CN110362900A - A kind of prediction technique of LED life - Google Patents

A kind of prediction technique of LED life Download PDF

Info

Publication number
CN110362900A
CN110362900A CN201910587767.1A CN201910587767A CN110362900A CN 110362900 A CN110362900 A CN 110362900A CN 201910587767 A CN201910587767 A CN 201910587767A CN 110362900 A CN110362900 A CN 110362900A
Authority
CN
China
Prior art keywords
firefly
formula
ant
follows
value
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
Application number
CN201910587767.1A
Other languages
Chinese (zh)
Inventor
张志洁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Lingnan Normal University
Original Assignee
Lingnan Normal University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Lingnan Normal University filed Critical Lingnan Normal University
Priority to CN201910587767.1A priority Critical patent/CN110362900A/en
Publication of CN110362900A publication Critical patent/CN110362900A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/26—Testing of individual semiconductor devices
    • G01R31/2642—Testing semiconductor operation lifetime or reliability, e.g. by accelerated life tests
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/44—Testing lamps
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • 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
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/04—Ageing analysis or optimisation against ageing

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The present invention relates to a kind of prediction techniques of LED life, comprising the following steps: data set is divided into training set and test set；ACO (mixing ant colony) and GSO (firefly) are combined and to form the optimization algorithm (ACO+GSO) based on mixing ant colony and firefly；It is optimized using initial parameter of the ACO+GSO mixing intelligent optimizing algorithm to WNN (wavelet neural network), detection accuracy problem caused by avoiding initial parameter selection improper learns training dataset using WNN, forms Optimized model；Wherein, input using parameters such as temperature, electric current, initial luminous flux and initial color coordinates as WNN, output of the LED life as WNN, the LED life in test set is detected using established Optimized model, solve the problems, such as that predicted time is long and precision is bad, the problems such as promoting the computational accuracy and convergence rate of traditional intelligent algorithm, avoid the occurrence of local optimal searching, reduce manual intervention, promotes the reliability of LED life detection.

Description

A kind of prediction technique of LED life
Technical field
The invention belongs to LED the field of test technology, and in particular to a kind of prediction technique of LED life.
Background technique
In recent years, semiconductor lighting in industrial circle and daily life using more and more extensive, light emitting diode (LightEmittingDiode, LED) is a kind of electroluminescent light emitting semiconductor device, belongs to New Solid cold light source, has Voltage driving it is low, efficiency is high, the service life is long and it is at low cost the advantages that.With the development of LED technology, do not have in the research and development of LED one not Negligible limiting factor, the i.e. gap of its actual life and theoretical value.Therefore, corresponding life test is carried out, and according to LED Indices estimate service life of LED, improve its reliability, extending its service life becomes a kind of necessary strategy how Accurately prediction LED life also becomes an important project.
Currently, mainly including statistical regression methods and classical machine learning by the LED life prediction technique of data-driven Algorithm.It is influenced by model selection, traditional statistical method can not push away LED life under the conditions of multidimensional influence factor It is disconnected；And although classical machine learning algorithm has preferable data-handling capacity, its own at runtime between and prediction essence Degree aspect still has certain limitation, cannot achieve LED life forecasting reliability.
Summary of the invention
The purpose of the invention is to provide a kind of prediction techniques of LED life, pass through the multidimensional index parameter pair of LED LED reliability is evaluated, and mixing intelligent optimizing algorithm and wavelet neural method carry out effectively precisely prediction, base to LED life In mixing ant colony (AntColony Optimization, ACO) and firefly (Glowworm Swarm Optimization, GSO) the LED life prediction technique (ACO+GSO-WNN) of Optimization of Wavelet neural network (WaveletNeuralNetwork, WNN), Optimizing is scanned for parameter, the shortcomings that this method avoid both ACO and GSO, and realize mutual supplement with each other's advantages, reduce simultaneously The manual intervention of WNN algorithm parameter tuning, precision of prediction height, strong robustness, with good application prospect.
For realize foregoing invention purpose, the technical solution adopted by the present invention is as follows:
A kind of prediction technique of LED life, including mixing ant colony (ACO), firefly (GSO) and wavelet neural network (WNN), which is characterized in that the method includes the following steps:
S1, LED data are obtained, data set is divided into training set and test set；
S2, setting ant colony advise NACO, greatest iteration number TACO, information volatility coefficient ρ ∈ [0,1] and pheromone concentration Q；If Initial bit of one group of argument sequence (σ, γ) as ant is randomly generated in the range for setting kernel functional parameter σ and regularization parameter γ Set vector；
S3, the fitness value for calculating each ant individual present position；It calculates ant and is presently in the letter at the j of position Plain concentration is ceased, fitness is smaller, and pheromone concentration is bigger, fitness is defined with mean square deviation:
S4, by the pheromone concentration size of every ant, determine the smallest position of ant fitness value；Carry out pheromones The iteration of concentration updates, and will meet fitness condition f≤fdAnt position vector (σ, γ) be put into set XACO, until iteration Terminate.
S5, the population scale that firefly is arranged are NGSO, maximum number of iterations TGSO, optimization that ant group algorithm is obtained XACOInitialize firefly NGSOA insect position, and the position of other remaining insects is randomly choosed, each firefly individual is taken With identical fluorescein concentration loWith the perception radius ro；
S6, the fluorescein for updating firefly；
S7, the neighbours for finding firefly i；
S8, firefly i moving direction is determined；When the fluorescein value of neighbours firefly j is bigger than firefly i, and two light of firefly The distance between worm is in sensing range rsWithin when, firefly i will be with Probability pij(t) neighbours firefly j is selected, and to neighbours firefly The direction of fireworm j is mobile；
S9, firefly i location updating is carried out, then carries out decision area update, each firefly individual extreme value is determined by iteration pbestAnd optimal location；
S10, the individual extreme value p by each fireflybestWith group optimal location fitness value gbestCompare, when more excellent, then Using the optimal location of the firefly as the optimal location of group, the individual extreme value p of the fireflybestAs group extreme value gbest； It checks whether and meets iteration optimizing termination condition, terminate optimizing if meeting, find out optimal solution (σ, γ)；Otherwise, S3 is returned；
S11, the weight W of optimal solution (σ, γ) initialization WNN is obtained using in S10kWith threshold value B, design parameter setting Are as follows:
X=(x1,x2,...,xn)TIt is inputted for the data of input node；
Wk=(wk1,wk2,...,wkn) be hidden layer k-th of hidden node and input layer connection weight；
Wo=(w1,w2,...,wk) be output layer and hidden node connection weight；
Ho=(h1,h2,...,hk) be each hidden node of hidden layer output valve；
B=(b1,b2,...,bk) be each hidden node of hidden layer threshold value；
boFor the threshold value for exporting node layer；
f1For the activation primitive of hidden layer,
Wavelet mother function indicates are as follows:
Hidden layer output are as follows:
In formula, bk,akThe respectively shift factor and contraction-expansion factor of wavelet basis function；
Output layer are as follows:
O=f2(WoHo-b0) (12)
S12, the WNN model that optimization is established according to output layer, predict LED life, obtain LED light life prediction knot Fruit.
Further, the specific formula of fitness value of each ant individual present position is calculated in the step S3 Are as follows:
In formula, yiFor the actual value of test set；
y'iFor the predicted value of test set；
N is the number of test set；
The calculating ant is presently in the specific formula of pheromone concentration at the j of position are as follows:
In formula,For the pheromones left at the j of position in this circulation of kth ant；
T is current time；
skFor the fitness value with kth ant in this position.
Further, the concrete operations formula that the iteration of pheromone concentration updates is carried out in the step S4 are as follows:
τj(t+m)=(1- ρ) τj(t)+Δτj(t) (4)
In formula, ρ (0 < ρ < 1) is pheromones volatility coefficient,
M is the time needed for each pheromones iteration；
J is at parameter optimization spatial position；
In an iterative process, t moment kth ant randomly chooses the new probability formula of position j from solution space I are as follows:
Further, the concrete operations formula of the fluorescein of firefly is updated in the step S6 are as follows:
li(t)=(1- ρ) li(t-1)+γJ(xi(t)) (5)
In formula, J (xiIt (t)) is position x of the every firefly i in t iterationi(t) corresponding target function value；
liIt (t) is the fluorescein value of current firefly；
γ is fluorescein turnover rate；
ρ is fluorescein volatilization factor.
Further, the concrete operations formula of the neighbours of firefly i is found in the step S7 are as follows:
In formula, NiIt (t) is the neighborhood of t i-th of firefly of generation；
| | x | | it is the norm of x；
xjIt (t) is the position of t j-th of firefly of generation；
lj(t) the fluorescein value of t j-th of firefly of generation；
Dynamic decision domain rangeDetermine the number of neighbours；
Its upper bound sensing range is
Further, Probability p in the step S8ij(t) its formula are as follows:
Further, the concrete operations formula of firefly i location updating is carried out in the S9 are as follows:
In formula, S is moving step length；
Carry out decision area update concrete operations formula are as follows:
In formula, | Ni(t) | for firefly number in firefly i neighborhood.
Compared with prior art:
1, LED based working environment considers the factor for influencing LED life from multidimensional index parametric synthesis, to The reliability characteristic of accurate description LED.
2, novel ACO+GSO mixing intelligent optimizing algorithm is devised, to promote the calculating essence of traditional intelligent algorithm Degree and convergence rate, the problems such as avoiding the occurrence of local optimal searching.
3, WNN is more sensitive to initial parameter, generally require skilled engineer according to the characteristics of business scenario into The experiment tune ginseng of row repeatedly, present invention uses initial parameter progress of the ACO+GSO mixing intelligent optimizing algorithm to WNN is automatic excellent Change and adjust, reduces manual intervention.
4, the WNN after optimizing can effectively predict LED life, promote the detection performance of LED reliability.
Detailed description of the invention
Fig. 1 is the flow chart of the prediction technique of LED life in the embodiment of the present invention；
Fig. 2 is the flow chart of the WNN algorithm in the embodiment of the present invention based on ACO+GSO；
Fig. 3 is the network topological diagram of WNN in the embodiment of the present invention.
Specific embodiment
With reference to the accompanying drawings and examples, specific embodiments of the present invention will be described in further detail.Implement below Example is not intended to limit the scope of the invention for illustrating the present invention.
In conjunction with shown in Fig. 1, Fig. 2, Fig. 3,
Step 1: using LED manufacturer LM-80-08 test report as data source, use TM-21-11 method Every kind of LED life is calculated, and data set is divided into training set and test set.(data source is collected, data set is divided into training Collection and test set)
Step 2: setting ant colony advises NACO, greatest iteration number TACO, information volatility coefficient ρ ∈ [0,1] and pheromone concentration Q；The range of kernel functional parameter σ and regularization parameter γ are set, one group of argument sequence (σ, γ) is randomly generated as the first of ant Beginning position vector.
Step 3: the fitness value of current location where calculating each ant individual with formula (1) is calculated using formula (2) Pheromone concentration at ant current time present position j, fitness is smaller, and pheromone concentration is bigger.It is suitable with mean square deviation definition Response:
Wherein, yiAnd y'iThe actual value and predicted value of test set are respectively indicated, N is the number of test set.
Pheromone concentration at t moment ant present position j:
Wherein,Indicate the pheromones left at the j of position in this circulation of kth ant, skIt indicates Fitness value with kth ant in this position.
Step 4: by the pheromone concentration size of every ant, determining the smallest position of ant fitness value.Pass through formula (4) iteration for carrying out pheromone concentration updates, and will meet fitness condition f≤fd(fdIt is worth size depending on checking computations experience) Ant position vector (σ, γ) is put into set XACO, until iteration terminates.In an iterative process, t moment kth ant is empty from solution Between the new probability formula of position j is randomly choosed in I:
Pheromone concentration more new formula is as follows at the j of parameter optimization spatial position:
τj(t+m)=(1- ρ) τj(t)+Δτj(t) (4)
Wherein, ρ (0 < ρ < 1) indicates that pheromones volatility coefficient, m are the time needed for each pheromones iteration.
Step 5: the population scale that firefly is arranged is NGSO, maximum number of iterations TGSO.The optimization that ant group algorithm is obtained XACOInitialize firefly NGSOA insect position, and randomly choose the position of other remaining insects, each firefly individual Carry identical fluorescein concentration loWith the perception radius ro。
Step 6: the fluorescein of firefly is updated using formula (5).
li(t)=(1- ρ) li(t-1)+γJ(xi(t)) (5)
Wherein, J (xiIt (t)) is position x of the every firefly i in t iterationi(t) corresponding target function value, li(t) table Show the fluorescein value of current firefly, γ is fluorescein turnover rate, and ρ is fluorescein volatilization factor.
Step 7: find the neighbours of firefly i:
Wherein, Ni(t) neighborhood of t i-th of firefly of generation is indicated, | | x | | indicate the norm of x, xjIt (t) is t generation The position of j-th of firefly, lj(t) the fluorescein value of t j-th of firefly of generation；Dynamic decision domain rangeDetermine neighbours Number, the upper bound be sensing range
Step 8: determining firefly i moving direction.When the fluorescein value of neighbours firefly j is bigger than firefly i, and two fireflies The distance between fireworm is in sensing range rsWithin when, firefly i will be with certain Probability pij(t) neighbours firefly j is selected, and It is mobile to the direction of neighbours firefly j.
Step 9: carrying out firefly i location updating using formula (8), decision area update is carried out using formula (9), by repeatedly In generation, determines each firefly individual extreme value pbestAnd optimal location:
Wherein, S is moving step length, | Ni(t) | indicate firefly number in firefly i neighborhood.
Step 10: by the individual extreme value p of each fireflybestWith group optimal location fitness value gbestCompare, if more It is excellent, then using the optimal location of the firefly as the optimal location of group, the individual extreme value p of the fireflybestAs group pole Value gbest.Maximum number of iterations is set, checks whether and meets iteration optimizing termination condition, terminates optimizing if meeting, finds out most Excellent solution (σ, γ).Otherwise, return step 3.
Step 11: obtaining the weight W of optimal solution (σ, γ) initialization WNN using in step (10)kWith threshold value B, specifically Parameter setting are as follows:
X=(x1,x2,...,xn)TIt is inputted for the data of input node；
Wk=(wk1,wk2,...,wkn) be hidden layer k-th of hidden node and input layer connection weight；
Wo=(w1,w2,...,wk) be output layer and hidden node connection weight；
Ho=(h1,h2,...,hk) be each hidden node of hidden layer output valve；
B=(b1,b2,...,bk) be each hidden node of hidden layer threshold value；
boFor the threshold value for exporting node layer.
f1For the activation primitive of hidden layer, it is wavelet mother function, indicates are as follows:
Hidden layer output are as follows:
Wherein, bk,akThe respectively shift factor and contraction-expansion factor of wavelet basis function.
Output layer are as follows:
O=f2(WoHo-b0) (12)
Step 12: establishing the WNN model of optimization using formula (12), LED life is predicted, LED reliability is obtained Testing result.
Obviously, the above embodiment of the present invention and is not to this hair only to clearly illustrate example of the present invention The restriction of bright embodiment.For those of ordinary skill in the art, it can also do on the basis of the above description Other various forms of variations or variation out.There is no necessity and possibility to exhaust all the enbodiments.It is all in the present invention Spirit and principle within made any modifications, equivalent replacements, and improvements etc., should be included in the guarantor of the claims in the present invention Within the scope of shield.

Claims (7)

1. a kind of prediction technique of LED life, including mixing ant colony (ACO), firefly (GSO) and wavelet neural network (WNN), It is characterized in that, the method includes the following steps:
S1, LED data are obtained, data set is divided into training set and test set；
S2, setting ant colony advise NACO, greatest iteration number TACO, information volatility coefficient ρ ∈ [0,1] and pheromone concentration Q；Core is set The range of function parameter σ and regularization parameter γ, be randomly generated one group of argument sequence (σ, γ) as ant initial position to Amount；
S3, the fitness value for calculating each ant individual present position；It calculates ant and is presently in the pheromones at the j of position Concentration, fitness is smaller, and pheromone concentration is bigger, defines fitness with mean square deviation:
S4, by the pheromone concentration size of every ant, determine the smallest position of ant fitness value；Carry out pheromone concentration Iteration update, fitness condition f≤f will be metdAnt position vector (σ, γ) be put into set XACO, until iteration terminates.
S5, the population scale that firefly is arranged are NGSO, maximum number of iterations TGSO, the X for the optimization that ant group algorithm is obtainedACOJust Beginningization firefly NGSOA insect position, and the position of other remaining insects is randomly choosed, each firefly individual carries phase Same fluorescein concentration loWith the perception radius ro；
S6, the fluorescein for updating firefly；
S7, the neighbours for finding firefly i；
S8, firefly i moving direction is determined；When the fluorescein value of neighbours firefly j is bigger than firefly i, and two fireflies it Between distance in sensing range rsWithin when, firefly i will be with Probability pij(t) neighbours firefly j is selected, and to neighbours firefly The direction of j is mobile；
S9, firefly i location updating is carried out, then carries out decision area update, each firefly individual extreme value p is determined by iterationbestWith Optimal location；
S10, the individual extreme value p by each fireflybestWith group optimal location fitness value gbestCompare, when more excellent, then should Optimal location of the optimal location of firefly as group, the individual extreme value p of the fireflybestAs group extreme value gbest；It checks Whether meet iteration optimizing termination condition, terminates optimizing if meeting, find out optimal solution (σ, γ)；Otherwise, S3 is returned；
S11, the weight W of optimal solution (σ, γ) initialization WNN is obtained using in S10kWith threshold value B, design parameter setting are as follows:
X=(x1,x2,...,xn)TIt is inputted for the data of input node；
Wk=(wk1,wk2,...,wkn) be hidden layer k-th of hidden node and input layer connection weight；
Wo=(w1,w2,...,wk) be output layer and hidden node connection weight；
Ho=(h1,h2,...,hk) be each hidden node of hidden layer output valve；
B=(b1,b2,...,bk) be each hidden node of hidden layer threshold value；
boFor the threshold value for exporting node layer；
f1For the activation primitive of hidden layer,
Wavelet mother function indicates are as follows:
Hidden layer output are as follows:
In formula, bk,akThe respectively shift factor and contraction-expansion factor of wavelet basis function；
Output layer are as follows:
O=f2(WoHo-b0) (12)
S12, the WNN model for establishing optimization, predict LED life, obtain LED light life prediction result.
2. a kind of prediction technique of LED life according to claim 1, which is characterized in that calculated in the step S3 every The specific formula of fitness value of a ant individual present position are as follows:
In formula, yiFor the actual value of test set；
y′iFor the predicted value of test set；
N is the number of test set；
The calculating ant is presently in the specific formula of pheromone concentration at the j of position are as follows:
In formula,For the pheromones left at the j of position in this circulation of kth ant；
T is current time；
skFor the fitness value with kth ant in this position.
3. a kind of prediction technique of LED life according to claim 1, which is characterized in that carry out letter in the step S4 Cease the concrete operations formula that the iteration of plain concentration updates are as follows:
τj(t+m)=(1- ρ) τj(t)+Δτj(t) (4)
In formula, ρ (0 < ρ < 1) is pheromones volatility coefficient,
M is the time needed for each pheromones iteration；
J is at parameter optimization spatial position；
In an iterative process, t moment kth ant randomly chooses the new probability formula of position j from solution space I are as follows:
4. a kind of prediction technique of LED life according to claim 1, which is characterized in that update firefly in the step S6 The concrete operations formula of the fluorescein of fireworm are as follows:
li(t)=(1- ρ) li(t-1)+γJ(xi(t)) (5)
In formula, J (xiIt (t)) is position x of the every firefly i in t iterationi(t) corresponding target function value；
liIt (t) is the fluorescein value of current firefly；
γ is fluorescein turnover rate；
ρ is fluorescein volatilization factor.
5. a kind of prediction technique of LED life according to claim 1, which is characterized in that find firefly in the step S7 The concrete operations formula of the neighbours of fireworm i are as follows:
In formula, NiIt (t) is the neighborhood of t i-th of firefly of generation；
| | x | | it is the norm of x；
xjIt (t) is the position of t j-th of firefly of generation；
lj(t) the fluorescein value of t j-th of firefly of generation；
Dynamic decision domain rangeDetermine the number of neighbours；
Its upper bound sensing range is
6. a kind of prediction technique of LED life according to claim 1, which is characterized in that Probability p in the step S8ij (t) its formula are as follows:
7. a kind of prediction technique of LED life according to claim 1, which is characterized in that carry out firefly i in the S9 The concrete operations formula of location updating are as follows:
In formula, S is moving step length；
Carry out decision area update concrete operations formula are as follows:
In formula, | Ni(t) | for firefly number in firefly i neighborhood.
CN201910587767.1A 2019-07-02 2019-07-02 A kind of prediction technique of LED life Pending CN110362900A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910587767.1A CN110362900A (en) 2019-07-02 2019-07-02 A kind of prediction technique of LED life

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910587767.1A CN110362900A (en) 2019-07-02 2019-07-02 A kind of prediction technique of LED life

Publications (1)

Publication Number Publication Date
CN110362900A true CN110362900A (en) 2019-10-22

Family

ID=68217829

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910587767.1A Pending CN110362900A (en) 2019-07-02 2019-07-02 A kind of prediction technique of LED life

Country Status (1)

Country Link
CN (1) CN110362900A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111914493A (en) * 2020-09-29 2020-11-10 北京中设光环境科技研究院有限公司 System and method for simulating service life of LED street lamp
CN113505875A (en) * 2021-07-20 2021-10-15 珠海格力电器股份有限公司 Fault prediction method, device and storage medium
CN113805060A (en) * 2021-05-21 2021-12-17 电子科技大学 Lithium battery residual life detection method based on relevance vector regression
CN115639456A (en) * 2022-12-08 2023-01-24 深圳市粉紫实业有限公司 Method, system and medium for predicting service life of light emitting diode

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111914493A (en) * 2020-09-29 2020-11-10 北京中设光环境科技研究院有限公司 System and method for simulating service life of LED street lamp
CN111914493B (en) * 2020-09-29 2021-02-02 北京中设光环境科技研究院有限公司 System and method for simulating service life of LED street lamp
CN113805060A (en) * 2021-05-21 2021-12-17 电子科技大学 Lithium battery residual life detection method based on relevance vector regression
CN113505875A (en) * 2021-07-20 2021-10-15 珠海格力电器股份有限公司 Fault prediction method, device and storage medium
CN115639456A (en) * 2022-12-08 2023-01-24 深圳市粉紫实业有限公司 Method, system and medium for predicting service life of light emitting diode

Similar Documents

Publication Publication Date Title
CN110362900A (en) A kind of prediction technique of LED life
CN109685277A (en) Electricity demand forecasting method and device
CN109241313A (en) A kind of image search method based on the study of high-order depth Hash
CN109598936A (en) A kind of prediction of short-term traffic volume method based on dynamic STKNN model
CN109543043A (en) A kind of automatic monitoring method of the power communication big data mistake of knowledge based map reasoning
CN108985342A (en) A kind of uneven classification method based on depth enhancing study
CN109325580A (en) A kind of adaptive cuckoo searching method for Services Composition global optimization
CN109657147A (en) Microblogging abnormal user detection method based on firefly and weighting extreme learning machine
Chen et al. LOGER: A learned optimizer towards generating efficient and robust query execution plans
CN114936307A (en) Method for constructing normal graph model
CN113032613B (en) Three-dimensional model retrieval method based on interactive attention convolution neural network
CN110232397A (en) A kind of multi-tag classification method of combination supporting vector machine and projection matrix
CN111291582B (en) RFID antenna optimal deployment method based on jellyfish group algorithm
Stojanova et al. Network regression with predictive clustering trees
CN104732522A (en) Image segmentation method based on polymorphic ant colony algorithm
CN115273645B (en) Map making method for automatically clustering indoor surface elements
CN111353525A (en) Modeling and missing value filling method for unbalanced incomplete data set
CN109754075A (en) Dispatching method, equipment, storage medium and the device of wireless sensor network node
CN112487723B (en) Multi-parameter-based optimized design method for reclaimed water heat source pump system
Cervone et al. Experimental validations of the learnable evolution model
Xu Multi-objective intelligent algorithm model design for housing environment optimization
He et al. Service discovery method based on two-step clustering
CN110702121A (en) Optimal path fuzzy planning method for hillside orchard machine
CN106709510A (en) Gaussian mixture model searching method based on immune clonal selection algorithm
CN114485699B (en) Self-adaptive path optimization method and system for unmanned aerial vehicle self-organizing network

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