Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

J. Cent. South Univ. Technol.

(2009) 16: 0683−0689


DOI: 10.1007/s11771−009−0113−4

Multi-objective coordination optimal model for


new power intelligence center based on hybrid algorithm

LIU Ji-cheng(刘吉成), NIU Dong-xiao(牛东晓), QI Jian-xun(乞建勋)


(School of Business Administration, North China Electric Power University, Beijing 102206, China)

Abstract: In order to resolve the coordination and optimization of the power network planning effectively, on the basis of
introducing the concept of power intelligence center (PIC), the key factor power flow, line investment and load that impact
generation sector, transmission sector and dispatching center in PIC were analyzed and a multi-objective coordination optimal model
for new power intelligence center (NPIC) was established. To ensure the reliability and coordination of power grid and reduce
investment cost, two aspects were optimized. The evolutionary algorithm was introduced to solve optimal power flow problem and
the fitness function was improved to ensure the minimum cost of power generation. The gray particle swarm optimization (GPSO)
algorithm was used to forecast load accurately, which can ensure the network with high reliability. On this basis, the multi-objective
coordination optimal model which was more practical and in line with the need of the electricity market was proposed, then the
coordination model was effectively solved through the improved particle swarm optimization algorithm, and the corresponding
algorithm was obtained. The optimization of IEEE30 node system shows that the evolutionary algorithm can effectively solve the
problem of optimal power flow. The average load forecasting of GPSO is 26.97 MW, which has an error of 0.34 MW compared with
the actual load. The algorithm has higher forecasting accuracy. The multi-objective coordination optimal model for NPIC can
effectively process the coordination and optimization problem of power network.

Key words: power intelligence center (PIC); coordination optimal model; power network planning; hybrid algorithm

However, they often cannot achieve satisfactory results


1 Introduction because the problem is a multi-objective coordination
optimal one.
The coordination and optimization problem of LIU et al [15], LIU and NIU [16] proposed the
power network planning is a key point in power system concept of power intelligence center (PIC) and
planning. However, solving power planning problem performed some basic researches. In this work, on the
from the overall situation of generation, transmission and basis of previous study, through processing three key
distribution is the difficulty of planning. So far, there factors that affect power grid, the corresponding model
have not many researches on the power network base was established by the improved intelligence
planning from the overall perspective. Most of them are algorithm, and then the solution to multi-objective
focused on the optimization and forecasting of key coordination optimal problem was obtained by calling
factors, such as the simple solution to the issue of the model base.
optimal power flow and load forecasting. On the one
hand, some traditional methods such as nonlinear 2 New power intelligence center (NPIC)
programming, linear programming and mixed-planning, model
have been used for solving the simple planning and
coordination problems [1]. On the other hand, many In order to make the model simple and convenient
intelligent optimization methods have been applied to for calculation and simulation, the impact of random
power network planning, such as the genetic algorithm factors on the NPIC is not considered. The basic idea of
(GA) [2−3], neural network (NN) [4−5], particle swarm NPIC model is established as follows: under the
optimization (PSO) [6−7], evolutionary algorithm (EA) circumstance of PIC, the function departments of power
[8−9], fuzzy theory [10−11], and immune algorithm system are divided into generation sector, transmission
[12−14]. These optimization methods have resolved the sector and distribution sector. The key factors
planning and coordination problems to some extent. corresponding to each department are different, which

Foundation item: Project (70671039) supported by the National Natural Science Foundation of China
Received date: 2009−02−25; Accepted date: 2009−04−20
Corresponding author: LIU Ji-cheng, Doctoral candidate; Tel: +86−13601030970; E-mail: ljc29@163.com
684 J. Cent. South Univ. Technol. (2009) 16: 0683−0689
are respectively power flow, line investment and load. It Regarding the minimum generation cost as the
is necessary to establish a corresponding model base in optimized objective function, the mathematical model of
order to solve the three departments objective optimal power flow is as follows.
optimization problem coordinately from the whole. (1) Objective function
Initially, a variety of implementation methods are listed, NG NG
such as genetic algorithm, neural network, and PSO. min F = ∑ F ( PGi ) = ∑ (ai + bi PGi + ci PG2i ) (1)
According to the characteristics of different algorithms, i =1 i =1

the problems corresponding to the key factors can be where F is the overall generation cost, PGi is the active
solved. Reselect if the method is improper. Once the power output of generator i, NG is the collection of all the
optimization result is better, it is identified as the method generators, ai, bi and ci are the cost factors of generator
of model base and stored in background model base of G.
the NPIC for the following process. In this way, the (2) Equality constraints
multi-objective coordination optimal problem can be NG
solved better by means of the dynamic coordination of PGi − PDi − Vi ∑ V j (Gij cos θij + Bij sin θij ) = 0 (2)
j =1
model base and power network. The NPIC model is
shown in Fig.1. NG
QGi − QDi − Vi ∑ V j (Gij sin θij − Bij cos θij ) = 0 (3)
j =1
3 NPIC model library based on hybrid
optimal algorithm where QGi is the reactive power output of generator i;
PDi and QDi are the active load and reactive load of node i;
3.1 Optimization model for solving optimal power Vi and θi are the voltage amplitude and phase angle of
flow problem node i, θij=θi−θj; Gij and Bij are respectively the real part
3.1.1 Optimal power flow model and imaginary part of the element in row i and column

Fig.1 Schematic diagram of NPIC model


J. Cent. South Univ. Technol. (2009) 16: 0683−0689 685
j in the node admittance matrix. as follows:
(3) Inequality constraints
X i′(G ) = X a(G ) + Z ( X b(G ) − X c(G ) ) (i=1, 2, ···, Np) (12)
The cap and floor constrain of PGi is
PGmin ≤ PGi ≤ PGmax (i=1, 2, ···, NG) (4) where a, b and c are randomly selected indicators, a, b,
i i
c∈{1, ···, Np}, and a≠b≠c≠i; scaling parameter Z is
The cap and floor constrain of QGi is the control variable of the algorithm, its range is [0, 2],
QGmin ≤ QGi ≤ QGmax (i=1, 2, ···, NG) (5) and it is often used to adjust the amplitude of the
i i
variation operation in order to strengthen the
The cap and floor constrain of VGi is convergence of the algorithm.
(3) Crossover
VGmin ≤ VGi ≤ VGmax (i=1, 2, ···, NG) (6)
i i
The vector generated from crossover is obtained
The cap and floor constrain of Ti is from the following formula:

Ti min ≤ Ti ≤ Ti max (i=1, 2, ···, NT) (7) ⎧⎪ X ′j(,Gi ) , if ρ j ≤ CR or j =q


X ″j (, Gi ) =⎨ (13)
(G )
where Ti is the transformation ratio of on-load ⎪⎩ X j , i , others
transformer of node i.
The cap and floor constrain of QCi is where i=1, 2, ···, Np, and j=1, 2, ···, D; ρj represents a
random number of uniformed distribution and its range is
QCmin
i
≤ QCi ≤ QCmax
i
(i=1, 2, ···, NC) (8) [0, 1], which is used to produce a new individual j; the
where QC is the alternative reactive power of node i. crossover vector CR is used to control differences of
The cap and floor constrain of VLi is population and control algorithm jump local optimum,
and its range is [0,1]; q is a randomly selected index and
VLmin
i
≤ VLi ≤ VLmax
i
(i=1, 2, ···, NL) (9) its range is q∈{1, 2, ···, D}.
where VL is the load node voltage of node i. (4) Selection
The cap and floor constrain of SLi is The formula of selection operation is as follows:
S Lmin
i
≤ S Li ≤ S Lmax
i
(i=1, 2, ···, NL) (10) ⎧⎪ X i′′(G ) , if f ( X i′′(G ) ) ≤ f ( X i(G ) )
X i(G +1) = ⎨ (14)
(G )
where SLi is the line transmission power of node i. ⎪⎩ X i , others
3.1.2 Optimal power flow model based on evolutionary
algorithm where i=1, 2, ···, Np.
Evolutionary algorithm (EA) is similar to genetic Formula (14) shows that when the individual after
algorithm, including initialization, variation, crossover crossover is substituted into the fitness function, and the
and evaluation processes. The main operation processes value is smaller than the designed standard value, the
to solve the optimal power flow problem using conditions are met, and the power flow is the optimal one.
evolutionary algorithm are as follows. Otherwise, the variation operation is returned and the
(1) Initialization repeated evolution is made until the termination
In initialization data, each decision parameter in conditions are satisfied. And the fitness function is
initialization population vector is transformed into a calculated according to Formula (1).
randomly selected parameter value through the following
formula: 3.2 Load forecasting model
GM(1,1) is the most commonly used as grey system
j, i = X j
X (0) min
+ µ j ( X max
j − X min
j ) model and its algorithm can be referred in Ref.[17]. The
(i=1, 2, ···, Np; j=1, 2, ···, D) (11) simple grey forecasting method can bring about larger
where µj presents the random number within [0,1], errors and is easy to generate noise and heavy-tailed
which is used to produce a new individual j; X max
j and phenomenon. In order to increase the forecasting
Xjmin
are respectively the upper and lower limits of accuracy and computational efficiency, the PSO
decision parameter j. algorithm is used to improve the simple grey forecasting
(2) Variation method and establish the grey forecasting method.
The variation operation is the combination of a Let t be any number, t∈{0, N−1}. The original load
randomly selected vector Xa and the deviation of the sequence is x(0)(t), the load sequence after
other two different random vectors Xb and Xc. The one-accumulation is x(1)(t), and the load sequence after
producing formula of a random variation vector X i′ is forecasting is xˆ (1) (t ) .
686 J. Cent. South Univ. Technol. (2009) 16: 0683−0689
We make sum of square error of x(0)(t) and From the above analysis we can see that when Q is
forecasting value xˆ ( 0) (t ) minimum, that is the minimum, the model forecasting accuracy is the
n highest. Thus, fitness function f(Ui) of individual Ui can
ε = ∑ [ xˆ (0) (t ) − x(0) (t )]2 (15) be expressed as follows:
t =1
n
Through the continuous optimization of PSO, the f (U i ) = ∑ [ xˆ (0) (t ) − x (0) (t )]2 (18)
t =1
forecasting accuracy of the gray neural network can be
improved. where f(Ui) changes reversely with the model
PSO expresses every possible solution as a particle forecasting accuracy, that is, when f(Ui) is smaller, the
of the group. Every particle has its own position vector, model accuracy is higher.
velocity vector and a fitness value decided by the (3) Update the location and speed of particles
objective function. All the particles fly at a certain according to Formulae (16) and (17).
velocity in the search space, and then find the global (4) When optimization reaches the maximum
optimum by following the optimal value of the current evolutionary generation Smax, the optimization is finished;
search. otherwise return to Eqn.(2).
Pretending in a D-dimensional objective search
space, the position vector of particle i is Xi=(xil, xi2, ···, 4 Multi-objective coordination optimal
xiD), where the arbitrary particle position is xid; the model for NPIC
individual extreme point is Pi=(Pil, Pi2, ···, PiD), and the According to the model base established, the
global extreme point is Pg=(Pgl, Pg2, ···, PgD); and the multi-objective coordination optimal model for NPIC is
velocity vector is Vi=(vil, vi2, ···, viD). The velocity established. There are three objective functions: the goal
changed by particles is decided by the following of optimal power flow, which is the power system
formula: condition required by Formula (1); the goal of load
forecasting, which makes Formula (18) achieve the
Vi +1 = wVi + c1r1 ( Pi − X i ) + c2 r2 ( Pg − X i ) (16) optimal sequence of load forecasting; the investment cost
X i +1 = X i + Vi +1 (17) of line, a route planning that makes investment cost
minimum. Thus, the multi-objective coordination
where d=1, 2, ···, D, i=1, 2, ···, N; N is the population optimal model is as follows:
size; w is the inertia weight; Xi is the current particle
⎧ NG NG
position; Vi is the current particle velocity and is limited ⎪ F = ∑ F ( PGi ) = ∑ (ai + bi PGi + ci PGi )
2

by a maximum velocity Vmax; r1 and r2 are the random ⎪ i =1 i =1


⎪ n
numbers distributed in the range [0,1]; c1 and c2 are the ⎪
min ⎨ f (U i ) = ∑ [ xˆ (0) (t ) − x (0) (t )]2 (19)
limited constants related to the position. ⎪ t =1
Combining the basic algorithm of PSO and the ⎪ n
characteristics of GM (1,1) model, and taking Formula ⎪ f ′ = ∑ si xi li
⎪⎩ i =1
(15) as the fitness value of PSO, a new grey forecasting
model based on PSO can be created. The basic algorithm where the parameters and constraints of the first and the
is as follows. second objective function have been described in the
(1) Randomly selecting m particles, the limited former section; the third objective function f ′ is the
constants with changed positions c1 and c2, the inertia investment cost for the line; si is the investment cost per
weight w and the maximum evolutionary generations unit length for the new line; xi is the number of loop of
Smax are given. Let the particles’ position vector and line i for selection; n is the total number of lines to be
velocity vector be respectively: Ui=(uil, ui2, ···, uim), chosen; and li is the length of line i.
Vi=(vil, vi2, ···, vim), where i=1, 2, ···, m, uij and vij are the The multi-objective coordination optimal model is
numbers in the range [0, 1], and n is the number of the solved by improving PSO algorithm. The specific
original load sequence. algorithm is as follows.
(2) Let the sum of square error of grey neural Step 1 Because the three objective functions in
network’s forecasting value and input value be Q, Formula (19) are more independent and they all seek the
according to Formula (15), we have minimum, the penalty factor is introduced, then the
n
multi-objective function is transformed into a
Q= ε = ∑ [ xˆ (0) (t ) − x(0) (t )]2 . single-objective problem. On the one hand the indicators
t =1 of the power generator can be controlled in the scope of
J. Cent. South Univ. Technol. (2009) 16: 0683−0689 687
the corresponding variables’ lower and upper limits, on factor is 10. In load forecasting, the particle number of
the other hand the single-objective function can be the gray particle swarm optimization (GPSO) is 200, the
processed as PSO fitness function. The objective limit constants with changed location c1 and c2 are 1.5,
function after transformation is as follows: the inertia weight factor w is 0.5+0.5σ, where σ is a
random number in the range [0, 1], and mutation
NG
min F ′ = λ1 ∑ (ai + bi PGi + ci PG2i ) + probability is 0.1, which is the same as that of GA. In the
i =1 calculation, the construction cost per unit length of each
n n line is assumed to be the same. So the length of such
λ2 ∑ [ xˆ (0) (t ) − x(0) (t )]2 + λ3 ∑ ci xi li (20) lines can be used to replace the cost to be calculated. The
t =1 i =1
evolutionary algorithm is used for optimal power flow
where F′is the transformed objective function, λ1, λ2 optimization on the IEEE30 bus system. The parameters
and λ3 are the penalty factors. are as follows: population size Np is 18, crossover
Step 2 Set the parameters of PSO and the constant CR is 0.5, the generator’s scaling parameter Z is
parameters of the three objective functions, and 0.9, the maximum number Gmax is 180. The penalty
randomly generate particles. factors λ1, λ2 and λ3 of multi-objective coordination
Step 3 Set Formula (20) as the fitness function of optimal model are all equal to 1.
PSO, set the corresponding penalty factor, calculate the Because there are 6 generator units, 4 on-load tap
fitness value, and then determine whether the minimum changers and 41 branch circuits in the IEEE30 node
of F′ is obtained. If F′min is gained, the algorithm is system, the fuel cost functions of generators are all
finished; otherwise, continue to the next step. quadratic functions. The parameters are shown in Table 1.
Step 4 According to the location and velocity IEEE30 nodes are optimized, and the calculation
formula, calculate the corresponding fitness value and results of four different algorithms are shown in Table 2.
update the particles position and velocity. We can see from Table 2 that the largest total
Step 5 When optimization reaches the maximum generating capacity of the four algorithms belongs to the
evolutionary generation or the fitness value achieves the evolutionary algorithm, and its value is 35.404 MW; and
minimum, the algorithm is finished; otherwise return to the smallest line investment also belongs to evolutionary
Step 2. algorithm, which is 15.672 k¥(RMB). However, the
traditional NR algorithm has the smallest total generating
5 Example capacity of 33.671 MW and it has the largest line
investment, which explains that intelligent optimization
In this example, the multi-objective coordination algorithm has certain advantage in solving optimal power
optimal model is used for calculation of optimal power
flow problem over the traditional algorithm. It can also
flow and load forecasting of IEEE30 node system. In
be noticed that the optimal results of GA and P-Q are
order to verify the validity of the algorithm, the
close, while optimal results by the evolutionary
simulation results of the algorithm proposed this work,
algorithm are superior to those of the two methods.
genetic algorithm(GA), Newton-Raphson algorithm (NR)
The key factors of multi-objective coordination
and P-Q analysis method of power flow calculation are
optimal model are obtained by short-term load
compared. The reason why the NR method is compared
forecasting in 24 h. The used method is the gray particle
with P-Q method is that P-Q method is derived from
swarm optimization method. The load forecasting data of
Newton-Raphson method which is expressed in polar
coordinates and is made the following simplifications. IEEE30 node system are shown in Table 3, and the load
The first simplification is that the reactance of every forecasting results of IEEE30 node system are shown in
component in power network is far greater than the Fig.2.
resistance, and the second one is that the constraint The load forecasting curves of Fig.2 and the
condition of state variable δ i is not too big. The test on calculation results of Table 3 show that the average load
bus system is based on Matlab7.0 platform, the CPU of forecasting of IEEE30 node system is 26.97 MW, which
computer is Intel Pentium IV (2.6 GHz), and the memory has an error of 0.34 MW compared with the actual load.
is 1 024 Mbit. The accuracy is high and the algorithm reaches
Before the optimization, the parameters of the convergence.
algorithm are needed to be set. The parameters of GA are The above results show that it is feasible for the
as follows: the population size is the same as that of multi-objective coordination optimal model to optimize
evolutionary algorithm in this work, crossover three key factors that affect the power network. The
probability is 0.9, mutation probability is 0.1, learning model can achieve convergence and make the total
probability is 0.5, learning system is 0.6, and penalty objective function minimum. Thereby, it lays a theoretical
688 J. Cent. South Univ. Technol. (2009) 16: 0683−0689
Table 1 Parameters of generator’s quadratic cost function
Cost factors of generator
Node number PGmin /MW PGmax /MW QGmin /MW QGmax /MW
a b c
1 50 200 −20 200 0 2.00 0.003 75
2 20 80 −20 100 0 1.75 0.017 50
5 15 50 −15 80 0 1.00 0.062 50
8 10 35 −15 60 0 3.25 0.008 34
11 10 30 −10 50 0 3.00 0.025 00
13 12 40 −15 60 0 3.00 0.025 00

Table 2 Calculation results of optimal power flow of four different algorithms (IEEE30 node system)

PG1/MW PG2/MW PG5/MW PG8/MW PG11/MW PG13/MW Total generating Line investment/
Algorithm
capacity/MW k¥(RMB)

GA 127.66 35.360 15.636 15.991 9.021 7 8.710 2 35.395 15.961

NR 125.57 35.201 13.946 9.289 8 8.441 3 9.579 6 33.671 17.683

P-Q 127.61 35.451 15.599 16.128 8.884 2 8.702 4 35.396 15.975

EA 127.81 35.542 15.594 15.817 8.947 9 8.711 7 35.404 15.672

Table 3 Load forecasting results of GPSO algorithm (IEEE30 node system)


Generating set/MW
Time/h
1 2 5 8 11 13
1 15.467 12.284 23.408 38.502 22.062 21.756
2 16.563 13.154 25.068 41.232 23.627 23.299
3 15.948 12.666 24.136 39.701 22.749 22.433
4 14.644 11.630 22.161 36.453 20.888 20.598
5 11.615 9.225 17.579 28.914 16.568 16.338
6 13.413 10.653 20.299 33.390 19.133 18.867
7 15.211 12.081 23.020 37.865 21.698 21.396
8 17.009 13.508 25.742 42.341 24.262 23.925
9 18.462 14.662 27.941 45.959 26.335 25.970
10 19.226 15.269 29.097 47.860 27.424 27.043
11 20.334 16.149 30.775 50.620 29.005 28.603
12 20.581 16.345 31.148 51.233 29.357 28.949
13 19.793 15.719 29.956 49.273 28.234 27.841
14 19.177 15.231 29.024 47.741 27.356 26.976
15 18.563 14.742 28.093 46.210 26.478 26.111
16 20.705 16.444 31.336 51.544 29.535 29.125
17 21.124 16.777 31.970 52.586 30.133 29.714
18 20.336 16.152 30.778 50.625 29.009 28.606
19 21.100 16.758 31.934 52.527 30.098 29.680
20 22.036 17.501 33.350 54.857 31.434 30.998
21 22.800 18.108 34.506 56.758 32.523 32.072
22 23.219 18.441 35.141 57.801 33.121 32.661
23 21.569 17.130 32.645 53.695 30.768 30.341
24 21.127 16.779 31.974 52.592 30.136 29.718
J. Cent. South Univ. Technol. (2009) 16: 0683−0689 689
the model established in this work.
(4) Merely considering the model with one key
factor is a special case of multi-objective coordination
optimal model. And multi-objective model can be
extended to taking into account three or more key
factors.

References
[1] BAI X Q, WEI H, FUJISAWA K, WANG Y. Semidefinite
programming for optimal power flow problems [J]. International
Journal of Electrical Power and Energy Systems, 2008, 30(6/7):
383−392.
[2] PAI P F, HONG W C. Forecasting regional electricity load based on
recurrent support vector machines with genetic algorithms [J].
Electric Power Systems Research, 2005, 74(3): 417−425.
[3] BAKIRTZIS A G, BISKAS P N, ZOUMAS C E, PETRIDIS V.
Optimal power flow by enhanced genetic algorithm [J]. IEEE Trans
on Power Systems, 2002, 17(2): 229−236.
[4] KANDIL N, WAMKEUE R, SAAD M, GEORGES S. An efficient
approach for short term load forecasting using artificial neural
networks [J]. Electrical Power and Energy Systems, 2006, 28(8):
525−530.
[5] LIU Ling, YAN Deng-jun, GONG Deng-cai, ZHANG Hong-mei, LI
Da-peng. New method for short term load forecasting based on
particle swarm optimization and fuzzy neural network [J].
Proceedings of the CSU-EPSA, 2006, 18(3): 47−50. (in Chinese)
[6] LIN Y L, CHANG W D, HSIEH J G. A particle swarm optimization
approach to nonlinear rational filter modeling [J]. Expert Systems
with Applications, 2008, 34(2): 1194−1199.
[7] MUKHERJEE V, GHOSHAL S P. Intelligent particle swarm
optimized fuzzy PID controller for AVR system [J]. Electric Power
Systems Research, 2007, 77(12): 1689−1698.
Fig.2 Load forecasting results of IEEE30 node system for [8] YOG R S. Evolutionary programming based optimal power flow and
its validation for deregulated power system analysis[J]. International
nodes 1, 2 and 5 (a) and for nodes 8, 11 and 13 (b)
Journal of Electrical Power and Energy Systems, 2007, 29(1): 65−75.
[9] BASU M. Optimal power flow with FACTS devices using
foundation for building NPIC in local areas to deal with differential evolution [J]. International Journal of Electrical Power
the key factors of the power network and optimization and Energy Systems, 2008, 30: 150−156.
[10] TAO Xiao-hu, HUANG Min-xiang. A load forecasting method based
network. on fuzzy regulation and neural network [J]. Proceedings of the
CSU-EPSA, 2000, 12(5): 37−41. (in Chinese)
6 Conclusions [11] PAI T Y, TSAI Y P, LO H M, TSAI C H, LIN C Y. Grey and neural
network prediction of suspended solids and chemical oxygen demand
in hospital wastewater treatment plant effluent [J]. Computers and
(1) The key factors in the three kinds of power Chemical Engineering, 2007, 31(10): 1272−1281.
network planning are considered. The intelligent [12] XIONG Hu-gang, CHENG Hao-zhong, LI Hai-yu. Optimal reactive
power flow incorporating static voltage stability based on
optimization method is proposed to solve the problems
multi-objective adaptive immune algorithm [J]. Energy Conversion
arising from the key factors, which lays the foundation and Management, 2008, 49(5): 1175−1181.
for cooperation and decision-making of power supply [13] HONG W C. Chaotic particle swarm optimization algorithm in a
chain alliance system. support vector regression electric load forecasting model [J]. Energy
Conversion and Management, 2009, 50(1): 105−117.
(2) The optimal power flow problem is effectively [14] LIAO G C. Short-term thermal generation scheduling using
solved and the minimum cost of line investment can be improved immune algorithm [J]. Electric Power Systems Research,
calculated by using evolutionary algorithm. The high 2006, 76(5): 360−373.
[15] LIU Ji-cheng, YAN Su-li, QI Jian-xun. Research on building business
precision load sequence is got by using the GPSO intelligence center in dynamic alliance [J]. China Market, 2008, 1(2):
method, which provides a model reserve for building 66−67. (in Chinese)
multi-objective coordination optimal model. Thereby, the [16] LIU Ji-cheng, NIU Dong-xiao. A novel recurrent neural network
forecasting model for power intelligence center [J]. Journal of
model base of the NPIC is established, which provides
Central South University of Technology, 2008, 15(5): 726−732.
the convenience for solving the same problem. [17] LI Cun-bin, WANG Ke-cheng. A new grey forecasting model based
(3) The examples show that the intelligent on BP neural network and Markov chain [J]. Journal of Central South
optimization method has a greater advantage in resolving University of Technology, 2007, 14(5): 713−718.
(Edited by CHEN Wei-ping)
the issue of multi-objective coordination optimal than the
traditional planning methods and verifies the validity of

You might also like