CN107316099A - Ammunition Storage Reliability Forecasting Methodology based on particle group optimizing BP neural network - Google Patents
Ammunition Storage Reliability Forecasting Methodology based on particle group optimizing BP neural network Download PDFInfo
- Publication number
- CN107316099A CN107316099A CN201710361921.4A CN201710361921A CN107316099A CN 107316099 A CN107316099 A CN 107316099A CN 201710361921 A CN201710361921 A CN 201710361921A CN 107316099 A CN107316099 A CN 107316099A
- Authority
- CN
- China
- Prior art keywords
- particle
- value
- neural network
- network
- storage reliability
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of Ammunition Storage Reliability Forecasting Methodology based on particle group optimizing BP neural network, mainly particle swarm optimization algorithm is combined with BP neural network, and apply to Ammunition Storage Reliability prediction field, avoid intelligent algorithm self-defect and solve existing BP neural network convergence rate slowly, the problem of being easily trapped into locally optimal solution.Its planning step is：Using ammunition as research object, goal in research is predicted as with Ammunition Storage Reliability, according to the changing rule of Ammunition Storage Reliability data, neural network prediction model is set up and Ammunition Storage Reliability is predicted；Global intelligent optimization algorithm PSO is further chosen, PSO BP neural network forecast models, the weights and threshold value of Optimized BP Neural Network are set up, so as to reach the precision of Ammunition Storage Reliability prediction.
Description
Technical field
The present invention relates to Ammunition Storage Reliability technical field, and in particular to one kind is based on particle group optimizing BP neural network
Ammunition Storage Reliability Forecasting Methodology.
Background technology
Storage life is a very important technical indicator of ammunition, and the assessment to the ammunition storage life-span is also one non-
Often important work, in order to accurately obtain the situation of Ammunition Storage Reliability, it is ensured that ammunition safe storage and provide performance at any time
Good ammunition, must just carry out the research in ammunition storage life-span, therefore carry out scientific forecasting right and wrong to Ammunition Storage Reliability
It is often necessary.Early in the beginning of the seventies, countries in the world have begun to the research to Ammunition Storage Reliability；Domestic and foreign scholars in recent years
Forecasting Methodology to Ammunition Storage Reliability has carried out numerous studies, including based on normal stress storage life test data assessment
Method, based on accelerated life test data assessment method, based on Markov process, based on model prediction, based on fuzzy number prediction
Method, the appraisal procedure based on artificial neural network and Forecasting Methodology based on neural network ensemble model etc., have scholar to weapon
The Storage Life Prediction method of system studied, and result verification neural network prediction model is more excellent, but is due to nerve
Network convergence speed is slow, is easily trapped into local minimum, proposes the Forecasting Methodology based on neural network ensemble model；Meanwhile, have
Many scholars have carried out substantial amounts of improvement to BP algorithm, but resulting convergence rate and precision are limited, are easily trapped into office
Portion's optimal solution.
In recent years, the colony intelligence optimized algorithm iteration BP nerves of many scholars and expert both domestic and external global optimization performance
Network, such as genetic algorithm, particle cluster algorithm, these algorithms can overcome the defect of BP algorithm.But, intelligent algorithm is in itself again
There is the defect of inherence.Most ripe genetic algorithm is studied in such as intelligent optimization method, theoretically it can solve various multiple
Miscellaneous problem, in fact genetic algorithm with complicated genetic manipulation because causing net training time and complexity occur exponential
Increase, while algorithm also lacks effective search mechanisms of regional area, may slowly there is stagnation behavior to go out in late convergence
It is existing.
At present, in other areas, scholar proposed using particle swarm optimization algorithm with BP neural network be combined into
Row prediction, PSO algorithms have the characteristic of only speed-displacement model simple operations, if according to the speed of oneself can since
The direction of search is determined, the defect in genetic algorithm can be avoided, and available for multiple target, task distribution, pattern classification and height
Tie up the optimization problems such as complicated data processing.
Because the PSO algorithm development times are shorter, theoretical foundation and application also need to further further investigation, both at home and abroad
PSO algorithms are combined with BP neural network and are applied to the prediction of storage reliability, relevant document is also seldom；Particularly by PSO
Algorithm optimization neural network model applies to this special dimension of Ammunition Storage Reliability prediction, does not find also both at home and abroad at present
Relevant document.
It is desirable to have a kind of PSO algorithm optimization BP networks that are based on to Ammunition Storage Reliability Forecasting Methodology, it can overcome
Or at least mitigate the drawbacks described above of prior art.
The content of the invention
The technical problem to be solved in the present invention is to be directed to deficiency of the prior art, it is proposed that one kind is based on particle group optimizing
The Ammunition Storage Reliability Forecasting Methodology of BP neural network, global intelligent optimization algorithm PSO optimizations are chosen in the algorithm to be had entirely
The network parameter of office, makes up BP neural network convergence rate and is absorbed in the defects such as local minimum slowly, easily, reach Ammunition Storage Reliability
The precision of prediction.
In order to achieve the above object, it is of the invention based on a kind of Ammunition Storage Reliability of particle group optimizing BP neural network
Forecasting Methodology.First, using the changing rule of Ammunition Storage Reliability data, neural network prediction model is set up；Further, it is sharp
With the weights and threshold value of global intelligent optimization algorithm PSO Optimized BP Neural Networks, Ammunition Storage Reliability prediction is carried out.Specific mistake
Journey includes following several steps：
1. parameter initialization is set, it is divided into two parts：Part I is the parameter setting of BP networks, according to input data and defeated
Go out data, determine input, output node number and the implicit number of plies and its nodes of BP networks；Part II is particle cluster algorithm
Parameter setting, particle number is set, particle initial position and speed and their scope, maximum iteration, study because
Sub- c1And c2Etc. parameter；Inertia weight uses the linear decrease Weight Algorithm proposed by shi, i.e.,
, in formula, ,
In formulaIt is inertia weight,For maximum iteration in PSO algorithms, t is current iteration number of times；
2. calculate the initial fitness value of particle.The error obtained during network forward-propagating is mean square error, i.e. population
The fitness function of algorithm is that situation that network output layer nodes are 1 is only discussed in mean square error formula, the present invention, and it is adapted to
Spend functional form：
,
In formula, N represents the sample group number of training；Represent the desired output of the network output node of i-th of sample；Represent
The real output value of the network output node of i-th of sample.Fitness function according to providing calculates each particle in solution space
Interior adaptation value matrix, obtains individual adaptive optimal control angle value and global optimum's fitness value, finds the position of optimal particle.
3. the fitness value of each particle is made comparisons with individual adaptive optimal control angle value and global optimum's fitness value,
To judge the quality of current location, and determine the individual optimal and global optimum of particle position；（1）The fitness value difference of particle
Compared with individual optimal value pbest, global optimum gbest, if the current fitness value of particle is better than individual adaptive optimal control degree
Value, then using currency as individual adaptive optimal control angle value, and replace with the current location of particle the personal best particle of history；
（2）If the current individual adaptive optimal control angle value in particle is better than the global optimum of history, current individual optimal location is made
For the global optimum position of the particle, i.e. global optimum be individual it is optimal in best values.
4. speed and position of each particle in every dimension are updated, shown in the mode equation below of renewal：
Wherein, i=1,2 ..., N；N is population scale；d=1,2,…,D；For particle i personal best particles pbest d
Tie up component；Component is tieed up for the d of optimal value gbest in population；C1 is the accelerated factor of cognitive part；C2 is social part
Accelerated factor；Rand () is the random number between [0,1]；Inertia weight w is a nonnegative number；Represent after the t times iteration
Particle i circle in the air speed d dimension component；Represent the d dimension components of particle i positions after the t times iteration.
5. above-mentioned particle cluster algorithm is updated to initial weight and threshold that obtained global optimum is mapped to BP neural network
Value, then carries out the training of network；Using the weights and threshold value of PSO algorithm optimizations in BP neural network back-propagation process,
Iterative cycles training further adjusts weights and threshold value, reaches that maximum or training error are set less than expectation when meeting iterations
Definite value, then end loop training, output result；Otherwise 2 are gone to, execution above-mentioned steps are continued cycling through.
6. carrying out the test of BP neural network using test sample, input test sample utilizes the above-mentioned network trained
Emulated, by output result with it is expected that setting value is made comparisons, so as to draw test effect.
The Ammunition Storage Reliability Forecasting Methodology based on particle group optimizing BP neural network of the present invention is for basic BP god
The problem of existing through network algorithm is improved from following 2 points：（1）The training functions of BP learning algorithms is respectively adopted dynamic
BP methods and learning rate changing BP methods are measured, accelerates the pace of learning of BP neural network, the learning efficiency of BP neural network is improved.
（2）Adjusted and optimized with global network parameter, i.e. weights and threshold value, with nerve using particle swarm optimization algorithm
Parameter of the online learning methods optimization with locality, it is to avoid be easily absorbed in the defect of local minimum, improve convergence rate, with up to
To precision of prediction.
Brief description of the drawings
Fig. 1 is the flow of the Ammunition Storage Reliability Forecasting Methodology based on particle group optimizing BP neural network of the present invention
Figure；
Fig. 2 is learning rate changing BP network error change curves；
Fig. 3 is PSO- learning rate changing BP network error change curves；
Fig. 4 is Momentum BP Neutral Network error change curve；
Fig. 5 is PSO- Momentum BP Neutral Network error change curves；
Fig. 6 is learning rate changing BP network accuracy rate curves；
Fig. 7 is PSO- learning rate changing BP network accuracy rate curves；
Fig. 8 is Momentum BP Neutral Network accuracy rate curve；
Fig. 9 is PSO- Momentum BP Neutral Network accuracy rate curves；
Figure 10 is each neural network forecast error amount comparison curves.
Specific implementation method
To make the purpose, technical scheme and advantage of the invention implemented clearer, below in conjunction with by attached in inventive embodiments
Figure, the technical scheme in the embodiment of the present invention is further described in more detail.The present embodiment based on particle group optimizing BP god
Ammunition Storage Reliability Forecasting Methodology through network, specifically includes following steps.
1. parameter initialization is set, point two parts：Part I is the parameter setting of BP networks, according to input data and
Output data, determines input, output node number and the implicit number of plies and its nodes of BP networks；Part II is that population is calculated
The parameter setting of method, sets particle number, particle initial position and speed and their scope, maximum iteration, study
Factor c1And c2Etc. parameter；Inertia weight uses the linear decrease Weight Algorithm proposed by shi, i.e.,
, in formula, ,
2. calculate the initial fitness value of particle.The error obtained during network forward-propagating is mean square error, i.e. population
The fitness function of algorithm is that situation that network output layer nodes are 1 is only discussed in mean square error formula, the present invention, and it is adapted to
Spend functional form：
,
In formula, N represents the sample group number of training；Represent the desired output of the network output node of i-th of sample；Represent
The real output value of the network output node of i-th of sample.Fitness function according to providing calculates each particle in solution space
Interior adaptation value matrix, obtains individual adaptive optimal control angle value and global optimum's fitness value, finds the position of optimal particle.
3. the fitness value of each particle is made comparisons with individual adaptive optimal control angle value and global optimum's fitness value,
To judge the quality of current location, and determine the individual optimal and global optimum of particle position；（1）The fitness value difference of particle
Compared with individual optimal value pbest, global optimum gbest, if the current fitness value of particle is better than individual adaptive optimal control degree
Value, then using currency as individual adaptive optimal control angle value, and replace with the current location of particle the personal best particle of history；
（2）If the current individual adaptive optimal control angle value in particle is better than the global optimum of history, current individual optimal location is made
For the global optimum position of the particle, i.e. global optimum be individual it is optimal in best values.
4. updating speed and position of each particle in every dimension, the mode of renewal is as shown in following equation：
Wherein, i=1,2 ..., N；N is population scale；d=1,2,…,D；For particle i personal best particles pbest d
Tie up component；Component is tieed up for the d of optimal value gbest in population；C1 is the accelerated factor of cognitive part；C2 is social part
Accelerated factor；Rand () is the random number between [0,1]；Inertia weight w is a nonnegative number；Represent after the t times iteration
Particle i circle in the air speed d dimension component；Represent the d dimension components of particle i positions after the t times iteration.
5. above-mentioned particle cluster algorithm is updated to initial weight and threshold that obtained global optimum is mapped to BP neural network
Value, then carries out the training of network；Using the weights and threshold value of PSO algorithm optimizations in BP neural network back-propagation process,
Iterative cycles training further adjusts weights and threshold value, reaches that maximum or training error are set less than expectation when meeting iterations
Definite value, then end loop training, output result；Otherwise 2 are gone to, execution above-mentioned steps are continued cycling through.
6. carrying out the test of BP neural network using test sample, input test sample utilizes the above-mentioned network trained
Emulated, by output result with it is expected that setting value is made comparisons, so as to draw test effect.
For the validity of verification method, that realizes three-layer neural network training function herein adds momentum BP methods
' traingdm ' and learning rate changing BP methods ' traingda ', and with PSO algorithms to being optimized respectively to it.Input layer and output
The neuron number of layer is directly determined that node in hidden layer utilizes " trial and error procedure " to be defined as 11 by problem in itself, according to by mistake
Difference and convergence number of times determine that learning rate is 0.1.
E-learning is carried out by identical training sample set, and verified by test sample set pair algorithm, each model
Network convergence speed ratio compared with：From Fig. 2 ~ Fig. 5, anticipation error is in learning rate changing BP networks by being reached after iteration 67 times；
Just reached for 50 times by iteration in learning rate changing BP networks after the PSO optimizations of the present invention；BP net of the anticipation error in momentum
By being reached after 194 successive ignitions in network；Just reached for 174 times by iteration in Momentum BP Neutral Network after the PSO optimizations of the present invention；
The problems such as independent BP networks are due to being easily absorbed in local minimum, convergence rate is slow, shock range is relatively large in an iterative process；
The BP networks of PSO optimizations through the present invention not only increase in convergence rate, and curve is also more smooth；Each model prediction
Accuracy rate compares：From Fig. 6 ~ Fig. 9, learning rate changing BP networks and Momentum BP Neutral Network training precision that PSO of the invention optimizes
More than 99% is reached, compared with single learning rate changing BP networks and Momentum BP Neutral Network, the BP networks optimized through PSO are in training
Slightly improved in precision.
The prediction error value of each model compares under Figure 10 test sets, as seen from the figure, no matter training function chooses momentum BP methods
Or learning rate changing BP methods, the BP network models after PSO algorithm optimizations predict the outcome better than two kinds independent BP network models
The result of prediction, had both reduced the extreme value of error, and the excursion of error is reduced again.BP neural network of the present invention to standard
The advantage that optimizes is done it is clear that the BP neural network after optimization effectively avoids inherent defects, algorithm is improved
Convergence rate and precision of prediction.
Obviously, above-described embodiment is only intended to clearly illustrate example of the present invention, is not the reality to the present invention
Apply the restriction of mode.For those of ordinary skill in the field, other can also be made on the basis of the above description
Various forms of changes or variation, there is no necessity and possibility to exhaust all the enbodiments.And these belong to the present invention
The obvious changes or variations amplified out of spirit still in protection scope of the present invention among.
Claims (1)
1. one kind is based on particle cluster algorithm Optimizing BP Network to Ammunition Storage Reliability Forecasting Methodology, it is characterised in that including as follows
Several steps：
（1）Parameter initialization is set, and is divided into two parts：
Part I is the parameter setting of BP networks, according to input data and output data, determines input, the output section of BP networks
Points and the implicit number of plies and its nodes；
Part II is the parameter setting of particle cluster algorithm, sets particle number, particle initial position and speed and they
Scope, maximum iteration, Studying factors c1And c2Etc. parameter；Inertia weight uses the linear decrease weight plan proposed by shi
Slightly, i.e. inertia weight W：
5, in formula,6, 7
（2）Calculate the initial fitness value of particle.The error obtained during network forward-propagating is mean square error, i.e. population
The fitness function of algorithm is that situation that network output layer nodes are 1 is only discussed in mean square error formula, the present invention, and it is adapted to
Spend functional form：
2,
In formula, N represents the sample group number of training；3Represent the desired output of the network output node of i-th of sample；4Represent
The real output value of the network output node of i-th of sample；Fitness function according to providing calculates each particle in solution space
Interior adaptation value matrix, obtains individual adaptive optimal control angle value and global optimum's fitness value, finds the position of optimal particle.
（3）The fitness value of each particle is made comparisons with individual adaptive optimal control angle value and global optimum's fitness value, to sentence
The quality of disconnected current location, and determine the individual optimal and global optimum of particle position；
1. the fitness value of particle is respectively compared with individual optimal value pbest, global optimum gbest, if particle is currently fitted
Angle value is answered to be better than individual adaptive optimal control angle value, then using currency as individual adaptive optimal control angle value, and with the current location of particle
Instead of the personal best particle of history；
If the current individual adaptive optimal control angle value 2. in particle is better than the global optimum of history, by current individual optimal location
As the global optimum position of the particle, i.e. global optimum be individual it is optimal in best values；
（4）Speed and position of each particle in every dimension are updated, shown in the mode equation below of renewal：
2
9
Wherein, i=1,2 ..., N；N is population scale；d=1,2,…,D；10Tieed up for particle i personal best particles pbest d
Component；11Component is tieed up for the d of optimal value gbest in population；C1 is the accelerated factor of cognitive part；C2 is social part
Accelerated factor；Rand () is the random number between [0,1]；Inertia weight w is a nonnegative number；12Represent grain after the t times iteration
Sub- i circle in the air speed d dimension component；13Represent the d dimension components of particle i positions after the t times iteration.
（5）Above-mentioned particle cluster algorithm is updated to initial weight and threshold value that obtained global optimum is mapped to BP neural network, so
The training of network is carried out afterwards；Using the weights and threshold value of PSO algorithm optimizations in BP neural network back-propagation process, iteration is followed
Ring training further adjusts weights and threshold value, reaches that maximum or training error are less than expectation setting value when meeting iterations,
Then end loop is trained, output result；Otherwise go to（2）, continue cycling through execution above-mentioned steps；
（6）The test of BP neural network is carried out using test sample, input test sample is carried out using the above-mentioned network trained
Emulation, by output result with it is expected that setting value is made comparisons, so as to draw test effect.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710361921.4A CN107316099A (en) | 2017-05-22 | 2017-05-22 | Ammunition Storage Reliability Forecasting Methodology based on particle group optimizing BP neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710361921.4A CN107316099A (en) | 2017-05-22 | 2017-05-22 | Ammunition Storage Reliability Forecasting Methodology based on particle group optimizing BP neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107316099A true CN107316099A (en) | 2017-11-03 |
Family
ID=60182058
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710361921.4A Pending CN107316099A (en) | 2017-05-22 | 2017-05-22 | Ammunition Storage Reliability Forecasting Methodology based on particle group optimizing BP neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107316099A (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108120521A (en) * | 2017-12-08 | 2018-06-05 | 囯网河北省电力有限公司电力科学研究院 | Coiling hot point of transformer temperature predicting method and terminal device |
| CN108182490A (en) * | 2017-12-27 | 2018-06-19 | 南京工程学院 | A kind of short-term load forecasting method under big data environment |
| CN108596260A (en) * | 2018-04-27 | 2018-09-28 | 安徽建筑大学 | Water pipe network leakage positioning method and device |
| CN109359469A (en) * | 2018-10-16 | 2019-02-19 | 上海电力学院 | A kind of Information Security Risk Assessment Methods of industrial control system |
| CN109359738A (en) * | 2018-10-19 | 2019-02-19 | 西南交通大学 | A kind of Landslide hazard appraisal procedure based on QPSO-BP neural network |
| CN109739980A (en) * | 2018-12-11 | 2019-05-10 | 中科恒运股份有限公司 | Emotion classifiers are carried out with the method, apparatus and terminal of tuning |
| CN109919489A (en) * | 2019-03-08 | 2019-06-21 | 北京工业大学 | Business equipment life-span prediction method based on enterprise asset management and GA-BP |
| CN109998495A (en) * | 2019-05-23 | 2019-07-12 | 河南工业大学 | A kind of electrocardiosignal classification method based on particle group optimizing BP neural network |
| CN110162807A (en) * | 2018-02-12 | 2019-08-23 | 沈阳理工大学 | Based on improvement ant group optimization BP neural network Ammunition Storage Reliability Evaluation method |
| CN110232432A (en) * | 2018-03-05 | 2019-09-13 | 重庆邮电大学 | A kind of lithium battery group SOC prediction technique based on artificial life model |
| CN110533484A (en) * | 2019-09-05 | 2019-12-03 | 四川长虹电器股份有限公司 | A kind of product Method for Sales Forecast method based on PCA and improved BP |
| CN110668276A (en) * | 2019-08-29 | 2020-01-10 | 浙江理工大学 | Method for predicting elevator fault based on BP neural network optimized by PSO |
| CN111222736A (en) * | 2019-09-29 | 2020-06-02 | 中国人民解放军空军工程大学 | Ammunition storage reliability evaluation method based on mixed correlation vector machine model |
| CN111859562A (en) * | 2020-07-10 | 2020-10-30 | 大连理工大学 | Pressure-independent intelligent regulating valve design method based on dynamic resistance characteristic |
| CN112100918A (en) * | 2020-09-15 | 2020-12-18 | 中国人民解放军陆军装甲兵学院 | GA-BP prediction model research method for ablation wear loss of gun barrel |
| CN108090629B (en) * | 2018-01-16 | 2021-02-09 | 广州大学 | Load prediction method and system based on nonlinear autoregressive neural network |
| CN114298758A (en) * | 2021-12-30 | 2022-04-08 | 北京航天智造科技发展有限公司 | Neural network prediction method based on particle swarm optimization |
| CN114818479A (en) * | 2022-04-11 | 2022-07-29 | 哈尔滨工程大学 | Online identification method for circulating air injection amount of direct injection engine in high-pressure natural gas cylinder based on machine learning |
| CN116579245A (en) * | 2023-05-16 | 2023-08-11 | 中国兵器装备集团自动化研究所有限公司 | Grain density prediction method and device |
| CN117951996A (en) * | 2024-01-11 | 2024-04-30 | 西南交通大学 | Intelligent design method, system and storage medium for energy dissipater based on neural network |
-
2017
- 2017-05-22 CN CN201710361921.4A patent/CN107316099A/en active Pending
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108120521B (en) * | 2017-12-08 | 2020-06-26 | 囯网河北省电力有限公司电力科学研究院 | Transformer winding hot spot temperature prediction method and terminal equipment |
| CN108120521A (en) * | 2017-12-08 | 2018-06-05 | 囯网河北省电力有限公司电力科学研究院 | Coiling hot point of transformer temperature predicting method and terminal device |
| CN108182490A (en) * | 2017-12-27 | 2018-06-19 | 南京工程学院 | A kind of short-term load forecasting method under big data environment |
| CN108090629B (en) * | 2018-01-16 | 2021-02-09 | 广州大学 | Load prediction method and system based on nonlinear autoregressive neural network |
| CN110162807B (en) * | 2018-02-12 | 2023-09-12 | 沈阳理工大学 | Ammunition storage reliability assessment method based on improved ant colony optimization BP neural network |
| CN110162807A (en) * | 2018-02-12 | 2019-08-23 | 沈阳理工大学 | Based on improvement ant group optimization BP neural network Ammunition Storage Reliability Evaluation method |
| CN110232432B (en) * | 2018-03-05 | 2022-09-20 | 重庆邮电大学 | Lithium battery pack SOC prediction method based on artificial life model |
| CN110232432A (en) * | 2018-03-05 | 2019-09-13 | 重庆邮电大学 | A kind of lithium battery group SOC prediction technique based on artificial life model |
| CN108596260A (en) * | 2018-04-27 | 2018-09-28 | 安徽建筑大学 | Water pipe network leakage positioning method and device |
| CN109359469A (en) * | 2018-10-16 | 2019-02-19 | 上海电力学院 | A kind of Information Security Risk Assessment Methods of industrial control system |
| CN109359738A (en) * | 2018-10-19 | 2019-02-19 | 西南交通大学 | A kind of Landslide hazard appraisal procedure based on QPSO-BP neural network |
| CN109739980A (en) * | 2018-12-11 | 2019-05-10 | 中科恒运股份有限公司 | Emotion classifiers are carried out with the method, apparatus and terminal of tuning |
| CN109919489B (en) * | 2019-03-08 | 2022-04-22 | 北京工业大学 | Enterprise asset management system and GA-BP-based enterprise equipment life prediction method |
| CN109919489A (en) * | 2019-03-08 | 2019-06-21 | 北京工业大学 | Business equipment life-span prediction method based on enterprise asset management and GA-BP |
| CN109998495A (en) * | 2019-05-23 | 2019-07-12 | 河南工业大学 | A kind of electrocardiosignal classification method based on particle group optimizing BP neural network |
| CN110668276A (en) * | 2019-08-29 | 2020-01-10 | 浙江理工大学 | Method for predicting elevator fault based on BP neural network optimized by PSO |
| CN110533484A (en) * | 2019-09-05 | 2019-12-03 | 四川长虹电器股份有限公司 | A kind of product Method for Sales Forecast method based on PCA and improved BP |
| CN111222736A (en) * | 2019-09-29 | 2020-06-02 | 中国人民解放军空军工程大学 | Ammunition storage reliability evaluation method based on mixed correlation vector machine model |
| CN111859562A (en) * | 2020-07-10 | 2020-10-30 | 大连理工大学 | Pressure-independent intelligent regulating valve design method based on dynamic resistance characteristic |
| CN112100918A (en) * | 2020-09-15 | 2020-12-18 | 中国人民解放军陆军装甲兵学院 | GA-BP prediction model research method for ablation wear loss of gun barrel |
| CN114298758A (en) * | 2021-12-30 | 2022-04-08 | 北京航天智造科技发展有限公司 | Neural network prediction method based on particle swarm optimization |
| CN114818479A (en) * | 2022-04-11 | 2022-07-29 | 哈尔滨工程大学 | Online identification method for circulating air injection amount of direct injection engine in high-pressure natural gas cylinder based on machine learning |
| CN116579245A (en) * | 2023-05-16 | 2023-08-11 | 中国兵器装备集团自动化研究所有限公司 | Grain density prediction method and device |
| CN116579245B (en) * | 2023-05-16 | 2024-05-14 | 中国兵器装备集团自动化研究所有限公司 | Grain density prediction method and device |
| CN117951996A (en) * | 2024-01-11 | 2024-04-30 | 西南交通大学 | Intelligent design method, system and storage medium for energy dissipater based on neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107316099A (en) | Ammunition Storage Reliability Forecasting Methodology based on particle group optimizing BP neural network | |
| Xiao et al. | Artificial bee colony algorithm based on adaptive neighborhood search and Gaussian perturbation | |
| CN111353582B (en) | Particle swarm algorithm-based distributed deep learning parameter updating method | |
| CN107704875A (en) | Based on the building load Forecasting Methodology and device for improving IHCMAC neutral nets | |
| CN106920008A (en) | A kind of wind power forecasting method based on Modified particle swarm optimization BP neural network | |
| CN104899431B (en) | Based on ant colony and swarm of particles into LSSVM fluctuating wind speed Forecasting Methodologies | |
| CN106910337A (en) | A kind of traffic flow forecasting method based on glowworm swarm algorithm Yu RBF neural | |
| CN103646278B (en) | Application of particle swarm algorithm based on adaptive strategy in robot path planning | |
| CN105512745A (en) | Wind power section prediction method based on particle swarm-BP neural network | |
| CN103105246A (en) | Greenhouse environment forecasting feedback method of back propagation (BP) neural network based on improvement of genetic algorithm | |
| CN108805268A (en) | Deeply learning strategy network training method based on evolution algorithm | |
| CN108038507A (en) | Local receptor field extreme learning machine image classification method based on particle group optimizing | |
| CN104636801A (en) | Transmission line audible noise prediction method based on BP neural network optimization | |
| CN108564592A (en) | Based on a variety of image partition methods for being clustered to differential evolution algorithm of dynamic | |
| CN105427241B (en) | Distortion correction method for large-view-field display equipment | |
| CN113722980B (en) | Ocean wave height prediction method, ocean wave height prediction system, computer equipment, storage medium and terminal | |
| CN105447567A (en) | BP neural network and MPSO algorithm-based aluminium electrolysis energy-saving and emission-reduction control method | |
| CN109800517B (en) | Improved reverse modeling method for magnetorheological damper | |
| CN113095477B (en) | Wind power prediction method based on DE-BP neural network | |
| CN111709511A (en) | Harris eagle optimization algorithm based on random unscented Sigma point variation | |
| CN105631516A (en) | Historical experience and real-time adjustment combination-based particle swarm optimization algorithm | |
| CN112149883A (en) | Photovoltaic power prediction method based on FWA-BP neural network | |
| CN111008790A (en) | Hydropower station group power generation electric scheduling rule extraction method | |
| CN107016239A (en) | ASCA and RBF neural network model a kind of steam turbine valve traffic characteristics analysis method | |
| CN107392315B (en) | Breast cancer data classification method for optimizing brain emotion learning model |
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: 20171103 |