out for 200 times after each inheritance, and the optimal network parameters are obtained after 20 times of inheritance.
CN202111639593.2A 2021-12-29 2021-12-29 Fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization Withdrawn CN114202063A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111639593.2A CN114202063A (en) 2021-12-29 2021-12-29 Fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111639593.2A CN114202063A (en) 2021-12-29 2021-12-29 Fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization

Publications (1)

Publication Number Publication Date
CN114202063A true CN114202063A (en) 2022-03-18

Family

ID=80657225

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111639593.2A Withdrawn CN114202063A (en) 2021-12-29 2021-12-29 Fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization

Country Status (1)

Country Link
CN (1) CN114202063A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116822328A (en) * 2023-05-11 2023-09-29 中南大学 Determination method for mine goaf earth surface subsidence prediction parameters

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116822328A (en) * 2023-05-11 2023-09-29 中南大学 Determination method for mine goaf earth surface subsidence prediction parameters

Similar Documents

Publication Publication Date Title
CN106022521B (en) Short-term load prediction method of distributed BP neural network based on Hadoop architecture
CN106529818B (en) Water quality assessment Forecasting Methodology based on Fuzzy Wavelet Network
Zhang et al. Enhanced ELITE-load: A novel CMPSOATT methodology constructing short-term load forecasting model for industrial applications
CN109214579B (en) BP neural network-based saline-alkali soil stability prediction method and system
CN104636801A (en) Transmission line audible noise prediction method based on BP neural network optimization
Dhahri et al. Hierarchical multi-dimensional differential evolution for the design of beta basis function neural network
CN110674965A (en) Multi-time step wind power prediction method based on dynamic feature selection
CN105608295A (en) Multi-objective evolutionary algorithm (MOEA) and radial basis function (RBF) neural network optimization modeling method of coking furnace pressure
Jaddi et al. Taguchi-based parameter designing of genetic algorithm for artificial neural network training
CN104732067A (en) Industrial process modeling forecasting method oriented at flow object
CN114492922A (en) Medium-and-long-term power generation capacity prediction method
CN116526473A (en) Particle swarm optimization LSTM-based electrothermal load prediction method
CN112036651A (en) Electricity price prediction method based on quantum immune optimization BP neural network algorithm
Guenounou et al. TSK fuzzy model with minimal parameters
Zhou et al. A self-organizing fuzzy neural network modeling approach using an adaptive quantum particle swarm optimization
CN114202063A (en) Fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization
CN109408896B (en) Multi-element intelligent real-time monitoring method for anaerobic sewage treatment gas production
CN110738363A (en) photovoltaic power generation power prediction model and construction method and application thereof
Sang et al. Ensembles of gradient boosting recurrent neural network for time series data prediction
Chan et al. Modelling and optimization of fluid dispensing for electronic packaging using neural fuzzy networks and genetic algorithms
CN113128666A (en) Mo-S-LSTMs model-based time series multi-step prediction method
CN111524348A (en) Long-short term traffic flow prediction model and method
Lee et al. A genetic algorithm based robust learning credit assignment cerebellar model articulation controller
Anuradha et al. Direct inverse neural network control of a continuous stirred tank reactor (CSTR)
CN115660347A (en) Energy storage system optimal configuration method facing multiple constraints and nonlinear targets

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication

Application publication date: 20220318

WW01 Invention patent application withdrawn after publication