CN106951983A - Injector performance Forecasting Methodology based on the artificial neural network using many parent genetic algorithms - Google Patents
Injector performance Forecasting Methodology based on the artificial neural network using many parent genetic algorithms Download PDFInfo
- Publication number
- CN106951983A CN106951983A CN201710107962.0A CN201710107962A CN106951983A CN 106951983 A CN106951983 A CN 106951983A CN 201710107962 A CN201710107962 A CN 201710107962A CN 106951983 A CN106951983 A CN 106951983A
- Authority
- CN
- China
- Prior art keywords
- neural network
- output
- chromosome
- node
- individual
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M99/00—Subject matter not provided for in other groups of this subclass
- G01M99/008—Subject matter not provided for in other groups of this subclass by doing functionality tests
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Genetics & Genomics (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Development Economics (AREA)
- General Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Physiology (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of injector performance Forecasting Methodology of the artificial neural network based on using many parent genetic algorithms, for given injector, collects relevant parameter;According to input/output argument, neutral net is set up;Genetic algorithm initial parameter is set;Set up evaluation function;Genetic algorithm processing, is constantly selected individual in population, many parents intersection, mutation operation and records fitness value, reach the evolution number of times upper limit;The corresponding chromosome of fitness optimal solution is the threshold value and weights corresponding to many parent BP GA neutral nets set up, and obtained many parent BP GA neutral nets are trained, until the error of prediction output variable is less than set-point;The input variable under virtual condition is measured, predicts that many parent BP GA neutral nets obtain Prediction Parameters by injector performance, by pcCarry out renormalization and obtain many parent BP GA neural network prediction outlet back pressures pcActual value.Precision of prediction of the present invention is higher, time-consuming shorter.
Description
Technical field
The present invention relates to a kind of injector performance Forecasting Methodology, especially a kind of people based on using many parent genetic algorithms
The injector performance Forecasting Methodology of artificial neural networks.
Background technology
When injector works, vacuum is produced by main jet by one pressure higher fluid, and suction pressure is relatively low
Fluid, after mixing by diffuser improve fluid pressure, finally give the fluid of middle pressure, i.e., by low pressure fluid pressure
Lifting, realize compression effect.Injector can be low-grade using industrial overbottom pressure, waste heat, used heat, solar heat, underground heat etc.
The energy is as driving, without consumption electric power, with good effects of energy saving and emission reduction;Be widely used to chemical industry, heat energy, refrigeration,
The fields such as HVAC.
Mass ratio of induced-to-inducing air (ε) and outlet back pressure (pc) under critical excitation are the performance parameters of injector most critical.But
Due to injector interior flow it is extremely complex, including be jammed twice, Supersonic Flow, all kinds of shock waves, the phenomenon such as fan-shaped diffusion, adopt
The parameters precision obtained with one-dimensional physical model simulation is relatively low, and effect is poor, such as (the W.Chen et of document 1
al.Theoretical analysis of ejector refrigeration system performance under
overall modes.Applied energy,185-2:System is sprayed under the full working scopes such as 2074-2084,2016., i.e. W.Chen
The theory analysis application energy of cooling system performance, 185-2:2074-2084,2016.) and (the JM.Cardemil et of document 2
al.A general model for evaluation of vapor ejectors performance for
application in refrigeration.Energy Conversion and Management,64:79-86,2012,
That is the conversion of model energy and management, 64 that the such as JM.Cardemil mono- are used for cooling steam jet ejector Performance Evaluation:79-
86,2012.) shown in, the mean error of conventional model is more in 5-10%, and worst error is up to more than 15%.As used
The method of Fluid Mechanics Computation then takes long, also labor intensive material resources, is not suitable for design and the research of associated cyclic.It is above-mentioned existing
Shape brings problem to work such as design application, the associated cyclic researchs of injector.
The content of the invention
In order to which the precision of prediction for overcoming the shortcomings of existing injector performance Forecasting Methodology is relatively low, time-consuming long, the present invention is carried
The injector that a kind of precision of prediction is higher, take the shorter artificial neural network based on using many parent genetic algorithms is supplied
Can Forecasting Methodology.
The technical solution adopted for the present invention to solve the technical problems is:
A kind of injector performance Forecasting Methodology of the artificial neural network based on using many parent genetic algorithms, including it is following
Step:
Step one, the collection and processing of data
For given injector, it is citation jet body pressure Pe, working fluid pressure Pp, outlet back pressure to collect relevant parameter
Pc and mass ratio of induced-to-inducing air ε, citation jet body pressure, working fluid pressure and outlet back pressure are normalized, and it is arrived [0,1]
Between, formula is as follows:
Wherein, k is the data after normalization, and x is to be normalized data, xminTo be normalized the minimum value in data,
xmaxTo be normalized maximum in data;
Step 2, the foundation of neural network structure
According to input/output argument, the input and output layer nodal point number of neutral net is determined, the hidden layer number of plies is one layer, node
It is several according to empirical equation:Draw, wherein l, n are input, output node number, a takes 1~10 for constant.Judge
The input layer number of neutral net, output layer nodes choose node in hidden layer, learning rate, training objective, cycle-index
Set and each layer weights initial value;
Step 3, sets genetic algorithm initial parameter
Population scale, evolution number of times, crossover probability and mutation probability are set, the weights of BP neural network and threshold value are constituted
The neutral net set up in genetic coding, the step 2 that reruns simultaneously records the genetic coding of each run completion and sets up population;
Step 4, sets up evaluation function, and the result that the neutral net constructed using every group chromosome is predicted is foundation, meter
Calculate and record corresponding fitness value;
Step 5, genetic algorithm processing
A. selection operation:With certain probability selection individual for producing the next generation in old colony, it is general that individual is selected
Rate is relevant with fitness value, and the value of fitness is better, and selected probability is bigger;
B. many parent crossover operations:B. many parent crossover operations:Refer to select a plurality of parent chromosome to carry out from population
The foundation of chromosome of future generation, by the combined crosswise of chromosome, produces new individual;
C. mutation operation:Refer to an optional individual from colony, certain section coding in selected chromosome is become
It is different to produce more excellent individual;
Step 6, the genetic coding in the chromosome newly obtained is decoded, and calculates fitness, and group enters with original seed
Row compares, and selects the superior and eliminates the inferior;Four-step 5 of repeat step, constantly individual in population is selected, intersected, mutation operation simultaneously
Fitness value is recorded, the evolution number of times upper limit is reached;The corresponding chromosome of fitness optimal solution is many parent BP-GA set up
Obtained many parent BP-GA neutral nets are trained by threshold value and weights corresponding to neutral net, until prediction output becomes
The error of amount is less than set-point;
Step 7, in practical engineering application, is carried out with many parent BP-GA neutral nets to certain given injector performance
Input variable under prediction, measurement virtual condition, i.e. citation jet body pressure pe, working fluid pressure pp, it is pre- by injector performance
Survey many parent BP-GA neutral nets and obtain Prediction Parameters, Prediction Parameters are outlet back pressure pc, mass ratio of induced-to-inducing air ε, by pcIt is counter to be returned
One change obtains many parent BP-GA neural network prediction outlet back pressures pcActual value.
Further, in the step 2, in the step 2, the selection of node in hidden layer, learning rate takes 0.1~0.2,
Training objective takes 10-3~10-6, each layer weights initial value takes the random number between [- 1,1].
Further, in the step 3, genetic algorithm initial parameter is set, makes population scale take N1, evolution number of times takes N2
Secondary, crossover probability is set to 20%~50%, and mutation probability is set to 1%~10%, each individual weights and threshold by neutral net
Value composition, BP-GA neural network topology structures have l node for l-m-n, i.e. input layer, and hidden layer has m node, output layer
There is n node, have l × m × n weights, m+n threshold value, so genetic algorithm individual chromosome code length is l × m × n
+ m+n;Initial training is carried out to the artificial neural network that step 2 is set up, chromosome is constituted with obtained weights, threshold value
Body, repeats n times and obtains the individual of N group chromosomes and constitute population.
Further, in the step 4, the result of the neural network prediction constructed using every group chromosome is foundation, meter
Calculate and record corresponding fitness value, predicting the outcome, its more accurate corresponding fitness value is smaller, and fitness computing formula is such as
Under:
In formula, F is fitness value, and k is neutral net output node number;yiFor the expectation of i-th of node of BP neural network
Output;oiFor the prediction output of i-th of node;C is coefficient.
In the step 5, selection operation is:Using roulette method selective staining body, make the chromosome that fitness value is more excellent
Selected probability is bigger.
In formula:piIt is the probability for selecting i-th of chromosome;
Crossover operation is:Crossover operation is carried out using real number interior extrapolation method, four parent chromosomes are chosen herein and are intersected
Operation.
Mutation operation is:Using j-th of gene a of i-th of individualijMutation operation is carried out, variation method is as follows:
aij=aij+(aij-amax) * f (g) r > 0.5
aij=aij+(amin-aij)*f(g) r≤0.5
In formula, r is the random number between [0,1], amaxFor the previous of gene, aminFor the next time of gene;F (g)=rand ×
(1-g/Gmax)2;Rand is the random number between [0,1];G is current iteration number of times;GmaxFor maximum evolution number of times.
In the step 6, the process that input training sample is trained to obtained many parent BP-GA neutral nets is such as
Under:
6.1) output of hidden layer is calculated;
Hj=f (Sj)
Wherein l, m, n represent input layer number, node in hidden layer and output layer nodes, H respectivelyjIt is defeated for hidden layer
Go out value, f (x) takes S types (Sigmoid) function for transmission function;
6.2) output of output layer is calculated
Yb=f (U)
YbFor prediction output;
6.3), when network output is not waited with desired output, there is output error e in error calculationk, formula is as follows:
ek=Yk-Ybk
YkFor desired output;
6.4) renewal of weights;
ωjk=ωjk+ηHjek
6.5) renewal of threshold value, more predicated error e update network node threshold value A, B;
Bk=Bk+ek
6.6) weights after adjusting assign many parent BP-GA neutral nets, repeat step 6.1 again) -6.5) until by mistake
Difference exceedes untill setting the upper limit less than training objective or cycle-index, and many parent BP-GA neural metwork trainings are completed.
The present invention technical concept be:Hereditary (GA) algorithm be a kind of simulation nature genetic mechanism and theory of biological evolution and
Into a kind of parallel Stochastic search optimization method, this method has good ability of searching optimum, quickly approaches optimal result;BP
Neutral net then possesses good adaptivity and self-learning capability and good local optimal searching ability.Therefore using heredity calculation
The BP neural network forecast model that method is set up, can make full use of both advantages, set up a higher injector of precision
Can forecast model.
But traditional genetic algorithm, which is intragroup two chromosome of selection, carries out the dye after cross and variation is evolved
Colour solid, the offspring obtained in this way is only possible to have the advantages that two parent chromosomes, present invention further propose that adopting
With the artificial neural network of many parent genetic algorithms, it can be obtained by the chromosome of future generation synthesized to choosing a plurality of chromosome
The advantage of a plurality of parent chromosome simultaneously therefrom selects foundation of the optimum individual for neural network prediction model, therefore can obtain
More accurate result of calculation.
Based on this, the present invention proposes to predict drawing for injector using the method for the artificial neural network of many parent genetic algorithms
Penetrate coefficient (ε) and outlet back pressure (pc) etc. key parameter, using this method without considering complicated flow mechanism, you can convenient fast
Obtain fastly it is high-precision predict the outcome, be injector it is related manufacture and design, circulating research etc. provides foundation.
Beneficial effects of the present invention are mainly manifested in:Precision of prediction is higher, prediction is time-consuming shorter.
Brief description of the drawings
Fig. 1 is BP neural network topological structure;
Fig. 2 is many parent BP-GA neural network predictions injector performance flow charts;
Fig. 3 is many parent BP-GA neural network predictions resultant error figures.
Embodiment
The invention will be further described below in conjunction with the accompanying drawings.
1~Fig. 3 of reference picture, one kind is based on the injector performance of the artificial neural network using many parent genetic algorithms (GA)
Forecasting Methodology, comprises the following steps:
Step one, the collection and processing of data
For given injector, it is citation jet body pressure (Pe), working fluid pressure (Pp), outlet to collect relevant parameter
Back pressure (Pc) and mass ratio of induced-to-inducing air (ε), for accelerate neutral net convergence and reduce the training time, it is necessary to citation jet body pressure,
Working fluid pressure is normalized with outlet back pressure, it is arrived between [0,1], formula is as follows:
Wherein, k is the data after normalization, and x is to be normalized data, xminTo be normalized the minimum value in data,
xmaxTo be normalized maximum in data;
Step 2, the foundation of neural network structure
According to input/output argument, the input layer number of neutral net is judged as 2, and output layer nodes are also 2,
The basis for selecting empirical equation of node in hidden layer takes 5, and learning rate takes 0.1, and training objective takes 0.0001, and cycle-index is set
For 100 times.Each layer weights initial value takes the random number between [- 1,1].
Step 3, sets genetic algorithm initial parameter
Population scale is 10, and evolution number of times takes 50 times, and crossover probability is 40%, and mutation probability is 10%, by BP nerves
The weights of network constitute the neutral net set up in genetic coding, the step 2 that reruns with threshold value and record each run completion
Genetic coding set up population.
Step 4, sets up evaluation function, and the result that the neutral net constructed using every group chromosome is predicted is foundation, meter
Calculate and record corresponding fitness value.
Step 5, genetic algorithm processing
A. selection operation:With certain probability selection individual for producing the next generation in old colony, it is general that individual is selected
Rate is relevant with fitness value, and the value of fitness is better, and selected probability is bigger.
B. many parent crossover operations:Refer to select four parent chromosomes to carry out building for chromosome of future generation from population
It is vertical, by the combined crosswise of chromosome, produce new individual.
C. mutation operation:Refer to an optional individual from colony, certain section coding in selected chromosome is become
It is different to produce more excellent individual.
Step 6, the genetic coding in the chromosome newly obtained is decoded, and calculates fitness, and group enters with original seed
Row compares, and selects the superior and eliminates the inferior.Four-step 5 of repeat step, is constantly selected individual in population, many parents intersect, variation
Operate and record fitness value, reach the evolution number of times upper limit.The corresponding chromosome of fitness optimal solution is many fathers set up
For the threshold value and weights corresponding to BP-GA neutral nets, obtained many parent BP-GA neutral nets are trained, until pre-
The error for surveying output variable is less than set-point.
Step 7, in practical engineering application, is carried out with many parent BP-GA neutral nets to certain given injector performance
Input variable (citation jet body pressure p under prediction, measurement virtual conditione, working fluid pressure pp), it is pre- by injector performance
Survey many parent BP-GA neutral nets and obtain Prediction Parameters (outlet back pressure pc, mass ratio of induced-to-inducing air ε), by pcRenormalization is carried out to be changed
Enter BP-GA neural network prediction outlet back pressures pcActual value, formula is as follows:
X=k (xmax-xmin)+xmin。
In the step 3, genetic algorithm initial parameter is set, makes population scale take 10, evolution number of times takes 50 times, intersected
Probability is set to 40%, and mutation probability is set to 10%.Each individual is made up of the weights of neutral net with threshold value, due to this BP-GA
Neural network topology structure is 2-5-2, i.e., input layer has 2 nodes, and output layer has 2 nodes, and hidden layer has 5 nodes, institute
To have 2 × 5 × 2=20 weights, 5+2 threshold value, so this genetic algorithm individual chromosome code length is 27.To step
Rapid two artificial neural networks set up carry out initial training, constitute chromosome with obtained weights, threshold value, repeat ten times
To ten group chromosomes individual and constitute population.
In the step 4, evaluation function is set up.The result of the neural network prediction constructed using every group chromosome as according to
According to, calculate and record corresponding fitness value, its more accurate corresponding fitness value that predicts the outcome is smaller, fitness calculating public affairs
Formula is as follows:
In formula, F is fitness value, and k is neutral net output node number;yiFor the expectation of i-th of node of BP neural network
Output;oiFor the prediction output of i-th of node;C is coefficient.
In the step 5, selection operation is:Using roulette method selective staining body, make the chromosome that fitness value is more excellent
Selected probability is bigger.
In formula:piIt is the probability for selecting i-th of chromosome.
Many parent crossover operations are:Crossover operation is carried out using real number interior extrapolation method, four parent chromosomes are chosen herein and are entered
Row crossover operation, by kth in colony1Individual chromosome ak1, kth2Individual chromosome ak2、k3Individual chromosome ak3And kth4Individual chromosome ak4
Crossover operation is carried out, method is as follows:
ak1=ak1p1+ak2p2+ak3p3+ak4p4
ak2=ak2p1+ak1p2+ak3p3+ak4p4
ak3=ak3p1+ak1p2+ak2p3+ak4p4
ak4=ak4p1+ak1p2+ak2p3+ak3p4
Wherein, p1、p2、p3And p4Be random number between [0,1] and they and be 1.
Mutation operation is:Using j-th of gene a of i-th of individualijMutation operation is carried out, variation method is as follows:
aij=aij+(aij-amax) * f (g) r > 0.5
aij=aij+(amin-aij)*f(g) r≤0.5
In formula, r is the random number between [0,1], amaxFor the previous of gene, aminFor the next time of gene;F (g)=rand ×
(1-g/Gmax)2;Rand is the random number between [0,1];G is current iteration number of times;GmaxFor maximum evolution number of times.
In the step 6, the genetic coding in the chromosome newly obtained is decoded, genetic algorithm meter is obtained
The weights and threshold value calculated are set up neutral net and trained, and the adaptation of new individual is obtained according to the fitness computing formula of step 4
Degree, group is compared with original seed, selects the superior and eliminates the inferior.Repeat step four-five, constantly carries out cross and variation behaviour to individual in population
Make, and the new fitness for producing individual is compared with former population's fitness, select the superior and eliminate the inferior, until reaching on iterations
Limit.Chromosome corresponding to fitness optimal solution is decoded, weights is obtained and is used for carrying out BP-GA nerve nets with threshold value
The foundation of network.
In the step 6, the process that input training sample is trained to obtained BP-GA neutral nets is as follows:
6.1) output of hidden layer is calculated.
Hj=f (Sj)
Wherein i, j, k represent input layer number, node in hidden layer and output layer nodes respectively.HjIt is defeated for hidden layer
Go out value.F (x) takes S types (Sigmoid) function for transmission function.
6.2) output of output layer is calculated
Yb=f (U)
YbFor prediction output.
6.3), when network output is not waited with desired output, there is output error e in error calculationk, formula is as follows:
ek=Yk-Ybk
YkFor desired output.
6.4) renewal of weights.
ωjk=ωjk+ηHjek
6.5) renewal of threshold value, more predicated error e update network node threshold value A, B.
Bk=Bk+ek
6.6) weights after adjusting assign BP-GA neutral nets, repeat step 6.1 again) -6.5) until error is less than instruction
Practice target or cycle-index exceedes untill setting the upper limit, many parent BP-GA neural metwork trainings are completed.
Example:Preferably to embody the effect of the present invention, the method for the present invention is now subjected to actual motion.Using document 3
(IW.Eames et al.A theoretical and experimental study of a small-scale steam
jet refrigerator.International journal of refrigeration,18(6):378-386,1995;I.e.
The theoretical and experimental study International refrigeration journals of the small steam ejector refrigeration machines such as IW.Eames, 18 (6):378-386,
1995) method in obtains 110 groups of data, randomly selects 80 groups of data according to certain rules therefrom as training sample, uses
The above method carries out successive ignition training and obtains neutral net.To remaining 30 groups of data as test data, input sample is chosen
This (citation jet body pressure pe, working fluid pressure pp) carry out outlet back pressure p using obtained many parent BP-GA neutral netscWith
Mass ratio of induced-to-inducing air ε prediction, obtained result and result by references are compared, and assess the precision of prediction of the neutral net.
Specifically predict the outcome with literature value as shown in table 1 and Fig. 3.
Table 1
The mean error finally predicted the outcome is 0.29%, and worst error is 1.1%.Predict the outcome more accurate, use
Injector mass ratio of induced-to-inducing air and Forecasting Methodology and the conventional method such as document 1 of outlet back pressure that many parent GA-BP neutral nets are completed
And the method in document 2 be compared it can be found that using conventional model mean error more than in 5-10%, and worst error
Up to more than 15%.It can be seen that the method for use this patent can greatly promote prediction on the premise of quick be predicted is ensured
Precision.
Claims (6)
1. a kind of injector performance Forecasting Methodology of the artificial neural network based on using many parent genetic algorithms, its feature exists
In:Comprise the following steps:
Step one, the collection and processing of data
For given injector, collect relevant parameter be citation jet body pressure Pe, working fluid pressure Pp, outlet back pressure Pc and
Mass ratio of induced-to-inducing airCitation jet body pressure, working fluid pressure and outlet back pressure are normalized, make its to [0,1] it
Between, formula is as follows:
Wherein, k is the data after normalization, and x is to be normalized data, xminTo be normalized the minimum value in data, xmaxFor
It is normalized maximum in data;
Step 2, the foundation of neural network structure
According to input/output argument, the input and output layer nodal point number of neutral net is determined, the hidden layer number of plies is one layer, and node is several
According to empirical equation:Draw, wherein l, n are input, output node number, and a is constant, judge the defeated of neutral net
Enter node layer number, output layer nodes are chosen node in hidden layer, learning rate, training objective, cycle-index and set and each layer power
It is worth initial value;
Step 3, sets genetic algorithm initial parameter
Population scale, evolution number of times, crossover probability and mutation probability are set, the weights of BP neural network and threshold value are constituted into heredity
Coding, the neutral net set up in the step 2 that reruns simultaneously records the genetic coding of each run completion and sets up population;
Step 4, sets up evaluation function, and the result that the neutral net constructed using every group chromosome is predicted is calculated simultaneously as foundation
Record corresponding fitness value;
Step 5, genetic algorithm processing
A. selection operation:Be used for producing the next generation with certain probability selection individual in old colony, the selected probability of individual with
Fitness value is relevant, and the value of fitness is better, and selected probability is bigger;
B. many parent crossover operations:Refer to the foundation for selecting a plurality of parent chromosome to carry out chromosome of future generation from population, lead to
The combined crosswise of chromosome is crossed, new individual is produced;
C. mutation operation:Refer to an optional individual from colony, certain section in selected chromosome is encoded into row variation with
Produce more excellent individual;
Step 6, the genetic coding in the chromosome newly obtained is decoded, and calculates fitness, and group is compared with original seed
Compared with the survival of the fittest;Four-step 5 of repeat step, is constantly selected, is intersected to individual in population, mutation operation and recorded
Fitness value, reaches the evolution number of times upper limit;The corresponding chromosome of fitness optimal solution is many parent BP-GA nerves set up
Obtained many parent BP-GA neutral nets are trained by threshold value and weights corresponding to network, until prediction output variable
Error is less than set-point;
Step 7, in practical engineering application, is carried out in advance with many parent BP-GA neutral nets to certain given injector performance
Survey, the input variable under measurement virtual condition, i.e. citation jet body pressure pe, working fluid pressure pp, pass through many parent BP-GA god
Prediction Parameters are obtained through network injector performance forecast model, Prediction Parameters are outlet back pressure pc, mass ratio of induced-to-inducing airBy pcCarry out anti-
Normalization obtains improving BP-GA neural network prediction outlet back pressures pcActual value.
2. the injector performance prediction side of the artificial neural network as claimed in claim 1 based on using many parent genetic algorithms
Method, it is characterised in that:In the step 2, the selection of node in hidden layer, learning rate takes 0.1~0.2, and training objective takes 10-3
~10-6, each layer weights initial value takes the random number between [- 1,1].
3. the injector performance of the artificial neural network as claimed in claim 1 or 2 based on using many parent genetic algorithms is pre-
Survey method, it is characterised in that:In the step 3, genetic algorithm initial parameter is set, makes population scale take N1, evolution number of times takes
N2Secondary, crossover probability is set to 20%~50%, and mutation probability is set to 1%~10%, each individual by neutral net weights with
Threshold value is constituted, and BP-GA neural network topology structures have l node for l-m-n, i.e. input layer, and hidden layer has m node, output
Layer has n node, has l × m × n weights, m+n threshold value, so genetic algorithm individual chromosome code length is l × m
× n+m+n;Initial training is carried out to the artificial neural network that step 2 is set up, chromosome is constituted with obtained weights, threshold value
Individual, repeats n times and obtains the individual of N group chromosomes and constitute population.
4. the injector performance of the artificial neural network as claimed in claim 1 or 2 based on using many parent genetic algorithms is pre-
Survey method, it is characterised in that:In the step 4, the result of the neural network prediction constructed using every group chromosome as foundation,
Calculate and record corresponding fitness value, its more accurate corresponding fitness value that predicts the outcome is smaller, fitness computing formula
It is as follows:
In formula, F is fitness value, and k is neutral net output node number;yiFor the desired output of i-th of node of BP neural network;
oiFor the prediction output of i-th of node;C is coefficient.
5. the injector performance of the artificial neural network as claimed in claim 1 or 2 based on using many parent genetic algorithms is pre-
Survey method, it is characterised in that:In the step 5, selection operation is:Using roulette method selective staining body,
In formula:piIt is the probability for selecting i-th of chromosome;N is population at individual number;
Many parent crossover operations are:Crossover operation is carried out using real number interior extrapolation method, X bar parent chromosomes are chosen herein and are intersected
Operation;
Mutation operation is:Using j-th of gene a of i-th of individualijMutation operation is carried out, variation method is as follows:
aij=aij+(aij-amax) * f (g) r > 0.5
aij=aij+(amin-aij)*f(g) r≤0.5
In formula, r is the random number between [0,1], amaxFor the previous of gene, aminFor the next time of gene;F (g)=rand × (1-
g/Gmax)2;Rand is the random number between [0,1];G is current iteration number of times;GmaxFor maximum evolution number of times.
6. the injector performance of the artificial neural network as claimed in claim 1 or 2 based on using many parent genetic algorithms is pre-
Survey method, it is characterised in that:In the step 6, the mistake that input training sample is trained to obtained BP-GA neutral nets
Journey is as follows:
6.1) output of hidden layer is calculated;
Hj=f (Sj)
Wherein l, m, n represent input layer number, node in hidden layer and output layer nodes, H respectivelyjFor hidden layer output valve,
F (x) takes S types (Sigmoid) function for transmission function;
6.2) output of output layer is calculated
Yb=f (U)
YbFor prediction output;
6.3), when network output is not waited with desired output, there is output error e in error calculationk, formula is as follows:
ek=Yk-Ybk
YkFor desired output;
6.4) renewal of weights;
ωjk=ωjk+ηHjek
6.5) renewal of threshold value, more predicated error e update network node threshold value A, B;
Bk=Bk+ek
6.6) weights after adjusting assign BP-GA neutral nets, repeat step 6.1 again) -6.5) until error is less than training mesh
Mark or cycle-index exceed untill setting the upper limit, and many parent BP-GA neural metwork trainings are completed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710107962.0A CN106951983A (en) | 2017-02-27 | 2017-02-27 | Injector performance Forecasting Methodology based on the artificial neural network using many parent genetic algorithms |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710107962.0A CN106951983A (en) | 2017-02-27 | 2017-02-27 | Injector performance Forecasting Methodology based on the artificial neural network using many parent genetic algorithms |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106951983A true CN106951983A (en) | 2017-07-14 |
Family
ID=59467772
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710107962.0A Pending CN106951983A (en) | 2017-02-27 | 2017-02-27 | Injector performance Forecasting Methodology based on the artificial neural network using many parent genetic algorithms |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106951983A (en) |
Cited By (34)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107909149A (en) * | 2017-10-26 | 2018-04-13 | 西北农林科技大学 | A kind of Temperature in Greenhouse Forecasting Methodology based on Genetic BP Neutral Network |
CN107909220A (en) * | 2017-12-08 | 2018-04-13 | 天津天大求实电力新技术股份有限公司 | Electric heating load prediction method |
CN108038570A (en) * | 2017-12-08 | 2018-05-15 | 云南电网有限责任公司 | Load forecasting method during ultrashort based on Descartes's genetic programming Recognition with Recurrent Neural Network |
CN108280791A (en) * | 2018-02-01 | 2018-07-13 | 交通运输部天津水运工程科学研究所 | Method is determined based on the sewage deep-sea optimization discharge response relation of power flow changing |
CN108417032A (en) * | 2018-03-19 | 2018-08-17 | 中景博道城市规划发展有限公司 | A kind of downtown area curb parking demand analysis prediction technique |
CN108460461A (en) * | 2018-02-06 | 2018-08-28 | 吉林大学 | Mars earth shear parameters prediction technique based on GA-BP neural networks |
CN108872508A (en) * | 2018-05-08 | 2018-11-23 | 苏州科技大学 | A kind of eutrophy quality evaluation method of GA-BP optimization TSFNN |
CN108876054A (en) * | 2018-07-06 | 2018-11-23 | 国网河南省电力公司郑州供电公司 | Short-Term Load Forecasting Method based on improved adaptive GA-IAGA optimization extreme learning machine |
CN109117491A (en) * | 2018-06-15 | 2019-01-01 | 北京理工大学 | A kind of agent model construction method for the higher-dimension small data merging expertise |
CN109242142A (en) * | 2018-07-25 | 2019-01-18 | 浙江工业大学 | A kind of spatio-temporal segmentation parameter optimization method towards infrastructure networks |
CN109284818A (en) * | 2018-09-07 | 2019-01-29 | 北方爆破科技有限公司 | A kind of blasting vibration control prediction technique based on accident tree and genetic algorithm |
CN109508488A (en) * | 2018-11-07 | 2019-03-22 | 西北工业大学 | Contour peening technological parameter prediction technique based on genetic algorithm optimization BP neural network |
CN109598092A (en) * | 2018-12-28 | 2019-04-09 | 浙江工业大学 | Merge the air source heat pump multi-objective optimization design of power method of BP neural network and more parent genetic algorithms |
CN109634121A (en) * | 2018-12-28 | 2019-04-16 | 浙江工业大学 | More parent genetic algorithm air source heat pump multiobjective optimization control methods based on radial basis function neural network |
CN109829244A (en) * | 2019-02-25 | 2019-05-31 | 浙江工业大学 | The blower optimum design method of algorithm optimization depth network and three generations's genetic algorithm |
CN109858093A (en) * | 2018-12-28 | 2019-06-07 | 浙江工业大学 | The air source heat pump multi-objective optimization design of power method of the non-dominated sorted genetic algorithm of SVR neural network aiding |
CN109918749A (en) * | 2019-02-25 | 2019-06-21 | 北京妙微科技有限公司 | The fan design two generations algorithm Multipurpose Optimal Method of learning rate changing network modelling |
CN109932903A (en) * | 2019-02-25 | 2019-06-25 | 北京妙微科技有限公司 | The air-blower control Multipurpose Optimal Method of more parent optimization networks and genetic algorithm |
CN110033118A (en) * | 2019-02-25 | 2019-07-19 | 浙江工业大学 | Elastomeric network modeling and the blower multiobjective optimization control method based on genetic algorithm |
CN110113836A (en) * | 2018-12-29 | 2019-08-09 | 中国计量大学 | Scene-type intelligent classroom lighting system, control device and optimization and control method |
CN110457758A (en) * | 2019-07-16 | 2019-11-15 | 江西理工大学 | Prediction technique, device, system and the storage medium in Instability of Rock Body stage |
CN110533109A (en) * | 2019-09-03 | 2019-12-03 | 内蒙古大学 | A kind of storage spraying production monitoring data and characteristic analysis method and its device |
CN110837223A (en) * | 2018-08-15 | 2020-02-25 | 大唐南京发电厂 | Combustion optimization control method and system for gas turbine |
CN110889495A (en) * | 2019-12-04 | 2020-03-17 | 河南中烟工业有限责任公司 | State maintenance analysis method for silk making equipment based on active parameters |
CN111105027A (en) * | 2018-10-25 | 2020-05-05 | 航天科工惯性技术有限公司 | Landslide deformation prediction method based on GA algorithm and BP neural network |
CN111126707A (en) * | 2019-12-26 | 2020-05-08 | 华自科技股份有限公司 | Energy consumption equation construction and energy consumption prediction method and device |
CN111324989A (en) * | 2020-03-19 | 2020-06-23 | 重庆大学 | GA-BP neural network-based gear contact fatigue life prediction method |
CN111382842A (en) * | 2020-03-06 | 2020-07-07 | 佳源科技有限公司 | High-speed carrier communication dynamic routing method and system |
CN111461286A (en) * | 2020-01-15 | 2020-07-28 | 华中科技大学 | Spark parameter automatic optimization system and method based on evolutionary neural network |
CN111814401A (en) * | 2020-07-08 | 2020-10-23 | 重庆大学 | LED life prediction method of BP neural network based on genetic algorithm |
CN112330435A (en) * | 2020-09-29 | 2021-02-05 | 百维金科(上海)信息科技有限公司 | Credit risk prediction method and system for optimizing Elman neural network based on genetic algorithm |
CN112766548A (en) * | 2021-01-07 | 2021-05-07 | 南京航空航天大学 | Order completion time prediction method based on GASA-BP neural network |
CN113012814A (en) * | 2021-03-10 | 2021-06-22 | 浙江大学医学院附属邵逸夫医院 | Acute kidney injury volume responsiveness prediction method and system |
CN113449930A (en) * | 2021-07-27 | 2021-09-28 | 威海长和光导科技有限公司 | Optical fiber preform preparation quality prediction method based on BP neural network |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103105246A (en) * | 2012-12-31 | 2013-05-15 | 北京京鹏环球科技股份有限公司 | Greenhouse environment forecasting feedback method of back propagation (BP) neural network based on improvement of genetic algorithm |
CN104820977A (en) * | 2015-05-22 | 2015-08-05 | 无锡职业技术学院 | BP neural network image restoration algorithm based on self-adaption genetic algorithm |
-
2017
- 2017-02-27 CN CN201710107962.0A patent/CN106951983A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103105246A (en) * | 2012-12-31 | 2013-05-15 | 北京京鹏环球科技股份有限公司 | Greenhouse environment forecasting feedback method of back propagation (BP) neural network based on improvement of genetic algorithm |
CN104820977A (en) * | 2015-05-22 | 2015-08-05 | 无锡职业技术学院 | BP neural network image restoration algorithm based on self-adaption genetic algorithm |
Non-Patent Citations (3)
Title |
---|
付晓明: "计算机仿真", 《基于多子代遗传算法优化BP神经网络》 * |
孙俊 等: "基于MEA-BP神经网络的大米水分含量高光谱技术检测", 《食品科学》 * |
黄亮亮: "基于人工神经网络的喷射器性能预测及优化研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Cited By (47)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107909149A (en) * | 2017-10-26 | 2018-04-13 | 西北农林科技大学 | A kind of Temperature in Greenhouse Forecasting Methodology based on Genetic BP Neutral Network |
CN107909220A (en) * | 2017-12-08 | 2018-04-13 | 天津天大求实电力新技术股份有限公司 | Electric heating load prediction method |
CN108038570A (en) * | 2017-12-08 | 2018-05-15 | 云南电网有限责任公司 | Load forecasting method during ultrashort based on Descartes's genetic programming Recognition with Recurrent Neural Network |
CN108280791A (en) * | 2018-02-01 | 2018-07-13 | 交通运输部天津水运工程科学研究所 | Method is determined based on the sewage deep-sea optimization discharge response relation of power flow changing |
CN108280791B (en) * | 2018-02-01 | 2021-07-27 | 交通运输部天津水运工程科学研究所 | Sewage deep sea optimized discharge response relation determination method based on tidal current change |
CN108460461A (en) * | 2018-02-06 | 2018-08-28 | 吉林大学 | Mars earth shear parameters prediction technique based on GA-BP neural networks |
CN108417032A (en) * | 2018-03-19 | 2018-08-17 | 中景博道城市规划发展有限公司 | A kind of downtown area curb parking demand analysis prediction technique |
CN108872508A (en) * | 2018-05-08 | 2018-11-23 | 苏州科技大学 | A kind of eutrophy quality evaluation method of GA-BP optimization TSFNN |
CN109117491A (en) * | 2018-06-15 | 2019-01-01 | 北京理工大学 | A kind of agent model construction method for the higher-dimension small data merging expertise |
CN109117491B (en) * | 2018-06-15 | 2023-04-07 | 北京理工大学 | Agent model construction method of high-dimensional small data fusing expert experience |
CN108876054A (en) * | 2018-07-06 | 2018-11-23 | 国网河南省电力公司郑州供电公司 | Short-Term Load Forecasting Method based on improved adaptive GA-IAGA optimization extreme learning machine |
CN109242142A (en) * | 2018-07-25 | 2019-01-18 | 浙江工业大学 | A kind of spatio-temporal segmentation parameter optimization method towards infrastructure networks |
CN109242142B (en) * | 2018-07-25 | 2020-06-26 | 浙江工业大学 | Space-time prediction model parameter optimization method for infrastructure network |
CN110837223A (en) * | 2018-08-15 | 2020-02-25 | 大唐南京发电厂 | Combustion optimization control method and system for gas turbine |
CN109284818A (en) * | 2018-09-07 | 2019-01-29 | 北方爆破科技有限公司 | A kind of blasting vibration control prediction technique based on accident tree and genetic algorithm |
CN111105027A (en) * | 2018-10-25 | 2020-05-05 | 航天科工惯性技术有限公司 | Landslide deformation prediction method based on GA algorithm and BP neural network |
CN109508488A (en) * | 2018-11-07 | 2019-03-22 | 西北工业大学 | Contour peening technological parameter prediction technique based on genetic algorithm optimization BP neural network |
CN109508488B (en) * | 2018-11-07 | 2022-08-02 | 西北工业大学 | Shot peening forming process parameter prediction method based on genetic algorithm optimization BP neural network |
CN109634121A (en) * | 2018-12-28 | 2019-04-16 | 浙江工业大学 | More parent genetic algorithm air source heat pump multiobjective optimization control methods based on radial basis function neural network |
CN109598092A (en) * | 2018-12-28 | 2019-04-09 | 浙江工业大学 | Merge the air source heat pump multi-objective optimization design of power method of BP neural network and more parent genetic algorithms |
CN109634121B (en) * | 2018-12-28 | 2021-08-03 | 浙江工业大学 | Multi-parent genetic algorithm air source heat pump multi-objective optimization control method based on radial basis function neural network |
CN109858093A (en) * | 2018-12-28 | 2019-06-07 | 浙江工业大学 | The air source heat pump multi-objective optimization design of power method of the non-dominated sorted genetic algorithm of SVR neural network aiding |
CN112533317B (en) * | 2018-12-29 | 2023-09-08 | 中国计量大学 | Scene type classroom intelligent illumination optimization method |
CN110113836A (en) * | 2018-12-29 | 2019-08-09 | 中国计量大学 | Scene-type intelligent classroom lighting system, control device and optimization and control method |
CN112533317A (en) * | 2018-12-29 | 2021-03-19 | 中国计量大学 | Scene type classroom intelligent lighting optimization method |
CN109918749A (en) * | 2019-02-25 | 2019-06-21 | 北京妙微科技有限公司 | The fan design two generations algorithm Multipurpose Optimal Method of learning rate changing network modelling |
CN110033118A (en) * | 2019-02-25 | 2019-07-19 | 浙江工业大学 | Elastomeric network modeling and the blower multiobjective optimization control method based on genetic algorithm |
CN109829244A (en) * | 2019-02-25 | 2019-05-31 | 浙江工业大学 | The blower optimum design method of algorithm optimization depth network and three generations's genetic algorithm |
CN109932903A (en) * | 2019-02-25 | 2019-06-25 | 北京妙微科技有限公司 | The air-blower control Multipurpose Optimal Method of more parent optimization networks and genetic algorithm |
CN110457758A (en) * | 2019-07-16 | 2019-11-15 | 江西理工大学 | Prediction technique, device, system and the storage medium in Instability of Rock Body stage |
CN110457758B (en) * | 2019-07-16 | 2023-03-24 | 江西理工大学 | Method, device and system for predicting unstable phase of rock mass and storage medium |
CN110533109A (en) * | 2019-09-03 | 2019-12-03 | 内蒙古大学 | A kind of storage spraying production monitoring data and characteristic analysis method and its device |
CN110889495B (en) * | 2019-12-04 | 2023-06-23 | 河南中烟工业有限责任公司 | State maintenance analysis method of filament making equipment based on active parameters |
CN110889495A (en) * | 2019-12-04 | 2020-03-17 | 河南中烟工业有限责任公司 | State maintenance analysis method for silk making equipment based on active parameters |
CN111126707B (en) * | 2019-12-26 | 2023-10-27 | 华自科技股份有限公司 | Energy consumption equation construction and energy consumption prediction method and device |
CN111126707A (en) * | 2019-12-26 | 2020-05-08 | 华自科技股份有限公司 | Energy consumption equation construction and energy consumption prediction method and device |
CN111461286B (en) * | 2020-01-15 | 2022-03-29 | 华中科技大学 | Spark parameter automatic optimization system and method based on evolutionary neural network |
CN111461286A (en) * | 2020-01-15 | 2020-07-28 | 华中科技大学 | Spark parameter automatic optimization system and method based on evolutionary neural network |
CN111382842A (en) * | 2020-03-06 | 2020-07-07 | 佳源科技有限公司 | High-speed carrier communication dynamic routing method and system |
CN111324989A (en) * | 2020-03-19 | 2020-06-23 | 重庆大学 | GA-BP neural network-based gear contact fatigue life prediction method |
CN111324989B (en) * | 2020-03-19 | 2024-01-30 | 重庆大学 | Gear contact fatigue life prediction method based on GA-BP neural network |
CN111814401A (en) * | 2020-07-08 | 2020-10-23 | 重庆大学 | LED life prediction method of BP neural network based on genetic algorithm |
CN111814401B (en) * | 2020-07-08 | 2023-10-27 | 重庆大学 | LED life prediction method of BP neural network based on genetic algorithm |
CN112330435A (en) * | 2020-09-29 | 2021-02-05 | 百维金科(上海)信息科技有限公司 | Credit risk prediction method and system for optimizing Elman neural network based on genetic algorithm |
CN112766548A (en) * | 2021-01-07 | 2021-05-07 | 南京航空航天大学 | Order completion time prediction method based on GASA-BP neural network |
CN113012814A (en) * | 2021-03-10 | 2021-06-22 | 浙江大学医学院附属邵逸夫医院 | Acute kidney injury volume responsiveness prediction method and system |
CN113449930A (en) * | 2021-07-27 | 2021-09-28 | 威海长和光导科技有限公司 | Optical fiber preform preparation quality prediction method based on BP neural network |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106951983A (en) | Injector performance Forecasting Methodology based on the artificial neural network using many parent genetic algorithms | |
CN106980897A (en) | A kind of injector performance parameter prediction method of the BP artificial neural networks based on learning rate changing | |
CN109932903A (en) | The air-blower control Multipurpose Optimal Method of more parent optimization networks and genetic algorithm | |
CN109634121B (en) | Multi-parent genetic algorithm air source heat pump multi-objective optimization control method based on radial basis function neural network | |
CN106960217A (en) | The Forecasting Methodology of injector performance based on the BP artificial neural networks using depth Adaboost algorithm | |
CN114692265B (en) | Zero-carbon building optimization design method based on deep reinforcement learning | |
CN109084415B (en) | Central air conditioner operation parameter optimizing method based on neural network and genetic algorithm | |
CN109598092A (en) | Merge the air source heat pump multi-objective optimization design of power method of BP neural network and more parent genetic algorithms | |
CN107067121A (en) | A kind of improvement grey wolf optimized algorithm based on multiple target | |
CN105404926B (en) | Aluminum electrolysis production technique optimization method based on BP neural network Yu MBFO algorithms | |
CN107092255A (en) | A kind of multi-robots path-planning method based on improved adaptive GA-IAGA | |
CN106960075A (en) | The Forecasting Methodology of the injector performance of RBF artificial neural network based on linear direct-connected method | |
CN104616215A (en) | Energy efficiency comprehensive evaluation method for thermal power plant | |
CN105809297A (en) | Thermal power plant environment economic dispatching method based on multi-target differential evolution algorithm | |
CN104765690A (en) | Embedded software test data generating method based on fuzzy-genetic algorithm | |
CN106909986A (en) | A kind of soil re-development plan method of use ant colony multiple target layout optimization model | |
CN106845012A (en) | A kind of blast furnace gas system model membership function based on multiple target Density Clustering determines method | |
CN106651628A (en) | Regional cool and thermal power comprehensive energy optimizing configuration method and apparatus based on graph theory | |
CN104680025B (en) | Oil pumper parameter optimization method based on genetic algorithm extreme learning machine | |
CN108224446A (en) | A kind of automatic combustion Study on Decision-making Method for Optimization of Refuse Incineration Process | |
CN109886448A (en) | Using learning rate changing BP neural network and the heat pump multiobjective optimization control method of NSGA-II algorithm | |
CN107400935A (en) | Adjusting method based on the melt-spinning technology for improving ELM | |
CN105046453A (en) | Construction engineering project cluster establishment method introducing cloud model for evaluation and selection | |
CN106202753A (en) | Shield mortar performance optimization method is realized based on BP neutral net and genetic algorithm | |
CN104656620A (en) | Comprehensive evaluation system for remanufacturing of heavy-duty machine tool |
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: 20170714 |