SPECIAL SECTION ON KEY ENABLING TECHNOLOGIES FOR PROSUMER ENERGY MANAGEMENT

Received September 3, 2020, accepted September 7, 2020, date of publication September 10, 2020,
date of current version September 24, 2020.
Digital Object Identifier 10.1109/ACCESS.2020.3023184

An Optimal Generation Scheduling

Approach Based on Linear Relaxation
and Mixed Integer Programming
YUNKAI LEI1 , (Member, IEEE), FANG LIU1 , AO LI1 , YUNCHE SU1 ,
XINTING YANG1 , AND JIEHUI ZHENG 2 , (Member, IEEE)
1 State Grid Sichuan Economic Research Institute, Chengdu 610000, China
2 School of Electric Power Engineering, South China University of Technology, Guangzhou 510640, China
Corresponding author: Jiehui Zheng ([email protected])
This work was supported in part by the State Grid of China under Grant SGSCJY00GHJS2000034.

ABSTRACT This paper proposes an optimal generation scheduling approach based on linear relaxation
and mixed integer programming, which is used to solve the generation dispatch problem. The quadratic
transmission loss constraint of each transmission line is converted into linear constraints by using the
linear relaxation and mixed integer programming technique. Consequently, the original optimal generation
scheduling problem is formulated as a quadratic programming or mixed integer quadratic programming
problem that can be solved by commercial optimization solver. In order to improve the efficiency of
algorithm, this paper further analyses the generation scheduling model and deletes the redundant variables
and constraints. Three test systems, including IEEE 30-node system, IEEE 118-node system, and Polish
2746-node system, are employed to examine the effectiveness of the proposed method. The comparative
results obtained by the proposed method, quadratically constrained quadratic programming method (QCQP),
and solving constraint integer programs solver (SCIP) verify the effectiveness of the proposed method in
solving the optimal generation scheduling problem.

INDEX TERMS Optimal generation scheduling, transmission losses, quadratic programming, mixed integer
programming, linear relaxation, prosumer energy management.

I. INTRODUCTION dispatching strategy is still the main dispatching method,

As a major energy consumer, the power industry’s slight when dispatching the generation plan, the dispatching depart-
improvement in generation scheduling will have a significant ment simply converts network losses into forecasted loads in
impact on the national economy and social environment. The a certain proportion, but lacks of accurate network loss model
energy distribution in China has obvious regional character- [3]. In order to determine an accurate generation scheduling
istics [1]. High-voltage long-distance transmission plays an plan, it is necessary to take network losses into account in the
important role in the transmission structure, and transmission generation scheduling model [4], [5].
loss becomes a key factor of generation scheduling. On the Many researches have been done on the generation
other hand, one important direction of renewable energy scheduling of traditional power grids considering renewable
utilization-power generation and grid connection, has a grow- energy generation and grid connection [6]–[11]. In this paper,
ing proportion in the traditional power system, which brings the above problems will be classified and explained from
great challenges to optimal the generation scheduling prob- two aspects: model and algorithm. In terms of generation
lem [2]. The growing penetration of DERs has made it pos- scheduling optimization models, the DC optimal power flow
sible for traditional passive consumers to evolve into active model and the AC optimal power flow model are often
prosumers. Compared with traditional consumers, prosumers used. (1) The first scheduling model is focus on reason-
are capable of managing their energy generation, storage ably arrange the power generation of different power plant
and consumption simultaneously. In China, the centralized generators on the premise of satisfied the load demand to
achieve the effect of allocating electricity and saving energy
The associate editor coordinating the review of this manuscript and and reducing emissions [7], [8], [10]. However, this model
approving it for publication was Huai-Zhi Wang . lacks consideration of the influence of network topology on

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
VOLUME 8, 2020 168625
 Y. Lei et al.: Optimal Generation Scheduling Approach Based on Linear Relaxation and Mixed Integer Programming

the scheduling scheme. For this scheduling model, the com- II. GENERAL FRAMEWORK
monly used optimization methods include equal increment The real power leaving out of a node by using DC power flow
rate criterion, dynamic programming, linear and quadratic can be shown as (1).
programming methods [11]. (2) The DC optimal power flow (
Plij = Vi2 Gij − Vi Vj Gij cos θij + Bij sin θij


model is a simple and efficient generation scheduling model (1)
Plji = Vj2 Gij − Vj Vi Gij cos θji + Bij sin θji


[12], [13]. This model considers the transmission constraints
of each line, but the model is based on a lossless network where lij is the end bus i and j of a transmission line, Vi and
and lacks consideration of transmission loss constraints of Vj are voltage magnitudes of bus i and j while θij = θi − θj
each transmission line on the generation scheduling plan. and θi and θj are voltage angles of bus i and j. Gij and Bij
Therefore, this scheduling plan often has a large deviation. are the conductance and susceptance of a transmission line,
(3) The AC optimal power flow model is a more compre- respectively.
hensive and widely used generation scheduling model [14], However, in real engineering, the injection active power
[15], which aims at minimizing total power generation cost, flow of the line is should be calculated as follows:
total coal consumption, or network loss, the constraints the
transmission capacity and voltage amplitude of each line are
2
Plij − Plji ≈ −2Vij,N Bij sin θij
taken into consideration. However, it is difficult to accu- 2 θ
Vij,N
xij ij
rately predict the reactive load in the day-ahead scheduling
2
≈ 2Vij,N sin θij ≈ 2 (2)
rij2 + xij2 xij
of generation plan, which limits the application of the AC
optimal power flow model to a certain extent [16]. Moreover, where Vij,N is rated voltage, rij and xij are the resistance and
the model often needs to be solved by non-linear program- reactance of the transmission line, respectively and rij  xij .
ming methods such as the interior point method which the However, the real power flow of transmission line lij can be
efficiency and robustness still need to be strengthened. approximately expressed as the average of the real power
In view of the influence of the network loss of renewable difference between the two ends of the transmission line
energy generation on the generation scheduling problem, [17] that can be expressed as (3). The transmission loss can be
and [18] used the Kron formula to calculate the network calculate by adding the both ends of a transmission line lij ,
loss, but this method can only roughly calculate the value which is show as (4). When we introduce a transmission loss
of the network loss, and the transmission loss coefficient variable Plij into the (4) that can be expressed as (5).
of Kron formula needs to be recalculated as the operating Ply − Plji 2 θ
Vij,N
f ij
conditions of the system change. Reference [4] proposed Plij ≈ ≈ (3)
a dynamic linear segment method to solve the problem of 2 xij
 
generation scheduling considering network loss. This method Plij + Plji ≈ Gij Vi2 + Vj2 − 2Vi Vj cos θij
linearizes the transmission loss curve of each transmission
line in a dynamic way until the mismatch satisfies the con-
2
≈ Vij,N Gij θij2 (4)
vergence conditions. Reference [5] relaxed the transmission Pllij = 2
Vij,N Gij θij2 (5)
line loss constraints with quadratic equality to the quadratic
inequality constraints based on the relaxation technique. For III. PROBLEM STATEMENT
the transmission line loss constraints that do not satisfied A. OBJECTIVE FUNCTIONS
the equality constraints, the corresponding transmission line The generator cost curve Fi (Pi ) are represented by quadratic
loss variables are dealt with by adding the terms of penalty. function and the total fuel cost FT that can be expressed as (6)
By employing of the terms of penalty, this method tends and (7)
to minimize network loss, which changes the nature of the Fi (Pi ) = ai P2i + bi Pi + ci (6)
original problem [19]. N
X
With the proportion of renewable energy generation min FT = Fi (Pi ) (7)
increasing, in order to effectively solve the genera- i=1
tion scheduling problem considering network losses, this
Fi (Pi ) is ith generator cost, ai , bi and ci are the cost coeffi-
paper proposes an optimal generation scheduling approach
cients of the ith generator, and Pi is the real power output of
based on linear relaxation and mixed integer program-
the ith generator, where N is the number of generators.
ming. Three test systems, including IEEE 30-node system,
IEEE 118-node system, and Polish 2746-node system, B. CONSTRAINTS
are employed to test the effectiveness of the proposed 1) EQUALITY CONSTRAINTS
method. The comparative results obtained by the proposed These constraints represent typical load flow equations as
method, quadratically constrained quadratic programming follows [1].
method (QCQP), and solving constraint integer programs X X f
X 1X
solver (SCIP) verify the effectiveness of the proposed method Pg − Pj = Pdk + Pl
g∈Gi j∈Li k∈Di 2 j∈Li j
in solving the optimal generation scheduling problem. i = 1, 2, · · · , NB (8)

168626 VOLUME 8, 2020

Y. Lei et al.: Optimal Generation Scheduling Approach Based on Linear Relaxation and Mixed Integer Programming

2 θ
Vl,N
f l
Pl = l = 1, 2, · · · , NL (9)
xl
2
Plj = Vl,N Gl θl2 l = 1, 2, · · · , NL
(10)
θref = 0 (11)
while (8) and (9) represent the balance limits of real power
and the constraint of the DC power flow. Gi , Li and Di are
the sets of generators, transmission lines and loads. Pg is the
real power output of the gth generator, Pdj is the real power
f
of the kth consumer, Pj and Plj denote the real power flow FIGURE 1. Linear relaxation technique for transmission loss of each line.

and the transmission loss of the line. θl = θi − θj , is the

phase difference between the voltages at both ends of the line, the quadratic equation constraints are relaxed into a
transmission line l. NB is the total number of buses whereas series of linear inequality constraints, thereby mathematically
the total number of transmission lines is defined by NL . The transforming the original optimization scheduling problem
transmission loss constraint of the transmission line l can be into a quadratic convex programming problem, which is
illustrated in (10), where Gl is conductance of transmission expressed as:
line l. The reference bus voltage phase angle constraint is 1
min x T Hx + f T x
presented by (13). x 2
s.t. Ax ≤ b
2) INEQUALITY CONSTRAINTS
Aeq x = beq
These constraints represent the system operating constraints
lb ≤ x ≤ ub (18)
as follows.
f
For secure operation, the transmission line loading Pl is where H and f are the coefficient matrix and coefficient
restricted by its upper limit as: vector of the objective function. A and b represent the coef-
f
Pl ≤ Pl ≤ P̄l l = 1, 2, · · · , NL (12) ficient matrix and coefficient vector of the linear inequality
constraints. Aeq and beq represent the coefficient matrix and
The node voltage phase difference constraint at both ends coefficient vector of the linear equality constraints. lb and ub
of the transmission line l is shown in the following formula: are upper and lower bound of the independent variable,
θl,min ≤ θl ≤ θl,max (13) respectively.

The ramp-up/down rate of the ith generator limited by its B. MIXED INTEGER PROGRAMMING FOR TRANSMISSION
physical characteristics [20]. It must satisfy the following LINE LOSS
constraint: In order to deal effectively with the situation that the marginal
cost of negative nodes exists in power grid, independent vari-
−DRi ≤ Pi − P0i ≤ URi i = 1, 2, · · · , N (14)
ables, the relaxed linear inequality constraints, which do not
The ith generator real power output Pi is restricted by its satisfied the given threshold after solving by quadratic pro-
lower and upper limits as follows: gramming, the following linear processing is further carried
Pi,min ≤ Pi ≤ Pi,max i = 1, 2, · · · , N (15) out in this paper.
The network loss constraints of each transmission line
treated is handled by linear approximation, i.e. transmission
IV. PROPOSED APPROACH
line loss constraints in the form of quadratic equations is
A. LINEARIZATION RELAXATION FOR TRANSMISSION
transformed into linear constraints as shown in Figure 2. The
LOSS CONSTRAINT
linearized transmission line loss constraints are as follows:
Let mlij is the total number of points that are taken on the
Rl
lj
network loss curve of the transmission line lij , which is show X
Pllij 2
Uij,r θij,r + Wij,r Zij,r


in Figure 1. The kth point is linearized to obtain the following = Vij,N Gij (19)
relaxed linear inequality: r=1

Pllij ≥ Klkij θij + Bklij k = 1, 2, . . . , mlij (16) Rlij represents the number of linear segments of the voltage
phase difference between the node i and j at the transmission
where Klkij and Bklij can be calculate as follows: line lij . Uij,r and Wij,r are the slope and intercept of the
(
2 G θk
linear segment r, respectively. Zij,r is the binary variable that
Klkij = 2Vij,N ij ij corresponding to the continuous sub-variable θij,r .
(17)
Bklij = Pl,k
lij − Klij θij
k k   
ij θij,r + θij,r
2 G
Uij,r = Vij,N ub 1b
After performing the above-mentioned linear relaxation (20)
W = V 2 G θ lb 2 − U θ lb,
 
processing on the transmission line loss constraints of each ij,r ij,N ij ij,r ij,r ij,

VOLUME 8, 2020 168627

 Y. Lei et al.: Optimal Generation Scheduling Approach Based on Linear Relaxation and Mixed Integer Programming

both ends of each line, which is shown below:

( ) ( )
Pl xl P̄l xl
max 2
, θl,min ≤ θl ≤ min 2
, θl,max (24)
Vl,N Vl,N
By merging the DC power flow constraint (9) into the node
power balance constraint (8) to remove the DC power flow
constraint, and the following node power balance constraint
is obtained as follows:
X Vj,N2 θ
X j X 1X l
Pg − = Pdk + Pj (25)
xj 2
g∈Gi j∈Li k∈Di j∈Li
FIGURE 2. Linear relaxation technique for transmission loss of each line. After the above processing, the real power flow variables,
the DC power flow constraint of the network and the security
θij,r
lb and θ ub represent the lower and upper bounds of the
ij,r transmission capacity constraint are deleted, so the efficiency
voltage phase difference of linear segment r, respectively. of the algorithm can be effectively improved.
After linearizing the transmission line loss constraints,
D. PROCEDURE SOLVING FOR GENERATION
since the voltage phase difference between the nodes at both
SCHEDULING PROBLEM
ends of the transmission line can only be located in one of
the linear segments, binary variables need to be introduced As a consequence, according to the description stated in the
and the following constraints are added: previous section, the procedure for the generation dispatch
Rl
problem is summarized as follows:
ij
X • Step1: Simplify the real optimal scheduling model by
Zij,r = 1 (21) using the deletion method to reduced model variables
r=1 and constraints of the network.
Constrained by (21), when the voltage phase difference • Step2: Linearize the relaxed transmission line loss con-
between the two ends of the line is in a linear segment, straints and use a commercial optimization solver to
the continuous sub-variable corresponding to the other linear solve the quadratic convex programming model. If all
segments is set to zero as follows: transmission line loss constraints satisfy the equality
θij,r
lb
Zij,r ≤ θij,r ≤ θij,r
ub
Zij,r (22) constraints or a given threshold valves, and go to step5;
otherwise, go to step3.
Constrained by (21) and (22), the voltage phase difference • Step3: The mixed integer programming technique is
between the two ends of the transmission line should be used to deal with the transmission line loss constraints
the sum of each continuous sub-variable θij,r , which can be that do not satisfy the equality constraints or given
expressed as follows: thresholds, then employ a commercial optimization
Rlj
X solver solve this mixed integer quadratic programming
θij = θij,r (23) model.
r=1 • Step4: If all transmission line loss constraints meet the
If some transmission line loss constraints solving by the equality constraints or a given threshold, and go to step5;
quadratic programming do not satisfy the constraints or otherwise, skip to step3;
exceed the threshold values after relaxing the transmission • Step5: Stop and output the results.
line loss constraints, we need to modify the transmission line It is noteworthy that the mixed integer quadratic programming
loss constraints into (19), (20) and (22) forms to obtain a model needs to be handled only when the marginal cost of
mixed integer quadratic programming problem. negative nodes exists in network. In most cases, in the process
of optimizing the unit’s power generation cost, the transmis-
C. DELETION METHOD OF MODEL VARIABLES AND
sion loss is also minimized, and the transmission loss con-
CONSTRAINTS
straint usually satisfies the equality constraints or less than
According to (9), the real power flow of each transmission a given threshold, so there is only a quadratic programming
line is expressed as a function of the voltage phase difference problem needs to be solved.
between the nodes at both ends of the transmission line.
f
Therefore, the real power flow Pl (l = 1, 2, · · · , NL ) of each V. SIMULATION STUDIES
transmission line is not taken as an independent variable in the In this paper, MATLAB R2017b is used and the MIP opti-
real power optimal scheduling model in this paper. Accord- mizer of CPLEX software is used to solve the proposed
ingly, the security transmission capacity constraints (10) of quadratic programming model and mixed integer quadratic
each transmission lines can be deleted. Moreover, the real programming model [21]. All methods in this paper are
power flow of each transmission line can be rewritten by implemented on a computer with 4G memory and 3.10-GHz.
correcting the voltage phase difference between the nodes at The examine time is averaged by 30 independent operations.

168628 VOLUME 8, 2020

Y. Lei et al.: Optimal Generation Scheduling Approach Based on Linear Relaxation and Mixed Integer Programming

TABLE 2. Comparison of results obtained by three methods on IEEE 30-

and 118-node systems.

TABLE 3. Comparison of results obtained by three methods on polish

2746-node system.

and IEEE 118-node system. These system-related datas,

IEEE30-node system and IEEE118-node system are come
from Matpower5.1.
The results obtained by QCQP, SCIP and the proposed
approach in this paper is show in Table 2. As shown in this
table, from the perspective of solution quality, both QCQP
and SCIP have the same global optimum solution on the two
test systems. Due to the proposed approach in this paper solv-
ing the line loss constraints in a linear relaxation technique,
there are little minor differences in the total generator cost
and network loss. However, from the point of view of total
generation cost and network loss, the simulation comparison
FIGURE 3. Flowchart of the proposed method. with QCQP and SCIP shows that the solution obtained by
proposed approach is close to the global optimum within the
TABLE 1. Essential characteristics of three test systems. range of error tolerance.
In terms of the time, for small-scale system, like
IEEE30-node test system, QCQP, SCIP and the proposed
approach all can obtain satisfactory solutions efficiently, and
there is little difference in the calculating time Compared
A. TEST SYSTEMS DESCRIPTION with the IEEE30-node system, the time of the IEEE118-node
Three test systems, including IEEE 30-node system, system of the three methods increases. As a sequence, it can
IEEE 118-node system, and Polish 2746-node system, are conclude that the calculation time of the proposed approach
employed to verify the effectiveness of the proposed method in this paper is much better that the QCQP and SCIP. Com-
in this paper, which is shown in Table 1. In order to examine pared with QCQP and SCIP, the advantage of the calculation
the effectiveness of the proposed approach, the following two efficiency of the proposed approach is mainly due to the
representative algorithms are selected for comparison: linear relaxation of transmission line loss constraints. After
1) QCQP: The transmission loss constraint is mathemati- linear relaxation of secondary linear transmission loss con-
cally relaxed into a quadratic constraint with convex. straints, the original problem is converted into a quadratic
In the process, the transmission line loss variables programming problem, which can be solved quickly by
which do not satisfy the equality constraints are added CPLEX.
to the objective function by the term of penalty fac-
C. CASE 2: POLISH 2746-NODE SYSTEMS
tor, and that will be solved by solving the quadratic
constraint quadratic programming problem [5]. To validate the extensibility of the proposed approach, we use
2) SCIP: This is a commercial optimization solver the Polish 2746-node systems. The system parameters are
for solving optimization problems with quadratic derived from the 2746-node system in Matpower 5.1, which
constraints, and it solver can find a global optimal corresponds to a load peak on a certain day in winter. Notes
solution quickly [22]. that this system is much larger than that of the aforementioned
IEEE30-node system and IEEE118-node system, and the
B. CASE 1: IEEE 30- AND 118-NODE SYSTEMS difficulty of solving the problem also increases.
In order to verify the effectiveness of the propose approach The results obtained by QCQP, SCIP and the proposed
in this paper, we first tests in IEEE30-node system method are shown in Table 3. From the Table 3, QCQP

VOLUME 8, 2020 168629

 Y. Lei et al.: Optimal Generation Scheduling Approach Based on Linear Relaxation and Mixed Integer Programming

and SCIP are very close in terms of total generation cost [8] Z. W. C. Bin, ‘‘A security-constrained energy-saving dispatch algorithm
and network loss, respectively. There is a little difference for optimal integrated coal consumption,’’ Power Syst. Protection Control,
vol. 38, no. 9, pp. 18–22, 2010.
between the proposed method in this paper and the two [9] N. Wang, L. Zhang, B. Liu, Q. Tang, and N. Cui, ‘‘Coordination mode
methods mentioned above. This is mainly due to the small of day-ahead generation schedule considering energy saving and real-time
mismatch between the transmission line loss variable and control,’’ Autom. Electr. Power Syst., vol. 36, no. 17, pp. 45–49, 2012.
[10] S. Lue, C. Li, Y. Cao, X. Li, Z. Zhang, Z. Cheng, and Y. Tan, ‘‘An energy-
the actual transmission line loss after the linear relaxation saving generation dispatch algorithm for thermal generators based on
of transmission line loss constraints. As the scale of system the integrated equal coal-consumption increment,’’ Proc. CSEE, vol. 32,
is enlarged, the error accumulates step by step and which no. 32, pp. 1–7, 2012.
[11] A. J. Wood and B. F. Wollenberg, Power Generation, Operation, and
will reach a very larger value. At the same time, in the opti- Control. New York, NY, USA: Wiley, 1984.
mization process, the value of transmission line loss variable [12] A. J. Conejo and J. A. Aguado, ‘‘Multi-area coordinated decentralized
is generally close to the relaxed linear constraints, which DC optimal power flow,’’ IEEE Trans. Power Syst., vol. 13, no. 4,
pp. 1272–1278, Nov. 1998.
makes the transmission line loss variable Plj less than or equal [13] X. Lai, L. Xie, Q. Xia, H. Zhong, and C. Kang, ‘‘Decentralized multi-area
2 G θ 2 , and it results in the total
to the actual line loss Vl,N economic dispatch via dynamic multiplier-based Lagrangian relaxation,’’
l l
IEEE Trans. Power Syst., vol. 30, no. 6, pp. 3225–3233, Nov. 2015.
output of proposed method is slightly less than the QCQP and [14] F. Capitanescu, M. Glavic, D. Ernst, and L. Wehenkel, ‘‘Interior-point
SCIP. When properly increasing the number of linear slack based algorithms for the solution of optimal power flow problems,’’ Electr.
segments of the line loss curve that will help to reduce the Power Syst. Res., vol. 77, nos. 5–6, pp. 508–517, Apr. 2007.
[15] R. A. Jabr, ‘‘Optimal power flow using an extended conic quadratic formu-
mismatch between the proposed method and QCQP and SCIP lation,’’ IEEE Trans. Power Syst., vol. 23, no. 3, pp. 1000–1008, Aug. 2008.
in the total power generation cost and network loss, but the [16] H.-Z. Wang, G.-Q. Li, G.-B. Wang, J.-C. Peng, H. Jiang, and Y.-T. Liu,
calculation time will increase. ‘‘Deep learning based ensemble approach for probabilistic wind power
forecasting,’’ Appl. Energy, vol. 188, pp. 56–70, Feb. 2017.
[17] A. I. Selvakumar and K. Thanushkodi, ‘‘A new particle swarm optimization
VI. CONCLUSION solution to nonconvex economic dispatch problems,’’ IEEE Trans. Power
This paper proposed a generation scheduling algorithm based Syst., vol. 22, no. 1, pp. 42–51, Feb. 2007.
[18] A. Rabiee, B. Mohammadi-Ivatloo, and M. Moradi-Dalvand, ‘‘Fast
on linear relaxation and mixed integer programming. This dynamic economic power dispatch problems solution via optimality condi-
method using linear relaxation and mixed integer program- tion decomposition,’’ IEEE Trans. Power Syst., vol. 29, no. 2, pp. 982–983,
ming model is used to deal with the transmission line Mar. 2014.
[19] H. Wang, Z. Lei, X. Zhang, B. Zhou, and J. Peng, ‘‘A review of deep
loss constraints, transforms the original power optimization learning for renewable energy forecasting,’’ Energy Convers. Manage.,
scheduling problem into a quadratic planning or mixed inte- vol. 198, Oct. 2019, Art. no. 111799.
ger quadratic programming problem, and solved by a com- [20] H. Z. Wang, G. B. Wang, G. Q. Li, J. C. Peng, and Y. T. Liu, ‘‘Deep
belief network based deterministic and probabilistic wind speed forecast-
mercial optimization solver, CPLEX. IEEE 30-node system, ing approach,’’ Appl. Energy, vol. 182, pp. 80–93, Nov. 2016.
IEEE 118-node system and Polish 2746-node system are used [21] The IBM ILOG CPLEX Optimization Studio CPLEX Userąŕs Manual,
to test the proposed problem. The validity and expansibility IBM, New York, NY, USA, 2016.
[22] T. Achterberg, ‘‘SCIP: Solving constraint integer programs,’’ Math.
of the proposed method are verified, and compared with the Program. Comput., vol. 1, no. 1, pp. 1–41, Jul. 2009.
two algorithms, QCQP and SCIP. The final results shows
that the method proposed in this paper can obtain an optimal YUNKAI LEI (Member, IEEE) received the Ph.D. degree in electrical engi-
scheduling solution efficiently. neering from Tianjin University, Tianjin, China, in 2018.
He is currently working as the Grid Planner with the State Grid Sichuan
Economic Research Institute. His research interests include reliability and
REFERENCES
risk assessments of power system and the integrated energy systems.
[1] H. Wang, Y. Liu, B. Zhou, C. Li, G. Cao, N. Voropai, and E. Barakhtenko,
‘‘Taxonomy research of artificial intelligence for deterministic solar
FANG LIU is an engineer. Her research interests include optimization and
power forecasting,’’ Energy Convers. Manage., vol. 214, Jun. 2020,
control methods and their applications in power systems.
Art. no. 112909.
[2] X. Fu, Q. Guo, and H. Sun, ‘‘Statistical machine learning model for
stochastic optimal planning of distribution networks considering a dynamic AO LI is an engineer. His research interests include optimization and control
correlation and dimension reduction,’’ IEEE Trans. Smart Grid, vol. 11, methods, and the coordinated operation of integrated energy systems.
no. 4, pp. 2904–2917, Jul. 2020.
[3] D. Xu, Q. Wu, B. Zhou, C. Li, L. Bai, and S. Huang, ‘‘Distributed YUNCHE SU is an engineer. Her research interests include the areas of power
multi-energy operation of coupled electricity, heating and natural gas system stability analysis and control, the control of wind power generation
networks,’’ IEEE Trans. Sustain. Energy, early access, Dec. 23, 2019, doi: systems, and nonlinear control theory.
10.1109/TSTE.2019.2961432.
[4] T. N. dos Santos and A. L. Diniz, ‘‘A dynamic piecewise linear model XINTING YANG is an engineer. His research interests include nonlinear
for DC transmission losses in optimal scheduling problems,’’ IEEE Trans. adaptive control, power quality, and power system control and its operation.
Power Syst., vol. 26, no. 2, pp. 508–519, May 2011.
[5] H. Zhong, Q. Xia, Y. Wang, and C. Kang, ‘‘Dynamic economic dispatch
JIEHUI ZHENG (Member, IEEE) received the B.E. degree in electrical engi-
considering transmission losses using quadratically constrained quadratic
neering from the Huazhong University of Science and Technology, Wuhan,
program method,’’ IEEE Trans. Power Syst., vol. 28, no. 3, pp. 2232–2241,
Aug. 2013. China, in 2012, and the Ph.D. degree in electrical engineering from the South
[6] S. Jianhua and T. Sumei, ‘‘Algorithm of energy saving generation dis- China University of Technology (SCUT), Guangzhou, China, in 2017. He is
patch scheduling,’’ Autom. Electr. Power Syst., vol. 32, no. 24, pp. 48–51, currently a Lecturer with SCUT. He has authored or coauthored more than 40
2008. peer-reviewed SCI journal articles. His research interests include the area of
[7] C. Zhixu, X. Kai, and Z. Jing, ‘‘Optimal model and algorithm for day-ahead optimization algorithms, and decision making methods and their applications
generation scheduling of transmission grid security constrained convention in integrated energy systems.
dispatch,’’ Autom. Electr. Power Syst., vol. 33, no. 1, pp. 10–13, 2009.

168630 VOLUME 8, 2020