CN107274038A - A kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion - Google Patents
A kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion Download PDFInfo
- Publication number
- CN107274038A CN107274038A CN201710641743.0A CN201710641743A CN107274038A CN 107274038 A CN107274038 A CN 107274038A CN 201710641743 A CN201710641743 A CN 201710641743A CN 107274038 A CN107274038 A CN 107274038A
- Authority
- CN
- China
- Prior art keywords
- ant
- lion
- ant lion
- value
- lssvm
- 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
- 241001206881 Myrmeleon inconspicuus Species 0.000 title claims abstract description 105
- 238000000034 method Methods 0.000 title claims abstract description 39
- 230000005611 electricity Effects 0.000 title abstract description 10
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 29
- 230000000717 retained effect Effects 0.000 claims abstract 2
- 238000005457 optimization Methods 0.000 claims description 26
- 239000011159 matrix material Substances 0.000 claims description 16
- 241000282320 Panthera leo Species 0.000 claims description 13
- 238000005295 random walk Methods 0.000 claims description 11
- 241000257303 Hymenoptera Species 0.000 claims description 10
- 238000012360 testing method Methods 0.000 claims description 10
- 238000010219 correlation analysis Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 4
- 206010038743 Restlessness Diseases 0.000 claims description 2
- 238000012843 least square support vector machine Methods 0.000 abstract description 41
- 241000258923 Neuroptera Species 0.000 abstract description 5
- 238000012549 training Methods 0.000 description 9
- 239000003795 chemical substances by application Substances 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 238000010606 normalization Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 241000283690 Bos taurus Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000019637 foraging behavior Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012067 mathematical method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 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
- 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)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Operations Research (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Biomedical Technology (AREA)
- Entrepreneurship & Innovation (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion, Prediction of annual electricity consumption method comprises the following steps:Determine the input variable of least square method supporting vector machine (Least Square Support Vector Machines, LSSVM) forecast model;Initialize ant lion optimized algorithm;The fitness value of initial ant lion is calculated, initial elite ant lion is obtained;The position of ant is updated, the fitness value of current ant is calculated, and the fitness value of ant lion corresponding to its is compared, and judges whether to update ant lion position;By the fitness value of the ant lion after more new position, the fitness value with previous generation elite ant lions is compared one by one, is retained the corresponding ant lion of smaller fitness value, is obtained the elite ant lion of current iteration;Judge whether to reach maximum iteration, if it has, then position and the corresponding Prediction of annual electricity consumption value of output elite ant lion, if it has not, continuing iteration.Compared with prior art, the present invention has the advantages that precision of prediction is higher and forecasting efficiency is higher.
Description
Technical Field
The invention relates to the technical field of annual power consumption prediction, in particular to an LSSVM annual power consumption prediction method based on ant lion optimization.
Background
Annual power consumption prediction is crucial to the planning, operation and maintenance of power systems, and can also reflect the economic development of a country or region to a certain extent. Accurate annual power usage predictions may provide valuable references for power system operators and economic managers. Therefore, the prediction of the power load is one of the most important basic works of the power system, and has important significance on energy planning, operation and control of the power system and research on economic development strategy. The prediction method comprises the following steps: regression analysis, analog-to-digital, elastic coefficient, time series, neural network, and the like.
In the modern economic society, the power consumption is in abnormal close relation with economy, society, population and ecological environment, namely, an electric system is a complex system influenced by a plurality of factors, a prediction model is established by a conventional mathematical method, the workload is large, and the prediction precision is difficult to ensure.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide an LSSVM annual power consumption prediction method based on ant lion optimization.
The purpose of the invention can be realized by the following technical scheme:
an LSSVM annual power consumption prediction method based on ant lion optimization comprises the following steps:
s1, determining input variables of the LSSVM prediction model;
s2, initializing a ant lion optimization algorithm, substituting the initial ant lion position as a nuclear parameter and a regularization parameter into the LSSVM model, and obtaining a corresponding annual power consumption predicted value;
s3, establishing a fitness function, calculating the fitness value of the position of the initial ant lion to obtain an initial fitness value, and keeping the ant lion corresponding to the minimum initial fitness value as the initial elite ant lion;
s4, updating the ant positions, calculating the fitness value of the current ant, comparing the fitness value with the fitness value of the ant lion position corresponding to the current ant, and judging whether the ant lion positions are updated or not;
s5, comparing the fitness value of the ant lion position obtained in the previous step with the fitness value of the position of the previous generation elite ant lion one by one, and reserving the ant lion position corresponding to a smaller fitness value to obtain the position of the current iteration elite ant lion;
and S6, judging whether the maximum iteration times is reached, if so, outputting the position of the elite lion and the corresponding predicted value of the annual power consumption, and if not, returning to S4.
Step S1 specifically includes: and obtaining a correlation value between the annual power consumption influence factors and the annual power consumption by adopting a grey correlation analysis method, and selecting the corresponding annual power consumption influence factors as input variables of the LSSVM prediction model according to the correlation value.
The ant lion optimization algorithm initialized in the step S2 comprises the following steps:
s201, initializing parameters including population size Agents _ no, solving dimension d as 2 and solving upper bound b of spaceupLower bound of solution space blowMax iteration number Max _ iter;
s202, initializing positions, and randomly generating a position matrix M of antsAntPosition matrix M of HelloAntlion。
The upper bound b of the solution space described in S201up=[1000,1000]Lower bound blow=[0.1,0.01]。
Position matrix M of ants in S202AntPosition matrix M of HelloAntlionComprises the following steps:
wherein M isAntAnd MAntlionThe value of (A) is represented by formula A*jOr AL*j=rand*(bupj-blowj)+blowjTo obtain a*jAnd AL*jThe j-th column values of the ant position matrix and the ant lion position matrix are respectively represented, rand is a random number between 0 and 1, and n is Agents _ no, bupjAnd blowjRespectively of column jUpper and lower bounds, MAntAnd MAntlionEach row of (a) corresponds to a set of kernel parameters and regularization parameters, i.e., (σ, γ), of the LSSVM.
The fitness function in step S3 is:wherein,andthe actual value and the predicted value of the s-th group of test samples are respectively, and N is the number of the samples.
The update formula of the ant position in step S4 isWherein,the t-th iteration of roulette around the ant lion is shown to bet the selected ant position,the ant positions of the t iteration are shown as randomly wandering around the elite ant lion,the position of the ith ant at the t iteration.
Wherein, aiRepresents the minimum value of the random walk position of the ith ant, hiRepresents the maximum value of the position of the random walk of the ith ant,represents the actual position of the ith ant at the t-th iterationThe minimum value of the sum of the values of,represents the maximum value of the actual position of the ith ant at the t iteration, Wi tRepresents the random walk position, W, of the ith ant at the t-th iterationi t=[0,cumsum(2r1(t)-1),…,cumsum(2rMax_iter(t)-1)]Cumsum is a function for calculating the accumulated value of the array, t represents the current iteration number, Max _ iter represents the maximum iteration number, r1(t),…,rMax_iter(t) are random functions and independent of each other.
The random function calculation formula is as follows:
r (t) is r1(t),…,rMax_iter(t), rand (t) is a random number between 0 and 1.
And step S4, comparing the fitness value of the current ant with the corresponding fitness value of the ant lion position, judging whether the ant lion position is updated, if the fitness value of the current ant is smaller than the fitness value of the ant lion, replacing the ant lion position with the current ant position, otherwise, keeping the original ant lion position.
Compared with the prior art, the invention has the following advantages:
1. compared with the existing prediction method, the prediction precision is higher: the prediction method of the LSSVM is optimized by using the ant lion optimization algorithm, the prediction result is closer to the actual power consumption, and the relative error of prediction is reduced;
2. the prediction efficiency is improved: automatic cycle iteration is realized, the prediction efficiency is high, and the iteration running time is short.
Drawings
FIG. 1 is a flow chart of the annual power consumption prediction method of LSSVM based on ant lion optimization according to the present invention;
FIG. 2 is a detailed flow chart of ALO optimization of LSSVM parameters;
FIG. 3 is a comparison graph of results predicted by the ALO-LSSVM and GCA-ALO-LSSVM methods;
FIG. 4 is a comparison of predicted values obtained using other methods and using the method of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, shall fall within the scope of protection of the present invention.
The invention provides an LSSVM annual power consumption prediction method based on ant lion optimization, which mainly comprises two parts: firstly, Correlation Analysis is carried out on power consumption influence factors and power consumption by adopting Grey Correlation Analysis (GCA), LSSVM input variables are determined according to Grey Correlation values, the LSSVM structure is simplified, and then the optimal parameters of a Least Square Support Vector Machine (LSSVM) are obtained through an Ant Lion optimization Algorithm (ALO).
(1)GCA
The GCA is one of the main contents of grey system theory, and the basic principle is to distinguish the closeness of the relationship of multiple factors in the system by comparing the statistical sequence set relationship, and the closer the set shapes of the sequence curves are, the greater the correlation degree between the sequence curves is, and vice versa. The specific process is as follows:
1) the electricity consumption sequence is X0=(x0(1),x0(2),…,x0(k),…,x0(p)), the factor influencing the amount of electricity used is Xg=(xg(1),xg(2),…,xg(k),…xg(p)), (g ═ 1,2, …, m), p represents the total number of years of selected electricity usage, and m represents the number of electricity usage influencing factors;
2) using the initialized operator to X0And XiInitializing to obtain initial images of Y0=(y0(1),y0(2),…,y0(k),…,y0(p)) and Yg=(yg(1),yg(2),…,yg(k),…,yg(p)), (g ═ 1,2, …, m), m represents the number of factors affecting the electricity consumption;
3) determining Y0And YgCoefficient of correlation between gamma0g(k),In the formula: y is0kI.e. y0(k),ygkI.e. yg(k),ξ∈(0,1]For the resolution factor, 0.5 is generally adopted;
4) calculating X0And XgDegree of association γ of0g,In the formula: gamma ray0gkNamely gamma0g(k),γ0g∈(0,1],γ0gThe larger, indicates XgTo X0The greater the effect of (a), if gamma0g≥γ0qThen factor XgIs superior to factor Xq,XqRepresents the q-th power consumption influence factor, q is 1,2, …, m, gamma0qIs X0And XqThe degree of association of (c).
(2) LSSVM prediction model based on ALO
The ALO algorithm is a new intelligent evolutionary algorithm proposed by s.mirjarlii in 2015, and the main inspiration of the ALO algorithm comes from foraging behavior of lion larvae. The algorithm follows the following criteria in the optimization process:
criterion 1: ants adopt different paths to randomly walk in a search space, the random walking is suitable for ants in all dimensions, and the random walking of the ants is influenced by ant lion traps.
Criterion 2: the ant lion can establish a trap which is matched with the fitness value, and the ant lion with the larger trap has higher probability of capturing ants.
Criterion 3: the random walk range decreases adaptively as ants move about the ant lion.
Criterion 4: if the fitness value of an ant is smaller than that of a lion, it means that it will be captured by the lion.
Criterion 5: the lion relocates to a new position near the capturing ant and the trap is rebuilt to accommodate the changes after the capturing ant.
The specific steps for solving the kernel parameter and the regularization parameter in the LSSVM based on the ALO are as follows:
1) carrying out normalization processing on input variables and output variables of the LSSVM, and dividing 25 groups of data into two parts after normalization to [0,1 ]: the first 20 groups of data were used as training sample set L and the last 5 groups of data were used as validation sample set Z. Then, selecting one group of data from a training sample set L (total N is 20 groups of data) to form a training sample subset U, establishing and training an LSSVM model, using the residual data in L as a test sample subset V, and testing the trained model.
2) Parameter initialization of the ALO algorithm: the method comprises population scale Agents _ no, wherein the position of each ant lion is a solution, namely two parameters in the LSSVM: kernel parameter σ and regularization parameter γ, so that the dimension d of the solution is 2, the upper bound b of the solution spaceup=[bup1,bup2]Lower bound blow=[blow1,blow2]Max iteration number Max _ iter. At the same time, a fitness function F is established,wherein,andrespectively the actual value and the predicted value of the s-th group of test samples, and N is the number of the samples;
3) and (3) initializing the position of the ALO algorithm: initial position of antInitial position of ant lionWherein, n is Agents _ no, MAntAnd MAntlionThe value of (A) is represented by formula A*jOr AL*j=rand*(bupj-blowj)+blowjTo obtain a*jAnd AL*jThe value representing the jth column of the position matrix, rand being a random number between 0 and 1, bupjAnd blowjUpper and lower bounds, M, respectively, for the jth column variableAntAnd MAntlionEach row of (a) corresponds to a set of kernel parameters and regularization parameters (σ, γ) of the LSSVM;
4) substituting the ant lion positions, namely the nuclear parameters and the regularization parameters into an LSSVM prediction model to obtain corresponding prediction values, then calculating the fitness value of the initial ant lion according to a fitness function F, wherein the ant lion with the minimum initial fitness value is the initial elite ant lion elite(0);
5) Keeping the position of the previous generation of elite lion, updating the ant position according to the ant lion position and the elite lion position, calculating the fitness value of the current generation of ants, comparing the fitness value with the fitness value of the corresponding ant lion, and when the fitness value of the ants is smaller than the fitness value of the ant lion, capturing the ants;
6) updating the ant lion position to the ant catching and killing position, calculating the adaptability value of the ant lion after updating the position, sequentially comparing the adaptability value with the adaptability value of the previous generation elite ant lion, and reserving the ant lion with smaller adaptability value to obtain the elite ant lion elite of the iteration(t);
7) And if the maximum iteration times are not reached, returning to the step 5), otherwise, outputting the position of the elite lion, namely the nuclear parameter and the regularization parameter (sigma, gamma) of the LSSVM.
The annual power consumption prediction method of the LSSVM based on ant lion optimization comprises the following steps, as shown in FIG. 2:
1) determining input variables of a Least Square Support Vector Machine (LSSVM);
2) initializing Ant Lion optimization algorithm (Ant Lion Optimizer, ALO);
3) establishing a fitness function, calculating an initial fitness value, and obtaining an initial elite ant lion position;
4) updating the ant position, calculating the fitness value of the current ant position, comparing the fitness value with the corresponding ant lion, and judging whether the ant lion position is updated;
5) calculating the fitness value corresponding to the updated ant lion position, comparing the fitness value with the fitness value of the previous generation elite ant lion one by one, and reserving the ant lion with smaller fitness value to obtain the elite ant lion of the iteration;
6) judging whether the maximum iteration frequency is reached, if so, turning to the next step, and if not, adding 1 to the iteration frequency and turning to 4);
7) and finishing the iterative process to obtain the predicted annual power consumption value.
The input variable determination method of the LSSVM in the step 1) comprises the following steps: and performing grey correlation analysis on the power consumption influence factors and the power consumption by adopting a grey correlation analysis method to obtain a correlation value between the power consumption influence factors and the power consumption, and selecting the annual power consumption influence factors as input variables of the LSSVM according to the correlation value.
The step 2) is specifically as follows:
201) ALO parameters are initialized, including population size Agents _ no, each ant lionThe position of (a) is a solution, i.e. two parameters in LSSVM: kernel parameter σ and regularization parameter γ, so that the dimension d of the solution is 2, the upper bound b of the solution spaceup=[bup1,bup2]Lower bound blow=[blow1,blow2]Max iteration number Max _ iter;
202) ALO position initialization, randomly generating position matrix M of antsAntPosition matrix M of HelloAntlion:Wherein, n is Agents _ no, MAntAnd MAntlionThe value of (A) is represented by formula A*jOr AL*j=rand*(bupj-blowj)+blowjTo obtain a*jAnd AL*jThe value representing the jth column of the position matrix, rand being a random number between 0 and 1, bupjAnd blowjUpper and lower bounds, M, of the jth column, respectivelyAntAnd MAntlionEach row of (a) corresponds to a set of kernel parameters and regularization parameters (σ, γ) of the LSSVM.
Upper bound b of the solution spaceup=[1000,1000]Lower bound blow=[0.1,0.01]。
The step 3) is specifically as follows:
301) carrying out normalization processing on input variables and output variables of the LSSVM, and dividing 25 groups of data into two parts after normalization to [0,1 ]: the first 20 groups of data were used as training sample set L and the last 5 groups of data were used as validation sample set Z. Then, selecting one group of data from a training sample set L (total N is 20 groups of data) to form a training sample subset U, establishing and training an LSSVM model, taking the residual data in L as a test sample subset V, and testing the trained model;
302) mixing M in step 2)AntlionSubstituting the LSSVM into the LSSVM, training the LSSVM, and predicting by using the trained LSSVM to obtain a predicted value;
303) according to the fitness function F,wherein,andand respectively testing the actual value and the predicted value of the sample of the s-th group, wherein N is the number of the samples, calculating the initial fitness value of each ant lion, comparing all the initial fitness values one by one to obtain and record the position of the initial elite ant lion and the fitness value thereof, wherein the position of the elite ant lion is the optimal (sigma, gamma) of the LSSVM.
The updating formula of the ant position in the step 4) is as follows:wherein,indicating the ant positions that the tth iteration randomly walked around the lion selected for roulette,representing the ant positions that randomly walked around the elite ant lion at the t-th iteration,the position of the ith ant at the t iteration.
Wherein, aiRepresents the minimum value of the random walk position of the ith ant, hiRepresents the maximum value of the position of the random walk of the ith ant,represents the minimum value of the actual position of the ith ant at the time of the t iteration,represents the maximum value of the actual position of the ith ant at the t iteration, Wi tThe random walk position of the ith ant at the t-th iteration is shown. Specifically, the ant itself random walk formula is: wi t=[0,cumsum(2r1(t)-1),…,cumsum(2rMax_iter(t)-1)]Cumsum is a function for calculating the accumulated value of the array, t represents the current iteration number, Max _ iter represents the maximum iteration number, r1(t),…,rMax_iter(t) are random functions and independent of each other.
When calculatingWhen the ant swims around the ant lion randomly, represents the minimum value of the actual position of the ith ant at the t iteration, ctIs the minimum of all ant positions at the t iteration,represents the maximum value of the actual position of the ith ant at the time of the t iteration, dtIs the maximum of all ant positions at the t iteration,indicating the ith ant lion position selected at the t-th iteration.
When calculatingWhen the ant swims around the Elite lion randomly, represents the minimum value of the actual position of the ith ant at the t iteration, ctIs the minimum of all ant positions at the t iteration,represents the maximum value of the actual position of the ith ant at the time of the t iteration, dtIs the maximum value of all ant positions at the t iteration, ElitetShowing the position of the elite lion at the t-th iteration.
ct,dtThe calculation formula is as follows:where t denotes the current number of iterations, I is the ratio, ct' represents the minimum of all ant positions at the last moment of the t-th iteration, dt' represents the maximum of all ant positions at the last moment in the tth iteration.
The ratio I is calculated as:
in the formula, t represents the current iteration number, and Max _ iter represents the maximum iteration number.
Random function r1(t),…,rMax_iter(t) the calculation formula is:
r (t) is r1(t),…,rMax_iter(t), rand is a random number between 0 and 1.
And then according to the F, calculating the fitness value of the current ant position, comparing the fitness value with the fitness value of the corresponding ant lion, and judging whether the ant lion position is updated or not.
And when the adaptability value of the ant is smaller than that of the ant lion, updating the ant lion position.
The step 5) is specifically as follows:
501) the ant lion position update formula is as follows:in the formula,indicating the position of the ith lion selected at the t-th iteration,represents the position of the ith ant at the t iteration;
502) and according to the F, calculating the fitness values of the ant lions after the positions are updated, comparing the fitness values with the fitness values of the previous generation elite ant lions one by one, and reserving the ant lions with smaller fitness values to obtain the position of the iteration elite ant lions.
Examples
The invention uses the measured data (unit: hundred million kilowatt hours) of the electricity consumption in a certain market to carry out the test. According to the graph 1, a model is trained by using the power consumption and the power consumption influence factor sequence from 1990 to 2009, and the trained model is used for predicting the power consumption from 2010 to 2014 in a test set. The ALO algorithm comprises the following parameters: the population size entries _ no is 30, the dimension d of the variable is 2, the maximum number of iterations Max _ iter is 200, and the upper bound b of the solution spaceup=[1000,1000]Lower bound blow=[0.1,0.01]. The LSSVM optimal parameter obtained according to the ALO algorithm is [1000,1000]. The ALO-LSSVM and the GCA-ALO-LSSVM are used for representing models before and after the GCA is applied to determine the input variable, the predicted values obtained by the two methods are shown in the table 1 and the figure 3, and the ALO-LSSVM is compared with the predicted results after the ALO-LSSVM is determined in the table 1.
TABLE 1
Meanwhile, a prediction model is established by adopting a GCA-Bayes BP neural network, and a comparison graph with the predicted value of the invention is shown in figure 4.
The predicted effects of several methods are shown in table 2, and table 2 shows the predicted effect evaluation of the three prediction methods used.
TABLE 2
According to the evaluation standard of a prediction model provided on page 15 in a book of 'power load prediction technology and application thereof' authored by dawn of cattle, caohua, zhao yue, zhang literary, and the like, in version 1 published by the power publishing society in 10 months in 1998, several prediction methods achieve good prediction effects. However, the specific evaluation standard value shows that the annual power consumption prediction method based on the ant lion optimized LSSVM has higher accuracy rate in the prediction of the annual power consumption, and can meet the requirement.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. The method for predicting the annual power consumption of the LSSVM based on ant lion optimization is characterized by comprising the following steps of:
s1, determining input variables of the LSSVM prediction model;
s2, initializing a ant lion optimization algorithm, substituting the initial ant lion position as a nuclear parameter and a regularization parameter into the LSSVM model, and obtaining a corresponding annual power consumption predicted value;
s3, establishing a fitness function, calculating the fitness value of the position of the initial ant lion to obtain an initial fitness value, and keeping the ant lion corresponding to the minimum initial fitness value as the initial elite ant lion;
s4, updating the ant positions, calculating the fitness value of the current ant, comparing the fitness value with the fitness value of the ant lion position corresponding to the current ant, and judging whether the ant lion positions are updated or not;
s5, comparing the fitness value of the ant lion position obtained in the previous step with the fitness value of the position of the previous generation elite ant lion one by one, and reserving the ant lion position corresponding to a smaller fitness value to obtain the position of the current iteration elite ant lion;
and S6, judging whether the maximum iteration times is reached, if so, outputting the position of the elite lion and the corresponding predicted value of the annual power consumption, and if not, returning to S4.
2. The method for predicting annual power consumption of LSSVM based on ant lion optimization as claimed in claim 1, wherein the step S1 is specifically as follows: and obtaining a correlation value between the annual power consumption influence factors and the annual power consumption by adopting a grey correlation analysis method, and selecting the corresponding annual power consumption influence factors as input variables of the LSSVM prediction model according to the correlation value.
3. The method for forecasting annual power consumption of LSSVM based on ant lion optimization as claimed in claim 1, wherein the step of initializing ant lion optimization algorithm in step S2 comprises the following steps:
s201, initializing parameters including population size Agents _ no, solving dimension d as 2 and solving upper bound b of spaceupLower bound of solution space blowMax iteration number Max _ iter;
s202, initializing positions, and randomly generating a position matrix M of antsAntPosition matrix M of HelloAntlion。
4. The method for forecasting annual power consumption of LSSVM based on ant lion optimization as claimed in claim 3, wherein the upper bound b of the solution space in S201up=[1000,1000]Lower bound blow=[0.1,0.01]。
5. The method for forecasting annual power consumption of LSSVM based on ant lion optimization as claimed in claim 3, wherein the position matrix M of ants in S202AntPosition matrix M of HelloAntlionComprises the following steps:
wherein M isAntAnd MAntlionThe value of (A) is represented by formula A*jOr AL*j=rand*(bupj-blowj)+blowjTo obtain a*jAnd AL*jThe j-th column values of the ant position matrix and the ant lion position matrix are respectively represented, rand is a random number between 0 and 1, and n is Agents _ no, bupjAnd blowjUpper and lower bounds, M, of the jth column, respectivelyAntAnd MAntlionEach row of (a) corresponds to a set of kernel parameters and regularization parameters, i.e., (σ, γ), of the LSSVM.
6. The method for forecasting annual power consumption of LSSVM based on ant lion optimization as claimed in claim 1, wherein the fitness function in step S3 is:wherein,andthe actual value and the predicted value of the s-th group of test samples are respectively, and N is the number of the samples.
7. The method for predicting annual power consumption of LSSVM based on ant lion optimization as claimed in claim 1, wherein the updating formula of ant position in step S4 isWherein,the t-th iteration of roulette around the ant lion is shown to bet the selected ant position,the ant positions of the t iteration are shown as randomly wandering around the elite ant lion,the position of the ith ant at the t iteration.
8. The method for prediction of annual power consumption of LSSVM based on ant lion optimization as claimed in claim 7,wherein, aiRepresents the minimum value of the random walk position of the ith ant, hiRepresents the maximum value of the position of the random walk of the ith ant,represents the minimum value of the actual position of the ith ant at the time of the t iteration,represents the maximum value of the actual position of the ith ant at the time of the t-th iteration,represents the random walk position of the ith ant at the t-th iteration,cumsum is a function for calculating the accumulated value of an array, t represents the current iteration number, and a Max _ iter tableShows the maximum number of iterations, r1(t),…,rMax_iter(t) are random functions and independent of each other.
9. The method for forecasting annual power consumption of LSSVM based on ant lion optimization as claimed in claim 8, wherein the stochastic function calculation formula is:
r (t) is r1(t),…,rMax_iter(t), rand (t) is a random number between 0 and 1.
10. The method for predicting annual power consumption of LSSVM based on ant lion optimization as claimed in claim 1, wherein in step S4, the fitness value of the current ant is compared with the fitness value of the ant lion position corresponding to the current ant, and whether the ant lion position is updated is determined, and if the fitness value of the current ant is smaller than the fitness value of the ant lion, the ant lion position is replaced by the current ant position, otherwise the original ant lion position is retained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710641743.0A CN107274038A (en) | 2017-07-31 | 2017-07-31 | A kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710641743.0A CN107274038A (en) | 2017-07-31 | 2017-07-31 | A kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107274038A true CN107274038A (en) | 2017-10-20 |
Family
ID=60076052
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710641743.0A Pending CN107274038A (en) | 2017-07-31 | 2017-07-31 | A kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107274038A (en) |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107798199A (en) * | 2017-11-09 | 2018-03-13 | 华中科技大学 | A kind of Hydropower Unit parameter closed-loop identification method |
CN108614242A (en) * | 2018-03-25 | 2018-10-02 | 哈尔滨工程大学 | A kind of radar-communication integration waveform design method based on the optimization of multiple target ant lion |
CN109062664A (en) * | 2018-07-25 | 2018-12-21 | 南京邮电大学 | Cloud computing method for scheduling task based on ant lion optimization algorithm |
CN109145464A (en) * | 2018-08-28 | 2019-01-04 | 暨南大学 | Merge the Structural Damage Identification of multiple target ant lion optimization and the sparse regularization of mark |
CN109212465A (en) * | 2018-09-01 | 2019-01-15 | 哈尔滨工程大学 | A kind of particular array dynamic direction-finding method based on cultural ant lion mechanism |
CN109284543A (en) * | 2018-09-04 | 2019-01-29 | 河北工业大学 | IGBT method for predicting residual useful life based on optimal scale Gaussian process model |
CN109767036A (en) * | 2018-12-28 | 2019-05-17 | 北京航空航天大学 | Support vector machines failure prediction method based on the optimization of adaptive ant lion |
CN109829420A (en) * | 2019-01-18 | 2019-05-31 | 湖北工业大学 | A kind of feature selection approach based on the high spectrum image for improving ant lion optimization algorithm |
CN109858816A (en) * | 2019-02-02 | 2019-06-07 | 湖北工业大学 | A method of production scheduling is carried out using ant lion algorithm |
CN110006649A (en) * | 2018-12-24 | 2019-07-12 | 湖南科技大学 | A kind of Method for Bearing Fault Diagnosis based on improvement ant lion algorithm and support vector machines |
CN110487519A (en) * | 2019-06-28 | 2019-11-22 | 暨南大学 | Structural Damage Identification based on ALO-INM and weighting trace norm |
CN111598255A (en) * | 2020-04-27 | 2020-08-28 | 同济大学 | Method for compensating cogging effect of suspension gap sensor of maglev train |
CN111985678A (en) * | 2020-07-06 | 2020-11-24 | 上海交通大学 | Photovoltaic power short-term prediction method |
CN113253709A (en) * | 2021-06-07 | 2021-08-13 | 江苏中车数字科技有限公司 | Health diagnosis method and device suitable for rail transit vehicle |
CN113449462A (en) * | 2021-06-09 | 2021-09-28 | 淮阴工学院 | Ultra-short-term wind speed prediction method and system based on MALO-BiGRU |
CN116205554A (en) * | 2023-04-26 | 2023-06-02 | 浙江天柜科技有限公司 | Mobile self-service vending equipment and vending control method |
CN117233540A (en) * | 2023-11-15 | 2023-12-15 | 广东电网有限责任公司江门供电局 | Metering pipeline fault detection method and system based on deep learning |
-
2017
- 2017-07-31 CN CN201710641743.0A patent/CN107274038A/en active Pending
Cited By (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107798199B (en) * | 2017-11-09 | 2020-07-07 | 华中科技大学 | Hydroelectric generating set parameter closed-loop identification method |
CN107798199A (en) * | 2017-11-09 | 2018-03-13 | 华中科技大学 | A kind of Hydropower Unit parameter closed-loop identification method |
CN108614242A (en) * | 2018-03-25 | 2018-10-02 | 哈尔滨工程大学 | A kind of radar-communication integration waveform design method based on the optimization of multiple target ant lion |
CN108614242B (en) * | 2018-03-25 | 2022-04-05 | 哈尔滨工程大学 | Radar communication integrated waveform design method based on multi-objective ant lion optimization |
CN109062664A (en) * | 2018-07-25 | 2018-12-21 | 南京邮电大学 | Cloud computing method for scheduling task based on ant lion optimization algorithm |
CN109145464A (en) * | 2018-08-28 | 2019-01-04 | 暨南大学 | Merge the Structural Damage Identification of multiple target ant lion optimization and the sparse regularization of mark |
CN109145464B (en) * | 2018-08-28 | 2022-11-01 | 暨南大学 | Structural damage identification method integrating multi-target ant lion optimization and trace sparse regularization |
CN109212465A (en) * | 2018-09-01 | 2019-01-15 | 哈尔滨工程大学 | A kind of particular array dynamic direction-finding method based on cultural ant lion mechanism |
CN109212465B (en) * | 2018-09-01 | 2024-01-30 | 哈尔滨工程大学 | Special array dynamic direction finding method based on cultural ant lion mechanism |
CN109284543A (en) * | 2018-09-04 | 2019-01-29 | 河北工业大学 | IGBT method for predicting residual useful life based on optimal scale Gaussian process model |
CN110006649A (en) * | 2018-12-24 | 2019-07-12 | 湖南科技大学 | A kind of Method for Bearing Fault Diagnosis based on improvement ant lion algorithm and support vector machines |
CN109767036A (en) * | 2018-12-28 | 2019-05-17 | 北京航空航天大学 | Support vector machines failure prediction method based on the optimization of adaptive ant lion |
CN109829420A (en) * | 2019-01-18 | 2019-05-31 | 湖北工业大学 | A kind of feature selection approach based on the high spectrum image for improving ant lion optimization algorithm |
CN109829420B (en) * | 2019-01-18 | 2022-12-02 | 湖北工业大学 | Hyperspectral image feature selection method based on improved ant lion optimization algorithm |
CN109858816A (en) * | 2019-02-02 | 2019-06-07 | 湖北工业大学 | A method of production scheduling is carried out using ant lion algorithm |
CN110487519A (en) * | 2019-06-28 | 2019-11-22 | 暨南大学 | Structural Damage Identification based on ALO-INM and weighting trace norm |
CN111598255A (en) * | 2020-04-27 | 2020-08-28 | 同济大学 | Method for compensating cogging effect of suspension gap sensor of maglev train |
CN111985678A (en) * | 2020-07-06 | 2020-11-24 | 上海交通大学 | Photovoltaic power short-term prediction method |
CN113253709B (en) * | 2021-06-07 | 2021-09-21 | 江苏中车数字科技有限公司 | Health diagnosis method and device suitable for rail transit vehicle |
CN113253709A (en) * | 2021-06-07 | 2021-08-13 | 江苏中车数字科技有限公司 | Health diagnosis method and device suitable for rail transit vehicle |
CN113449462A (en) * | 2021-06-09 | 2021-09-28 | 淮阴工学院 | Ultra-short-term wind speed prediction method and system based on MALO-BiGRU |
CN113449462B (en) * | 2021-06-09 | 2023-06-20 | 淮阴工学院 | Ultra-short-term wind speed prediction method and system based on MALO-BiGRU |
CN116205554A (en) * | 2023-04-26 | 2023-06-02 | 浙江天柜科技有限公司 | Mobile self-service vending equipment and vending control method |
CN116205554B (en) * | 2023-04-26 | 2024-02-09 | 浙江天柜科技有限公司 | Mobile self-service vending equipment and vending control method |
CN117233540A (en) * | 2023-11-15 | 2023-12-15 | 广东电网有限责任公司江门供电局 | Metering pipeline fault detection method and system based on deep learning |
CN117233540B (en) * | 2023-11-15 | 2024-02-20 | 广东电网有限责任公司江门供电局 | Metering pipeline fault detection method and system based on deep learning |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107274038A (en) | A kind of LSSVM Prediction of annual electricity consumption methods optimized based on ant lion | |
Tikhonov et al. | Using joint species distribution models for evaluating how species‐to‐species associations depend on the environmental context | |
Amid et al. | Prediction of output energies for broiler production using linear regression, ANN (MLP, RBF), and ANFIS models | |
Jaquiéry et al. | Inferring landscape effects on dispersal from genetic distances: how far can we go? | |
Welham et al. | A comparison of analysis methods for late‐stage variety evaluation trials | |
Martínez-Frutos et al. | Kriging-based infill sampling criterion for constraint handling in multi-objective optimization | |
Thornton et al. | Body size and spatial scales in avian response to landscapes: a meta‐analysis | |
CN103577694B (en) | Aquaculture water quality short-time combination forecast method on basis of multi-scale analysis | |
CN104765690B (en) | Embedded software test data generation method based on fuzzy genetic algorithm | |
Morales‐Castilla et al. | Combining phylogeny and co‐occurrence to improve single species distribution models | |
CN107222333A (en) | A kind of network node safety situation evaluation method based on BP neural network | |
Xuesong et al. | Research on contaminant sources identification of uncertainty water demand using genetic algorithm | |
Zhang et al. | Measurement of lumber moisture content based on PCA and GS-SVM | |
CN113466710B (en) | SOC and SOH collaborative estimation method for energy storage battery in receiving-end power grid containing new energy | |
Chattopadhyay et al. | Fluctuating fortunes: genomes and habitat reconstructions reveal global climate-mediated changes in bats' genetic diversity | |
McCulloch et al. | Calibrating agent-based models using uncertainty quantification methods | |
CN107122869A (en) | The analysis method and device of Network Situation | |
CN106296434B (en) | Grain yield prediction method based on PSO-LSSVM algorithm | |
Lin et al. | Parameters optimization of GM (1, 1) model based on artificial fish swarm algorithm | |
Li et al. | A hybrid framework for predicting the remaining useful life of battery using Gaussian process regression | |
Tabatabaee et al. | Bayesian approach to updating Markov-based models for predicting pavement performance | |
Chao et al. | Estimation of species richness and shared species richness | |
CN116794547A (en) | Lithium ion battery residual service life prediction method based on AFSA-GRU | |
Horng et al. | Bat algorithm assisted by ordinal optimization for solving discrete probabilistic bicriteria optimization problems | |
CN114881343A (en) | Short-term load prediction method and device of power system based on feature selection |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171020 |