CN108280510A - Safe early warning model based on genetic wavelet neural network - Google Patents
Safe early warning model based on genetic wavelet neural network Download PDFInfo
- Publication number
- CN108280510A CN108280510A CN201711273522.9A CN201711273522A CN108280510A CN 108280510 A CN108280510 A CN 108280510A CN 201711273522 A CN201711273522 A CN 201711273522A CN 108280510 A CN108280510 A CN 108280510A
- Authority
- CN
- China
- Prior art keywords
- neural network
- genetic
- wavelet neural
- early warning
- input
- 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
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
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Fuzzy Systems (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention discloses a kind of safe early warning model based on genetic wavelet neural network, includes the following steps：1) blurring of input data；2) number of plies and input, the output and node in hidden layer of genetic wavelet neural network are determined；3) information coding；4) initialization of population；5) fitness function calculates；6) selection operation；7) genetic manipulation；8) selection optimum individual decoding；9) it is trained according to the parameter of the BP networks of first step setting；10) Optimization of Wavelet neural network；11) optimize genetic function.The safe early warning model of the present invention, wavelet neural network structure is simpler, convergence rate is faster, being used as neural network neuron by wavelet basis function makes neural network have stronger learning ability, precision higher, realize immediately, efficiently to chemical producing system science, accurate early warning, avoid Chemical Manufacture from accident occur.
Description
Technical field
The present invention relates to Chemical Manufacture early warning fields, pre- more particularly to a kind of safety based on genetic wavelet neural network
Alert model.
Background technology
Chemical process is due to being related to that hazardous chemical quantity is more, manufacturing technique requirent is harsh and process units
Enlargement, serialization and automation, once accident occurs, consequence will be extremely serious, and chemical process is one typical huge
System is difficult to one science of the security system row, accurate early warning, BP nerves using general mathematical model construction method
Network is used widely as a kind of typical feed-forward type neural network in multiple fields, with non-linear, self study, certainly
The advantages that organizing and be adaptive, but in practical applications, there is also certain limitations for neural network：Convergence rate is slower, and
And the setting of the other factors of network such as various parameters also influences convergent speed, what this was obviously required with safety pre-warning system
Immediately efficiently incompatible.
The prior art such as, Chinese invention granted patent document, Authorization Notice No.：CN 103077408B, the invention are provided
The Sonar image based on wavelet neural network be converted to acoustic seafloor classification method, utilize genetic wavelet neural network
Algorithm can carry out partial analysis, and by genetic algorithm optimization network initial parameter, it is smaller to avoid being absorbed in part, is effectively prevented from
Noise and local extremum so that acoustic seafloor classification more accurate reliable is converted to by Sonar image, at seabed bottom
There is important practical value in qualitative classification, but it is complicated for same task wavelet neural network, and convergence rate is not
It is too ideal.
Invention content
The purpose of the present invention is to provide a kind of safe early warning model based on genetic wavelet neural network, Wavelet Neural Network
Network structure is simpler, and convergence rate faster, by wavelet basis function is used as neural network neuron, and so that neural network is had stronger
Learning ability, precision higher are realized immediately, efficiently to safety problem science, accurate early warning in chemical process.
The technical solution that the present invention is taken to achieve the above object is：Safe early warning based on genetic wavelet neural network
Model includes the following steps：
1) blurring of input data：Warning index is standardized by fuzzy mathematics and membership function and nondimensionalization
Processing；
2) number of plies and input, the output and node in hidden layer of genetic wavelet neural network are determined；
1. single BP wavelet neural networks are established：Determine network inputs, output neuron, the input of network is that the interior of system changes
Amount, output is the exogenous variable of system, for the early warning of production process subsystem and system, input number of nodes warning index
Number, output node number are divided into 5 ranks of early warning according to warning grade；
2. planned network implies the number of plies：Hidden layer number decides the precision of prediction and convergence rate of network, can by increasing hidden layer number
To reduce network error, but also make network more complicated, increases the learning time of network, hereditary small echo BP neural network design
Activation primitive for three layers of BP networks of Bao Fanyi hidden layer, hidden node is Morlet morther wavelet basic functions；
3. determining the number of hidden nodes：Since there is presently no ripe theories and method to determine the number of hidden nodes, network is needed to set
Meter person is according to the experience of itself and attempts to determine；
3) information coding：The length of coding is determined according to coding formula:
N=nk×i+nk×j+j
Wherein i is input layer number, and j is output layer number of nodes, mkFor the number of hidden nodes；
4) initialization of population：Initializega function pair populations in being converged by the tool boxes MATLAB are initialized, and are terminated
Condition refers to error E and tends towards stability less than a certain given value or group's fitness, or training has reached scheduled evolutionary generation.
5) fitness function calculates：It is fitness function according to the inverse of the error sum of squares of test set data,
In formulaI-th of predicted value in gathering for test, riFor i-th in test set of actual value, n is number of samples；
6) selection operation:
1. calculating the sum of fitness individual in population
2. calculating the fitness of each individual in population, the probability being genetic in next-generation population is selected as individual,
7) genetic manipulation carries out dimensionality reduction operation, to parent using single-point crossover operator and single-point mutation operator to input independent variable
After being arranged in order with the fitness value of filial generation, the larger individual of N number of fitness value is therefrom selected as follow-on sample
This, repetition training is until reaching trained end condition；
8) selection optimum individual decoding, obtains the initial weight and thresholds of BP neural network；
9) it is trained according to the parameter of the BP networks of first step setting, frequency of training reaches predetermined value or error is less than mesh
Scale value, then network training terminate, be otherwise transferred to previous step；
10) Optimization of Wavelet neural network, according to optimization calculate obtain as a result, will select participation modeling input independent variable
Corresponding training set and test set data extract, and re-establishing model using wavelet neural network carries out emulation experiment, and
Carry out interpretation of result；
11) optimize genetic function, changing genetic function, the specific method is as follows：Use the ga letters in MATLAB GAs Toolboxes
Number is realized, wherein containing the operation of selection, intersection and variation, method of calling：[x, endPop, bPop, traceInfo]=
Ga (bounds, evaLFN, eva10ps, startPop, opts, termFN, termOps, sellectFN, selectOps,
XOverOps, mutFNs, mutOps0).
Compared with prior art, beneficial effects of the present invention are：The safety based on genetic wavelet neural network of the present invention
Early-warning Model replaces neuron nesting type wavelet neural network with Wavelet Element so that the weights of input layer to hidden layer and
Hidden layer threshold value is replaced by the scale of wavelet function and translation parameters respectively, to same learning tasks, makes Wavelet Neural Network
Network structure is simpler, and convergence rate faster, by wavelet basis function is used as neural network neuron, and so that neural network is had stronger
Learning ability, precision higher, realize immediately, efficiently to chemical producing system science, accurate early warning, avoid Chemical Manufacture
Appearance accident.
Present invention employs the safe early warning models based on genetic wavelet neural network that above-mentioned technical proposal provides, and make up
The deficiencies in the prior art, reasonable design, easy operation.
Description of the drawings
Fig. 1 is Genetic-fuzzy wavelet neural network structure of the present invention；
Fig. 2 is neural metwork training flow of the present invention.
Specific implementation mode
The present invention is described in detail with attached drawing with reference to embodiments：
Embodiment 1：
As shown in Figure 1, 2, the safe early warning model based on genetic wavelet neural network, includes the following steps：
1) blurring of input data：Warning index is standardized by fuzzy mathematics and membership function and nondimensionalization
Processing；
2) number of plies and input, the output and node in hidden layer of genetic wavelet neural network are determined；
1. single BP wavelet neural networks are established：Determine network inputs, output neuron, the input of network is that the interior of system changes
Amount, output is the exogenous variable of system, for the early warning of production process subsystem and system, input number of nodes warning index
Number, output node number are divided into 5 ranks of early warning according to warning grade；
2. planned network implies the number of plies：Hidden layer number decides the precision of prediction and convergence rate of network, can by increasing hidden layer number
To reduce network error, but also make network more complicated, increases the learning time of network, hereditary small echo BP neural network design
Activation primitive for three layers of BP networks of Bao Fanyi hidden layer, hidden node is Morlet morther wavelet basic functions；
3. determining the number of hidden nodes：Since there is presently no ripe theories and method to determine the number of hidden nodes, network is needed to set
Meter person is according to the experience of itself and attempts to determine；
3) information coding：The length of coding is determined according to coding formula:
N=nk×i+nk×j+j
Wherein i is input layer number, and j is output layer number of nodes, mkFor the number of hidden nodes；
4) initialization of population：Initializega function pair populations in being converged by the tool boxes MATLAB are initialized, and are terminated
Condition refers to error E and tends towards stability less than a certain given value or group's fitness, or training has reached scheduled evolutionary generation.
5) fitness function calculates：It is fitness function according to the inverse of the error sum of squares of test set data,
In formulaI-th of predicted value in gathering for test, riFor i-th in test set of actual value, n is number of samples；
6) selection operation:
1. calculating the sum of fitness individual in population
2. calculating the fitness of each individual in population, the probability being genetic in next-generation population is selected as individual,
7) genetic manipulation carries out dimensionality reduction operation, to parent using single-point crossover operator and single-point mutation operator to input independent variable
After being arranged in order with the fitness value of filial generation, the larger individual of N number of fitness value is therefrom selected as follow-on sample
This, repetition training is until reaching trained end condition；
8) selection optimum individual decoding, obtains the initial weight and thresholds of BP neural network；
9) it is trained according to the parameter of the BP networks of first step setting, frequency of training reaches predetermined value or error is less than mesh
Scale value, then network training terminate, be otherwise transferred to previous step；
10) Optimization of Wavelet neural network, according to optimization calculate obtain as a result, will select participation modeling input independent variable
Corresponding training set and test set data extract, and re-establishing model using wavelet neural network carries out emulation experiment, and
Carry out interpretation of result；
11) optimize genetic function：Changing genetic function, the specific method is as follows：Use the ga letters in MATLAB GAs Toolboxes
Number is realized, wherein containing the operation of selection, intersection and variation, method of calling：[x, endPop, bPop, traceInfo]=
Ga (bounds, evaLFN, eva10ps, startPop, opts, termFN, termOps, sellectFN, selectOps,
XOverOps, mutFNs, mutOps0).
Embodiment 2：
The Early-warning Model of the present invention can be mounted in CD or chip and be realized to the safe pre- of entire Chemical Manufacture by industrial personal computer
It is alert, since Chemical Manufacture environment is complicated, corrosion-inhibiting coating is set on industrial personal computer surface, avoids industrial personal computer damage that from can not alarming, anti-corrosion
Coating is made of following component and parts by weight：It is 70-76.3 parts of aqueous acrylic emulsion, 1-2 parts of ethyl naphthol, 8-18 parts of zinc powder, double
Ten 1-4 parts of four carbon alcohols esters, 0.1-0.2 parts of lithium chromate, 2-3 parts of polyether Glycols, 0.6-1 parts of polyacrylamide, dodecyl Portugal
0.3-1.2 parts of polyglycoside, 0.6-1 parts of DBPC 2,6 ditertiary butyl p cresol, 1-5 parts of oxidized polyethylene wax, Potassium dodecylbenzenesulfonate
1-2 parts, 0.02-0.04 parts of Antimony pentachloride, 0.4-1.5 parts of N-Methyl pyrrolidone, 0.3-0.5 parts of ricinoleic acid, expanded pearlite
6-8.5 parts of rock, 5-8 parts of sericite in powder, 6-8 parts of coalescents, 2-3 parts of accelerating agent, 10-16 parts of deionized water, by industry control
Machine surface coating corrosion-inhibiting coating can enhance its antiseptic property, and industrial personal computer is avoided to be damaged by chemical attack, and coating also has excellent
Anti-aging property, stretching resistance are strong, bonding force is good, since the zinc powder in coating excessively can vivaciously make coating premature failure, by
Antimony pentachloride is added in coating, it can be achieved that Antimony pentachloride generates inhibiting effect to zinc powder activity, and the mechanism of action is still not clear, also
It needs to be studied, avoids zinc powder excessively active, coating effective time can be made to be up to 20 years or more, and do not crack.
Coalescents in corrosion-inhibiting coating are bought for market, and accelerating agent is made of following component and parts by weight：Solid thermoplastic
It is 17~28 parts of poly- hydrocarbon resin, 5~10 parts of solid-state gum dammar, 0.01-0.03 parts of chiral epichlorohydrin, 4~12 parts of butyl ester, different
2~5 parts of propyl alcohol, 44~60 parts of hexahydrotoluene, solid state modification rosin resin 5~10, wherein left-handed epoxychloropropane and dextrorotation
The ratio of epoxychloropropane is 1:1~1.5, make connecing for each ingredient of coating and station machine by adding accelerating agent in coating formula
It touches mouth and forms mutual tolerance, so that interface is disappeared, the effect that is adsorbed in conducive to coating on station machine surface, accelerating agent can be obviously improved coating
With the adhesive force of station machine, pass through the ratio of left-handed epoxychloropropane and dextrorotation epoxychloropropane in control chiral epichlorohydrin
Example is conducive to shorten the action time that accelerating agent promotes coating attachment after coating spraying, can also enhance accelerating agent in the coating
Stability avoids the adherency of accelerating agent failure effect coating.
Solid material in above-mentioned corrosion-inhibiting coating carries out pulverization process after being dried at 40-60 DEG C, and is sieved, and will crush
Material and surplus material be put into reaction kettle stirring to complete fusion, whipping temp is 55~94 DEG C, obtains anticorrosive paint wherein
The mass ratio of zinc powder and polyether Glycols is 2.6~9:1, solid content is 10wt%~40wt% in obtained anticorrosive paint, is glued
Degree is 9s~15s, most of unified left-handed at the antifouling paint being grouped as by above method acquisition, controls zinc powder and polyethers
The mass ratio of dihydric alcohol is remarkably improved the adhesion strength of coating, and coating is avoided to fall off or crack in use, the anti-corrosion
Coating can not only have excellent anti-corrosion, anti-aging, water resistance, also have excellent bonding force, stretching resistance strong, low temperature is not opened
It splits, high temperature does not play bulging effect, and coating can also significantly increase the impact resistance ability of station machine after coating station machine, be conducive to protect
Protect the normal operation of safety pre-warning system.
The prior art of routine techniques dawn known to those skilled in the art in above-described embodiment, is not chatted in detail herein
It states.
Embodiment of above is merely to illustrate the present invention, and not limitation of the present invention, the ordinary skill people of this field
Member can also make a variety of changes and modification without departing from the spirit and scope of the present invention.Therefore, all equivalent
Technical solution also belong to scope of the invention, scope of patent protection of the invention should be defined by the claims.
Claims (5)
1. the safe early warning model based on genetic wavelet neural network, which is characterized in that include the following steps：
1) blurring of input data：Warning index is standardized by fuzzy mathematics and membership function and nondimensionalization
Processing；
2) number of plies and input, the output and node in hidden layer of genetic wavelet neural network are determined；
3) information coding：The length of coding is determined according to coding formula:
N=nk×i+nk×j+j
Wherein i is input layer number, and j is output layer number of nodes, mkFor the number of hidden nodes；
4) initialization of population：Initializega function pair populations in being converged by the tool boxes MATLAB are initialized；
5) fitness function calculates：It is fitness function according to the inverse of the error sum of squares of test set data,
In formulaI-th of predicted value in gathering for test, riFor i-th in test set of actual value, n is number of samples；
6) selection operation:
1. calculating the sum of fitness individual in population
2. calculating the fitness of each individual in population, the probability being genetic in next-generation population is selected as individual,
7) genetic manipulation carries out dimensionality reduction operation, to parent using single-point crossover operator and single-point mutation operator to input independent variable
After being arranged in order with the fitness value of filial generation, the larger individual of N number of fitness value is therefrom selected as follow-on sample
This, repetition training is until reaching trained end condition；
8) selection optimum individual decoding, obtains the initial weight and thresholds of BP neural network；
9) it is trained according to the parameter of the BP networks of first step setting, frequency of training reaches predetermined value or error is less than mesh
Scale value, then network training terminate, be otherwise transferred to previous step；
10) Optimization of Wavelet neural network, according to optimization calculate obtain as a result, will select participation modeling input independent variable
Corresponding training set and test set data extract, and re-establishing model using wavelet neural network carries out emulation experiment, and
Carry out interpretation of result；
11) optimize genetic function.
2. the safe early warning model according to claim 1 based on genetic wavelet neural network, it is characterised in that：The step
In rapid 2 determine genetic wavelet neural network the number of plies and input, output the specific steps are：
1. single BP wavelet neural networks are established：Determine network inputs, output neuron, the input of network is that the interior of system changes
Amount, the output i.e. exogenous variable of system, input number of nodes warning index number, output node number are divided into early warning according to warning grade
5 ranks；
2. planned network implies the number of plies：The hereditary small echo BP neural network of design is three layers of BP networks of Bao Fanyi hidden layer, hidden layer
The activation primitive of node is Morlet morther wavelet basic functions.
3. the safe early warning model according to claim 1 based on genetic wavelet neural network, it is characterised in that：The step
Its call format of initialization of population function initializega functions is in rapid 4：Pop=initializega
(populatiaonSize, variableBounds, evalfn, evalops, optiaons).
4. the safe early warning model according to claim 1 based on genetic wavelet neural network, it is characterised in that：The step
End condition refers to error E and tends towards stability less than a certain given value or group's fitness in rapid 7, or training have reached it is scheduled into
Change algebraically.
5. the safe early warning model according to claim 1 based on genetic wavelet neural network, it is characterised in that：Described 11
The specific method is as follows for middle optimization genetic function：It is realized using the ga functions in MATLAB GAs Toolboxes, wherein including
Selection intersects and the operation of variation, method of calling：[x, endPop, bPop, traceInfo]=ga (bounds,
EvaLFN, eva10ps, startPop, opts, termFN, termOps, sellectFN, selectOps, xOverOps,
MutFNs, mutOps0).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711273522.9A CN108280510A (en) | 2017-12-05 | 2017-12-05 | Safe early warning model based on genetic wavelet neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711273522.9A CN108280510A (en) | 2017-12-05 | 2017-12-05 | Safe early warning model based on genetic wavelet neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108280510A true CN108280510A (en) | 2018-07-13 |
Family
ID=62801441
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711273522.9A Pending CN108280510A (en) | 2017-12-05 | 2017-12-05 | Safe early warning model based on genetic wavelet neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108280510A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109688154A (en) * | 2019-01-08 | 2019-04-26 | 上海海事大学 | A kind of Internet Intrusion Detection Model method for building up and network inbreak detection method |
| CN110824915A (en) * | 2019-09-30 | 2020-02-21 | 华南师范大学 | GA-DBN network-based intelligent monitoring method and system for wastewater treatment |
| CN114279914A (en) * | 2021-11-22 | 2022-04-05 | 中国地质大学（武汉） | Method and equipment for measuring sand content of drilling fluid based on neural network |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102618130A (en) * | 2012-03-26 | 2012-08-01 | 江门市苏比化工有限公司 | Chlorine-free environment-friendly polyethylene plastic coating accelerator |
| CN104130647A (en) * | 2014-06-26 | 2014-11-05 | 安徽快来防水防腐有限公司 | Corrosion-resistant and waterproof coating |
-
2017
- 2017-12-05 CN CN201711273522.9A patent/CN108280510A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102618130A (en) * | 2012-03-26 | 2012-08-01 | 江门市苏比化工有限公司 | Chlorine-free environment-friendly polyethylene plastic coating accelerator |
| CN104130647A (en) * | 2014-06-26 | 2014-11-05 | 安徽快来防水防腐有限公司 | Corrosion-resistant and waterproof coating |
Non-Patent Citations (1)
| Title |
|---|
| 郭健: "井工金矿采选过程安全预警系统构建及应用研究", 《中国博士学位论文全文数据库 工程科技Ⅰ辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109688154A (en) * | 2019-01-08 | 2019-04-26 | 上海海事大学 | A kind of Internet Intrusion Detection Model method for building up and network inbreak detection method |
| CN109688154B (en) * | 2019-01-08 | 2021-10-22 | 上海海事大学 | Network intrusion detection model establishing method and network intrusion detection method |
| CN110824915A (en) * | 2019-09-30 | 2020-02-21 | 华南师范大学 | GA-DBN network-based intelligent monitoring method and system for wastewater treatment |
| CN110824915B (en) * | 2019-09-30 | 2022-06-07 | 华南师范大学 | GA-DBN network-based intelligent monitoring method and system for wastewater treatment |
| CN114279914A (en) * | 2021-11-22 | 2022-04-05 | 中国地质大学（武汉） | Method and equipment for measuring sand content of drilling fluid based on neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Ship motion attitude prediction based on an adaptive dynamic particle swarm optimization algorithm and bidirectional LSTM neural network | |
| CN108280510A (en) | Safe early warning model based on genetic wavelet neural network | |
| CN110928189B (en) | Robust control method based on reinforcement learning and Lyapunov function | |
| CN113095481B (en) | Air combat maneuver method based on parallel self-game | |
| CN100458607C (en) | Method of soft measuring fusion index of producing propylene through polymerization in industrialization | |
| CN108920812A (en) | A kind of machining surface roughness prediction technique | |
| Hao et al. | Dynamic path planning of a three-dimensional underwater AUV based on an adaptive genetic algorithm | |
| CN101520857B (en) | Inverse kinematics resolution method of permanent magnetism spherical electric motor on the basis of neural network | |
| Wang et al. | Simulation with RBF neural network model for reservoir operation rules | |
| Lee et al. | UAS mission path planning system (MPPS) using hybrid-game coupled to multi-objective optimizer | |
| CN112084727A (en) | Transition prediction method based on neural network | |
| CN111007724A (en) | Hypersonic aircraft designated performance quantitative tracking control method based on interval II type fuzzy neural network | |
| CN102054199A (en) | BP (Back Propagation) neural network algorithm based method for analyzing coating aging | |
| CN105976029B (en) | A kind of overhead crane neural network modeling approach of cuckoo behavior RNA-GA | |
| CN110309909A (en) | A kind of intelligent real-time predicting method of a wide range of maneuvering target track of high speed | |
| CN103675006B (en) | The industrial melt index soft measurement instrument of least square and method | |
| Lyu et al. | An Embedded Linear Model Three‐Dimensional Fuzzy PID Control System for a Bionic AUV under Wave Disturbance | |
| CN117590867B (en) | Underwater autonomous vehicle connection control method and system based on deep reinforcement learning | |
| CN117268391B (en) | Intelligent planning method and system for deformed aircraft based on target layered architecture | |
| CN117452827A (en) | Under-actuated unmanned ship track tracking control method | |
| CN114943168B (en) | Method and system for combining floating bridges on water | |
| CN111695195A (en) | Spatial physical moving body modeling method based on long-time memory network | |
| CN105260556B (en) | The overhead crane modeling method of hair clip mutation operation RNA genetic algorithm | |
| CN116224798A (en) | Autonomous underwater vehicle track tracking control method based on event triggering | |
| CN115903820A (en) | Multi-unmanned-boat pursuit and escape game control method |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180713 |