Article

Power Transmission Scheduling for Generators in a

Deregulated Environment Based on a
Game-Theoretic Approach
Bingtuan Gao 1,2, *, Tingting Ma 1 and Yi Tang 1
Received: 10 September 2015; Accepted: 2 December 2015; Published: 5 December 2015
Academic Editor: David Wood
1
2

School of Electrical Engineering, Southeast University, Nanjing 210096, China;

matingting9414@163.com (T.M.); tangyi@seu.edu.cn (Y.T.)
Jiangsu Key Laboratory of Smart Grid Technology and Equipment, Nanjing 210096, China
Correspondence: gaobingtuan@seu.edu.cn; Tel. +86-25-8379-4163; Fax: +86-25-8379-0617

Abstract: In a deregulated environment of the power market, in order to lower their energy
price and guarantee the stability of the power network, appropriate transmission lines have to
be considered for electricity generators to sell their energy to the end users. This paper proposes
a game-theoretic power transmission scheduling for multiple generators to lower their wheeling
cost. Based on the embedded cost method, a wheeling cost model consisting of congestion cost,
cost of losses and cost of transmission capacity is presented. By assuming each generator behaves
in a selfish and rational way, the competition among the multiple generators is formulated as
a non-cooperative game, where the players are the generators and the strategies are their daily
schedules of power transmission. We will prove that there exists at least one pure-strategy Nash
equilibrium of the formulated power transmission game. Moreover, a distributed algorithm
will be provided to realize the optimization in terms of minimizing the wheeling cost. Finally,
simulations were performed and discussed to verify the feasibility and effectiveness of the proposed
non-cooperative game approach for the generators in a deregulated environment.
Keywords: power transmission scheduling; non-cooperative game; wheeling cost; congestion cost

1. Introduction
Over the past few years, the electric power industry in China has experienced a strong drive
toward deregulation. In 2014, transmission and distribution electricity pricing reform has been
carried out in Shenzhen as the document The notice about the reform pilot of transmission
and distribution price carried out in Shenzhen issued by the National Development and Reform
Commission of China [1], which would deregulate the power generation side market and the demand
side market. Once both sides of the market are opening, the role of the electric power utilities in China
will change from traditionally buying and selling electricity to only providing transmission service
for generators and end users, which also means that the electricity pricing mechanism in China
will be changed dramatically. Reasonable pricing mechanisms based on open electricity markets
is essential to achieve optimal allocation of resources, lower energy cost, higher service quality, and
so on. Moreover, establishing appropriate independent transmission electricity pricing mechanism is
still an open problem for deregulated electricity markets all over the world [2].
Independent transmission electricity price is based on wheeling cost. Wheeling is defined as the
use of a utilitys transmission facilities to transmit power for other buyers and sellers [3]. The buyer
or seller pays transmission cost for the usage of utilitys transmission facilitates. This transmission
cost is also called wheeling cost [4]. Wheeling cost is mainly used for recovery of investment and costs
Energies 2015, 8, 1387913893; doi:10.3390/en81212401

www.mdpi.com/journal/energies

Energies 2015, 8, 1387913893

of operation, management and staff, in order to maintain good operation and development of the
transmission system. Wheeling cost is reasonable and affects not only the survival and development
of electric utilities but also the benefit of generators and end users. Various algorithms and methods
have been used and proposed to evaluate the cost of transmission services. The current methods
can be mainly classified into two categories [2]: embedded cost of wheeling methods [3,57] and
marginal cost pricing method [810]. Embedded cost of wheeling methods, including the postage
stamp method, contract path method, boundary flow methods and so on, recover the embedded
capital costs and the average annual operating and maintenance costs of existing facilities from a
particular wheeling transaction. The marginal cost pricing method is based on microeconomics.
Additionally, Nojeng et al. [11] proposed an improved MW-Mile method by considering not only the
changes in MW flows but also the quality of the load, i.e., the power factor. Monsef et al. [12] proposed
a reliability-based method for allocating the cost of transmission networks. Moreover, a two-step
method based on the perfect coupling of the circuit theory with the AumannShapley method is
proposed in [13]. The method can calculate allocation costs for each branch of the transmission system
to identify and quantify the individual responsibility of generators and loads.
Although the aforementioned papers can calculate wheeling cost under several different
conditions, most of them missed studying the effect of congestion cost, which should be covered
by wheeling cost. Congestion is defined as the overloading of one or more transmission lines
and/or transformers in the power system [14]. In order to solve the problem of congestion,
particle swarm optimization (PSO) methods [14,15] and bacterial foraging algorithms [16,17] were
successfully implemented. Kanwardeep Singh et al. [18] presented an effective methodology for
congestion management in deregulated power system networks considering optimal placement
of a distributed generator based on bus impedance matrix considering contribution factors.
Harry Singh et al. [19] proposed a pool model and a bilateral model to deal with the transmission
congestion costs in competitive market and evaluated some aspects of the models based on game
theory. Erli et al. [20] proposed a cooperative game based scheme for cost allocation of transmission
line expansion to tackle with congestion problem in the network.
This paper deals with a scenario in a deregulated environment of power market having multiple
generators, multiple transmission lines, and one aggregated end user. The generators have to
compete with each other to optimize their profit. By considering the game theory-based/double
auction-based approaches and control theory-based approaches [2124], we formulate wheeling
pricing including congestion cost for the players in the deregulated environment of power market.
To make their electricity more competitive in the end users market, a game-theoretic approach is
proposed for generators to schedule their power transmission effectively to lower the wheeling costs.
The main contributions of this paper can be summarized as: (1) by assuming generators are selfish
and rational, a non-cooperative game approach is proposed for multiple generators to lower their
wheeling costs, where strategies of the non-cooperative game are the power transmission scheduling
of the generators; (2) a distributed algorithm is presented to realize the optimization in terms of
minimizing the wheeling cost, which can be guaranteed at the Nash equilibriums of the formulated
non-cooperative games; (3) simulations are performed to verify the effectiveness of the proposed
approach, and discussions show that all the generators, electric utilities, and aggregated end users
can benefit from the game.
The rest of the paper is organized as follows. Section 2 presents the method that this paper
adopts for the pricing model of the wheeling cost. In Section 3, the novel proposed method of this
paper for optimizing wheeling cost based on the game-theoretic method is presented. Simulation
results are presented and discussed in Section 4. Finally, conclusions are provided in Section 5.

13880

Energies 2015, 8, 1387913893

2. System Model
2.1. Wheeling Cost Model
As shown in Figure 1, we consider a typical power system consisting of a set of N = {1, . . . , N }
generators, a set of M = {1, . . . , M } transmission lines managed by different wheeling utilities and
different end users. Each generator n N can supply electrical power to the end users in a certain
area via a single transmission line or several transmission lines. Generator n N can be a classical
thermal power plant, a renewable solar power station or a wind farm, and other new energy power
plant. Transmission lines belongs to utility companies who provide the power transmission service.
End users are taken as a whole in this paper. Generators and/or users have to pay the wheeling cost
to electric utilities.

Generator 1

Power
Line
LAN

Wheeling utility 1

Generator n

Wheeling utility M

Users

Generator N

Figure 1. Figure of wheeling transaction.

Based on the embedded cost method [11,25], wheeling cost can be calculated as
G (q) = R + W (q) + Z (q)

(1)

where q denotes transmitted electricity energy, R is the cost of transmission capacity which reflects
investment and maintenance cost of electric utilities, W (q) is the cost of losses, Z (q) is the
congestion cost.
Electric utilities make profits only from wheeling cost paid by generators. Accounting for a large
proportion of wheeling cost, cost of transmission capacity directly relates to whether the investment
of electric utilities can be recycled. Cost of transmission capacity generators need to pay can be
calculated as:
!
r
Tm
1 M
(1+r ) t 1
R=
(2)
M m
Qm,year
=1
where Tm is investment cost of transmission line m M, r is discount rate, t is recovery period,
Qm,year is the predicted transmission capacity of transmission line m during one year. For the line loss
in Equation (1), it can be calculated as:
W (q) = C (q)

(3)

where C (q) is the generation cost, and is line loss rate, which is widely used by electric utilities to
solve the allocation problem of loss cost.
13881

Energies 2015, 8, 1387913893

If user loads are stable, load curve can be modified into a polygonal line. Each segment denotes
a period of time, and load is invariable during this time. Inspired by [25], congestion cost can be
described as:


(q)
Z (q) = R ln
(4)
e (q)
(q)
where
(q) =

qm,all
qn,m
, (q) =
qm,max
qm,max

(5)

and qm,max is maximum transmission capacity of transmission line m, qm,all is power flow on
transmission line m of all generators, qn,m is power flow related to generator n on transmission line
N

m. Therefore, one can see qm,all = qn,m straightforwardly. Logarithmic term and exponential term
n =1

of Equation (4) denote two mathematical models: the former is congestion extent in branch caused
by generator n, and the latter is the proportion of power flow caused by generator n in power flow on
transmission line m. During the period under heavy load, congestion cost will be high. The additional
cost caused by congestion can be regarded as the fund used to improve the transmission capacity of
transmission lines. Accordingly, congestion cost will be low and even be negative during the period
under light load. In other words, congestion cost can be effectively reduced by alleviating congestion.
Under this condition, more generators can be encouraged to use transmission lines with light-loads
to transmit power. It should be noted that wheeling cost cannot be negative practically even though
there is no congestion on transmission line. Therefore, constraint Z (q) 0 should be satisfied.
2.2. Cost Model of Generators
Assuming that the daily power consumption of users is predicted by electric utilities, and total
energy generated by all generators are determined by electric utilities. Thus, balance of supply and
demand in power system can be guaranteed. If the generated electric energy production is only
determined by generators or regulated by the market, an imbalance between supply and demand
may occur. However, the imbalance is harmful for stability of power system. Generation cost
of generators can be divided into two parts: the fixed cost and the variable cost. For example,
considering thermal power plants, the fixed cost mainly consists of capacity cost, which does not
change with different generations. Variable cost of thermal power plant mainly consists of fuel
consumption, which changes with different generations.
For each generator n, let qnh denote the generated energy at time slot h H = {1, , H }.
For a daily operation of the grid, each time slot may take 15 min and
h we have H = 96.
i Generation

scheduling for generator n in a whole day is denoted by qn = q1n , , qnh , , qnH . Let Cn (qn )
denote the cost of generating electricity by each generator n N , which can be modeled as the
following quadratic function [26]:
Cn (qn ) = an q2n + bn qn + cn

(6)

where an > 0, bn > 0, cn > 0 are fixed parameters; cn denotes fixed cost; an q2n + bn qn denotes
variable cost.
Although generators cannot decide the electric energy production, we assume that generators
can determine the quantity of electricity transported on each transmission line. For each transmission
line m M = {1, . . . , M}, we define a power transmission scheduling vector as:
h
i
1
h
H
xn,m = xn,m
, xn,m
, xn,m

(7)

h
h
where xn,m
denotes the proportionality coefficient decided by generator n, xn,m
[0, 1].
The proportionality coefficient is the proportion of energy transported by transmission line m in total

13882

Energies 2015, 8, 1387913893

energy generated by generator n at time slot h. Note that total electricity will be transmitted to users
by M transmission lines. Thus, for each generator n, the following constraint has to be satisfied:
M

h
xn,m
= 1, h H

(8)

m =1

The proportion of energy transported by transmission line m for each generator n needs to be
max
constrained. Let min
n,m denote the minimum proportion and n,m denote the maximum proportion.
h
max
min
n,m xn,m n,m , h H

(9)

In order to guarantee stability of power system, we suppose energy generation equal to energy
consumption at each time slot h H, namely,
N

qnh = Eh

(10)

n =1

where Eh denotes total energy consumption of end users at time slot h. By introducing vector xn as
power transmission scheduling on all transmission lines m M, a power transmission scheduling
set can be described as:
( 
)
 N

h
h min
h
max
Xn = xn  qn = E , n,m xn,m n,m , h H
(11)
 n =1
Let qn,m be electric quantity transported by transmission line m of each generator n N .
qn,m = xn,m qTn

(12)

Consequently, the total power transmission of transmission line m is calculated as:

N

qm,all =

xn,m qTn

(13)

n =1

We define a vector cn,m is electric quantity transported by transmission line m of all generators except
generator n.
N

cn,m =

xi,m qTn

(14)

i =1,i 6=n

In other words, cost of losses W and congestion cost Z in Equation (1) can be regarded as functions of
xn,m qTn . By assuming qn is constant, congestion cost in Equation (4) for generator n on transmission
line m can be written as:
!
qT
n
xn,m qTn + cn,m xqn,m
(15)
e m,max
Z (xn,m ) = R ln
xn,m qTn
Meanwhile, cost of line loss in Equation (3) for generator n on transmission line m can be written as:


W (xn,m ) = Cn xn,m qTn

(16)

Because the denominator cannot be 0, xn,m 6= 0, qn,m 6= 0. Note that xn,m 0, qn,m 0, where
generated energy cannot be negative, namely, xn,m > 0, qn,m > 0. The constraint xn,m 6= 0 will not
obviously influence the validity of our study, because the optimal choices for generators generally

13883

Energies 2015, 8, 1387913893

transport power through all transmission lines. Then, let G (xn,m ) denote wheeling cost of generator
n on transmission line m, which can be calculated as:
G (xn,m ) = R + W (xn,m ) + Z (xn,m )

(17)

Therefore, total cost of each generator n N is

M

Pn = Cn (qn ) +

G (xn,m )

(18)

m =1

When cost of generator is reduced, the electric price can be reduced for users. It can also make
generators be more competitive.
3. Non-cooperative Game-Theoretic Approach
In this section, based on the wheeling cost in Equation (18) for each generator, a non-cooperative
game among generators is proposed to reduce wheeling cost by optimizing power transmission
scheduling. And also a distributed optimization algorithm is provided to realize the optimization
in terms of minimizing the wheeling cost.
3.1. Non-Cooperative Game
Our goal is to find the optimal scheduling of generators that minimize the wheeling cost. The
optimization problem can be expressed as follows:
H

minimize

Pn

xn Xn ,nN h=1

!
h
xn,m

(19)

m M

The optimization problems defined in Equation (19) can be solved to obtain an optimal solution
for the given generators and their power transmission scheduling. Each generator selects strategy
to make probability that congestion occurs on transmission lines as low as possible. Accordingly,
wheeling cost of the generator will decrease and congestion can also be effectively alleviated and
even avoided.
Generally, each generator behaves in a selfish and rational way, and attempts to maximize the
steady-state value of payoff [27]. Therefore, the interaction among the generators can be modeled
as a non-cooperative game, whose solution is the well-known Nash equilibrium. We formulate a


non-cooperative game among generators, = N , {Xn }nN , {Sn }nN , which is characterized by
three main elements as following:
Players: Generators in the set N .
Strategies: Each generator n N selects its power transmission strategy to maximize its payoff.
Payoffs: Gains and costs from adjusting power transmission scheduling for each generator
n N . Given a certain strategy choice xn by any generator, the utility function can be
characterized by:
M

Sn (xn , xn ) = Pn = Cn (qn )

G (xn,m )

(20)

m =1

where xn = [xn , . . . , xn1 , xn+1 , . . . , xn ] denotes the power transmission scheduling for all
generators except the generator n.
Based on the definitions of the strategy and payoff in the game, all generators will try to adjust
their power transmission scheduling to minimizing their wheeling costs until a Nash equilibrium of
the game is reached:

13884

Energies 2015, 8, 1387913893

Nash equilibrium: Consider the game played among a set N of players. The power
transmission scheduling xn Xn for each generator n N form a unique Nash equilibrium if
and only if we have:
Sn (xn , xn ) Sn (xn , xn )

(21)

It means that once the power transmission game is at its unique Nash equilibrium, no one will benefit
by deviating from xn , n N . Although the existence of Nash equilibrium is not guaranteed for
any non-cooperative game in general, we can obtain the following existence result for the proposed
game [28].


Theorem 1: For the non-cooperative game = N , {Xn }nN , {Sn }nN , there exists a unique
pure-strategy Nash equilibrium.
Theorem 2: The unique Nash equilibrium of game is the optimal solution of problem in Equation (19).
The proof of Theorem 1 and Theorem 2 are given in Appendix A and Appendix B, respectively.
3.2. Algorithm for the Wheeling Cost Game
According to the proof in the Appendix, the problem in Equation (19) is concave and admits
an unique optimal solution. The algorithm basically allows each player to play his best response
strategy, announce only his transmission scheduling, and repeat until equilibrium is reached. The
algorithm is explained in detail in Algorithm 1. First, it start with some random power transmission
schedules. Each step will result in the total wheeling cost decreasing or remaining the same.
Since the objective function is bounded from below, the iterations will eventually converge to a
fixed point which wheeling cost Pn up to some accuracy . Once generators reach this point,
they will have no preference to change since the NE is unique, and convergence is achieved.
Algorithm 1: Executed by generator n.
Randomly initialize xn,m and pn
Choose (0, 1)
repeat
Calculate total transmission energy of each line according to Equation (13)
Update wheeling cost of each generator Pn
if Pn less than pn then
7:
Update pn equal to the new value Pn
8: end if
9: until | pn Pn | < is satisfied.
1:
2:
3:
4:
5:
6:

4. Simulation and Discussion

4.1. Simulation Results
This section presents several simulations of the method presented in the previous section on
the problem of wheeling cost optimization. We assume N = 8 in Figure 1, and eight generators
are two wind farms, five solar power stations and one thermal power plants, respectively. For
simplification of simulation, two wind farms are taken as a whole, meanwhile, five solar power
stations are taken as a whole. Installed capacity of thermal power plant is 5000 MW, and total installed
capacity of two wind farms and five solar power stations are both 1500 MW. The eight generators
supply power to users through two transmission lines, supposing that line 1 with 9000 MW of
power transmission capacity and line 2 with 5000 MW of power transmission capacity. In order
to promote generation of electricity by new energy, all electricity produced by wind farms and solar
power stations will be transported to the users preferentially.
For simplicity, it is assumed that generation cost function of thermal power plant is
C1 (q1 ) = 0.018q21 + 1.7q1 , generation cost function of wind farm is C2 (q2 ) = 0.022q22 + 1.7q2 ,
generation cost function of solar power station is C3 (q3 ) = 0.021q23 + 2.1q3 . Daily load curve of users
is shown in Figure 2. One day is divided into 96 time slots, electricity price updates every 15 min.
Suppose that the sum of electricity generated by three power plants is consistent with the daily load
13885

Energies 2015, 8, 1387913893

curve of users every time, which also means that supply and demand obtain a balance. Electricity
generated by solar power stations and wind farms are shown in Figure 3. Then electricity generated
by thermal power plant is the difference between loads of users and total electricity generated by
wind farms and solar power stations. Parameters of wheeling cost model are as follows: investment
cost of transmission line 1 T1 = 2 106 , investment cost of transmission line 2 T2 = 1.8 106 , discount
rate r = 12%, line loss rate = 5%, accuracy = 1. Before optimization, electricity of generated by
each plant will be divided equally to transmit through the two transmission lines at every moment,
h
hence, xn,m
= 0.5 for each generator.
103
5

Profile of users loads (MW)

4.5

3.5

2.5

2
0

10

20

30

40

50

60

70

80

90

100

Time of day (h)

Figure 2. Daily load curve of users.
3
1.6 10

3
1.6 10

1.5

Solar power stations (MW)

1.4

Wind farms (MW)

1.4
1.3
1.2
1.1
1
0.9
0.8

1
0.8
0.6
0.4
0.2

0.7
0.6

1.2

0 10 20 30 40 50 60 70 80 90 100
Time of day (h)
(a)

0 10 20 30 40 50 60 70 80 90 100
Time of day (h)
(b)

Figure 3. Generation of electricity: (a) Wind farms; (b) Solar power stations.

By using the proposed method, the non-cooperative game among generators can be
implemented automatically with the distributed algorithm. Figure 4 shows the quantity of electricity
transported through each transmission line after optimization, which generated by wind farms, solar
13886

Energies 2015, 8, 1387913893

power stations and thermal power plant, respectively. From the figure, one can see that quantity
of electricity transported through transmission line 1 is large because of large transmission capacity.
Table 1 shows daily wheeling cost of three power plants, including cost before and after optimization,
where Pn means the difference of wheeling cost between before and after optimization for each
typical generator. One can see that wheeling costs of generators are all decreased, which validates
the effectiveness and efficiency of the proposed optimal algorithm in this paper. Therefore, all the
generators will be benefited and willing to take part in the game. Figure 5ac denotes wheeling cost
of wind farms, solar power stations and thermal power plant with and without game, respectively.
Wheeling cost with game is lower than wheeling cost without game at each time, which can also
validate effectiveness of algorithm presented previously. According to Figures 3 and 5, note that the
wheeling cost increases obviously when quantity of transmission electricity is large. Meanwhile, the
wheeling cost is low when quantity of transmission electricity is small, which can promote power
transmission. It is important to highlight that when congestion is alleviated effectively, stability of
power system is guaranteed. Besides, service life of transmission lines will be extended, which is
beneficial for electric utilities because of the reduced fixed cost.
3

1.4 10

1.4 10

Line 1
Line 2

Solar power stations (MW)

Wind farms (MW)

1.2
1
0.8
0.6
0.4
0.2
0

Line 1
Line 2

1.2
1
0.8
0.6
0.4
0.2
0

0 10 20 30 40 50 60 70 80 90 100
Time of day (h)

0 10 20 30 40 50 60 70 80 90 100
Time of day (h)

(a)

(b)
3

Thermal power plant (MW)

4 10
3.5

Line 1
Line 2

3
2.5
2
1.5
1
0.5
0

0 10 20 30 40 50 60 70 80 90 100
Time of day (h)

(c)
Figure 4. Electricity transported through each transmission line: (a) Wind farms, (b) Solar power
stations and (c) Thermal power plant.

13887

Energies 2015, 8, 1387913893

Wind farms (dollars)

103

3
7 10
With game
No game

Solar power stations (dollars)

6
5
4
3
2
1

6
5
4
3
2
1
0

0 10 20 30 40 50 60 70 80 90 100
Time of day (h)

With game
No game

0 10 20 30 40 50 60 70 80 90 100
Time of day (h)

(a)

(b)

Thermal power plant (dollars)

2.5

104
With game
No game

2
1.5
1
0.5
0

0 10 20 30 40 50 60 70 80 90 100
Time of day (h)

(c)
Figure 5. Wheeling cost of generators: (a) Wind farms, (b) Solar power stations and (c) Thermal
power plant.
Table 1. Daily wheeling cost of each generator (dollars).
Generators

With Game

No Game

Pn

Wind farm
Solar power station
Thermal power plant

79,372
64,681
139,950

88,198
68,571
152,590

8826
3890
12,640

The game-based method proposed previously is one decentralized control approach, in contrast
with central control approach. In central control system, not only generated electric energy
production of generators but power transmission scheduling is determined by utility companies.
System parameters are the same as the previous case, and the total wheeling cost of generators during
one day with central control is shown in Table 2. P1 denotes the difference of wheeling cost between
scheduling with game and naive scheduling. Similarly, P2 denotes the difference of wheeling cost
between scheduling with central control and naive scheduling. One can see that the wheeling cost
with central control is slightly higher than the wheeling cost with decentralized control. Furthermore,
the decentralized control approach based on non-cooperative games can be implemented more easily,

13888

Energies 2015, 8, 1387913893

since the system with central control needs a center price setter. In decentralized control systems with
game, however, each generator is motivated to adjust scheduling to lower wheeling costs.
Table 2. Daily wheeling cost under different condition (dollars).
Condition

No Game

With Game

Central Control

P1

P2

Wheeling cost

309,359

284,003

285,675

25,356

23,684

4.2. Discussions
Previous simulation results show that wheeling cost and total cost of each generators, Pn will be
decreased once the generators participated the game. In other words, each generator will save some
costs P according to the game, as shown in Table 1. However, do the end users have any chances to
benefit from the game? Normally, different generators will have its own basic profit rate, which can
profit
be assumed as n
for each generator. Based on the total cost of each generators Pn , the electricity
payment from the end users to each generator Pneu can be written as:
profit

Pneu = (1 + n

) Pn

(22)
profit

In this case, the end users can save electricity payments as the amount of (1 + n )Pn for each
generator, which means the end users will save more electricity payment than the cost saved by the
generator once the generator does have profit. The total saved payment of the end users because the
N

profit

generators that participate in the game will be (1 + n

n =1

)Pn . As a result, the end users will be

benefited also according to the game.

profit
would be varied in deregulated
One has to notice that the profit rate of each generator n
environment of electricity market. However, regardless of the variance of the profit rate, the decrease
of the total cost could lead to lower electricity price for the generator, which will result in more
competitive in the market in practice.
5. Conclusions
In this paper, we have proposed a non-cooperative game based optimizing power transmission
scheduling for generators in a deregulated environment. The wheeling cost including congestion cost
has been selected as the key optimizing item. Detailed modeling, existence of the Nash equilibrium,
and algorithm to realize the optimization were presented. Simulations were performed to show
the effectiveness and feasibility of the proposed scheme and algorithm. Simulation results and
discussions show that all the generators, electric utilities, and aggregated end users can benefit from
the game: (1) for the generators participated in the game, the wheeling cost and total cost will be
decreased and, therefore, they will be more competitive in the market; (2) for the electric utilities,
the congestion can be alleviated effectively according to the game and the stability of the power
system can be improved; (3) for the end users, the electricity payments will be saved according to the
decreased total cost of the generators participated in the game.
Acknowledgments: The work is financially supported by the National Science Foundation of China (51577030),
the Excellent Young Teachers Programm of Southeast University (2242015R30024), and Six Talent Peaks Project
of Jiangsu Province (2014-ZBZZ-001).
Author Contributions: Bingtuan Gao contributed in developing the ideas of this research, Bingtuan Gao and
Tingting Ma performed this research. All of the authors were involved in preparing this manuscript.
Conflicts of Interest: The authors declare no conflict of interest.

13889

Energies 2015, 8, 1387913893

Nomenclature

N
N
M
M
q
R
W
Z
r
t
Tm
Qm,year
C (q)

qm,max
qm,all
qn,m
qnh
H
H
qn
xn,m
h
xn,m
min
n,m
max
n,m
Eh
Xn
qn,m
qm,all
cn,m
Pn

Set of indices of generators

Number of generators
Set of indices of transmission lines
Number of transmission lines
Transmitted electricity energy
Cost of transmission capacity
Cost of losses
Congestion cost
Discount rate
Recovery period
Investment cost of transmission line m
Predicted transmission capacity of transmission line m during one year
Generation cost
Line loss rate
Maximum transmission capacity of transmission line m
Power flow on transmission line m of all generators
Power flow related to generator n on transmission line m
Generated energy at time slot h
Set of indices of time slots
Number of time slots
Vector of generation scheduling for generator n in a whole day
Vector of power transmission scheduling
Proportion of energy transported by line m in total energy generated by generator n at time slot h
The minimum proportion
The maximum proportion
Total energy consumption of end users at each time slot h
Set of power transmission scheduling
Vector of electric quantity transported by transmission line m of each generator n
Vector of total power transmission of transmission line m
Electric quantity transported by transmission line m of all generators except generator n
Total cost of each generator n

Appendix A
Assume Xn R a (n N ) is a non-empty, convex and compact set. As shown in [29], if n,
Sn : X R1 is 1) graph-continuous, 2) upper semi-continuous in x, and 3) concave in xn , then the


game = N , {Xn }nN , {Sn }nN possesses a unique pure-strategy Nash equilibrium.
Note that Cn (qn ) is a quadratic function, it is strictly concave. Meantime, R + W (xn,m ) also is
strictly concave. Therefore, if Z (xn,m ) is concave, it can be guaranteed that function Sn is concave. In
order to simplify the calculation, we use xn,m in place of the vector xn,m , since xn,m and the vector xn,m
is same for process of proof. Similarly, cn,m is substituted for xn,m , and qn is substituted for qn . Let
f ( xn,m ) = ln

n,m qn
xn,m qn + cn,m qxm,max
e
xn,m qn

and
n,m =

qn
qm,max

13890

(A1)

(A2)

Energies 2015, 8, 1387913893

Therefore, we just need to prove f ( xn,m ) is concave. The partial derivative of f on xn,m is:


cn,m
xn,m qn + cn,m
f
= e xn,m n,m
+ n,m ln
xn,m
xn,m ( xn,m qn + cn,m )
xn,m qn
Then, we define:
g ( xn,m ) =

cn,m
xn,m qn + cn,m
+ n,m ln
xn,m ( xn,m qn + cn,m )
xn,m qn

thus,

f
= e xn,m n,m g ( xn,m )
xn,m

(A4)

(A5)

cn,m n,m
cn,m (2xn,m qn + cn,m )
g

=
2
2
xn,m
x
x
(
n,m
n,m qn + cn,m )
xn,m ( xn,m qn + cn,m )

(A3)

cn,m [2xn,m qn + cn,m n,m xn,m ( xn,m qn + cn,m )]

(A6)

2
2 (x
xn,m
n,m qn + cn,m )

Since the power transported by transmission line cannot be larger than the transmission capacity
of the line, n,m xn,m < 1. In addition, due to the fact that xn,m > 0, qn > 0,
2xn,m qn + cn,m n,m xn,m ( xn,m qn + cn,m ) > xn,m qn + cn,m n,m xn,m ( xn,m qn + cn,m ) > 0

(A7)

Namely,
g
>0
xn,m

(A8)

g
2 f
= n,m e xn,m n,m g ( xn,m ) + e xn,m n,m
2
xn,m
xn,m

(A9)

We can get:

It means that, f ( xn,m ) is concave when

n,m g ( xn,m ) +

g
>0
xn,m

(A10)

Hence, Equation (A10) is another constraint of the non-cooperative game, and the game


= N , {Xn }nN , {Sn }nN possesses a unique pure-strategy Nash equilibrium when it satisfies
all required conditions. In other words, the non-cooperative game among generators can obtain an
equilibrium when all constraints proposed in our paper are satisfied.
Appendix B
For notational simplicity, let x1 , , xN denote the optimal solution of the problem in
Equation (19). Besides, define:
!
P =

h =1

h
xn,m

(A11)

mM nN

By the definition of optimality, for each generator n N , one can see that
P

h =1

mM i N \{n}

h
xi,m
+

h
xn,m

(A12)

mM

together with the payoff definition of the game in Equation (20), multiplying both sides in
Equation (A12) by -1, one can get the necessary and sufficient condition of Nash equilibrium in

13891

Energies 2015, 8, 1387913893

Equation (21) directly. Note that game has a unique Nash equilibrium. Therefore, the Nash
equilibrium of game is the optimal solution of problem in Equation (19).
References
1.

2.
3.
4.
5.

6.

7.
8.
9.
10.
11.

12.
13.
14.
15.
16.
17.
18.
19.
20.
21.

22.

Electricity price reform project group of economic and technical research institute of Hubei electric power
company. Enlightenment of reform pilot of transmission and distribution price in Shenzhen. China Power
Enterp. Manag. 2015, 13, 5863. (in Chinese)
Lee, W.J.; Lin, C.H.; Swift, L.D. Wheeling charge under a deregulated environment. IEEE Trans. Ind. Appl.
2001, 37, 178183.
Happ, H.H. Cost of wheeling methodologies. IEEE Trans. Power Syst. 1994, 9, 147156.
Bashian, A.; Hojat, M.; Javidi, M.H.; Golmohamadi, H. Security-based tariff for wheeling contracts
considering fair congestion cost allocation. J. Control Autom. Electr. Syst. 2014, 25, 368380.
Kumar, N.; Reddy, Y.R.V.; Das, D.; Padhy, N.P. Transmission cost allocation by using MW-Mile approaches
in a restructured Indian power system. In Proceedings of the Power and Energy Society General Meeting,
Minneapolis, MN, USA, 2529 July 2010; pp. 18.
Orfanos, G.A.; Georgilakis, P.S.; Hatziargyriou, N.D. A more fair power flow based transmission cost
allocation scheme considering maximum line loading for N-1 security. IEEE Trans. Power Syst. 2013, 28,
33443352.
Shirmohammadi, D.; Gribik, P.R.; Law, E.T.K.; Malinowski, J.H.; ODonnell, R.E. Evaluation of transmission
network capacity use for wheeling transactions. IEEE Trans. Power Syst. 1989, 4, 14051413.
Gu, C.; Li, F. Long-run marginal cost pricing based on analytical method for revenue reconciliation.
IEEE Trans. Power Syst. 2011, 26, 103110.
Rao, M.S.S.; Soman, S.A. Marginal pricing of transmission services using min-max fairness policy.
IEEE Trans. Power Syst. 2015, 30, 573584.
Rubio-Oderiz, F.J.; Perez-Arriaga, I.J. Marginal pricing of transmission services: A comparative analysis of
network cost allocation methods. IEEE Trans. Power Syst. 2000, 15, 448454.
Nojeng, S.; Hassan, M.Y.; Said, D.M.; Abdullah, M.P.; Hussin, F. Improving the MW-Mile method using
the power factor-based approach for pricing the transmission services. IEEE Trans. Power Syst. 2014, 29,
20422048.
Monsef, H.; Jaefari, M. Transmission cost allocation based on use of reliability margin under contingency
conditions. IET Gener. Transm. Distrib. 2009, 3, 574585.
Molina, Y.P.; Saavedra, O.R.; Amaris, H. Transmission network cost allocation based on circuit theory and
the Aumann-Shapley method. IEEE Trans. Power Syst. 2013, 28, 45684577.
Pandya, K.S.; Joshi, S.K. Sensitivity and particle swarm optimization-based congestion management.
Electr. Power Compon. Syst. 2013, 41, 465484.
Hazra, J.; Sinha, A.K. Congestion management using multiobjective particle swarm optimization.
IEEE Trans. Power Syst. 2007, 22, 17261734.
Pandi, V.R.; Biswas, A.; Dasgupta, S.; Panigrahi, B.K. A hybrid bacterial foraging and differential evolution
algorithm for congestion management. Eur. Trans. Electr. Power 2010, 20, 862871.
Panigrahi, B.K.; Pandi, R.V. Congestion management using adaptive bacterial foraging algorithm.
Energy Convers. Manag. 2009, 50, 12021209.
Singh, K.; Yadav, V.K.; Padhy, N.P.; Sharma, J. Congestion management considering optimal placement of
distributed generator in deregulated power system networks. Electr. Power Compon. Syst. 2014, 42, 1322.
Singh, H.; Hao, S.Y.; Papalexopoulos, A. Transmission congestion management in competitive electricity
markets. IEEE Trans. Power Syst. 1998, 13, 672680.
Erli, G.; Takahasi, K.; Chen L.; Kurihara I. Transmission expansion cost allocation based on cooperative
game theory for congestion relief. Electr. Power Energy Syst. 2005, 27, 6167.
Kiani, A.; Annaswamy, A. Wholesale energy market in a smart grid: Dynamic modeling and stability.
In Proceedings of the 50th IEEE Conference on Decision and Control and European Control Conference,
Orlando, FL, USA, 1215 December 2011; pp. 22022207.
Bejestani, A.K.; Annaswamy, A. A dynamic mechanism for wholesale energy market: Stability and
robustness. IEEE Trans. Smart Grid 2014, 5, 28772888.

13892

Energies 2015, 8, 1387913893

23.

24.

25.
26.
27.
28.
29.

Papavasiliou, A.; Hindi, H.; Greene, D. Market-based control mechanisms for electric power demand
response. In Proceedings of the 49th IEEE Conference on Decision and Control, Atlanta, MS, USA,
1517 December 2010; pp. 18911898.
Tadahiro, T.; Koki, K.; Yoshiro, F.; Tomohiro, T.; Shiro, Y. Automated dinear function submission-based
double auction as bottom-up real-time pricing in a regional prosumers electricity network. Energies 2015,
8, 73817406.
Luo, Q.; Lu, L. Wheeling price calculating method in the price of direct power transmission to large
customers. Electr. Power Sci. Eng. 2008, 24, 3841. (in Chinese)
Kang, M.-S. Generation cost assessment of an isolated power system with a fuzzy wind power generation
model. IEEE Trans. Energy Convers. 2007, 22, 397404.
Frihauf, P.; Krstic, M.; Basar, T. Nash equilibrium seeking in noncooperative games. IEEE Trans. Autom.
Control 2012, 57, 11921207.
Wang, Y.P.; Saad, W.; Zhu, H.; Poor, H.V.; Basar, T. A game-theoretic approach to energy trading in the
smart grid. IEEE Trans. Smart Grid 2014, 5, 14391450.
Dasgupta, P.; Maskin, E. The existence of equilibrium in discontinuous economic games, I: Theory.
Rev. Econ. Stud. 1986, 53, 126.
c 2015 by the authors; licensee MDPI, Basel, Switzerland. This article is an open

access article distributed under the terms and conditions of the Creative Commons by
Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

13893