CN108537370A - Especially big basin water station group Optimization Scheduling based on hybrid intelligent dimension-reduction algorithm - Google Patents
Especially big basin water station group Optimization Scheduling based on hybrid intelligent dimension-reduction algorithm Download PDFInfo
- Publication number
- CN108537370A CN108537370A CN201810245918.0A CN201810245918A CN108537370A CN 108537370 A CN108537370 A CN 108537370A CN 201810245918 A CN201810245918 A CN 201810245918A CN 108537370 A CN108537370 A CN 108537370A
- Authority
- CN
- China
- Prior art keywords
- individual
- value
- population
- current
- indicate
- 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.)
- Granted
Links
Classifications
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- 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—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
Abstract
The invention discloses a kind of especially big basin water station group Optimization Scheduling based on integrated intelligent algorithm, selection participate in the power station calculated and corresponding constraints simultaneously are arranged, and encode individual using individual tandem coding method, and generate initial population；The fitness value for assessing each individual carries out mutation operation, then all a body positions in Population Regeneration after more new individual extreme value and global extremum to individual extreme value, then executes mixed search strategy to the individual in external archive set to improve diversity of individuals；It finally repeats the above process until meeting end condition.Optimizing superior performance of the present invention, strong robustness, the dimension calamity problem for fast convergence rate, being easily programmed realization, avoiding conventional scheduling algorithms.Wujiang River Basin application example shows that the method for the present invention effectively increases individual convergence rate and population ability of searching optimum, can be quickly obtained rationally effective GROUP OF HYDROPOWER STATIONS dispatching running way, improves the integrated scheduling benefit of especially big basin water station group scheduling.
Description
Technical field
The invention belongs to high efficient utilization of water resources and GROUP OF HYDROPOWER STATIONS Optimized Operation field, are based on more particularly, to one kind
The especially big basin water station group Optimization Scheduling of hybrid intelligent dimension-reduction algorithm.
Background technology
In recent years, China's hydropower high speed development, more and more power stations are continuously developed utilization, especially in west
The especially big Basin Hydropower bases such as the southern area Wujiang River, Hongsuihe River are gone into operation successively after operation, and China has formed unprecedented in the world
Ultra-large hydroelectric system.With the gradually expansion of system scale, " how especially big basin water station group is carried out scientific and reasonable
Scheduling, with realize hydraulic power potentials it is efficient using and hydroelectric system maximization of economic benefit " become as high efficient utilization of water resources and
The core content in GROUP OF HYDROPOWER STATIONS Optimum Scheduling Technology field.Therefore, there is an urgent need for further investigate suitable for especially big basin water station group
Optimization Scheduling, to promoting, China's efficiency of energy utilization is promoted and energy-saving and emission-reduction career development has highly important reality for this
Meaning and application value.
GROUP OF HYDROPOWER STATIONS Optimized Operation is usually up to object function with generated energy, and mathematical model is described as follows：Known
In schedule periods under the conditions of the reservoir inflow process in each power station and beginning, last water level, the complexity such as water level, output, flow are considered
Constraints is up to target with especially big Basin Hydropower system gross generation, so that it is determined that the output in each power station, water level are run
Process.
In formula：E is the annual electricity generating capacity in power station, kWh；N is the number in power station；I is the serial number in power station；T is dispatching cycle；
T is the serial number of period；Δ t is the hourage of each period, h；Pi,tI-th of power station is represented in the output of t-th of period, kW；
Need the constraints met as follows：
(1) water balance constrains：Vi,t+1=Vi,t+(qi,t-Qi.t-Si,t) Δ t, wherein Vi,tFor t-th of i-th power station
Initial reservoir storage (the m of period3)；qi,t、Qi,t、Si,tRespectively reservoir inflow (the m of t-th of period of i-th of power station3/ s), hair
The magnitude of current (m3/ s), abandon water flow (m3/s)。
(2) pondage constrains：Wherein,Respectively t-th of period of i-th of power station
The upper and lower limit of reservoir storage, m3。
(3) generating flow constrains： To be respectively i-th power station, t-th of period power generation
The upper and lower limit of flow, m3/s。
(4) reservoir letdown flow constrains： When respectively i-th of power station t-th
The letdown flow upper and lower limit of section, m3/s。
(5) power station units limits： The respectively output of t-th of period of i-th of power station
Upper and lower limit, kW.
GROUP OF HYDROPOWER STATIONS Optimal Operation Model makes its solution procedure have there is the problems such as in large scale, constraints is complicated
There are the features such as various dimensions, multistage, non-linear, strong constraint, the result and actual schedule that are obtained using conventional linear programming evaluation
Process variations are larger；And Dynamic Programming can encounter " dimension calamity " problem.Therefore, researching and developing appropriate method for solving seems particularly heavy
It wants.Swarm Intelligence Algorithm carries out optimizing using good coevolution mechanism, avoids the enumeration operation of traditional algorithm, has memory
Occupy less, dimensionality reduction effect protrude etc. advantages, be gradually used widely in GROUP OF HYDROPOWER STATIONS Optimized Operation field.
Sun etc. is after having extensively studied Swarm Evolution process, by quantum-mechanical inspiration, it is proposed that one kind is searched with the overall situation
The quantum groups of individuals algorithm of Suo Nengli.In QPSO algorithms, individual speed and position cannot accurately measure, and need to use
Wave function obtains the probability density letter that individual occurs in a certain position to describe individual state by solving Schrodinger equation
Number finally obtains the position equation of individual in space in the way of Monte-Carlo stochastic simulations.In individual evolution process
In, by optimal location center, certain attraction potential is fettered individual poly- scattered property, is realized by track individual extreme value and global extremum
The continuous renewal of a body position so that the individual in bound state can appear in the arbitrary of search space with certain probability
Position, to effectively increase the ability of searching optimum of individual.
Standard QPSO algorithms have many advantages, such as that calculating is simple, control parameter is few, is easily programmed, and cause the concern of scholars
And research, also some sharing of loads, the network optimization the problems such as in be widely applied.However, in the especially big of solving complexity
When basin water station group scheduling problem, standard QPSO still has diversity of individuals deficiency, easy Premature Convergence, is absorbed in local optimum
The problems such as.
Invention content
For the disadvantages described above or Improvement requirement of the prior art, the present invention provides one kind being based on hybrid intelligent dimension-reduction algorithm
Especially big basin water station group Optimization Scheduling, thus solve the especially big basin water station group scheduling problem in solving complexity
When, standard QPSO still has diversity of individuals deficiency, easy Premature Convergence, is absorbed in the problems such as local optimum.
To achieve the above object, the present invention provides a kind of especially big basin water station group based on hybrid intelligent dimension-reduction algorithm
Optimization Scheduling, including：
(1) each power station of coding is sequentially connected in series in the water level of different periods by the sequence in the power station for participating in calculating, obtain
Single individual UVR exposure value, and initial population is generated according to single individual UVR exposure value at random in preset feasible water level range, it will
Initial population is as current population；
(2) for any of current population individual, if the fitness value of the individual is less than its history adaptive optimal control degree
Value, then the individual extreme value of the individual remains unchanged, and otherwise, the individual pole of the individual is replaced with the current location residing for the individual
Value, and from picking out the maximum value of individual extreme value as global extremum in the individual extreme value of all individuals in current population, wherein
The desired positions that the individual extrema representation individual is undergone, global extremum indicate the best position of all Individual Experiences in current population
It sets；
(3) for all individual extreme values in current population, two different individual poles are randomly choosed from current population
Be worth and subtract each other generation differential vector, by the differential vector according to preset ratio be superimposed to global extremum using generate variation vector as
New individual extreme value, if the individual extreme value fitness value after variation is used new better than the fitness value of the individual extreme value before variation
Individual extreme value replace the individual individual extreme value, otherwise the individual extreme value of the individual remain unchanged；
(4) it by the individual extreme value of each individual in the global extremum of current population and current population, updates in current population
The current location of each individual；
(5) if δ >=Pa, then move out at random from current population several individual constitute external archive collection, wherein δ be [0,
1] random number of section random distribution,K indicates current iteration number,Indicate maximum iteration；
(6) negative value of the former fitness value of all individuals of external archive concentration is used to concentrate each individual as external archive
Target fitness value, and concentrate according to external archive the corresponding individual of maximum target fitness value of all individuals, secondary big mesh
The corresponding individual of fitness value and the corresponding individual of minimum target fitness value are marked, it is corresponding to maximum target fitness value a
The mapping point of body carries out expansion or shrinkage operation, and redefines the corresponding individual of maximum target fitness value, and secondary big target is suitable
The corresponding individual of angle value and the corresponding individual of minimum target fitness value are answered, to corresponding new of maximum target fitness value
The mapping point of body carries out expansion or scaling operation, until meeting default execution number, merges external archive collection and current population, from
Several individuals with preferable fitness replace the individual in current population before being chosen in population after merging, next to be formed
For population；
(7) increase population iterations, if current population iterations are less than maximum iteration, by next-generation population
As current population, and (2) are returned to step, otherwise, individual by the global optimum of the current population of last time iteration
To each power station different periods optimal scheduling process.
Preferably, step (1) includes：
(1.1) each power station of coding is sequentially connected in series in the water level of different periods by the sequence in the power station for participating in calculating, obtain
To single individual UVR exposure value, wherein single individual UVR exposure value is expressed as： Indicate n-th of electricity
It standing in the water position status of j-th of period, N is power station number, n=1,2 ..., N, and T is the fixed number in dispatching cycle, j=1,
2,…,T；
(1.2) initial value of setting k is 1, and byIt is random in preset feasible water level range
Generate initial population Uk, wherein UkKth is indicated for population,Indicate kth for population UkIn i-th body position, i=1,
2 ..., m, r are the random number of [0,1] section distribution,X is respectively the upper and lower limit of relevant variable, and m is population scale.
Preferably, step (2) includes：
(2.1) individual for any of current population, byCalculate the suitable of each individual
Answer angle value, whereinIndicate individualCorresponding fitness value, G are constraint number, Pn,jFor n-th of power station when
The output of section j, tjFor scheduling slot hourage, XXk,gIndicate particleThe corresponding value of g-th of constraint in gained scheduling process,
λgFor the destruction penalty coefficient of g-th of constraints,Respectively XXk,qUpper and lower limit；
(2.2) byMore new individual extreme value and global extremum, whereinTable
Show individual extreme value of the kth for individual i in population,Indicate the individual extreme value of individual i in kth -1 generation population, GBkIndicate kth
For the global extremum of population,Indicate the fitness value of the individual extreme value of individual i in kth -1 generation population.
Preferably, step (3) includes：
ByTo of any of current population individual
Body extreme value executes mutation operation, whereinIndicate the new individual extreme value generated after mutation operation,For [0,1] section point
The random number matched, Ind1 and Ind2 indicates the integer selected from set { 1,2 ..., m } at random respectively, and has Ind1 ≠ Ind2.
Preferably, step (4) includes：
ByUpdate in current population it is each individual it is current
Position, wherein k is current iteration number,The current location of individual i, mbest when iteration secondary for kthkWhen iteration secondary for kth
The optimal location center of population,When indicating kth time iteration betweenAnd GBkBetween position, v, u between [0,1]
The random number of even distribution, βkIndicate converging diverging coefficient when kth time iteration.
Preferably, step (6) includes：
(6.1) J=1 is enabled, by f (X)=- F (X) external archive collection SkIn each individual former fitness value negative value
The target fitness value of each individual is concentrated as external archive, wherein SkMiddle individual number is D+1, D=N × T；
(6.2) S is determinedkIn maximum target fitness value f (Xhigh) corresponding to individual Xhigh, secondary big target fitness
Value f (Xsec) corresponding to individual Xsec, minimum target fitness value f (Xlow) corresponding to individual Xlow, and calculate SkIn remove Xhigh
The mean place X of outer all individualscenter, and byCalculate XhighMapping point Xr, α is
Mapping coefficient, if f (Xlow)≤f(Xr)≤f(Xsec), then Xhigh=XrAnd step (6.5) is executed, if f (Xr) ＜ f (Xlow), then
Step (6.3) is executed, if f (Xr) ＞ f (Xsec), it thens follow the steps (6.4)；
(6.3) by Xe=Xcenter+β(Xr-Xcenter) to mapping point XrIt carries out dilation procedure and obtains inflexion point Xe, β is expansion
Coefficient, if f (Xe)≤f(Xlow), then enable Xhigh=XeAnd it executes step (6.5) and otherwise enables Xhigh=XrAnd execute step
(6.5)；
(6.4) if f (Xr) ＞ f (Xsec) and f (Xr)≤f(Xhigh), enable Xhigh=Xr, redefined according to step (6.2)
Xr, then by Xc=Xcenter+γ(Xhigh-Xcenter) carry out shrinkage operation obtain constriction point XcIf f (Xr) ＞ f (Xhigh), then directly
It connects by Xc=Xcenter+γ(Xhigh-Xcenter) carry out shrinkage operation obtain constriction point Xc, γ is constriction coefficient；
If f (Xc)≤f(Xhigh), then enable Xhigh=XcAnd execute step (6.5)；
(6.5) value of J is increased by 1, if(6.6) are thened follow the steps, (6.2) are otherwise returned to step,For
Preset maximum execution algebraically；
(6.6) merge external archive collection SkWith current population Uk, m is a with preferably suitable before being chosen from the population after merging
The individual of response replaces current population UkIn individual.
In general, through the invention it is contemplated above technical scheme is compared with the prior art, can obtain down and show
Beneficial effect：
Using effective individual tandem coding method and Complex Constraints processing strategy, electricity can be expressed visual and clearly
Stage space topological relation is conducive to improve computational efficiency；Establish external elite setIt is excellent during evolutionary process is found to store
Elegant individual effectively increases Evolution of Population direction, improves algorithm the convergence speed；Mutation search operation is carried out to individual extreme value, is shown
Work improves diversity of individuals, strengthens population overall situation producing capacity；Referred to using dynamic probability identification mechanism, mixed search strategy
It leads population and carries out neighborhood optimizing, algorithm exploration ability greatly improved, avoid being absorbed in local optimum；Machine is cooperateed with using swarm intelligence
System scans for, and effectively prevents the dimension calamity problem of conventional combination optimization algorithm, and EMS memory occupation is less, has good dimensionality reduction
Effect.The present invention realizes a kind of especially big basin water station group Optimization Scheduling based on hybrid intelligent dimension-reduction algorithm, has
Strong robustness, fast convergence rate, calculating is simple, is easily programmed realization and is combined the advantages that feasible effective with scheduling problem, shows
Write the integrated scheduling benefit for improving GROUP OF HYDROPOWER STATIONS.
Description of the drawings
Fig. 1 is a kind of flow signal solving especially big basin water station group Optimization Scheduling provided in an embodiment of the present invention
Figure；
Fig. 2 is calculated using the method for the present invention and other optimizations under the conditions of a kind of normal flow year water provided in an embodiment of the present invention
The comparison diagram of method convergence process；
Fig. 3 (a) is the Hong Jiadu that the method for the present invention is used under the conditions of a kind of normal flow year water provided in an embodiment of the present invention
Power station schematic diagram of calculation result；
Fig. 3 (b) is the east wind electricity using the method for the present invention under the conditions of a kind of normal flow year water provided in an embodiment of the present invention
It stands schematic diagram of calculation result；
Fig. 3 (c) is the Suofengying that the method for the present invention is used under the conditions of a kind of normal flow year water provided in an embodiment of the present invention
Power station schematic diagram of calculation result；
Fig. 3 (d) is the Wu Jiangdu that the method for the present invention is used under the conditions of a kind of normal flow year water provided in an embodiment of the present invention
Power station schematic diagram of calculation result；
Fig. 3 (e) is the Goupitan that the method for the present invention is used under the conditions of a kind of normal flow year water provided in an embodiment of the present invention
Power station schematic diagram of calculation result.
Specific implementation mode
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.As long as in addition, technical characteristic involved in the various embodiments of the present invention described below
It does not constitute a conflict with each other and can be combined with each other.
The present invention proposes a kind of based on mixing to overcome existing QPSO existing deficiencies when solving water power scheduling problem
The especially big basin water station group Optimization Scheduling of intelligent dimension-reduction algorithm.The invention is innovatively drawn on the basis of standard QPSO
Enter dynamic probability identification mechanism, mixed search strategy etc. to improve mechanism, to accelerate individual convergence rate and kind group hunting energy conscientiously
Power, to improve GROUP OF HYDROPOWER STATIONS integrated scheduling benefit.
It is a kind of stream solving especially big basin water station group Optimization Scheduling provided in an embodiment of the present invention as shown in Figure 1
Journey schematic diagram, including：
(1) selection participates in the power station calculated, and the corresponding constraints in each power station is arranged, including water level limits, goes out
Power limit, flow restriction, interval inflow situation and initial water level, end of term water level；
(2) it determines individual UVR exposure mode, individual tandem coding method may be used in embodiments of the present invention, it is suitable by power station
Sequence is sequentially connected in series the water level for encoding each power station in different periods, and equivalent water level is characterized using decimal floating point type data, then
Decision variable number D=N × T, single individual X are as follows：
Wherein,Indicate n-th of power station j-th of period water position status；N is power station number, n=1,2 ..., N；T is
Fixed number in dispatching cycle, j=1,2 ..., T；
(3) correlation computations parameter, including population scale m, maximum iteration are setAnd in mixed search strategy
Maximum executes algebraicallyMapping coefficient α, flare factor β, constriction coefficient γ；
(4) counter k=1 is set, then generates initial population U at random in the feasible water level range of settingk, wherein Uk
Kth is indicated for population, then single individual calculates as follows：
In formula,Indicate i-th body position in kth generation, i=1,2 ..., m；R is the random of [0,1] section distribution
Number, with the difference of r, the location of each individual is also different； XThe respectively upper and lower limit of relevant variable；
(5) in embodiments of the present invention, Means of Penalty Function Methods can be utilized to calculate the fitness value of each individual, calculating formula is such as
Under；
In formula：Indicate individualCorresponding fitness value；G is constraint number, g=1,2 ..., G；Pn,jIt is
N power station is in the output of period j, kW；tjFor scheduling slot hourage, h；XXk,gIndicate particleIn gained scheduling process
The corresponding value of g constraint；λgFor the destruction penalty coefficient of g-th of constraints；Respectively XXk,qIt is upper,
Lower limit；
(6) more new individual extreme value and global extremum：It is optimal suitable with itself history after the fitness of each individual is calculated
Response is made comparisons, if being less than itself history adaptive optimal control degree, history is optimal to be remained unchanged；Otherwise, most instead of the history of individual
It is excellent；Then from all individual history it is optimal in pick out the maximum individual of fitness as global optimum's individual, it is corresponding to express
Formula is as follows：
In formula,Indicate the desired positions that kth generation individual i is undergone, i.e., individual extreme value；GBkIndicate kth for all
The desired positions of body experience, i.e. global extremum；
(7) direction that preferable individual extreme value can scientifically guide individual to fly is considered, therefore in the embodiment of the present invention
In, individual extreme value Mutation Strategy is introduced to improve individual to the close speed of optimal solution, and concrete mode is：Two are randomly choosed first
A different individual extreme value simultaneously subtracts each other generation differential vector；Then this differential vector is superimposed to global extremum according to a certain percentage
To generate variation vector；Finally check whether the Individual Quality after variation is improved, if its fitness is better than individual extreme value,
Individual extreme value is directly replaced, is otherwise not processed.Corresponding expression formula is as follows：
In formula：Indicate the new individual extreme value generated after mutation operation；For [0,1] section distribution random number,
Ind1 and Ind2 indicates the integer selected from set { 1,2 ..., m } at random respectively, and has Ind1 ≠ Ind2；
(8) current location individual in Population Regeneration, specific formula are as follows：
In formula：M is population scale；K is current iteration number；For maximum iteration, It is
The position of individual i when k iteration；The history optimal location of individual i when iteration secondary for kth；GBkPopulation when iteration secondary for kth
Optimal location；mbestkThe optimal location center of population when iteration secondary for kth；When indicating kth time iteration betweenAnd GBk
Between position；v,u,r1The equally distributed random number between [0,1]；βkIndicate converging diverging coefficient when kth time iteration；
βs,βeThe respectively initial value and stop value of compressibility factor, it is preferable that take βs=1.0, βe=0.5；
(9) judge the condition that mixed search strategy starts：Mixed search strategy be conducive to assist algorithm carry out local exploration,
And then the defect of annual reporting law Premature Convergence is solved, therefore introduce dynamic probability identifier and be made as mixed search strategy entry condition, it moves
State probability PaCalculation is as follows：
Then, determine whether to meet mixed search strategy entry condition：If [0,1] the random number δ of section random distribution >=
Pa, then from current population Uk(D+1) individual of moving out at random constitutes external archive collection Sk, then go to step (10)；Otherwise, directly
It connects and enters step (11)；
(10) use mixed search strategy to external archive set SkIn individual carry out secondary optimization, the specific steps are：
A) J=1 is enabled, then to SkThe fitness value of all individuals takes negative value, enables f (X)=- F (X)；
B) map operation：Compare SkThe fitness value of all individuals then determines maximum adaptation angle value f (Xhigh) corresponding
Individual Xhigh, secondary big fitness value f (Xsec) corresponding to individual Xsec, minimum fitness value f (Xlow) corresponding to individual
Xlow, then calculate SkIn remove XhighThe mean place X of outer all individualscenter, and calculate XhighMapping point Xr, corresponding to calculate
Formula is as follows：
If f (Xlow)≤f(Xr)≤f(Xsec), then Xhigh=XrAnd it enters step e)；
If f (Xr) ＜ f (Xlow), then it enters step c)；
If f (Xr) ＞ f (Xsec), then it enters step d)；
C) dilation procedure：To mapping point XrIt carries out dilation procedure and obtains inflexion point Xe, calculation formula is as follows：
Xe=Xcenter+β(Xr-Xcenter)
If at this point, f (Xe)≤f(Xlow), then enable Xhigh=XeAnd it enters step e)；Otherwise, X is enabledhigh=XrAnd it enters step
e)；
D) shrinkage operation：As f (Xr) ＞ f (Xsec) and f (Xr)≤f(Xhigh) when, X is enabled firsthigh=Xr, according to step b)
Redefine Xr, then carry out shrinkage operation and obtain constriction point Xc；As f (Xr) ＞ f (Xhigh), then it directly carries out shrinkage operation and obtains
To constriction point Xc.Shrinkage operation calculation formula is as follows：
Xc=Xcenter+γ(Xhigh-Xcenter)
If at this point, f (Xc)≤f(Xhigh), then enable Xhigh=XcAnd it enters step e)；
E) J=J+1 is enabled, ifSatisfaction then goes to (f)；Otherwise (b) is gone to；
F) merge external archive collection SkWith current population Uk, therefrom choose the m individuals with preferable fitness and replace population
UkIn individual；
(11) k=k+1 is enabled, ifReturn to step (5), otherwise enters step (12)；
(12) stop calculating, export the global optimum individual GB of current populationk, you can each power station is obtained in different periods
Optimal scheduling process.
The invention will be further described with reference to the accompanying drawings and examples.
The validity and reasonability of the method for the present invention are now verified by taking Wujiang River Basin Optimal operation of cascade hydropower stations as an example.
M=500, D=55 in example of calculation of the present invention,α=1, β=2, γ=0.5, constraint destroy penalty coefficient
It is taken as 1000.
Table 1 is that DE, QPSO, IQPSO (the method for the present invention) is respectively adopted, according to the wujiangdu hydropower station group typical low flow year,
It is calculating 30 that the result of calculation that normal flow year, the water situation of high flow year obtain, wherein DE, QPSO, IQPSO, which optimize result of calculation,
Secondary obtained optimal value.Table 2 is that optimal value, worst-case value that DE, QPSO, IQPSO random walk obtains for 30 times, is respectively adopted
Value, standard deviation and very poor statistical conditions.By Tables 1 and 2 it is found that from electricity, knots of the IQPSO under different representative conditions
Fruit is all better than DE and QPSO；From the random standard deviation for calculating 30 times and it is very poor from the point of view of, IQPSO calculate 30 obtained standard deviations and
The very poor result that smaller than QPSO is calculated.It follows that IQPSO algorithms are when solving GROUP OF HYDROPOWER STATIONS Optimal Scheduling,
Optimizing superior performance may search for obtaining and stablize efficient scheduling process.
Table 1
Table 2
Fig. 2 is the convergence curve comparison diagram of each method when choosing normal flow year.As seen from the figure, IQPSO algorithms ratio QPSO algorithms
Convergence is fast, while restraining result and being better than QPSO, illustrates that the method for the present invention can increase the diversity of population, accelerates individual to most
The excellent search capability for solving convergent speed and enhancing population, to improve the integrated scheduling benefit of GROUP OF HYDROPOWER STATIONS.
Fig. 3 (a) to Fig. 3 (e) is the output and water of each power station for being obtained using the method for the present invention in normal flow year different periods
Position process schematic.As seen from the figure, the method for the present invention can obtain rationally effective GROUP OF HYDROPOWER STATIONS dispatching running way.Big vast family is crossed
When flood season, water was larger, water level is raised in power station, while being contributed and being run for the increase of reducing abandoned water power station, east wind, Suofengying
Hydropower station during flood period is kept substantially stable water storage, and dry season can also run compared with high water head, can reduce water consume in this way, be conducive to increase
Add generated energy.It follows that the method for the present invention can obtain the scheduling process of reasonable.
In conclusion the present invention have strong robustness, fast convergence rate, calculate it is simple, be easily programmed realization and with tune
Degree problem combines the advantages that feasible effective, to provide new approaches to solve especially big basin water station group Optimal Scheduling.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, not to
The limitation present invention, all within the spirits and principles of the present invention made by all any modification, equivalent and improvement etc., should all include
Within protection scope of the present invention.
Claims (6)
1. a kind of especially big basin water station group Optimization Scheduling based on hybrid intelligent dimension-reduction algorithm, which is characterized in that including：
(1) it is sequentially connected in series the water level for encoding each power station in different periods by the sequence in the power station for participating in calculating, obtains single
Individual UVR exposure value, and initial population is generated according to single individual UVR exposure value at random in preset feasible water level range, it will be initial
Population is as current population；
(2) for any of current population individual, if the fitness value of the individual is less than its history adaptive optimal control angle value,
The individual extreme value of the individual remains unchanged, and otherwise, replaces the individual extreme value of the individual with the current location residing for the individual, and from
The maximum value of individual extreme value is picked out as global extremum in the individual extreme value of all individuals in current population, wherein individual pole
Value indicates that the desired positions that the individual is undergone, global extremum indicate the desired positions of all Individual Experiences in current population；
(3) for all individual extreme values in current population, two different individual extreme values are randomly choosed simultaneously from current population
Subtract each other generation differential vector, by the differential vector according to preset ratio be superimposed to global extremum using generate variation vector as newly
Individual extreme value, if the individual extreme value fitness value after variation uses new better than the fitness value of the individual extreme value before variation
Body extreme value replaces the individual extreme value of the individual, and otherwise the individual extreme value of the individual remains unchanged；
(4) it by the individual extreme value of each individual in the global extremum of current population and current population, updates each in current population
The current location of body；
(5) if δ >=Pa, then several individual composition external archive collection of moving out at random from current population, wherein δ is [0,1] area
Between random distribution random number,K indicates current iteration number,Indicate maximum iteration；
(6) target for using external archive that the negative value of the former fitness value of each individual is concentrated to concentrate each individual as external archive
Fitness value, and the corresponding individual of maximum target fitness value of all individuals is concentrated according to external archive, secondary big target adapts to
The corresponding individual of angle value and the corresponding individual of minimum target fitness value, reflect the corresponding individual of maximum target fitness value
Exit point carries out expansion or shrinkage operation, and redefines the corresponding individual of maximum target fitness value, secondary big target fitness value
Corresponding individual and the corresponding individual of minimum target fitness value, to reflecting for the new corresponding individual of maximum target fitness value
Exit point carries out expansion or scaling operation, until meeting default execution number, merges external archive collection and current population, after merging
Population in choose before several individuals with preferable fitness replace the individual in current population, to form next-generation kind
Group；
(7) increase population iterations, if current population iterations are less than maximum iteration, using next-generation population as
Current population, and (2) are returned to step, otherwise, obtained respectively by global optimum's individual of the current population of last time iteration
Optimal scheduling process of the power station in different periods.
2. according to the method described in claim 1, it is characterized in that, step (1) includes：
(1.1) each power station of coding is sequentially connected in series in the water level of different periods by the sequence in the power station for participating in calculating, obtain list
Individual encoded radio, wherein single individual UVR exposure value is expressed as： Indicate n-th of electricity
It standing in the water position status of j-th of period, N is power station number, n=1,2 ..., N, and T is the fixed number in dispatching cycle, j=1,
2,…,T；
(1.2) initial value of setting k is 1, and byIt is generated at random in preset feasible water level range
Initial population Uk, wherein UkKth is indicated for population,Indicate kth for population UkIn i-th body position, i=1,2 ...,
M, r are the random number of [0,1] section distribution, XThe respectively upper and lower limit of relevant variable, m are population scale.
3. according to the method described in claim 2, it is characterized in that, step (2) includes：
(2.1) individual for any of current population, byCalculate the suitable of each individual
Answer angle value, whereinIndicate individualCorresponding fitness value, G are constraint number, Pn,jFor n-th of power station when
The output of section j, tjFor scheduling slot hourage, XXk,gIndicate particleThe corresponding value of g-th of constraint in gained scheduling process,
λgFor the destruction penalty coefficient of g-th of constraints,Respectively XXk,qUpper and lower limit；
(2.2) byMore new individual extreme value and global extremum, whereinIndicate the
K for individual i in population individual extreme value,Indicate the individual extreme value of individual i in kth -1 generation population, GBkIndicate kth generation kind
The global extremum of group,Indicate individual in kth -1 generation populationiIndividual extreme value fitness value.
4. according to the method described in claim 3, it is characterized in that, step (3) includes：
ByTo the individual pole of any of current population individual
Value executes mutation operation, whereinIndicate the new individual extreme value generated after mutation operation,For the distribution of [0,1] section
Random number, Ind1 and Ind2 indicates the integer selected from set { 1,2 ..., m } at random respectively, and has Ind1 ≠ Ind2.
5. according to the method described in claim 4, it is characterized in that, step (4) includes：
ByThe current location of each individual in current population is updated,
Wherein, k is current iteration number,The current location of individual i, mbest when iteration secondary for kthkPopulation when iteration secondary for kth
Optimal location center,When indicating kth time iteration betweenAnd GBkBetween position, v, u are uniformly distributed between [0,1]
Random number, βkIndicate converging diverging coefficient when kth time iteration.
6. according to the method described in claim 5, it is characterized in that, step (6) includes：
(6.1) J=1 is enabled, by f (X)=- F (X) external archive collection SkIn each individual former fitness value negative value conduct
External archive concentrates the target fitness value of each individual, wherein SkMiddle individual number is D+1, D=N × T；
(6.2) S is determinedkIn maximum target fitness value f (Xhigh) corresponding to individual Xhigh, secondary big target fitness value f
(Xsec) corresponding to individual Xsec, minimum target fitness value f (Xlow) corresponding to individual Xlow, and calculate SkIn remove XhighOutside
The mean place X of all individualscenter, and byCalculate XhighMapping point Xr, α is to reflect
Coefficient is penetrated, if f (Xlow)≤f(Xr)≤f(Xsec), then Xhigh=XrAnd step (6.5) is executed, if f (Xr) ＜ f (Xlow), then it holds
Row step (6.3), if f (Xr) ＞ f (Xsec), it thens follow the steps (6.4)；
(6.3) by Xe=Xcenter+β(Xr-Xcenter) to mapping point XrIt carries out dilation procedure and obtains inflexion point Xe, β is flare factor,
If f (Xe)≤f(Xlow), then enable Xhigh=XeAnd it executes step (6.5) and otherwise enables Xhigh=XrAnd execute step (6.5)；
(6.4) if f (Xr) ＞ f (Xsec) and f (Xr)≤f(Xhigh), enable Xhigh=Xr, X is redefined according to step (6.2)r, so
Afterwards by Xc=Xcenter+γ(Xhigh-Xcenter) carry out shrinkage operation obtain constriction point XcIf f (Xr) ＞ f (Xhigh), then directly by
Xc=Xcenter+γ(Xhigh-Xcenter) carry out shrinkage operation obtain constriction point Xc, γ is constriction coefficient；
If f (Xc)≤f(Xhigh), then enable Xhigh=XcAnd execute step (6.5)；
(6.5) value of J is increased by 1, if(6.6) are thened follow the steps, (6.2) are otherwise returned to step,It is preset
Maximum executes algebraically；
(6.6) merge external archive collection SkWith current population Uk, m have preferable fitness before being chosen from the population after merging
Individual replace current population UkIn individual.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810245918.0A CN108537370B (en) | 2018-03-23 | 2018-03-23 | Super-large basin hydropower station group optimal scheduling method based on hybrid intelligent dimensionality reduction algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810245918.0A CN108537370B (en) | 2018-03-23 | 2018-03-23 | Super-large basin hydropower station group optimal scheduling method based on hybrid intelligent dimensionality reduction algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108537370A true CN108537370A (en) | 2018-09-14 |
| CN108537370B CN108537370B (en) | 2021-07-27 |
Family
ID=63483733
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810245918.0A Active CN108537370B (en) | 2018-03-23 | 2018-03-23 | Super-large basin hydropower station group optimal scheduling method based on hybrid intelligent dimensionality reduction algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108537370B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109635999A (en) * | 2018-11-06 | 2019-04-16 | 华中科技大学 | A kind of power station dispatching method looked for food based on population-bacterium and system |
| CN109636043A (en) * | 2018-12-16 | 2019-04-16 | 华中科技大学 | A kind of Hydro Power Systems with Cascaded Reservoirs power generation dispatching adaptive optimization method and system |
| CN110222938A (en) * | 2019-05-10 | 2019-09-10 | 华中科技大学 | A kind of Hydropower Stations head relation cooperative optimization method and system |
| CN111915162A (en) * | 2020-07-15 | 2020-11-10 | 华中科技大学 | Intelligent matching layering coupling method and system for cascade reservoir group joint scheduling |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103971174A (en) * | 2014-05-06 | 2014-08-06 | 大连理工大学 | Hydropower station group optimized dispatching method based on improved quantum-behaved particle swarm algorithm |
| CN104091207A (en) * | 2014-06-19 | 2014-10-08 | 南方电网科学研究院有限责任公司 | Wind power plant included multiple-target unit commitment optimization method considering harmful gas discharge amount |
| CN104182806A (en) * | 2014-08-21 | 2014-12-03 | 大连理工大学 | Optimal operation method of hydropower station group on the basis of orthogonal dimensionality reduction search algorithm |
| CN106203689A (en) * | 2016-07-04 | 2016-12-07 | 大连理工大学 | A kind of Hydropower Stations cooperation Multiobjective Optimal Operation method |
| CN106355292A (en) * | 2016-09-21 | 2017-01-25 | 广东工业大学 | Method and system for optimally dispatching cascade reservoirs on basis of quantum-behaved particle swarm algorithms |
| CN107798379A (en) * | 2017-11-23 | 2018-03-13 | 东北大学 | Improve the method for quantum particle swarm optimization and the application based on innovatory algorithm |
-
2018
- 2018-03-23 CN CN201810245918.0A patent/CN108537370B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103971174A (en) * | 2014-05-06 | 2014-08-06 | 大连理工大学 | Hydropower station group optimized dispatching method based on improved quantum-behaved particle swarm algorithm |
| CN104091207A (en) * | 2014-06-19 | 2014-10-08 | 南方电网科学研究院有限责任公司 | Wind power plant included multiple-target unit commitment optimization method considering harmful gas discharge amount |
| CN104182806A (en) * | 2014-08-21 | 2014-12-03 | 大连理工大学 | Optimal operation method of hydropower station group on the basis of orthogonal dimensionality reduction search algorithm |
| CN106203689A (en) * | 2016-07-04 | 2016-12-07 | 大连理工大学 | A kind of Hydropower Stations cooperation Multiobjective Optimal Operation method |
| CN106355292A (en) * | 2016-09-21 | 2017-01-25 | 广东工业大学 | Method and system for optimally dispatching cascade reservoirs on basis of quantum-behaved particle swarm algorithms |
| CN107798379A (en) * | 2017-11-23 | 2018-03-13 | 东北大学 | Improve the method for quantum particle swarm optimization and the application based on innovatory algorithm |
Non-Patent Citations (3)
| Title |
|---|
| 冯仲恺等: "改进量子粒子群算法在水电站群优化调度中的应用", 《水科学进展》 * |
| 牛文静等: "梯级水电站群优化调度多目标量子粒子群算法", 《水力发电学报》 * |
| 陈功贵等: "基于改进量子粒子群算法的电力系统经济调度仿真研究", 《实验技术与管理》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109635999A (en) * | 2018-11-06 | 2019-04-16 | 华中科技大学 | A kind of power station dispatching method looked for food based on population-bacterium and system |
| CN109635999B (en) * | 2018-11-06 | 2023-06-20 | 华中科技大学 | Hydropower station scheduling method and system based on particle swarm-bacterial foraging |
| CN109636043A (en) * | 2018-12-16 | 2019-04-16 | 华中科技大学 | A kind of Hydro Power Systems with Cascaded Reservoirs power generation dispatching adaptive optimization method and system |
| CN109636043B (en) * | 2018-12-16 | 2020-10-30 | 华中科技大学 | Adaptive optimization method and system for power generation dispatching of cascade hydropower system |
| CN110222938A (en) * | 2019-05-10 | 2019-09-10 | 华中科技大学 | A kind of Hydropower Stations head relation cooperative optimization method and system |
| CN110222938B (en) * | 2019-05-10 | 2021-03-26 | 华中科技大学 | Short-term peak-load regulation scheduling collaborative optimization method and system for cascade hydropower station group |
| CN111915162A (en) * | 2020-07-15 | 2020-11-10 | 华中科技大学 | Intelligent matching layering coupling method and system for cascade reservoir group joint scheduling |
| CN111915162B (en) * | 2020-07-15 | 2022-05-31 | 华中科技大学 | Cascade reservoir group joint scheduling intelligent matching layered coupling method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108537370B (en) | 2021-07-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022193642A1 (en) | Reservoir scheduling multi-objective optimization method based on graph convolutional network and nsga-ii | |
| CN105809279B (en) | A kind of water resources optimal operation method based on multi-target quantum shuffled frog leaping algorithm | |
| CN108537370A (en) | Especially big basin water station group Optimization Scheduling based on hybrid intelligent dimension-reduction algorithm | |
| CN109670650B (en) | Multi-objective optimization algorithm-based solving method for cascade reservoir group scheduling model | |
| CN109636043B (en) | Adaptive optimization method and system for power generation dispatching of cascade hydropower system | |
| CN107527119A (en) | Water resources optimal operation method based on improved multi-target quantum genetic algorithm | |
| CN106651628B (en) | Regional cooling, heating and power comprehensive energy optimal allocation method and device based on graph theory | |
| CN105719091B (en) | A kind of parallel Multiobjective Optimal Operation method of Hydropower Stations | |
| CN105243458A (en) | Reservoir dispatching method based on multi-target shuffled frog leaping and differential algorithms | |
| CN106251001A (en) | A kind of based on the photovoltaic power Forecasting Methodology improving fuzzy clustering algorithm | |
| CN103971174A (en) | Hydropower station group optimized dispatching method based on improved quantum-behaved particle swarm algorithm | |
| CN111079983A (en) | Optimization method for vehicle path planning of assembly type construction site | |
| CN105447509A (en) | Short-term power prediction method for photovoltaic power generation system | |
| CN109886468A (en) | Charging station planing method based on improved self-adapted genetic algorithm | |
| CN104182806B (en) | A kind of GROUP OF HYDROPOWER STATIONS Optimization Scheduling based on orthogonal dimensionality reduction searching algorithm | |
| CN107609683A (en) | A kind of Cascade Reservoirs method for optimizing scheduling based on glowworm swarm algorithm | |
| CN112036633A (en) | Construction method and optimized scheduling method of reservoir ecological power generation multi-target medium-long term random scheduling model | |
| CN106779219A (en) | A kind of electricity demand forecasting method and system | |
| CN109492813A (en) | Short-term natural gas load combination forecasting method based on improved adaptive GA-IAGA | |
| CN110163420B (en) | Multi-target ecological scheduling method and system based on decomposition culture evolution algorithm | |
| Sheng et al. | An optimized prediction algorithm based on XGBoost | |
| CN109300058A (en) | A kind of especially big basin water station group Optimized Operation two stages direct search dimension reduction method | |
| CN115828726A (en) | Machine learning-based solar cross-season heat storage and supply system optimization control method | |
| CN109635999A (en) | A kind of power station dispatching method looked for food based on population-bacterium and system | |
| CN104408531B (en) | A kind of uniform dynamic programming method of multidimensional multistage complicated decision-making problems |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |