CN110188956A - Load forecasting method based on ant colony neural network and the system comprising this method, memory - Google Patents
Load forecasting method based on ant colony neural network and the system comprising this method, memory Download PDFInfo
- Publication number
- CN110188956A CN110188956A CN201910473081.XA CN201910473081A CN110188956A CN 110188956 A CN110188956 A CN 110188956A CN 201910473081 A CN201910473081 A CN 201910473081A CN 110188956 A CN110188956 A CN 110188956A
- Authority
- CN
- China
- Prior art keywords
- ant
- data
- neural network
- load
- historical
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 70
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000013277 forecasting method Methods 0.000 title claims abstract description 19
- 230000015654 memory Effects 0.000 title claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 238000005457 optimization Methods 0.000 claims abstract description 9
- 238000012549 training Methods 0.000 claims abstract description 8
- 239000003016 pheromone Substances 0.000 claims description 37
- 210000002569 neuron Anatomy 0.000 claims description 5
- 241000257303 Hymenoptera Species 0.000 claims description 4
- 238000004140 cleaning Methods 0.000 claims description 3
- 230000001186 cumulative effect Effects 0.000 claims description 2
- 238000006467 substitution reaction Methods 0.000 claims description 2
- 230000005611 electricity Effects 0.000 description 5
- 238000001514 detection method Methods 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910002056 binary alloy Inorganic materials 0.000 description 1
- 230000009193 crawling Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 210000005036 nerve Anatomy 0.000 description 1
- 235000020825 overweight Nutrition 0.000 description 1
- 230000005612 types of electricity Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Economics (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of load forecasting method based on ant colony neural network and the systems comprising this method, memory, the factor and the historical load data of corresponding period that this method will affect electric load are as training sample, after pre-processing to data, the BP neural network prediction model of ant group algorithm optimization is established.The present invention finds optimal value as the parameter of BP neural network using ant group algorithm, so that the convergence rate of BP neural network is promoted, while effectively improving the accuracy of prediction.And the present invention devises system and memory comprising this method according to this method.
Description
Technical field
The present invention relates to smart grid prediction field more particularly to a kind of load forecasting methods based on ant colony neural network
And the system comprising this method, memory.
Background technique
Since the 17th century electromagnetism comes out, electricity plays increasingly important role in people's lives.Electric energy has not
The characteristic easily stored, if the electric energy of production is more than the electric energy used, it will cause wastes, if the electric energy of production is less than use
Electric energy, it is in short supply that it will cause electric energy.Therefore, load forecast is an important research topic from beginning to end.
BP neural network is load forecasting method one of of the widely used electric power based on ant colony neural network in recent years,
It is more convenient to the processing of the factors such as temperature, humidity, wind scale number, and there is extremely strong adaptivity to unstructuredness data,
Big data field uses extensive.Using BP neural network carry out load forecast, calculating speed is very fast, but error also compared with
Greatly.
Meanwhile having a large amount of literature research ant group algorithm Optimized BP Neural Network, but ant group algorithm would generally make
Ant falls into locally optimal solution, can not obtain substantive optimal solution, and algorithm is divided into part by the ant group algorithm in the document of part
Search and two parts of global search solve the problems, such as that ant group algorithm falls into locally optimal solution to a certain extent, but generate
The search of two different levels, improves the complexity of algorithm, causes efficiency of algorithm low.
Summary of the invention
To solve the problems, such as that BP neural network prediction model precision of prediction is poor, solves ant group algorithm optimal solution and algorithm is multiple
The problem of miscellaneous degree matches, the purpose of the present invention use following technical scheme
A kind of load forecasting method based on ant colony neural network, it is characterised in that the following steps are included:
S1: choosing the historical data of historical date, the historical data include historical load data and with the history
The relevant history meteorological data of load data；
S2: establishing BP neural network, and using meteorological data as input quantity, load data sets implicit mind as output quantity
Through first number, and confirm learning rules；
S3: retrieval map is established according to the weight number n that the neural network includes, retrieved map is by n weight
The data acquisition system of generation is constituted；
S4: the probability of all elements on initialization retrieval map is equal, and initializes ant number；
S5: each ant is driven to creep between the element of n weight set by probability；
S6: the pheromones that each ant leaves in the element of the weight set are calculated；
S7: calculating the pheromones summation that the passed through ant of element of weight leaves, and obtains cumulative information element；
S8: according to the Pheromone update ant by the probability of the element of weight:
Wherein, α indicates that the relative importance of pheromones, β indicate the relative importance of heuristic greedy method；Ja(i) table
Show the weight element set that ant a is selected in next step, t is the moment；ηij(t) indicate t moment heuristic greedy method, reflection ant from
The inspiration degree of element i to element j；τis(t) pheromones in t moment in i set on s element are indicated；
S9: re-executeing the steps S5, until maximum number of iterations or termination condition；
S10: the ant for exporting optimal value climbs the weight element traveled across；
S11: corresponding to weight for weight element substitution BP neural network, and by historical load data and its relevant goes through
History meteorological data substitutes into neural network and completes training.
Further, in step S6, pheromones that the element that ant gets over leaves:
Wherein, Q indicate ant a complete one cycle after caused by pheromones and, for the constant of setting；eaFor element
When weight of the j at neural network, the maximum output error of all training samples, i.e.,N is sample
This number, x0With xkFor the reality output and desired output of neural network.
Further, in step S7, pheromones summation that passed through ant leaves:
Wherein, ρ indicates the volatility coefficient of pheromones,Expression circulates in weight set at t+1 timesThe pheromones summation that j-th of element leaves.
Further, the input quantity in step S2 further includes date type.
Further, in step S1 further include cleaning to the historical load data, the meteorological data.
It further, further include to the historical load data in step S1, the history meteorological data is normalized
Processing.
Further, the historical date in step S1 refers to prediction N days a few days ago, and the historical load data refers to pre-
The load data of N days a few days ago M periods is surveyed, when the history meteorological data includes first N days maximum temperatures, minimum temperature and M
Temperature, the humidity of section.
Further, termination condition described in step S9 is that all ants converge to same path.
The present invention also provides a kind of memories, and it includes the load forecasting methods based on ant colony neural network.
Meanwhile the present invention also provides a kind of load prediction systems, including data preprocessing module, BP neural network mould
Block, ant group optimization module and output module, the data preprocessing module is used to acquire the historical data of historical date, described to go through
History data include historical load data and history meteorological data relevant to the historical load data；
The BP neural network module is for receiving meteorological data, output load data；
The ant group optimization module is used to optimize the weight of the neural network module, and ant passes through the general of the element of weight
Rate:
Wherein, α indicates that the relative importance of pheromones, β indicate the relative importance of heuristic greedy method；Ja(i) table
Show the element set for the weight that ant a is selected in next step, t is the moment；ηij(t) heuristic greedy method for indicating t moment, reflects ant
Inspiration degree from element i to element j；τis(t) pheromones in t moment in i set on s element are indicated；
The output module is used for the output prediction load data after receiving prediction day meteorological data.
Compared with prior art, the beneficial effects of the present invention are:
(1) present invention predicts that electric load, algorithm is simple using BP neural network model, and calculating speed is fast, right
The prediction of electric load has good applicability；
(2) present invention combines ant group algorithm that BP neural network is optimized, and BP nerve has been determined using ant group algorithm
The optimal weight of network and threshold value, accelerate the convergence rate of neural network, and improve the accuracy of prediction model；
(3) present invention uses heuristic greedy method η when updating probability of the ant by the element of weightij(t), and believe
The relative importance of element is ceased, the relative importance index of heuristic greedy method can be in ant in opposing factors select probability
In lesser situation, ant further goes to explore unknown element, to reach the defect for improving ant group algorithm limitation.
Detailed description of the invention
Fig. 1: for the load forecasting method frame of the invention based on ant colony neural network；
Fig. 2: for the data preprocessing module of Fig. 1；
Fig. 3: for the BP neural network module of Fig. 1；
Fig. 4: for the ant group optimization module of Fig. 1.
Specific embodiment
To solve the problems, such as that BP neural network prediction model precision of prediction is poor, the present invention will use ant group algorithm to confirm mind
Through the initial parameter in network, accelerates the convergence rate of BP neural network and improve the precision of algorithm.Firstly, extracting influences electric power
The historical load data of the factor of load and corresponding period；Then, data are pre-processed, rejecting abnormalities data are laggard
Row normalized establishes BP neural network model；Finally, finding optimal value as the ginseng of BP neural network using ant group algorithm
Number, output calculate error as a result, compared with actual value.Improved prediction model convergence rate is promoted, and the accuracy of prediction also has
It is improved.In the following, being described further in conjunction with attached drawing and specific embodiment to the present invention:
Referring to attached drawing 2, the present invention is analyzed by data of the Principal Component Analysis to short-term effect electric load, is obtained
The important factor in order of electric load.And since the change of electric load is the number with the electrical equipment electricity consumption main body in region
It measures closely related, according to basis of the data as prediction data of long period, is then faced with historical time electricity consumption main body
Quantity and status electricity consumption main body quantity are inconsistent and cause the deviation of prediction, therefore prediction data was using prediction a few days ago N days
Historical data, wherein N is not more than 360, preferably 30 days, could influence by the variation of electricity consumption main body to prediction data reduce.Through
Principal component analysis is crossed, historical load data of the invention preferably predicts the load data of N days a few days ago M periods, the history gas
Image data includes first N days maximum temperature, temperature, the humidity of minimum temperature and M period.Further, due to working day with it is non-
On working day, quantity, the type of electricity consumption main body are different, in order to distinguish to working day and nonworkdays, in historical data also
It should include the type of historical date.Training sample of the invention includes the data such as load, temperature, humidity, it would be possible to be influenced
Exhausted big several factors of electric load are all considered in model, so that forecast result of model is more preferably.
After historical data acquisition, it is likely to occur data incompleteness in historical data, such as lacks load data, maximum temperature, most
It is one or more in low temperature, humidity, it is also possible to appear in that occur an exception in continuous historical date high or abnormal
Low electric load needs to reject the above abnormal data when obtaining historical data, i.e., carries out to historical data clear
It washes, obtains the historical data with reference value.The cleaning method of historical data includes but is not limited to following methods: value missing inspection
Survey method, error value detection method repeat record detection method, inconsistency detection method.
After history data collection, the dimension of temperature, humidity, load data etc. is inconsistent in historical data, and highest
Temperature, minimum temperature, humidity etc. are to directly affect the factor of neural network fitting, if directly dimension is removed, the highest temperature
Degree, minimum temperature, humidity maxima and minima between span by generate difference, cause the downward gradient of various data not
Unanimously, if larger difference occurs in downward gradient, influence of the data that downward gradient may be caused small to output result is unknown
It is aobvious, it is inconsistent with practical situation, neural network can not the influence to each factor accurately evaluated.It is therefore preferred that
Historical data after collection is normalized:
Wherein, LiIndicate the data of i-th of factor,Indicate the maximum value in the factor observation,It indicates
Maximum value in the factor observation.
BP neural network can be established after data acquisition, determines input quantity, output quantity, hidden neuron number, and really
Recognize learning rules.
The input quantity of neural network of the present invention is meteorological data, and output quantity is electric load.Further, input quantity may be used also
With the date type on the day of including meteorological data.
Preferably, historical load data preferably predicts the load data of N days a few days ago M periods, and history meteorological data includes
First N days maximum temperatures, temperature, the humidity of minimum temperature and M period.History meteorological data and its relevant historical load data
Training sample is constituted, BP neural network is trained.
Referring to attached drawing 3, BP neural network is established, determines hidden layer neuron numberWherein, m is
Input layer number, n are output layer neuron number, and k is hidden layer neuron number, and p is the constant between 1-10.Such as
Input quantity selects maximum temperature, minimum temperature, humidity, and output quantity is electric load, then m is equal to 3, n and is equal to 1.M and n can roots
According to the quantity of data and expression pattern (decimal system, binary system etc.) appropriate adjustment of data.
Referring to attached drawing 4, ant group algorithm is established according to arrangement above, neural network weight total ρ is obtained, is denoted as H1,
H2... Hρ, constitute ρ set The size of set can predict according to the magnitude range to weight, computational accuracy
Prediction is defined.And each is initialized in orderIn elementValue, and set the pheromones of each element
Amount isInitialization ant number is F, and maximum optimizing number is T.ρ weight setSequentially splice
Form the retrieval map of ant.Regulation is set, and each ant successively gets over weight setAnd it can and can only select every
One setIn 1 element obtained the path of creeping of a wherein ant.
By the ant creep path process element according toEstablish regulation, the correspondence weight of reverse setting BP neural network,
The meteorological data of sample is inputted into neural network input layer, obtaining output may compare the output and desired value (historical load number
According to) error, and the pheromones that the element that the ant gets over leaves can be obtained:
Wherein, Q indicate ant a complete one cycle after caused by pheromones and, for the constant of setting；eaFor element
When weight of the j at neural network, the maximum output error of all training samples, i.e.,N is sample
This number, x0With xkFor the reality output and desired output of neural network.
All ants are in weight set at this timeThe pheromones total amount that j-th of element leaves can be asked, each optimizing result
Afterwards, it is adjusted according to the volatilization of pheromones and new generation, formula is as follows:
Wherein, ρ indicates the volatility coefficient of pheromones,Expression circulates in weight set at t+1 timesThe pheromones summation that j-th of element leaves.
It can be obtained any one ant right to choose value set according to pheromonesThe probability of element j is the element information
Element accounts for the ratio of all elements total information element, i.e.,
Wherein, α indicates the relative importance of pheromones；The relative importance of β expression heuristic greedy method；Ja(i) table
Show the set for the weight element that ant a is selected in next step, t is the moment；ηij(t) heuristic greedy method for indicating t moment, reflects ant
Inspiration degree from element i to element j；τis(t) pheromones in t moment in i set on s element are indicated.The present invention increases
Heuristic greedy method ηij(t) and the relative importance of pheromones, the relative importance index of heuristic greedy method, Ke Yi
In the lesser situation of opposing factors select probability, ant further goes to explore unknown element ant, so that reaching improves ant
The defect of group's algorithm limitation.
All ants converge to same path or are optimal number stopping, and output optimal value is joined as neural network
Number, otherwise repeatedly ant crawling process.Bring optimized parameter into neural network model, be trained using sample, and with reality
Value compares, and calculates error amount.
Using BP neural network to load prediction when, it is thus only necessary to by predict day on the day of maximum temperature, minimum temperature,
It predicts that the temperature of day prediction period, humidity input BP neural network, can be obtained corresponding prediction load output.
The present invention also provides a kind of memories, and it includes the load forecasting methods based on ant colony neural network.
Meanwhile the present invention also provides a kind of load prediction system (attached drawing 1), including data preprocessing module, BP neural network module,
Ant group optimization module and output module.Referring to attached drawing 2, the data preprocessing module is used to acquire the history number of historical date
According to the historical data includes historical load data and history meteorological data relevant to the historical load data.
Referring to attached drawing 3, the BP neural network module is for receiving meteorological data, output load data.
Referring to attached drawing 4, the ant group optimization module is used to optimize the weight of the neural network module, and ant passes through weight
Element probability:
Wherein, α indicates that the relative importance of pheromones, β indicate the relative importance of heuristic greedy method；Ja(i) table
Show the city gather that ant a is selected in next step, t is the moment；ηij(t) heuristic greedy method for indicating t moment reflects ant from element
The inspiration degree of i to element j；τis(t) pheromones in t moment in i set on s element are indicated.
The output module is used for the output prediction load data after receiving prediction day meteorological data.
Further, the historical date refers to prediction N days a few days ago, and the historical load data refers to prediction N days a few days ago
The M period load data, the history meteorological data include before N days maximum temperature, the temperature of minimum temperature and M period,
Humidity.
Further, the data preprocessing module further includes date type, and the BP neural network module is also used to connect
Receive date type.
Further, the data preprocessing module is also used to that history meteorological data is normalized, specific to locate
Reason method can be found in the normalized in a kind of load forecasting method based on ant colony neural network of the present invention.It is worth note simultaneously
Meaning when, it is a kind of based on ant colony neural network that BP neural network module, the ant group optimization module of the system can be found in the present invention
About BP neural network, ant group algorithm part in load forecasting method.
It will be apparent to those skilled in the art that can make various other according to the above description of the technical scheme and ideas
Corresponding change and deformation, and all these changes and deformation all should belong to the protection scope of the claims in the present invention
Within.
Claims (10)
1. the load forecasting method based on ant colony neural network, it is characterised in that the following steps are included:
S1: choosing the historical data of historical date, the historical data include historical load data and with the historical load
The relevant history meteorological data of data；
S2: establishing BP neural network, and using meteorological data as input quantity, load data sets hidden neuron as output quantity
Number, and confirm learning rules；
S3: retrieval map is established according to the weight number n that the neural network includes, retrieved map is generated by n weight
Data acquisition system constitute；
S4: the probability of all elements on initialization retrieval map is equal, and initializes ant number；
S5: each ant is driven to creep between the element of n weight set by probability；
S6: the pheromones that each ant leaves in the element of the weight set are calculated；
S7: calculating the pheromones summation that the passed through ant of element of weight leaves, and obtains cumulative information element；
S8: according to the Pheromone update ant by the probability of the element of weight:
Wherein, α indicates that the relative importance of pheromones, β indicate the relative importance of heuristic greedy method；Ja(i) ant is indicated
The city gather that a is selected in next step, t is the moment；ηij(t) heuristic greedy method for indicating t moment reflects ant from element i to member
The inspiration degree of plain j；τis(t) pheromones in t moment in i set on s element are indicated；
S9: re-executeing the steps S5, until maximum number of iterations or termination condition；
S10: the ant for exporting optimal value climbs the weight element traveled across；
S11: corresponding to weight for weight element substitution BP neural network, and by historical load data and its relevant history gas
Image data substitutes into neural network and completes training.
2. the load forecasting method according to claim 1 based on ant colony neural network, it is characterised in that: in step S6,
The pheromones that the weight element that ant gets over leaves:
Wherein, Q indicate ant a complete one cycle after caused by pheromones and, for the constant of setting；eaIt is element j into mind
When weight through network, the maximum output error of all training samples, i.e.,N is sample number
Mesh, x0With xkFor the reality output and desired output of neural network.
3. the load forecasting method according to claim 1 based on ant colony neural network, it is characterised in that: in step S7,
The pheromones summation that passed through ant leaves:
Wherein, ρ indicates the volatility coefficient of pheromones,Expression circulates in weight set at t+1 timesThe
The pheromones summation that j element leaves.
4. the load forecasting method based on ant colony neural network according to claim 1 to 3, it is characterised in that:
Input quantity in step S2 further includes date type.
5. the load forecasting method based on ant colony neural network according to claim 1 to 3, it is characterised in that:
It further include the cleaning to the historical load data, the meteorological data in step S1.
6. the load forecasting method based on ant colony neural network according to claim 1 to 3, it is characterised in that:
It further include to the historical load data in step S1, the history meteorological data is normalized.
7. the load forecasting method based on ant colony neural network according to claim 1 to 3, it is characterised in that:
The historical date in step S1 refers to prediction N days a few days ago, and the historical load data refers to the M period of prediction N days a few days ago
Load data, the history meteorological data include before N days maximum temperature, temperature, the humidity of minimum temperature and M period.
8. the load forecasting method based on ant colony neural network according to claim 1 to 3, it is characterised in that:
Termination condition described in step S9 is that all ants converge to same path.
9. a kind of memory, it is characterised in that want the load based on ant colony neural network described in 1-8 any one comprising right
Prediction technique.
10. a kind of load prediction system, including data preprocessing module, BP neural network module, ant group optimization module and output
Module, it is characterised in that:
The data preprocessing module is used to acquire the historical data of historical date, and the historical data includes historical load data
And history meteorological data relevant to the historical load data；
The BP neural network module is for receiving meteorological data, output load data；
The ant group optimization module is used to optimize the weight of the neural network module, and ant passes through the probability of the element of weight:
Wherein, α indicates that the relative importance of pheromones, β indicate the relative importance of heuristic greedy method；Ja(i) ant is indicated
The set for the weight element that a is selected in next step, t is the moment；ηij(t) heuristic greedy method for indicating t moment reflects ant from element
The inspiration degree of i to element j；τis(t) pheromones in t moment in i set on s element are indicated；
The output module is used for the output prediction load data after receiving prediction day meteorological data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910473081.XA CN110188956A (en) | 2019-05-31 | 2019-05-31 | Load forecasting method based on ant colony neural network and the system comprising this method, memory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910473081.XA CN110188956A (en) | 2019-05-31 | 2019-05-31 | Load forecasting method based on ant colony neural network and the system comprising this method, memory |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110188956A true CN110188956A (en) | 2019-08-30 |
Family
ID=67719727
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910473081.XA Pending CN110188956A (en) | 2019-05-31 | 2019-05-31 | Load forecasting method based on ant colony neural network and the system comprising this method, memory |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110188956A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111144543A (en) * | 2019-12-30 | 2020-05-12 | 中国移动通信集团内蒙古有限公司 | Data center air conditioner tail end temperature control method, device and medium |
| CN113159265A (en) * | 2021-03-24 | 2021-07-23 | 国网河南省电力公司电力科学研究院 | Traction load parameter identification method and system based on SVM-ant colony algorithm |
| CN114251753A (en) * | 2021-12-29 | 2022-03-29 | 西安建筑科技大学 | Ice storage air conditioner cold load demand prediction distribution method and system |
| CN117420052A (en) * | 2023-10-09 | 2024-01-19 | 江苏海洋大学 | PM2.5 prediction method integrating multi-scale space-time information |
| CN118494257A (en) * | 2024-07-15 | 2024-08-16 | 北京这房行信息技术有限公司 | Charging station load optimization method, device, equipment and medium based on ant colony algorithm |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109583503A (en) * | 2018-12-03 | 2019-04-05 | 国网江苏省电力有限公司扬州供电分公司 | A kind of interruptible load prediction technique |
| CN109670628A (en) * | 2018-11-09 | 2019-04-23 | 昆明理工大学 | A kind of micro-grid load prediction technique of the neural network based on LSSVM |
-
2019
- 2019-05-31 CN CN201910473081.XA patent/CN110188956A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109670628A (en) * | 2018-11-09 | 2019-04-23 | 昆明理工大学 | A kind of micro-grid load prediction technique of the neural network based on LSSVM |
| CN109583503A (en) * | 2018-12-03 | 2019-04-05 | 国网江苏省电力有限公司扬州供电分公司 | A kind of interruptible load prediction technique |
Non-Patent Citations (4)
| Title |
|---|
| 翟帅华 等: "基于改进PSO-BP神经网络的短期电力负荷预测", 《工业控制计算机》 * |
| 赵亚光: "基于蚁群算法和BP神经网络的WSN数据融合算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 郝晶: "基于蚁群算法优化神经网络的包头地区中长期电力负荷预测", 《中国优秀硕士学位论文全文数据库 经济与管理科学辑》 * |
| 陈旭: "基于神经网络的电力系统短期负荷预测研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111144543A (en) * | 2019-12-30 | 2020-05-12 | 中国移动通信集团内蒙古有限公司 | Data center air conditioner tail end temperature control method, device and medium |
| CN113159265A (en) * | 2021-03-24 | 2021-07-23 | 国网河南省电力公司电力科学研究院 | Traction load parameter identification method and system based on SVM-ant colony algorithm |
| CN114251753A (en) * | 2021-12-29 | 2022-03-29 | 西安建筑科技大学 | Ice storage air conditioner cold load demand prediction distribution method and system |
| CN117420052A (en) * | 2023-10-09 | 2024-01-19 | 江苏海洋大学 | PM2.5 prediction method integrating multi-scale space-time information |
| CN118494257A (en) * | 2024-07-15 | 2024-08-16 | 北京这房行信息技术有限公司 | Charging station load optimization method, device, equipment and medium based on ant colony algorithm |
| CN118494257B (en) * | 2024-07-15 | 2024-10-18 | 北京这房行信息技术有限公司 | Charging station load optimization method, device, equipment and medium based on ant colony algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110188956A (en) | Load forecasting method based on ant colony neural network and the system comprising this method, memory | |
| CN111932402B (en) | Short-term power load bidirectional combination prediction method based on similar day and LSTM | |
| CN108022001A (en) | Short term probability density Forecasting Methodology based on PCA and quantile estimate forest | |
| CN109002915B (en) | Photovoltaic power station short-term power prediction method based on Kmeans-GRA-Elman model | |
| CN106485262A (en) | A kind of bus load Forecasting Methodology | |
| CN108846527A (en) | A kind of photovoltaic power generation power prediction method | |
| CN112215426A (en) | Short-term power load prediction method | |
| CN116911806B (en) | Internet + based power enterprise energy information management system | |
| CN112381673B (en) | Park electricity utilization information analysis method and device based on digital twin | |
| CN112418346B (en) | Numerical weather forecast total radiation system error classification calculation method | |
| CN112329990A (en) | User power load prediction method based on LSTM-BP neural network | |
| CN112330050A (en) | Power system load prediction method considering multiple features based on double-layer XGboost | |
| CN112508734B (en) | Method and device for predicting power generation capacity of power enterprise based on convolutional neural network | |
| CN115907131B (en) | Method and system for constructing electric heating load prediction model in northern area | |
| CN112883632A (en) | Lithium battery equivalent circuit model parameter identification method based on improved ant colony algorithm | |
| CN116611702A (en) | Integrated learning photovoltaic power generation prediction method for building integrated energy management | |
| CN115759389A (en) | Day-ahead photovoltaic power prediction method based on weather type similar day combination strategy | |
| CN114692981B (en) | Method and system for forecasting medium-long-term runoff based on Seq2Seq model | |
| CN114357670A (en) | Power distribution network power consumption data abnormity early warning method based on BLS and self-encoder | |
| CN114330934A (en) | Model parameter self-adaptive GRU new energy short-term power generation power prediction method | |
| CN111680818A (en) | Short-term reactive load prediction method and system | |
| CN117407681A (en) | Time sequence data prediction model establishment method based on vector clustering | |
| CN117374917A (en) | Multiple regression prediction method and device for ultra-short-term power load | |
| CN116933201A (en) | Method and system for identifying illegal electricity utilization behavior of low-voltage charging pile | |
| CN116523540A (en) | Ultra-short-term electricity price prediction method based on self-adaptive LGBM |
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 | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Room 509, building D, 39 Ruihe Road, Huangpu District, Guangzhou City, Guangdong Province Applicant after: Guangzhou Shuimu Qinghua Technology Co.,Ltd. Address before: Room 936, Room 333 Jiufojian Road, Zhongxin Guangzhou Knowledge City, Guangzhou, Guangdong Province Applicant before: Guangzhou Shuimu Qinghua Technology Co.,Ltd. |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190830 |