CN104573818B - A kind of office building room classes method based on neutral net - Google Patents
A kind of office building room classes method based on neutral net Download PDFInfo
- Publication number
- CN104573818B CN104573818B CN201410828012.3A CN201410828012A CN104573818B CN 104573818 B CN104573818 B CN 104573818B CN 201410828012 A CN201410828012 A CN 201410828012A CN 104573818 B CN104573818 B CN 104573818B
- Authority
- CN
- China
- Prior art keywords
- room
- electricity consumption
- data
- office building
- echo state
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000007935 neutral effect Effects 0.000 title claims abstract description 17
- 230000005611 electricity Effects 0.000 claims abstract description 80
- 238000012549 training Methods 0.000 claims abstract description 34
- 238000004378 air conditioning Methods 0.000 claims abstract description 16
- 241001269238 Data Species 0.000 claims abstract description 10
- 238000010276 construction Methods 0.000 claims abstract 2
- 238000012216 screening Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 abstract description 9
- 238000005286 illumination Methods 0.000 abstract description 8
- 239000011159 matrix material Substances 0.000 description 18
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000005265 energy consumption Methods 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 239000013589 supplement Substances 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000004134 energy conservation Methods 0.000 description 2
- 210000004218 nerve net Anatomy 0.000 description 2
- 238000000714 time series forecasting Methods 0.000 description 2
- 208000019901 Anxiety disease Diseases 0.000 description 1
- 230000036506 anxiety Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000002592 echocardiography Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000002203 pretreatment Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000005612 types of electricity Effects 0.000 description 1
Abstract
The invention discloses a kind of office building room classes method based on neutral net, comprise the following steps:Data prediction, socket, illumination and three kinds of electricity consumption datas of air-conditioning to office building room are screened, rejected, supplemented, and obtain complete electricity consumption data;Netinit, construction with pre-process the corresponding three echo state nets of three class electricity consumption datas and an extreme learning machine, and parameter to each neutral net is initialized;Network training, each echo state net is trained using the room electricity consumption data of pretreatment, rebuilds the use power mode in room, and using the room power mode after rebuilding and known room class training extreme learning machine;Room classes, give new three kinds of office building room electricity consumption data as grouped data, the use power mode in room are rebuild using the three echo state nets for training, and the type in room is obtained using the extreme learning machine for training.
Description
Technical field
The invention belongs to office electricity consumption modeling technique field, and in particular to a kind of office building room based on neutral net
Sorting technique.
Background technology
Building energy consumption is produced with the process such as the primary demands such as people's heating, air-conditioning, daylighting and amusement, work.With warp
Ji development and social progress, the requirement that people are used building is also by more and more higher;This change is embodied directly in building energy consumption
Increase year by year on.By 2012, building energy consumption accounted for the ratio of global total energy consumption up to 30%.With energy supply increasingly
Anxiety, the building energy conservation closely bound up with human lives, also by more and more extensive concern and attention, and Energy Saving of Office Building
There is vital meaning wherein.
In general, office building is made up of different types of room, including office, computer room, storeroom, meeting room
Deng being also not quite similar with power mode for, different type room, therefore launch energy-conservation behavior for different types of room and can be
Optimize the energy management of whole building and improve power consumption efficiency and reference is provided.
The content of the invention
Regarding to the issue above, the present invention proposes a kind of office building room classes method based on neutral net, is based on
The room electricity consumption data of office building, is classified using neural network model to the room of office building, is optimized and is entirely built
The energy management built simultaneously improves power consumption efficiency and provides reference.
A kind of office building room classes method based on neutral net proposed by the present invention, comprises the following steps:
Step S1:Data prediction, its method is:The electricity consumption data in office building room is divided into N classes, to N class electricity consumptions
Data carry out data prediction, obtain complete electricity consumption data;
Step S2:Netinit, its method is:Construct N number of echo state net corresponding with N class electricity consumption datas and one
Extreme learning machine, and parameter to each neutral net initializes;
Step S3:Network training, its method is:Each echo state net is trained using electricity consumption data complete in step S1,
The use power mode in room is rebuild, and using electricity consumption mode reconstruction result and room class training extreme learning machine;
Step S4:Room classes, its method is:The new office building room N class electricity consumption datas of input, using in step S3
Three echo state nets of training rebuild the use power mode in room, and are handled official business using the extreme learning machine trained in step S3
The room classes of building.
Preferably, the electricity consumption data in office building room is divided three classes in step S1, respectively socket electricity consumption data, illumination
Electricity consumption data and air conditioning electricity data.
The method of electricity consumption data pretreatment includes data screening, data rejecting, data filling.
Room electricity consumption data of the present invention based on office building, using two kinds of nerve nets of echo state net and extreme learning machine
Network model is classified to the room of office building, is to be carried out for different room class with classification accuracy higher
Energy saving optimizing provides reference.
Brief description of the drawings
Fig. 1 is the flow chart of the office building room classes method based on neutral net in the present invention;
Fig. 2 is the structure chart of echo state net in the present invention;
Fig. 3 is the structure chart of extreme learning machine in the present invention;
Fig. 4 a, 4b are respectively 5 working day interpolation block electricity consumptions and socket electricity consumption echo state net weight in the present embodiment office
Build result;
Fig. 4 c, 4d are respectively 5 working day intraoral illumination electricity consumptions and electric consumption on lighting echo state net weight in the present embodiment office
Build result;
Fig. 4 e, 4f are respectively in the present embodiment office air conditioning electricity and air conditioning electricity echo state net weight in 5 working days
Build result;
Fig. 5 a, 5b are respectively 5 working day interpolation block electricity consumptions and socket electricity consumption echo state net in the present embodiment computer room and rebuild
As a result;
Fig. 5 c, 5d are respectively 5 working day intraoral illumination electricity consumptions and electric consumption on lighting echo state net in the present embodiment computer room and rebuild
As a result;
Fig. 5 e, 5f are respectively in the present embodiment computer room air conditioning electricity and air conditioning electricity echo state net in 5 working days and rebuild
As a result;
Fig. 6 a, 6b are respectively 5 working day interpolation block electricity consumptions and socket electricity consumption echo state net weight in the present embodiment storeroom
Build result;
Fig. 6 c, 6d are respectively 5 working day intraoral illumination electricity consumptions and electric consumption on lighting echo state net weight in the present embodiment storeroom
Build result;
Fig. 6 e, 6f are respectively in the present embodiment storeroom air conditioning electricity and air conditioning electricity echo state net weight in 5 working days
Build result;
Fig. 7 a, 7b are respectively 5 working day interpolation block electricity consumptions and socket electricity consumption echo state net weight in the present embodiment meeting room
Build result;
Fig. 7 c, 7d are respectively 5 working day intraoral illumination electricity consumptions and electric consumption on lighting echo state net weight in the present embodiment meeting room
Build result;
Fig. 7 e, 7f are respectively in the present embodiment meeting room air conditioning electricity and air conditioning electricity echo state net weight in 5 working days
Build result.
Specific embodiment
In order to illustrate more clearly of the object, technical solutions and advantages of the present invention, below in conjunction with specific embodiment, and reference
Accompanying drawing, the present invention is described in further detail.
As shown in figure 1, the method includes the following steps for performing successively:
Step S1:Three kinds of electricity consumption datas (socket, illumination and air-conditioning) in office building room are carried out pre- by data prediction
Treatment, including data screening, rejecting, supplement etc., so as to obtain complete electricity consumption data;
Step S2:Netinit, constructs three echo state nets (Echo State Network, ESN) and a pole
Limit learning machine (Extreme Learning Machine, ELM), and the parameters such as the weights of each neutral net, threshold value are initialized, its
In three echo state nets correspond respectively to three kinds of electricity consumption datas;
Step S3:Network training, the room electricity consumption data pre-processed using S1 trains each echo state net, to rebuild room
Use power mode, and using rebuild after room power mode and known room class training extreme learning machine;
Step S4:Room classes, give new three kinds of office building room electricity consumption data, three echoes trained using S3
State net rebuilds the use power mode in room, and the extreme learning machine trained using S3 obtains the type in room.
Each step to the office building room classes method based on neutral net carries out launching to retouch in detail separately below
State:
Step S1:Three kinds of electricity consumption datas (socket, illumination and air-conditioning) in office building room are carried out pre- by data prediction
Treatment, including data screening, rejecting, supplement etc., so as to obtain complete electricity consumption data.
Office building room classes method based on neutral net of the invention is based on the electricity consumption data in office building room,
Data include 3 types, i.e. electric consumption on lighting data, socket electricity consumption data and air conditioning electricity data, are common office and use
Electric type and can completely represent office building room use power mode.
On the basis of initial data is obtained, initial data is screened according to certain period of time, obtain the time
All types of electricity consumption data in section.On this basis, total data is traveled through, redundant data therein is rejected and is supplemented missing number
According to so as to improve the quality of data, electricity consumption data when obtaining complete whole provides complete and accurate for ensuing data are calculated
Data.
Step S2:Netinit, constructs three echo state nets (Echo State Network, ESN) and a pole
Limit learning machine (Extreme Learning Machine, ELM), and the parameters such as the weights of each neutral net, threshold value are initialized, its
In three echo state nets correspond respectively to the electricity consumption data of three types.
The present invention is rebuild using the good time series forecasting performance of echo state net to room electricity consumption behavior, and profit
Room is classified with extreme learning machine quick network training, wherein echo state net and extreme learning machine are two kinds new
Neutral net, with good network training ability and generalization ability.
Echo state network is a kind of new neural network, its neuron circuit structure for imitating recurrence connection in brain,
The reserve pool for using the neuron connected by Random sparseness to constitute is unique in that as hidden layer, is used to carry out height to input
Dimension, nonlinear expression.The generating process of reserve pool independently of echo state network training process, therefore, only need to use line
Property method train reserve pool to the weights of output layer, be simplified the training process of network, and ensure weights determine the overall situation
Optimality and good generalization ability, it is to avoid training algorithm present in traditional neural network is complicated, be easily absorbed in it is local most
Small the problems such as.
Shown in the typical structure accompanying drawing 2 of echo state network, it is made up of input layer, reserve pool and output layer.When echo shape
When state network is used for time series forecasting problem, its output variable and input variable are time series, at this moment, can be by echo
State network sees nonlinear filter as, realizes the conversion of input to output.
Extreme learning machine is a kind of easy to use, effective Single hidden layer feedforward neural networks learning algorithm.Traditional nerve
Learning Algorithms (such as BP algorithm) need artificially to set substantial amounts of network training parameter, and are easy to produce local optimum
Solution.Extreme learning machine only needs to set the hidden node number of network, and the defeated of network need not be adjusted during algorithm performs
Enter the threshold value of weights and hidden neuron, and unique optimal solution can be obtained, thus it is fast and extensive with pace of learning
The good advantage of performance.
Shown in the typical structure accompanying drawing 3 of extreme learning machine, it is made up of input layer, hidden layer and output layer, is one simple
Single hidden layer feedforward neural networks.
In this step, for the electricity consumption data of three types, i.e. socket electricity consumption, electric consumption on lighting and air conditioning electricity, respectively
Three echo state nets are set up, is rebuild with to electricity consumption behavior, obtain three kinds of electricity consumption types uses power mode.Next, building
An extreme learning machine is found, the purpose is to being trained with power mode and known room class using three kinds of electricity consumption types, from
And realize the room classes of office building.After totally 4 neutral nets are set up, initialization network parameter, including each echo state
The input layer weights of net, reserve pool weights and deposit pool unit number, and extreme learning machine input layer weights, hidden layer threshold value
With hidden neuron number.
Step S3:Network training, the room electricity consumption data pre-processed using S1 trains each echo state net, to rebuild room
Use power mode, and using rebuild after room power mode and known room class training extreme learning machine.
In this step, room electricity consumption data and known room class based on S1 pretreatments are to echo state net and pole
Limit learning machine is trained.In training process, the input layer weights of each echo state net, reserve pool weights and extreme learning machine
Input layer weights, hidden layer threshold value keep it is constant, only adjust the output layer weights of each network.Two kinds of training process difference of network
Details are as follows.
(1) training of echo state net
For echo state net as shown in Figure 2, it is assumed that the echo state net is by K input block, N1Individual reserve pool list
Unit and L output unit composition.Input block is by being input into layer matrixIt is connected with deposit pool unit, reserve pool list
Pass through reserve pool weight matrix between unitConnection, deposit pool unit is by exporting layer matrixWith output
Unit is connected.At the j moment, input block, deposit pool unit and output unit are respectively by vectorial sj∈RK、And oj∈RL
Represent, then the fundamental equation of echo state network is
xj=f (Winsj+Wresxj-1),
oj=Woutxj,
Wherein, f () is the activation primitive for laying in pool unit.
For M1Individual training sample (sj,uj), j=1,2 ..., M1, wherein sj∈RK, uj∈RL, according to above-mentioned echo state
The fundamental equation of net can be obtained
Wherein, ojIt is the corresponding network output of j-th input sample,WithIt is weight matrix Win、WresWith
WoutIn corresponding weight vector.
Because echo state net can approach any nonlinear system with arbitrary accuracy, therefore for this M1Individual training sample,
We haveExistWithSo that
Above M1Individual equation can be merged into
BWout=U,
Wherein
Matrix B is referred to as the reserve pool output matrix of echo state net.
Final accounting equation BWoutThe Minimal Norm Least Square Solutions of=U as echo state net output weight matrix,
I.e.
Wherein,It is the generalized inverse matrix of matrix B.
(2) training of extreme learning machine
For extreme learning machine as shown in Figure 3, it is assumed that the extreme learning machine is by D input block, N2Individual Hidden unit
Constituted with Q output unit.Input block is by being input into layer matrixIt is connected with Hidden unit,For hidden
The threshold matrix of layer unit, Hidden unit is by exporting layer matrixIt is connected with output unit.Input block and output
Unit is respectively by vectorial uj∈RDAnd yj∈RQRepresent, the fundamental equation of extreme learning machine is
yj=β f (Wuj+ b), j=1,2 ..., M2,
Wherein, g () is the activation primitive of Hidden unit.
For M2Individual training sample (uj,yj), j=1,2 ..., M2, wherein uj∈RD, yj∈RQ, learnt according to the above-mentioned limit
The fundamental equation of machine can be obtained
Wherein, vjIt is the corresponding network output of j-th input sample, wi、biAnd βiIt is respectively corresponding in matrix W, b and β
Vector.
The extreme learning machine of global approximation properties due to to(for) nonlinear system, therefore for this M2Individual training sample we
HaveThere is wi、biAnd βiSo that
Above M2Individual equation can be merged into
H β=Y,
Wherein
Matrix H is referred to as the reserve pool output matrix of echo state net.
The Minimal Norm Least Square Solutions of final accounting equation H β=Y as echo state net output weight matrix, i.e.,
Wherein,It is the generalized inverse matrix of matrix H.
Three echo state nets and an extreme learning machine are trained respectively using above training process, wherein each time
The input of sound state net is each electricity consumption type electricity consumption data interior for a period of time, is output as electricity consumption mode reconstruction result;The limit
The input of habit machine is electricity consumption mode reconstruction result, is output as known room class.
Step S4:Room classes, give new office building room electricity consumption data as grouped data, are trained using S3
Three echo state nets rebuild the use power mode in room, and the extreme learning machine trained using S3 obtains the type in room.
Using grouped data as input, by the echo state net operation that S3 is trained, obtain corresponding room uses power mode
Reconstructed results, and then the extreme learning machine of S3 training is input into, the type in corresponding room is finally obtained, so that the office building realized
Room classes.
Implement citing:
Using certain office building in July, 2013 to the electricity consumption data in December, 2013 each room as data sample, wherein 7~
, used as training data, 10~December, data were used as grouped data for September data.Described in detail with reference to each step of the invention as follows.
Step S1:Data prediction, pre-processes to training data and grouped data respectively, including data screening, picks
Except, supplement etc., so as to obtain complete electricity consumption data.
Step S2:Netinit, constructs three echo state nets and an extreme learning machine, and initialize each nerve net
The parameters such as weights, the threshold value of network.
Step S3:Network training, the training data pre-processed using S1 is trained to each echo state net, to rebuild room
Between use power mode, and using rebuild after room power mode and known room class training extreme learning machine.
The room of the office building can be divided into office, computer room, four kinds of classifications of storeroom and meeting room, room of all categories 5
Result such as Fig. 4 a of the power consumption of three kinds of electricity consumption types and echo state net reconstruction power mode~4f, 5a in individual working day~
Shown in 5f, 6a~6f, 7a~7f.
Step S4:Room classes, give new grouped data, and after being pre-processed through step S1, three trained using S3 are returned
Sound state net rebuilds the use power mode in room, and then the extreme learning machine trained using S3 obtains the type in room.
The classification results in each room of the office building are as shown in the table.
Each room classes result in table 1 office building
★:Office △:Computer room zero:Storeroom ■:Meeting room
Particular embodiments described above, has been carried out further in detail to the purpose of the present invention, technical scheme and beneficial effect
Describe in detail bright, it should be understood that the foregoing is only specific embodiment of the invention, be not intended to limit the invention, it is all
Within the spirit and principles in the present invention, any modification, equivalent substitution and improvements done etc. should be included in protection of the invention
Within the scope of.
Claims (3)
1. a kind of office building room classes method based on neutral net, it is characterised in that comprise the following steps:
Step S1:Data prediction, its method is:The electricity consumption data in office building room is divided into N classes, to N class electricity consumption datas
Pre-processed, obtained complete electricity consumption data;
Step S2:Netinit, its method is:Construction N number of echo state net corresponding with N class electricity consumption datas and a limit
Learning machine, and parameter to each neutral net initializes;
Step S3:Network training, its method is:Each echo state net is trained using electricity consumption data complete in step S1, is rebuild
The use power mode in room, and using electricity consumption mode reconstruction result and room class training extreme learning machine;
Step S4:Room classes, its method is:The new office building room N class electricity consumption datas of input, train using in step S3
Three echo state nets rebuild room use power mode, and using in step S3 train extreme learning machine carry out office building
Room classes.
2. the method for claim 1, it is characterised in that the electricity consumption data in office building room is divided into three in step S1
Class, respectively socket electricity consumption data, electric consumption on lighting data and air conditioning electricity data.
3. method as claimed in claim 2, it is characterised in that the method for electricity consumption data pretreatment includes data screening, data
Rejecting, data filling.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410828012.3A CN104573818B (en) | 2014-12-25 | 2014-12-25 | A kind of office building room classes method based on neutral net |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410828012.3A CN104573818B (en) | 2014-12-25 | 2014-12-25 | A kind of office building room classes method based on neutral net |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104573818A CN104573818A (en) | 2015-04-29 |
CN104573818B true CN104573818B (en) | 2017-06-13 |
Family
ID=53089835
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410828012.3A Active CN104573818B (en) | 2014-12-25 | 2014-12-25 | A kind of office building room classes method based on neutral net |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104573818B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105071421B (en) * | 2015-08-17 | 2019-03-15 | 中国科学院自动化研究所 | Office Building Energy Consumption management method |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102075352A (en) * | 2010-12-17 | 2011-05-25 | 北京邮电大学 | Method and device for predicting network user behavior |
CN102262211A (en) * | 2011-04-20 | 2011-11-30 | 哈尔滨工业大学 | Analog circuit failure diagnosis method based on dynamic classification of echo state network |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103971690A (en) * | 2013-01-28 | 2014-08-06 | 腾讯科技(深圳)有限公司 | Voiceprint recognition method and device |
-
2014
- 2014-12-25 CN CN201410828012.3A patent/CN104573818B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102075352A (en) * | 2010-12-17 | 2011-05-25 | 北京邮电大学 | Method and device for predicting network user behavior |
CN102262211A (en) * | 2011-04-20 | 2011-11-30 | 哈尔滨工业大学 | Analog circuit failure diagnosis method based on dynamic classification of echo state network |
Non-Patent Citations (2)
Title |
---|
基于非线性读出器的回声状态网络及其应用研究;乌日娜;《中国优秀硕士学位论文全文数据库信息科技辑》;20141015(第10期);参见第14页、第22页3.3小节,图3.4 * |
建筑分类分项能耗计算方法的研究;马金星等;《建筑热能通风空调》;20120430;第31卷(第2期);第3.1-4.3小节,图7 * |
Also Published As
Publication number | Publication date |
---|---|
CN104573818A (en) | 2015-04-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104268682B (en) | Planning method and device for active power distribution network | |
CN110285532A (en) | Method for controlling machine room air conditioner, apparatus and system based on artificial intelligence | |
CN107579518B (en) | Power system environment economic load dispatching method and apparatus based on MHBA | |
CN109948833A (en) | A kind of Hydropower Unit degradation trend prediction technique based on shot and long term memory network | |
Girard | The role of cultural urban landscape towards a new urban economics: new structural assets for increasing economic productivity through hybrid processes | |
CN106230026A (en) | The power distribution network bilayer coordinated planning method containing distributed power source analyzed based on temporal characteristics | |
CN106803135A (en) | The Forecasting Methodology and device of a kind of photovoltaic power generation system output power | |
CN107681654A (en) | Consider the Multi-Objective Electric Power Network Planning method and apparatus of new energy consumption and environmental benefit | |
CN102629106A (en) | Water supply control method and water supply control system | |
CN104134097A (en) | Severe cold area office building form energy-saving design method based on GANN-BIM model | |
CN107706921A (en) | Micro-capacitance sensor voltage adjusting method and device based on Nash Game | |
CN105896578B (en) | A kind of random production analog method for wind-solar-storage joint electricity generation system | |
CN107609667A (en) | Heating load forecasting method and system based on Box_cox conversion and UFCNN | |
CN110163145A (en) | A kind of video teaching emotion feedback system based on convolutional neural networks | |
CN106296425A (en) | Attributed graph clustering method based on Weight associating Non-negative Matrix Factorization and system | |
CN110400234A (en) | A kind of urban water supply dispatching method and system based on BP neural network | |
CN108960288A (en) | Threedimensional model classification method and system based on convolutional neural networks | |
CN104751373A (en) | Environmental economic dispatching method considering polluted gas emission risk | |
CN109376941A (en) | Energy simulation and optimization method based on artificial neural network and BIM | |
CN107145968A (en) | Photovoltaic apparatus life cycle cost Forecasting Methodology and system based on BP neural network | |
CN104573818B (en) | A kind of office building room classes method based on neutral net | |
CN112182951A (en) | Land utilization planning method considering multiple uncertainties | |
Younes | A novel hybrid FFA-ACO algorithm for economic power dispatch | |
Subramanian et al. | An efficient firefly algorithm to solve economic dispatch problems | |
CN106503793B (en) | A kind of neural network short-term wind speed forecasting method based on improvement difference algorithm |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |