Swarm and Evolutionary Computation 38 (2018) 1–11

Contents lists available at ScienceDirect

Swarm and Evolutionary Computation

journal homepage: www.elsevier.com/locate/swevo

A survey on multi-objective evolutionary algorithms for the solution of the MARK

environmental/economic dispatch problems
⁎
B.Y. Qua,b, Y.S. Zhua,c, Y.C. Jiaoa, M.Y. Wua, P.N. Suganthand, J.J. Lianga,c,
a
School of Electrical and Information Engineering, Zhongyuan University of Technology, 450007, China
b
School of Information Engineering, Zhengzhou University, 450001, China
c
School of Electrical Engineering, Zhengzhou University, 450001, China
d
School of Electrical and Electronic Engineering, Nanyang Technological University, 639798, Singapore

A R T I C L E I N F O A BS T RAC T

Keywords: Development of efficient multi-objective evolutionary algorithms (MOEAs) has provided effective tools to solve
Evolutionary algorithms (EAs) environmental/economic dispatch (EED) problems. EED is a highly constrained complex bi-objective
Multi-objective evolutionary algorithms optimization problem. Since 1990s, numerous publications have reported the applications of MOEAs to solve
(MOEAs) the EED problems. This paper surveys the state-of-the-art of research related to this direction. It covers topics of
Environmental economic dispatch (EED)
typical MOEAs, classical EED problems, Dynamic EED problems, EED problems incorporating wind power,
Wind power
EED problems incorporating electric vehicles and EED problems within micro-grids. In addition, some
Electric vehicles
Multi-objective optimization potential directions for future research are also presented.

1. Introduction multi-objective EED optimization problems. Numerous multi-objective

In recent decades, the highly constrained nonlinear multi-objective
optimization problem known as environmental/economic dispatch
(EED) problem has attracted research efforts due to the increasing
concerns about environmental pollution. EED is a bi-objective problem
with two conflicting objectives which are the minimization of genera-
tion cost and pollution emission. Various approaches have been
reported in literature to handle the EED problem.
Initially, conventional optimization methods such as linear pro-
gramming techniques were mainly used as the optimizing tool for
solving the EED problem [1,2]. However, these methods are not
effective when the dispatch problem becomes complex. Hence, re-
searchers turned to artificial intelligent techniques especially evolu-
tionary algorithms (EAs) and swarm algorithms (SAs). These meta-
heuristics use mechanisms inspired by Darwinian Theory of biological
evolution and social interactions, respectively. The studies have shown
that EAs and SAs can effectively overcome most of the drawbacks of
classical method. EAs and SAs have been successfully adopted to solve
various kinds of power dispatch problems [3–11].
Since EAs and SAs use a population of solutions to conduct the
search process, multiple non-dominated solutions can be found in one
single run. Moreover, EAs and SAs require few domain information of
the given problem. These features are attractive for solving complex
evolutionary algorithms (MOEAs) have been suggested to solve the
EED problem [12–15]. The aim of this paper is to provide a broad view
of using MOEAs in EED applications and encourage researchers in
power application domains to benefit from further use of MOEAs. The
taxonomy adopted in this paper is based on the topics reviewed and it
is divided into 5 parts. The first one studies the classical EED problems
using MOEAs while the remaining sections reviews other types of the
EED applications.
The remainder of this paper is organized as follows. Section 2
provides a brief introduction of multi-objective optimization and the
state-of-the-art MOEAs. Sections 3–7 present MOEAs for the classical
EED problems, Dynamic EED problems, EED problems with wind
power, EED problems with electric vehicles and EED problems within a
micro-grid, respectively. The paper is concluded in Section 8.
2. Multi-objective optimization
In this part, the basic concepts of multi-objective optimization and
some typical MOEAs are introduced.
2.1. Formulation of multi-objective optimization problems
The multi-objective optimization problems (MOPs) can be mathe-

⁎
Corresponding author at: School of Electrical and Information Engineering, Zhongyuan University of Technology, 450007, China.
E-mail addresses: qby1984@hotmail.com (B.Y. Qu), zhuysy@163.com (Y.S. Zhu), yuechaojiao@163.com (Y.C. Jiao), mywu@zut.edu.cn (M.Y. Wu),
epsugan@ntu.edu.sg (P.N. Suganthan), liangjing@zzu.edu.cn (J.J. Liang).

http://dx.doi.org/10.1016/j.swevo.2017.06.002
Received 21 July 2016; Received in revised form 8 June 2017; Accepted 21 June 2017
Available online 23 June 2017
2210-6502/ © 2017 Elsevier B.V. All rights reserved.
B.Y. Qu et al.
matically formulated as follows assuming as minimization problems:
⎧ min y = F (x) = [ f (x ), f (x ), ..., f (x )]T
⎪ 1 2 m
⎨ s. t. gj (x ) ≤ 0, j = 1, 2, ..., p
⎪
⎩ hk (x ) = 0, k = 1, 2, ..., q
where x = (x1, x2, ..., xn ) ∈ X ⊂ Rn is the vector of decision variables,
which constitute the decision space X , and Rn is an n dimensional
Euclidean space; y = ( y1, y2, ..., ym) ∈ Y ⊂ Rm is the vector of objectives,
which constitute the objective space Y ;gj (x) ≤ 0( j = 1, 2, ..., p ) and
hk (x) = 0(k = 1, 2, ..., q ) are the constraint functions of the problem.
To solve multi-objective optimization problems, the following
concepts are essential.
Feasible solution set: For x ∈ X , if x satisfies all constraints, x
is a feasible solution and the set of all feasible solutions is the feasible
solution set, denoted as Xf , where Xf ⊆ X .
Domination: Given two feasible solutions x1 and x2 , we say that x1
dominates x2 (denoted as x1 ≺x2 ), if ∀ i ∈ {1, 2, ..., m}, fi (x1) ≤ fi (x2 ),
∃ i ∈ {1, 2, ..., m}, fi (x1) < fi (x2 ).
Pareto-optimal set: For a feasible solution x ⊆ X , if there does
not exist another feasible solution x′ ⊆ X satisfying x′≺x , we say that x
is non-dominated with respect to X , and this feasible solution x is
defined as a Pareto-optimal solution as x*. The set of all Pareto-optimal
solutions is defined as the Pareto-optimal set denoted as P*, i.
e.,P* = {x* ¬ ∃ x ∈ X : x≺x*}.
Pareto-optimal front: The Pareto-optimal front is defined as PF ,
where PF = {F (x*) = [ f1 (x*), f2 (x*), ..., fm (x*)]T x* ∈ P*}.
Based on the above concepts, obtaining the Pareto-optimal set is
the key task of multi-objective optimization algorithms.
2.2. Typical MOEAs for solving EED problem
EAs and SAs are stochastic optimization techniques inspired by the
natural evolutionary and swarming processes. Due to their own
properties, they are more suitable for solving MOPs than other
conventional mathematical techniques. Since early 1990s, researchers
proposed numerous Multi-objective Evolutionary Algorithms (MOEAs)
and used them to solve complex MOPs. In this section, we aim to
present a short review of some typical MOEAs especially those used for
solving the EED problems. These MOEAs are presented in the
chronological order.
(1) Vector Evaluated Genetic Algorithm (VEGA) [16]: VEGA is
commonly known as the first MOEA. It modifies the original
genetic algorithm to make it capable of handling multi-objective
optimization problems. In VEGA, the population is divided into
several subpopulations and the number of subpopulations is equal
to the number of objectives. Each subpopulation is responsible for
searching one objective. While the concept of this algorithm is
straightforward, the solutions obtained by this technique are
usually not uniformly distributed along the Pareto front especially
in the tradeoff regions.
(2) Non-dominated Sorting Genetic Algorithm (NSGA) [14,17]: NSGA
was introduced by Srinivas and Deb in 1994. This method uses
ranking selection and niching techniques to find the non-domi-
nated solutions and maintain the diversity of the population. Two
main steps are involved in this method known as fitness assign-
ment and fitness sharing. Fitness assignment helps fast conver-
gence while fitness sharing increase the diversity.
(3) Multi-Objective Stochastic Search Technique (MOSST) [13]: The
MOSST heuristic has been designed as a combination of real coded
genetic algorithm (GA) and simulated annealing (SA) techniques.
It incorporates a genetic crossover operator BLX-α and a problem
specific mutation operator with a local search heuristic to provide a
better search capability [13]. It can offer the advantages of both GA
and SA.
Swarm and Evolutionary Computation 38 (2018) 1–11
(4) Non-dominated Sorting Genetic Algorithm II (NSGA-II) [18]:
NSGA-II is the most popular multi-objective evolutionary algo-
rithm. The dominance concept is used in NSGA-II to sort/rank the
solutions. Moreover, it uses crowding distance to estimate the
(1) density of solutions near each solution. NSGA-II uses both the
non-domination rank and crowding distance to select individuals
to survive to the next generation.
(5) Niched Pareto Genetic Algorithm (NPGA) [19]: NPGA uses Pareto
domination tournaments selection scheme to find the good solu-
tions and remove the bad ones. Different from the method used in
[17], only two candidates are randomly picked for tournament
each time. To compare the two candidates, a randomly selected
comparison set is used. Then, the dominance of both individuals
with respect to the comparison set is checked. Sharing procedure
will be used if both individuals are dominated by the comparison
set.
(6) Multi-objective Particle Swarm Optimization (MOPSO) [20]:
MOPSO is a variation of the PSO to solve MOPs [21].
Determining global best (gbest) is the key issue of MOPSO.
MOPSO uses the non-dominated solutions as the basis of selecting
the gbest. The algorithm maintains two archives to save the global
best individuals found so far and the local bests, respectively. The
selection of a global best is based on roulette wheel selection of a
hypercube score [22].
(7) Multi-Objective Evolutionary Algorithm based on Decomposition
(MOEA/D) [23]: MOEA/D was introduced by Zhang et al. in 2007.
It provides a new framework to solve MOPs. MOEA/D handles an
MOP by decomposing it into numerous single objective sub-
problems and optimizes the sub-problems using evolutionary
approach collaboratively and concurrently.
(8) Multi-objective Differential Evolution (MODE) [24]: MODE is like
the basic DE algorithm expect the selection process. Generally,
MODE adopts the non-dominated sorting and ranking selection
methods developed by Deb et al. [18]. The non-dominated sorting
is performed on the combined population of new generated
offspring and parents and the selection is based on the non-
dominated rank and crowding distance. Other techniques like
summation based sorting and diversified selection were also
proposed and used in the literature [25].
3. Classical EED problems
3.1. Problem formulation
The prime target of traditionally electric power systems is to
schedule the outputs of the generators to meet the load requirement
with a minimum fuel cost regardless of emissions produced [1]. With
the increasing requirements for the environmental protection, alter-
native operational strategies are needed to reduce the pollution of the
electric power plants. Environmental/Economic Dispatch (EED) pro-
blems treat the pollution emissions and the fuel cost as two conflicting
objectives which are optimized simultaneously subjected to the prac-
tical constraints. Generally, the problem can be formulated as follows:
Fuel cost objective: The cost curves of the generators can be
represented by quadratic functions and the total fuel cost $/h can be
expressed as:
NG
min F (PG ) = ∑ (ai + bi PGi + ci PG2i )
i =1 (2)
th
where ai, bi, ci are the cost coefficients for the i thermal power
generator. F(PG) is the total fuel cost of the system while NG identifies
the number of thermal units. If the rippling effects produced by the
steam admission valve openings are considered, a sine component
needs to be added to Eq. (2) and expression becomes [12,26]:
2
B.Y. Qu et al.
NG
min F (PG ) = ∑ (ai + bi PGi + ci PG2i + di sin[ei (PGmin
i − PGi )] )
i =1
where di and ei are the cost coefficients of the ith generator while PGmin
i is
the minimum output of ith power generator.
Pollution emission objective: The total emission ton/h of various
pollutants are generally expressed as [2,26]:
NG
min E (PG ) = ∑ [10−2 (αi + βi PGi + γi (PGi )2) + ζi exp(φi PGi )]
i =1
where αi, βi, γi, ζi and φi are the emission coefficients of the ith power
generator and E(PG) is the total pollution emission of the system.
Power balance constraint: the total power generation must cover
the sum of total demands and the real power losses. This constraint can
be mathematically expressed as:
NG
∑ PGi − PD − Ploss = 0
i =1
Generation capacity constraint: The real power output of each
generator is restricted by the lower and upper limits to ensure stable
operation and these limits are also generally considered as the bounds
of the problems.
PGmin max
i ≤ PGi ≤ PGi
Security constraints: For a secure operation, the transmission line
loading Sl should be restricted by its upper limit [14,19]:
|Sli | ≤ Slmax , i = 1, 2, ..., NL
i
3.2. Early approaches and weighted sum approaches
3.2.1. Early approaches
In early days, the EED problems are generally handled by including
emissions either as a constraint or as a weighted function in the
objective of the overall dispatching problem [1]. In [1], several
methods/algorithms were introduced to solve the EED problems
including fuel switching techniques. Fuel switching is either switching
the types of the fuel or lowering the sulfur content in the fuel. In [2],
linear programming techniques which incorporated the techniques of
section reduction method and third simplex method were used to
tackle the EED problems. The methods were applied on the 10-bus
model system as well as the 30-bus model system and the results
showed that the proposed method could solve the problem effectively.
In [27,28], the EED problem was changed to a single objective problem
by considering the emission as a constraint. However, this method
cannot obtain the trade-off relations between the two objectives.
3.2.2. Weighted sum approaches
As mentioned in the previous section, linear combination of
different objectives as a weighted sum is another commonly used
approach [29–32] for solving multi-objective EED problems. In this
method, the objectives are transformed into a single-objective problem
by using suitable scalarization. This approach generally takes the
following form of objective function:
NG
min ∑ [ωFi (PGi ) + (1 − ω)∙λ∙Ei (PGi )]
i =1
where λ is the scaling factor which blends the emission objective with
the fuel objective and ω is the weight factor in the range of [0,1].
Although this approach is easy to implement, it requires multiple runs
by varying the weights to obtain the trade-off non-dominated solutions.
With the development of multi-objective evolutionary algorithms
(MOEAs), researchers frequently use MOEAs to solve the EED
problems as introduced in detail in the subsequent section.
Swarm and Evolutionary Computation 38 (2018) 1–11
3.3. MOEA approaches
(3)
In recent years, researchers have handled both objectives of the
EED problem simultaneously as competing objectives using MOEAs
[33]. The evolutionary algorithms have shown that they are able to
overcome most of the disadvantages of the classical methods [25–35].
Since evolutionary algorithms are population based optimization
techniques, multiple optimal solutions can be found in one single
run. This feature is attractive for solving the EED problems, as multiple
(4) solutions need to be located to form the non-dominated front. Using
MOEAs to solve the EED problems can be traced back to early 1990s
[12,13]. Since then MOEAs have become the main technique for
solving the EED problems. The aim of this section is to provide a
review of the works that used MOEAs to solve the EED problems in the
past few decades. These works are presented based on the search
techniques namely, Genetic Algorithm (GA) based approaches, Particle
Swarm Optimization (PSO) based approaches, Differential Evolution
(DE) based approaches, hybrid approaches and other evolutionary
(5)
approaches.
3.3.1. GA based approaches
In [14,36], a modified Non-dominated Sorting Genetic Algorithm
(NSGA) based method was proposed. It employed a diversity-preser-
ving technique to prevent premature convergence. This approach also
(6)
imposed a hierarchical clustering technique to generate a representa-
tive and manageable Pareto-optimal set. In 2003, similar diversity-
preserving and hierarchical clustering mechanisms were incorporated
into a Niched Pareto Genetic Algorithm (NPGA) [19]. This algorithm
(7)
used a Pareto domination tournament selection scheme to find the
non-dominated set of solutions. In 2006, Abido [33] performed a
comprehensive comparative discussion on using different MOEAs
(specifically, NSGA, NPGA and SPEA) to solve the EED problems. In
[37], a multi-objective genetic algorithm based on the concept of
epsilon-dominance was introduced to solve the EED problems. The
epsilon-dominance was applied to members in the archive where the
solutions are stored and it improved the solution quality and conver-
gence characteristics.
In [38], Harry and Robert used NSGA-II to overcome the weakness
of NSGA. Similarly, Guesmi et al. [39] also demonstrated the superior
performance of NSGA-II over NSGA and NPGA. In [40], a modified
NSGA-II that utilized the Dynamic Crowding Distance (DCD) and
Controlled Elitism (CE) techniques was proposed to overcome the
drawbacks of poor diversity of the original NSGA-II. The modified
NSGA-II was tested and compared on the IEEE 14-, 30-, 57- and 118-
bus systems. In [41], a similar NSGA-II method was used to solve the
EED problems.
3.3.2. PSO based approaches
PSO based methods are one of the most popular approaches for
solving the EED problems. In [20], an external memory based Multi-
objective Particle Swarm Optimization (MOPSO) was adopted. Two
external memories called repositories were maintained in addition to
the search population in this method. One was used to store the global
best individuals while the other one was for the local best of each
particle. This algorithm also used a geographically-based technique to
maintain the diversity of the obtained solutions. In 2009, Abiod also
adopted MOPSO technique to handle the EED problems [42]. In this
work, the global best and local best were redefined using the non-
(8)
domination concept which could lead the particles to locate the Pareto
front effectively. Similar methods can also be found in [43].
In 2007, Wang and Singh [44] proposed a Fuzzified Multi-Objective
Particle Swarm Optimization (FMOPSO) to preserve the diversity of
the solutions when dealing with the EED problems. This method
treated the global best not just as a point but as an area and each point
in this area had different possibilities to be selected as the global best. A
fuzzy clustering-based multi-objective particle swarm algorithm
3
B.Y. Qu et al.
(FCPSO) was proposed to solve the EED problem in 2008 [45]. This
algorithm incorporated niching mechanism to direct the particles to
search the less explored regions of the Pareto front. It also used a self-
adaptive mutation operator to enhance the diversity of the solutions.
In [47], a multi-objective chaotic particle swarm optimization
(MOCPSO) was developed and proven to be more effective than
conventional MOPSO for solving the EED problems. In 2012, a
parameter-free multi-objective particle swarm optimization algorithm
called bare-bones MOPSO (BB-MOPSO) was proposed to solve EED
problems [48]. The results showed BB-MOPSO outperforming seven
MOEAs and three well-known MOPSO techniques on the IEEE 30-bus
test system. Recently, Liu et al. [46] introduced a cultural quantum-
behaved particle swarm optimization (CMOQPSO) algorithm to solve
the EED problems. In this algorithm, the cultural evolution mechanism
was used to enhance its global searching ability while a local search
operator was used to enhance its local searching ability. CMOQPSO is
tested on the 6- and 40-generator systems, respectively to demonstrate
its effectiveness.
3.3.3. DE-based approaches
In [24], a differential evolution (DE) based approach was used to
solve the EED problem. To enable DE with the ability of handling
multi-objective optimization problems, non-domination sorting and
ranking selection procedures were adopted in the selection step. The
simulation results showed its superior performance over NSGA [33]
and SPEA [33]. Similarly, Basu [49] also demonstrated the effective-
ness of MODE algorithms for solving the EED problems. Wu et al.
presented a modified MODE that adopted an external elitist archive
and a crowding entropy strategy to store the non-dominated solutions
and maintain the diversity in 2010 [50]. This algorithm was used to
solve the IEEE 30-bus and 118-bus test systems and it was proven to be
effective in solving the EED problems. In [51], an enhanced multi-
objective differential evolution algorithm was proposed to solve the
EED problems. This algorithm adopted an elitist archive technique to
store the non-dominated solutions and the DE operators were modified
according to the characteristics of the problems. To prevent premature
convergence, this algorithm introduced a local random search operator
to improve its convergence performance.
3.3.4. Hybrid approaches
Hybridizing different EA techniques is a commonly used approach
to utilize the advantages of different EAs for dealing with complicated
optimization problems. Das and Patvardhan [13] introduced one of the
earliest hybrid multi-objective evolutionary algorithms called Multi-
Objective Stochastic Search Technique (MOSST) to solve the EED
problems. MOSST was designed as a combination of the real coded
genetic algorithms and simulated annealing. It also incorporated a local
search technique to provide a better search capability. However, this
approach suffers from severe premature convergence problem. In
2010, Gong et al. introduced a hybrid MOEA based on the techniques
of PSO and DE [52]. This algorithm utilized PSO as the global
searching tool while DE was used as the local search tool. The
combination of PSO and DE strengthens both the exploration and
exploitation of the algorithm when solving the EED problems. In 2010,
a modified NSGA-II which integrated a Convergence Accelerator
Operator (CAO) into the original NSGA-II was proposed by Alawode
to solve the EED problems [53]. The CAO consisted of two steps which
were a deterministic local improvement procedure and a neural net-
work based procedure. The modified NSGA-II was proven to be more
effective than the original NSGA-II. In [54,55] a combination of
differential evolution and biogeography-based optimization (BBO)
algorithm was developed to improve the convergence speed as well
as the solution quality when solving the EED problems. The algorithm
utilized the good exploration ability of DE and good exploitation ability
of BBO to improve its performance.
4
Swarm and Evolutionary Computation 38 (2018) 1–11
3.3.5. Other evolutionary methods
In [12], a heuristics-guided evolutionary method was introduced to
evaluate the economic impacts of environmental dispatching and fuel
switching. Although this method handled both objectives at the same
time, the stochastic competition between two individuals was only
based on one randomly chosen objective. This approach can produce
several alternatives along the Pareto-optimal front, but it cannot find
the entire non-dominated Front. In [26], Abido adopted a Strength
Pareto Evolutionary Algorithm (SPEA) based approach to tackle the
two conflicting objectives. This algorithm embedded the same diver-
sity-preserving mechanism and hierarchical clustering as in [14–19] to
improve the solutions’ quality.
In [42], a non-domination sorting based multi-objective evolution-
ary programming (MOEP) was proposed to show its superior perfor-
mance over NSGA-II. However, the non-domination sorting process is
computationally expensive and time consuming. To overcome this
problem, Qu et al. [56] used a fast sorting technique to replace the
complex non-domination sorting. In [57], MOEP was also used as the
optimization tool to solve the EED problems. However, the emissions
were split into two separated objectives known as NOx emission
objective and SO2 emission objective and the EED was changed to a
three-objective optimization problem.
In [58], a multi-objective chaotic ant swarm optimization
(MOCASO) method was developed to solve three different test power
systems. In 2010, Hota et al. presented a fuzzy based bacterial foraging
algorithm that utilized the successful foraging strategies of the global
optimum bacterium and showed its robustness and reliability in
handling the EED problems [59]. In the same year, Panigrahi et al.
[60] proposed a fuzzy dominance based bacterial foraging algorithm to
tackle the EED problems. This algorithm obtained the most recent
bacterial locations by using the chemotaxis process and the fuzzy
dominance based sorting technique was used to select the Pareto
optimal front. In 2011 [61], Panigrahi et al. further modified the multi-
objective bacteria foraging algorithm by using the intermediate bacteria
positions during chemotaxis.
In [62], Hemamalini and Simon developed a multi-objective
artificial bee colony algorithm which used the non-domination sort in
the ranking and selection process to solve the standard IEEE 30 bus
system [62]. In [63], a multi-objective directed bee colony optimization
algorithm was developed to handle the EED problems. This algorithm
is a hybrid version of deterministic search, multi-agent system and bee
decision-making process and it used a modified Nelder-Mead method
to find the optimal solutions. In [64], Niknam et al. handled the EED
problems using a modified shuffle frog leaping algorithm. This algo-
rithm used a strong mutation technique to increase the population
diversity. In 2011, Sivasubramani and Swarup developed a multi-
objective harmony search (MOHS) algorithm to solve the EED
problems [65]. The algorithm adopted a non-dominated sorting and
ranking procedure with dynamic crowding distance to locate and
maintain the Pareto front. IEEE 30 bus and 118 bus systems were
used to demonstrate the superior performance of the proposed algo-
rithm.
In addition to the above-mentioned methods, there are also many
other evolutionary and swarm methods proposed in the literature to
handle the EED problems in the recent years. These methods include
biogeography based optimization [66], Interactive honey bee mating
optimization [67], cultural algorithm [68], bacterial colony chemotaxis
optimization algorithm [69], adaptive clonal selection algorithm [70],
scatter search [71], quasi-oppositional teaching learning based opti-
mization [72], multi-objective backtracking search algorithm [73] and
so on.
3.3.6. Typical systems and results
In the works reviewed above, 7 systems are commonly used as case
studies and the descriptions of these systems are presented as follows:
System A: 6-generator 30-bus standard test system (Table 1).
B.Y. Qu et al. Swarm and Evolutionary Computation 38 (2018) 1–11

Table 1
Generator cost and emission coefficients of system A [36].

Generator a b c α β γ ζ φ PGmin PGmax

1 10 200 100 4.091 − 5.554 6.490 2.0e − 4 2.857 0.05 0.5

2 10 150 120 2.543 − 6.047 5.638 5.0e − 4 3.333 0.05 0.6
3 20 180 40 4.258 − 5.094 4.586 1.0e − 6 8.000 0.05 1.0
4 10 100 60 5.326 − 3.550 3.380 2.0e − 3 2.000 0.05 1.2
5 20 180 40 4.258 − 5.094 4.586 1.0e − 6 8.000 0.05 1.0
6 10 150 100 6.131 − 5.555 5.151 1.0e − 5 6.667 0.05 0.6

Table 2
Generator cost and emission coefficients of system B [13].

Generator a b c α β γ ζ φ PGmin PGmax

1 550 810 2.8 5.326 − 3.55 3.38 2.0e − 3 2.0 0 6.8

2 309 810 5.6 4.258 − 5.094 4.586 1.0e − 6 8.0 0 3.6
3 307 810 5.6 4.258 − 5.094 4.586 1.0e − 6 8.0 0 3.6
4 240 774 32.4 4.091 − 5.554 6.490 2.0e − 4 2.857 0.6 1.8
5 240 774 32.4 4.091 − 5.554 6.490 2.0e − 4 2.857 0.6 1.8
6 240 774 32.4 4.091 − 5.554 6.490 2.0e − 4 2.857 0.6 1.8
7 240 774 32.4 4.091 − 5.554 6.490 2.0e − 4 2.857 0.6 1.8
8 240 774 32.4 4.091 − 5.554 6.490 2.0e − 4 2.857 0.6 1.8
9 240 774 32.4 4.091 − 5.554 6.490 2.0e − 4 2.857 0.6 1.8
10 126 860 28.4 2.543 − 6.047 5.638 5.0e − 4 3.333 0.4 1.2
11 126 860 28.4 2.543 − 6.047 5.638 5.0e − 4 3.333 0.4 1.2
12 126 860 28.4 2.543 − 6.047 5.638 5.0e − 4 3.333 0.55 1.2
13 126 860 28.4 2.543 − 6.047 5.638 5.0e − 4 3.333 0.55 1.2

System B: 13-generator 57-bus system (Table 2). Table 4

System C: 3-generator system (Table 3). Generator cost and emission coefficients of system D [49].
System D: 6-generator system (Table 4).
Generator a b c α β γ PGmin PGmax
System E: 14-generator 118-bus system (Table 5).
System F: 40-generator system (Table 6). 1 756.80 3853.97 1524.7 1.3859 3.2767 4.19 0.1 1.25
System G: 10-generator system (Table 7). 2 451.33 4615.92 1058.7 1.3859 3.2767 4.19 0.1 1.5
To provide a clear view of the systems used in different works, 3 1049.33 4039.66 280.3 4.0266 − 5.4551 6.83 0.4 2.5
Table 8 summarizes the various methods mentioned in Section 3.3. As 4 1243.53 3830.55 354.6 4.0266 − 5.4551 6.83 0.35 2.1
5 1658.57 3632.78 211.1 4.2895 − 5.1116 4.61 1.3 3.25
can be seen from this table, system A (6-generator 30-bus standard test
6 1356.66 3827.04 179.9 4.2895 − 5.1116 4.61 1.25 3.15
system) is the most commonly used test system. Therefore, some
typical results on this system are provided in Tables 9–14. Tables 8–11
present the results on 6-generator 30-bus lossless system without Table 5
considering the security constraints while Tables 11–14 present the Generator cost and emission coefficients of system E [50].
results on 6-generator 30-bus considering losses and all constraints.
Generator a b c α β γ PGmin PGmax
Note that Tables 10, 14 present the best compromise results for both
cases. The best compromise results balance the two objectives and it is 1 150 189 0.50 0.016 − 1.500 23.333 300 50
commonly used to compare different algorithms on the EED problems. 2 115 200 0.55 0.031 − 1.820 21.022 300 50
The best compromise solution is obtained using a fuzzy-based mechan- 3 40 350 0.60 0.013 − 1.249 22.050 300 50
ism and membership value of each solution in the final non-dominated 4 122 315 0.50 0.012 − 1.355 22.983 300 50
5 125 305 0.50 0.020 − 1.900 21.313 300 50
set Fi computed using the following membership function [33]:
6 70 275 0.70 0.007 0.805 21.900 300 50
7 70 345 0.70 0.015 − 1.401 23.001 300 50
⎧ 1 if Fi ≤ Fimin 8 70 345 0.70 0.018 − 1.800 24.003 300 50
⎪
⎪ Fimax − Fi 9 130 245 0.50 0.019 − 2.000 25.121 300 50
μi = ⎨ max min if Fi max
< Fi < Fimin 10 130 245 0.50 0.012 − 1.360 22.990 300 50
⎪ Fi − Fi
⎪ Fi ≥ Fimax
11 135 235 0.55 0.033 − 2.100 27.010 300 50
⎩ 0 if (9) 12 200 130 0.45 0.018 − 1.800 25.101 300 50
13 70 345 0.70 0.018 − 1.810 24.313 300 50
where μi stands for the membership value of the ith function (Fi). For 14 45 389 0.60 0.030 − 1.921 27.119 300 50
each non-dominated solution k, the normalized membership value
( μ [k ]) is calculated using

Table 3
Generator cost and emission coefficients of system C [38].

Generator a b c α β γ PGmin PGmax

1 561 7.92 0.001562 0.04373254 − 9.4868099e − 5 1.4721848e − 7 150 600

2 310 7.85 0.00194 0.055821713 − 9.7252878e − 5 3.0207577e − 7 100 400
3 78 7.97 0.00482 0.027731524 − 3.5373734e4 1.9338531e − 6 50 200

5
B.Y. Qu et al. Swarm and Evolutionary Computation 38 (2018) 1–11

Table 6
Generator cost and emission coefficients of system F [59].

Generator ai bi ci αi βi γi ςi λi Pmin Pmax

1 0.00690 6.73 94.705 0.048 − 2.22 60 1.31 0.0569 36 114

2 0.00690 6.73 94.705 0.048 − 2.22 60 1.31 0.0569 36 114
3 0.02028 7.07 309.54 0.0762 − 2.36 100 1.31 0.0569 60 120
4 0.00942 8.18 369.03 0.054 − 3.14 120 0.9142 0.0454 80 190
5 0.01140 5.35 148.89 0.085 − 1.89 50 0.9936 0.0406 47 97
6 0.01142 8.05 222.33 0.0854 − 3.08 80 1.31 0.0569 68 140
7 0.00357 8.03 287.71 0.0242 − 3.06 100 0.655 0.02846 110 300
8 0.00492 6.99 391.98 0.031 − 2.32 130 0.655 0.02846 135 300
9 0.00573 6.6 455.76 0.0335 − 2.11 150 0.655 0.02846 135 300
10 0.00605 12.9 722.82 0.425 − 4.34 280 0.655 0.02846 130 300
11 0.00515 12.9 635.20 0.0322 − 4.34 220 0.655 0.02846 94 375
12 0.00569 12.8 654.69 0.0338 − 4.28 225 0.655 0.02846 94 375
13 0.00421 12.5 913.40 0.0296 − 4.18 300 0.5035 0.02075 125 500
14 0.00752 8.84 1760.4 0.0512 − 3.34 520 0.5035 0.02075 125 500
15 0.00708 9.15 1728.3 0.0496 − 3.55 510 0.5035 0.02075 125 500
16 0.00708 9.15 1728.3 0.0496 − 3.55 510 0.5035 0.02075 125 500
17 0.00313 7.97 647.85 0.0151 − 2.68 220 0.5035 0.02075 220 500
18 0.00313 7.95 649.69 0.0151 − 2.66 222 0.5035 0.02075 220 500
19 0.00313 7.97 647.83 0.0151 − 2.68 220 0.5035 0.02075 242 550
20 0.00313 7.97 647.81 0.0151 − 2.68 220 0.5035 0.02075 242 550
21 0.00298 6.63 785.96 0.0145 − 2.22 290 0.5035 0.02075 254 550
22 0.00298 6.63 785.96 0.0145 − 2.22 285 0.5035 0.02075 254 550
23 0.00284 6.66 794.53 0.0138 − 2.26 295 0.5035 0.02075 254 550
24 0.00284 6.66 794.53 0.0138 − 2.26 295 0.5035 0.02075 254 550
25 0.00277 7.10 801.32 0.0132 − 2.42 310 0.5035 0.02075 254 550
26 0.00277 7.10 801.32 0.0132 − 2.42 310 0.5035 0.02075 254 550
27 0.52124 3.33 1055.1 1.842 − 1.11 360 0.9936 0.0406 10 150
28 0.52124 3.33 1055.1 1.842 − 1.11 360 0.9936 0.0406 10 150
29 0.52124 3.33 1055.1 1.842 − 1.11 360 0.9936 0.0406 10 150
30 0.01140 5.35 148.89 0.085 − 1.89 50 0.9936 0.0406 47 97
31 0.00160 6.43 222.92 0.0121 − 2.08 80 0.9142 0.0454 60 190
32 0.00160 6.43 222.92 0.0121 − 2.08 80 0.9142 0.0454 60 190
33 0.00160 6.43 222.92 0.0121 − 2.08 80 0.9142 0.0454 60 190
34 0.00010 8.95 107.87 0.0012 − 3.48 65 0.655 0.02846 90 200
35 0.00010 8.62 116.58 0.0012 − 3.24 70 0.655 0.02846 90 200
36 0.00010 8.62 116.58 0.0012 − 3.24 70 0.655 0.02846 90 200
37 0.01610 5.88 307.45 0.095 − 1.98 100 1.42 0.0677 25 110
38 0.01610 5.88 307.45 0.095 − 1.98 100 1.42 0.0677 25 110
39 0.01610 5.88 307.45 0.095 − 1.98 100 1.42 0.0677 25 110
40 0.00313 7.97 647.83 0.0151 − 2.68 220 0.5035 0.02075 242 550

M
∑i =1 μi [k ] the DEED problem schedules the outputs of online generator with the
μ [k ] = Npareto M predicted load demands over a period [75]. The DEED problem is
∑ j =1 ∑i =1 μi [ j ] (10) much more complex than the classical EED problem due to the large
M is the number of objectives and Npareto is the number of solutions number of variables and constraints. This section presents an intro-
in the final non-dominated front. The best compromise solution is that duction to the DEED problem as well as the state-of-the-art MOEA
for which μ [k ] is the largest. approaches for solving the DEED problem.

4. Dynamic EED problem 4.1. Problem formulation

Dynamic Environmental/Economic Dispatch (DEED) is an exten- Differences between the DEED and the EED problem formulations
sion of the classical EED problem which takes into consideration of the are incorporation of the time variables. The detailed description of the
ramping rate constraints of the generators [75–79]. Optimization of DEED model is as follows:

Table 7
Generator cost and emission coefficients of system G [49].

Generator a b c α β γ ζ φ PGmin PGmax

1 1000.403 40.5407 0.12951 360.0012 − 3.9864 0.04702 0.25475 0.01234 10 55

2 950.606 39.5804 0.10908 350.0056 − 3.9524 0.04652 0.25475 0.01234 20 80
3 900.705 36.5104 0.12511 330.0056 − 3.9023 0.04652 0.25163 0.01215 47 120
4 800.705 39.5104 0.12111 330.0056 − 3.9023 0.04652 0.25163 0.01215 20 130
5 756.799 38.5390 0.15247 13.8593 0.3277 0.00420 0.24970 0.01200 50 160
6 451.325 46.1592 0.10587 13.8593 0.3277 0.00420 0.24970 0.01200 70 240
7 1243.531 38.3055 0.03546 40.2669 − 0.5455 0.00680 0.24800 0.01290 60 300
8 1049.998 40.3965 0.02803 40.2669 − 0.5455 0.00680 0.24990 0.01203 70 340
9 1658.569 36.3278 0.02111 42.8955 − 0.5112 0.00460 0.25470 0.01234 135 470
10 1356.659 38.2704 0.01799 42.8955 − 0.5112 0.00460 0.25470 0.01234 150 470

6
B.Y. Qu et al. Swarm and Evolutionary Computation 38 (2018) 1–11

Table 8
Overview on the systems used by various methods.

System A System B System C System D System E System F System G

GA based [14,19,33,36,39–41] [38] [38]

PSO based [20,42,43,45–48] [43,44] [46]
DE based [24,50] [49,51] [50] [49] [49]
Hybrid [13,52–55] [13] [54,55] [54,55]
Others [26,56–67,69–71,73] [66] [42,57,58] [37,42,57,68,72] [58,65,67,73] [37,59,66,67,72] [68,72,73]

Fuel cost objective: 4.3. MOEA approaches to solve DEED

T NG
min F (PG ) = ∑ ∑ {(ai + bi PGi,t + ci PG2i,t ) + di sin[ei (PGi, min − PGi, t )] } In 2008, Basu applied NSGA-II to solve the DEED problem as a
t =1 i =1 true multi-objective optimization problem with non-commensurable
(11) and competing objectives [76]. The algorithm was verified on the 10-
unit test system with non-smooth fuel cost and emission level func-
where T is the number of hours in the time horizon and PGi,t is the
tions. The predicted load demands of the system were divided into 24
output of the ith thermal unit at time t.
intervals (hourly generation schedule) and the simulation results
Pollution emission objective:
showed that NSGA-II was able to generate superior performance over
T N classical techniques. Guo et al. introduced a multi-objective evolu-
min E (PG ) = ∑ ∑ [αi + βi PGi,t + γi (PGi,t )2 + ζi exp(φi PGi )] tionary method called group search optimizer with multiple producers
t =1 i =1 (12)
(GSOMP) to handle the DEED problem [81]. The algorithm adopted
Power balance constraint: the constrained tournament method [82] and Lamarckian method [83]
N
to handle the inequality and equality constraints, respectively.
∑ PGi,t − PD,t − Ploss, t = 0, t = 1, ..., T Technique for Order Preference Similar to an ideal solution (TOPSIS)
i =1 (13) was developed to find the final solution from the non-dominated
solutions according to a decision maker's preference. The simulation
Where PD,t and Ploss,t are the total demand and system loss at time t.
results demonstrated that GSOMP could achieve better convergence
Ramping rate constraints: during conjoint dispatching time peri-
and diversity performance than MOPSO and NSGA-II. In [84], the
ods, each generator must satisfy certain ramping rate limit and this
DEED problem considering hourly load and wind power was solved
constraint can be model as:
using a multi-objective particle swarm optimization algorithm.
⎧ PGi, t − PGi, t −1 − URi × ΔT ≤ 0 Incorporation of the wind power changes the DEED problem into a
⎨ , i = 1, ..., N stochastic problem. A two-stage scenario-based method was used to
⎩ PGi, t −1 − PGi, t − D Ri × ΔT ≤ 0 (14)
convert the problem into its respective deterministic equivalents first
where URi and Dri are the up and down ramping rate limits for the ith and solved subsequently. Similar problems were also investigated by
thermal unit, respectively while ΔT is the dispatching time interval. Bahmani-Firouzi et al. using particle swarm optimization based algo-
Generation capacity constraint: rithms [85,86].
In [87], Niknam et al. proposed a theta-multi-objective teaching-
PGmin max
i ≤ PGi, t ≤ PGi (15)
learning-based optimization to solve DEED problem. The algorithm
incorporated several modifications such as phase angles based optimi-
zation process to improve its performance. Moreover, the algorithm
4.2. Weighted sum approaches
used a niching mechanism to allocate more searching resources to less
explored regions. In [88], a multi-objective differential evolution
Weighted sum approach is also commonly used to solve the DEED
algorithm with expanded double selection and adaptive random restart
problem. In [11], Nwulu and Xia combined the DEED problem with
was applied to solve the DEED. The expanded double selection was
game theory based demand (DR) response programs. This work
proposed to strengthen the global search ability while the adaptive
optimized three objectives (minimize the fuel and emissions costs
random restart was proposed to use the elitist information of the non-
and maximize the utility demand response benefit) simultaneously and
dominated solutions saved in the external archive. This work also
the results showed that the method was an efficient way to reduce
introduced a Dynamic Heuristic Constraint Handling (DHCH) techni-
demand in the power system. In [80], three versions of Differential
que to deal with the complicated constraints. DHCH was employed to
Harmony Search (DHS) algorithms were used to solve the DEED
handle the equality constraints through an iterative correction process
problem. In this work, the weights were varied in steps and each
of infeasible solutions. With the help of DHCH, the number of
variation generated one compromise solution. The best compromise
infeasible solutions was reduced significantly. A similar constraint
solution was selected by using a fuzzy based method.

Table 9
The best cost obtained by typical MOEAs on 6-generator 30-bus lossless system without considering the security constraints.

NSGA [33] NPGA [33] SPEA [33] MOPSO [42] BB-MOP SO [48] FMOEP [56] MBFA [59] NSGA-II [70] MOACSA [70] SMODE [74]

1 0.1038 0.1116 0.1009 0.1183 0.1090 0.0872 0.1133 0.1094 0.1090 0.1077
2 0.3228 0.3143 0.3186 0.3019 0.3005 0.2868 0.3005 0.2994 0.2989 0.2990
3 0.5123 0.5419 0.5400 0.5224 0.5234 0.5488 0.5202 0.5236 0.5262 0.5269
4 1.0387 1.0415 0.9903 1.0116 1.0170 1.0114 0.9882 1.0167 1.0183 1.0128
5 0.5324 0.4726 0.5336 0.5254 0.5238 0.5477 0.5409 0.5244 0.5227 0.5269
6 0.3241 0.3512 0.3507 0.3544 0.3603 0.3521 0.3709 0.3605 0.3589 1.0128
Cost 600.34 600.31 600.22 600.12 600.11 600.24 600.17 600.11 600.11 600.11
Emission 0.2241 0.2238 0.2206 0.2216 0.2222 0.2232 0.2200 0.2222 0.2223 0.2221

7
B.Y. Qu et al. Swarm and Evolutionary Computation 38 (2018) 1–11

Table 10
The best emission obtained by typical MOEAs on 6-generator 30-bus lossless system without considering the security constraints.

NSGA [33] NPGA [33] SPEA [33] MOPSO [42] BB-MOP SO [48] FMOEP [56] MBFA [59] NSGA-II [70] MOACSA [70] SMODE [74]

1 0.4072 0.4146 0.4240 0.4015 0.4071 0.3926 0.3943 0.4059 0.4062 0.4002
2 0.4536 0.4419 0.4577 0.4590 0.4591 0.4570 0.4627 0.4586 0.4577 0.4531
3 0.4888 0.5411 0.5301 0.5332 0.5374 0.5549 0.5423 0.5382 0.5373 0.5430
4 0.4302 0.4067 0.3721 0.3891 0.3838 0.3799 0.3946 0.3832 0.3821 0.4019
5 0.5836 0.5318 0.5311 0.5456 0.5369 0.5434 0.5346 0.5385 0.5404 0.5361
6 0.4707 0.4979 0.5190 0.5057 0.5098 0.5061 0.5056 0.5097 0.5105 0.4997
Cost 633.83 636.04 640.42 637.42 638.26 638.97 636.73 638.22 638.30 635.99
Emission 0.1946 0.1943 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942 0.1942

handling method was also incorporated into MOEA/D to solve the
DEED problem considering wind power [89].
5. EED problems incorporating wind power
Wind power is a clean energy source which plays an increasing role
in power generation [90]. It can reduce the pollution emissions as well
as the dependence on fossil fuels. However, the stochastic nature of
wind speed leads to an uncertainty on wind power output.
Consequently, the high penetrations of wind power complicate the
EED problems significantly.
In general, the research framework of the EED problems with wind
power is based on the models described in the Sections 3.1 and 4.1. The
difference is the wind power modeling methods which include deter-
ministic modeling method, probabilistic modeling method and fuzzy
modeling method. Note that most of the current works convert the
wind power EED problem to a single objective one due to the
complexity and uncertainty of wind power and its significant impact
on environmental and economic objectives. In [90], the multi-objective
EED problem was converted to a single objective optimization problem
by the concept of Price Penalty Factors (PPFs). The stochastic
characteristic of wind power was expressed by the penalty cost of
overestimating and underestimating for wind turbines. An optimiza-
tion technique, called Gravitational Search Algorithm (GSA), was
applied to solve the complex optimization problem. In [91], Yao
et al. solved the stochastic EED model considering wind power
probability and carbon emission tax by using the Quantum-inspired
Particle Swarm Optimization (QPSO) method.
The wind energy is intermittent and uncertain which is highly
coupled with time variable. Hence, compared with the traditional static
dispatching model, the dynamic dispatching model is more suitable for
the power system with wind integration. Liang et al. [92] established a
dynamic model considering wind power with emission constraint and
proposed a GA based fuzzy-optimization approach to handle it. Jadhav
et al. [93] used a single objective dynamic model to demonstrate the
overall operation cost of conventional and wind power generators. The
emission was converted to cost using carbon tax value, and the Gbest
guided Artificial Bee Colony (GABC) algorithm was applied. Lee et al.
[94] added the emission to the general operation cost by using
environmental coefficient and the wind power was treated as common
load. The Quantum Genetic Algorithm (QGA) was employed to solve
the problem. In [95], the spinning reserve compensation cost owing to
the random nature of the wind resources was included and an
optimization algorithm based Chaotic Quantum Genetic Algorithm
(CQGA) was proposed to solve the economic dispatch problem with
wind energy. In [96], the Weibull distribution and the probability
density functions were used to calculate the underestimated and
overestimated unbalance cost of wind power considering its uncer-
tainty, and a Bi-Population Chaotic Differential Evolution (BPCDE)
algorithm was proposed to solve the wind-thermal dispatching pro-
blem.
As mentioned before, most of the above methods converted the
multi-objective EED problem to a single objective one, thereby ignoring
the importance of Pareto-optimality in the dispatch decision making
process. In the recent few years, some works have reported the usage of
MOEAs to solve the EED problems considering wind power. Wafa [97]
established a multi-objective EED model that included the cost of
modern thermal units with valves point effect, the polluting gas
emission and the factor for both overestimation and underestimation
of available wind. NSGA-II was used as the optimizer to solve this
model. In [98], a multi-objective energy dispatch model considered
environmental and fuel cost under large wind energy was proposed.
The wind power was treated as a common load as in [94] and an
effective encoding/decoding scheme for the SPEA procedure was used
to improve the performance. In [99], a summation based multi-
objective differential evolution (SMODE) algorithm was adopted to
optimize the EED problem with stochastic wind power. This work used
Weibull probability distribution function to model the stochastic nature
of the wind power and the uncertainty is treated as system constraints.
In [100], an online learning meta-heuristic method was used to
optimize the EED problem with wind power units. The uncertainty of
wind power was estimated using a practical model known as 2 m-point.
In [101], the multi-objective dynamic economic emission dispatch
problem integrating wind power was researched with a scenario-based
stochastic programming framework to model the random nature of
load demand and wind forecast errors. Moreover, a novel Fuzzy
Adaptive Improved Particle Swarm Optimization (FAIPSO) algorithm
was used to obtain the best solution for the corresponding scenarios. In
[102], a stochastic approach based on scenarios for the uncertainties
associated with hourly load and wind power forecasts was built for
DEED problem incorporating wind power plant. A Fuzzy Self Adaptive
Learning Particle Swarm Optimization (FSALPSO) algorithm was

Table 11
The best compromise solutions obtained by typical MOEAs on 6-generator 30-bus lossless system without considering the security constraints.

NSGA [33] NPGA [33] SPEA [33] MOPSO [42] BB-MOP SO [48] FMOEP [56] MBFA [59] NSGA-II [70] MOACSA [70] SMODE [74]

1 0.2252 0.2663 0.2623 0.2516 0.2595 0.2652 0.2661 0.2426 0.2311 0.2630
2 0.3622 0.3700 0.3765 0.3770 0.3698 0.3634 0.3792 0.4021 0.3665 0.3684
3 0.5222 0.5222 0.5428 0.5283 0.5351 0.5185 0.5387 0.5207 0.5420 0.5417
4 0.7660 0.7202 0.6838 0.7124 0.6919 0.7293 0.6750 0.6936 0.7127 0.7004
5 0.5397 0.5256 0.5381 0.5566 0.5500 0.5571 0.5383 0.5640 0.5471 0.5356
6 0.4187 0.4296 0.4305 0.4081 0.4277 0.4005 0.4366 0.4110 0.4347 0.4250
Cost 606.03 608.90 610.30 608.65 609.75 608.16 610.91 609.71 608.24 609.45
Emission 0.2041 0.2015 0.2004 0.2017 0.2008 0.2023 0.2000 0.2011 0.2020 0.2010

8
B.Y. Qu et al. Swarm and Evolutionary Computation 38 (2018) 1–11

Table 12
The best cost obtained by typical MOEAs on 6-generator 30-bus system considering loss and all constraints.

NSGA [33] NPGA [33] SPEA [33] MOPSO [42] MODE [24] FMOEP [56] MBFA [59] NSGA-II [70] MOACSA [70] SMODE [74]

1 0.1358 0.1127 0.1319 0.1524 0.1361 0.1866 0.1175 0.1619 0.1696 0.1730
2 0.3151 0.3747 0.3654 0.3427 0.3455 0.3531 0.3617 0.3629 0.3491 0.3564
3 0.8418 0.8057 0.7791 0.7857 0.7573 0.7587 0.7899 0.6068 0.6047 0.7404
4 1.0431 0.9031 0.9282 1.0180 0.6016 0.5982 0.9591 0.6059 0.6059 0.5946
5 0.0631 0.1347 0.1308 0.0995 0.5998 0.5400 0.1457 0.7155 0.7144 0.5914
6 0.4664 0.5331 0.5292 0.4669 0.4162 0.4214 0.4916 0.4055 0.4149 0.4023
Cost 620.87 620.46 619.60 618.54 618.46 619.44 618.06 618.35 618.34 619.07
Emission 0.2368 0.2243 0.2244 0.2308 0.2051 0.2031 0.2264 0.2034 0.2032 0.2034

developed to solve the problem. In [103], the wind power generators
were modeled by a piecewise linear approximation while the 2 m-point
estimation and modified teaching-learning algorithm were used to
obtain the set of non-dominated optimal solutions.
6. EED problems considering electric vehicles (EVs)
Considering the requirements for energy saving and environmental
protection, some researchers investigated the EED problems with EVs.
Although EVs need to charge from the power system, in V2G mode,
EVs can also supply power to the system thereby increasing the
complexity of the EED problem.
In [104,105], Saber et al. proposed a typical EED model for both
cost and emission optimization with a single objective function. In this
model, the V2G behaviors were characterized by the “registered”
numbers and present/departure SoC (State of Charge) of EVs. The
PSO was used to generate the optimal dispatch schedule and control of
EVs. Using the similar methods, Gholami et al. [106] established an
EED model with a combined economic and emission objective by using
cost and emission coefficients. PSO was also adopted as the optimizer.
Based on the above work, Debnath et al. [107] presented a system
model considering EV's battery lifetime. In [108], the multi-objective
DEED problem was also transformed into a single objective problem
and a self-learning teaching-learning based optimization (TLBO)
algorithm was employed to obtain the dispatching scheme. In [109],
a dynamic economic dispatching model with EVs for smart grid was
constructed. In this model, the minimization of power generation costs,
the lowest charging cost of EVs’ owners, the least air pollution and
maximization of synthetic load ratio were taken as objectives and the
NSGA-II algorithm was used to solve the problem.
7. EED problems within a micro-grid
the dynamic EED problem where the emissions were converted to
costs. In [113], a two-stage method for micro-grid system was
presented to minimize the total net present cost and carbon dioxide
emission simultaneously in a life circle. On the first stage, NSGA-II was
applied to solve the optimal design problem while on the second stage;
the Mixed-Integer Linear Programming (MILP) algorithm was used to
solve the optimal dispatch problem. In [114], Motevasel et al. proposed
an efficient Modified Bacterial Foraging Optimization (MBFO) algo-
rithm to minimize both the total operation costs and the net emission
in a micro-grid system.
8. Conclusions and future directions
In this paper, research works aiming to solve the EED problem
especially using multi-objective evolutionary algorithms have been
surveyed. Typical MOEAs, classical EED problems, dynamic EED
problems, EED problems considering wind power, EED problems
considering electric vehicles and the EED problems within a micro-
grid have been covered.
Although there have been numerous publications on the classical
EED problems, the study of the EED problems using multi-objective
evolutionary algorithms is still in its early stages. At present, there have
been some achievements in solving the dynamic EED problem and the
EED problem with wind power, while research on EED associated with
micro-grids and EVs is limited. Modeling and solving of such EED
problems are worthy of investigating especially using multi-objective
evolutionary algorithms. Furthermore, the new scheduling modes
(such as security dispatch and risk dispatch, etc.) and new application
scenarios (such as virtual power plants and active distribution net-
works, etc.) continue to emerge in the recent years. Developing the
corresponding dispatch models and effective solution algorithms are
important future research directions.

Micro-grid is the key component of the Smart-grid of the future and
it has become an active research area in the recent years. In [110], a
Fuzzy Self Adaptive Particle Swarm Optimization (FSAPSO) algorithm
was proposed to solve the EED in a typical micro-grid considering
economy and emission as competing objectives. In [111], the Multi-
objective Mesh Adaptive Direct Search (MOMADS) was employed to
minimize the cost and emission objectives of a micro-grid system. In
[112], the chaotic quantum genetic algorithm method was used to solve
Acknowledgements
This research is partially supported by National Natural Science
Foundation of China (61305080, 61473266, 61673404) and
Postdoctoral Science Foundation of China (2014M552013) and the
Scientific and Technological Project of Henan Province
(132102210521, 152102210153). Dr. P. N. Suganthan acknowledges
the support offered by the National Research Foundation Singapore

Table 13
The best emission obtained by typical MOEAs on 6-generator 30-bus system considering loss and all constraints.

NSGA [33] NPGA [33] SPEA [33] MOPSO [42] MODE [24] FMOEP [56] MBFA [59] NSGA-II [70] MOACSA [70] SMODE [74]

1 0.4403 0.4753 0.4419 0.4589 0.4184 0.3980 0.4716 0.4103 0.4090 0.3983
2 0.4940 0.5162 0.4598 0.5121 0.4622 0.4778 0.5127 0.4637 0.4624 0.4601
3 0.7509 0.6513 0.6944 0.6524 0.5441 0.5628 0.6189 0.5459 0.5412 0.5423
4 0.5060 0.4363 0.4616 0.4331 0.3793 0.3795 0.5032 0.3881 0.3933 0.4045
5 0.1375 0.1896 0.1952 0.1981 0.5520 0.5403 0.1788 0.5425 0.5445 0.5448
6 0.5364 0.5988 0.6131 0.6129 0.5068 0.5049 0.5822 0.5146 0.5146 0.5139
Cost 649.24 657.59 651.71 656.87 645.74 645.24 651.93 645.39 644.84 643.01
Emission 0.2048 0.2017 0.2019 0.2014 0.1942 0.1942 0.2019 0.1942 0.1942 0.1942

9
B.Y. Qu et al. Swarm and Evolutionary Computation 38 (2018) 1–11

Table 14
The best compromise solutions obtained by typical MOEAs on 6-generator 30-bus system considering loss and all constraints.

NSGA [33] NPGA [33] SPEA [33] MOPSO [42] MODE [24] FMOEP [56] MBFA [59] NSGA-II [70] MOACSA [70] SMODE [74]

1 0.2712 0.2998 0.3052 0.2882 0.3017 0.3230 0.2983 0.3133 0.3004 0.3140
2 0.3670 0.4325 0.4389 0.3965 0.4019 0.4056 0.4332 0.3844 0.3873 0.4169
3 0.8099 0.7342 0.7163 0.7320 0.5815 0.5669 0.7350 0.5399 0.5659 0.5424
4 0.7550 0.6852 0.6978 0.7520 0.5967 0.5725 0.6899 0.6039 0.6023 0.5856
5 0.1357 0.1560 0.1552 0.1489 0.5352 0.5305 0.1569 0.5581 0.5481 0.5490
6 0.5239 0.5561 0.5507 0.5463 0.4436 0.4638 0.5503 0.4630 0.4588 0.4552
Cost 625.71 630.06 629.59 626.10 622.37 625.29 629.56 622.77 622.41 624.44
Emission 0.2136 0.2079 0.2079 0.2106 0.1975 0.1966 0.2080 0.1974 0.1976 0.1968

under its Campus for Research Excellence and Technological
Enterprise (CREATE) programme.
References
[1] J.H. Talaq, F.E. Hawary, M.E. Hawary, A summary of environmental/economic
dispatch algorithms, IEEE Trans. Power Syst. 9 (1994) 1508–1516.
[2] A. Farag, S.A. Baiyat, T.C. Cheng, Economic load dispatch multi-objective
optimization procedures using linear programming techniques, IEEE Trans.
Power Syst. 10 (1995) 731–738.
[3] L. Du, S. Grijalva, R.G. Harley, Game-theoretic formulation of power dispatch
with guaranteed convergence and prioritized best response, IEEE Trans. Sustain.
Energy 6 (2015) 51–59.
[4] T. Jayabarathi, T. Raghunathan, B.R. Adarsh, P.N. Suganthan, Economic dispatch
using hybrid grey wolf optimizer, Energy 111 (2016) 630–641.
[5] T. Sen, H.D. Mathur, A new approach to solve economic dispatch problem using a
hybrid ACO-ABC-HS optimization algorithm, Int. J. Electr. Power Energy Syst. 78
(2016) 735–744.
[6] K. Bhattacharjee, A. Bhattacharya, S.H.N. Dey, Backtracking search optimization
based economic environmental power dispatch problems, Int. J. Electr. Power
Energy Syst. 73 (2015) 830–842.
[7] A.Y. Abdelaziz, E.S. Ali, S.M.A. Elazim, Combined economic an emission dispatch
solution using flower pollination algorithm, Int. J. Electr. Power Energy Syst. 80
(2016) 264–274.
[8] H. Li, E.R. Carlos, Y. Zheng, Economic dispatch optimization algorithm based on
particle diffusion, Energy Convers. Manag. 105 (2015) 1251–1260.
[9] S. Sayah, A. Hamouda, A. Bekrar, Efficient hybrid optimization approach for
emission constrained economic dispatch with nonsmooth cost curves, Int. J.
Electr. Power Energy Syst. 56 (2014) 127–139.
[10] V.K. Jadoun, N. Gupta, K.R. Niazi, A. Swarnkar, Modulated particle swarm
optimization for economic emission dispatch, Int. J. Electr. Power Energy Syst. 73
(2015) 80–88.
[11] N.I. Nwulu, X. Xia, Multi-objective dynamic economic emission dispatch of
electric power generation integrated with game theory based demand response
programs, Energy Convers. Manag. 89 (2015) 963–974.
[12] D. Srinivasan, A.G.B. Tettamanzi, An evolutionary algorithm for evaluation of
emission compliance options in view of the clean air act amendments, IEEE
Trans. Power Syst. 12 (1997) 336–341.
[13] D.B. Das, C. Patvardhan, New multi-objective stochastic search technique for
economic load dispatch, IEE Proc. -Gener. Transm. Distrib. 145 (1998) 747–752.
[14] M.A. Abido, A new Multi-objective evolutionary algorithm for environmental/
economic power dispatch, in: Proceedings of IEEE PES Summer Meeting, vol. 65,
Vancouver, Canada, 2001, pp. 71–81.
[15] D.C. Walters, G.B. Sheble, Genetic algorithm solution of economic dispatch with
valve point loading, IEEE Trans. Power Syst. 8 (1993) 1325–1332.
[16] J. Schaffer, Multiple objective optimization with vector evaluated genetic algo-
rithms, in: Proceedings of the 1st International Conference Genetics Algorithms,
1985, pp. 93–100.
[17] N. Srinivas, K. Deb, Multi-objective function optimization using non-dominated
sorting genetic algorithm, Evol. Comput. 2 (1994) 221–248.
[18] K. Deb, A. Pratap, S. Agrawal, T. Meyarivan, A fast and elitist multi-objective
genetic algorithm: NSGA-II, IEEE Trans. Evol. Comput. 6 (2002) 182–197.
[19] M.A. Abido, A Niched Pareto genetic algorithm for multi-objective environmental/
economic dispatch, Int. J. Electr. Power Energy Syst. 25 (2003) 97–105.
[20] B. Zhao, Y.J. Cao, Multiple objective particle swarm optimization technique for
economic load dispatch, J. Zhejiang Univ.: Sci. A 6 (2005) 420–427.
[21] C.A.C. Coello, M.S. Lechuga, MOPSO: a proposal formultiple objective particle
swarm optimization, in: IEEE Proceedings World Congress on Computational
Intelligence, vol. 2, 2002, pp. 1051–1056.
[22] J.D. Knowles, D. Corne, Approximating the non-dominated front using the pareto
archived evolution strategy, Evol. Comput. 8 (2000) 149–172.
[23] Q.F. Zhang, H. Li, MOEA/D: a multi-objective evolutionary algorithm based on
decomposition, IEEE Trans. Evol. Comput. 11 (2007) 712–731.
[24] M. Varadarajan, K.S. Swarup, Solving multi-objective optimal power flow using
differential evolution, IET Gener. Transm. Distrib. 2 (2008) 720–730.
[25] B.Y. Qu, P.N. Suganthan, Multi-objective evolutionary algorithms based on the
summation of normalized objectives and diversified selection, Inform. Sci. 180
(2010) 3170–3181.
[26] M.A. Abido, environmental/economic power dispatch using multi-objective
evolutionary algorithm, IEEE Trans. Power Syst. 18 (2003) 1529–1537.
[27] S.F. Brodesky, R.W. Hahn, Assessing the influence of power pools on emission
constrained economic dispatch, IEEE Trans. Power Syst. 1 (1986) 57–62.
[28] G.P. Granelli, M. Montagna, G.L. Pasini, P. Marannino, Emission constrained
dynamic dispatch, Electr. Power Syst. Res. 24 (1992) 56–64.
[29] J. Zahavi, L. Eisenberg, Economic/environmental power dispatch, IEEE Trans.
Syst. Man, Cyber. 5 (1985) 485–489.
[30] J.S. Dhillon, S.C. Parti, D.P. Kothari, Stochastic economic emission load dispatch,
Electr. Power Syst. Res. 26 (1993) 179–186.
[31] D. Aydin, S. Özyön, C. Yasar, T. Liao, Artificial bee colony algorithm with dynamic
population size to combined economic and emission dispatch problem, Int. J.
Electr. Power Energy Syst. 54 (2014) 144–153.
[32] K. Bhattacharjee, A. Bhattacharya, S.H. nee Dey, Solution of economic emission
load dispatch problems of power systems by real coded chemical reaction
algorithm, Int. J. Electr. Power Energy Syst. 59 (2014) 176–187.
[33] M.A. Abido, Multiobjective evolutionary algorithms for electric power dispatch
problem, IEEE Trans. Evol. Comput. 10 (2006) 315–329.
[34] B.Y. Qu, P.N. Suganthan, J.J. Liang, Differential evolution with neighborhood
mutation for multimodal optimization, IEEE Trans. Evol. Comput. 16 (2012)
601–614.
[35] K.P. Wong, B. Fan, C.S. Chang, A.C. Liew, Multi-objective generation dispatch
using bi-criterion global optimization, IEEE Trans. Power Syst. 10 (1995)
1813–1819.
[36] M.A. Abido, A novel multi-objective evolutionary algorithm for environmental/
economic power dispatch, Electr. Power Syst. Res. 65 (2003) 71–81.
[37] E. Afzalan, M. Joorabian, Emission, reserve and economic load dispatch problem
with non-smooth and non-convex cost functions using epsilon-multi-objective
genetic algorithm variable, Int. J. Electr. Power Energy Syst. 52 (2013) 55–67.
[38] C.S.R. Harry, T.F.K. Robert, Environmental/economic dispatch of thermal units
of an elitist multi-objective evolutionary algorithm, ICIT 1 (2003) 48–53.
[39] T. Guesmi, H.H. Abdallah, A. Toumi, New approach to solve multi-objective
environmental/economic dispatch, J. Electr. Syst. 2 (2006) 64–81.
[40] S. Dhanalakshmi, S. Kannan, K. Mahadevan, S. Baskar, Application of modified
NSGA-II algorithm to combined economic and emission dispatch problem, Int. J.
Electr. Power Energy Syst. 33 (2011) 992–1002.
[41] R. Muthuswamy, M. Krishnan, K. Subramanian, B. Subramanian, Environmental
and economic power dispatch of thermal generators using modified NSGA-II
algorithm, Int. Trans. Electr. Energy Syst. 25 (2014) 1552–1569.
[42] M.A. Abido, Multi-objective particle swarm optimization for environmental/
economic dispatch problem, Electr. Power Syst. Res. 79 (2009) 1105–1113.
[43] J. Hazra, A.K. Sinha, A multi-objective optimal power flow using particle swarm
optimization, Eur. Trans. Electr. Power 21 (2010) 1028–1045.
[44] L.F. Wang, C. Singh, Environmental/economic power dispatch using a fuzzified
multi-objective particle swarm optimization algorithm, Electr. Power Syst. Res. 77
(2007) 1654–1664.
[45] S. Agrawal, B.K. Panigrahi, M.K. Tiwari, Multi-objective particle swarm algorithm
with fuzzy clustering for electrical power dispatch, IEEE Trans. Evol. Comput. 12
(2008) 529–541.
[46] T.Y. Liu, L.C. Jiao, W.P. Ma, J.J. Ma, R.H. Shang, Cultural quantum-behaved
particle swarm optimization for environmental/economic dispatch, Appl. Soft
Comput. 48 (2016) 597–611.
[47] J.J. Cai, X.Q. Ma, Q. Li, L.X. Li, H.P. Peng, A multi-objective chaotic particle
swarm optimization for environmental/economic dispatch, Energy Convers.
Manag. 50 (2009) 1318–1325.
[48] Y. Zhang, D.W. Gong, Z.H. Ding, A bare-bones multi-objective particle swarm
optimization algorithm for environmental/economic dispatch, Inform. Sci. 192
(2012) 213–227.
[49] M. Basu, Economic environmental dispatch using multi-objective differential
evolution, Appl. Soft Comput. 11 (2011) 2845–2853.
[50] L.H. Wu, Y.N. Wang, X.F. Yuan, S.W. Zhou, Environmental/economic power
dispatch problem using multi-objective differential evolution algorithm, Electr.
Power Syst. Res. 80 (2010) 1171–1181.
[51] Y.L. Lu, J.Z. Zhou, H. Qin, Y. Wang, Y.C. Zhang, Environmental/economic
dispatch problem of power system by using an enhanced multi-objective
differential evolution algorithm, Energy Convers. Manag. 52 (2011) 1175–1183.
[52] D.W. Gong, Y. Zhang, C.L. Qi, Environmental/economic power dispatch using a
hybrid multi-objective optimization algorithm, Int. J. Electr. Power Energy Syst.
32 (2010) 607–614.
[53] K.O. Alawode, A.M. Jubril, A.O. Komolafe, Multi-objective optimal power flow
using hybrid evolutionary algorithm, WASET (2010) 790–795.
[54] A. Bhattacharya, P.K. Chattopadhyay, Hybrid differential evolution with biogeo-
graphy-based optimization algorithm for solution of economic emission load
dispatch problems, Expert Syst. Appl. 38 (2011) 14001–14010.
[55] A. Bhattacharya, P.K. Chattopadhyay, Solving economic emission load dispatch
problems using hybrid differential evolution, Appl. Soft Comput. 11 (2011)

10
B.Y. Qu et al.
2526–2537.
[56] B.Y. Qu, P.N. Suganthan, V.R. Pandi, B.K. Panigrahi, Multi-objective evolutionary
programming to solve environmental economic dispatch problem, Int. Conf.
Control Autom. Robot. Vis. (2010) 673–1679.
[57] P. Venkatesh, K.Y. Lee, Multi-objective evolutionary programming for economic
emission dispatch problem, in: Proceedings of the IEEE Power and Energy Society
General Meeting, Pittsburgh, PA, USA, 2008.
[58] J.J. Cai, X.Q. Ma, Q. Li, L.X. Li, H.P. Peng, A multi-objective chaotic and swarm
optimization for environmental/economic dispatch, Int. J. Electr. Power Energy
Syst. 32 (2010) 337–344.
[59] P.K. Hota, A.K. Barisal, R. Chakrabarti, Economic emission load dispatch through
fuzzy based bacterial foraging algorithm, Int. J. Electr. Power Energy Syst. 32
(2010) 794–803.
[60] B.K. Panigrahi, V.R. Pandi, S. Das, S. Das, Multi-objective fuzzy dominance based
bacterial foraging algorithm to solve economic emission dispatch problem, Energy
35 (2010) 4761–4770.
[61] B.K. Panigrahi, V.R. Pandi, R. Sharma, S. Das, S. Das, Multi-objective bacteria
foraging algorithm for electrical load dispatch problem, Energy Convers. Manag.
52 (2011) 1334–1342.
[62] S. Hemamalini, S.P. Simon, Economic/emission load dispatch using artificial bee
colony algorithm, ACEEE Int. J. Electr. Power Eng. 1 (2010) 27–33.
[63] R. Kumar, A. Sadu, R. Kumar, S.K. Panda, A novel multi-objective directed bee
colony optimization algorithm for multi-objective emission constrained economic
power dispatch, Int. J. Electr. Power Energy Syst. 43 (2012) 1241–1250.
[64] T. Niknam, M.R. Narimani, M. Jabbari, A.R. Malekpour, A modified shuffle frog
leaping algorithm for multi-objective optimal power flow, Energy 36 (2011)
6420–6432.
[65] S. Sivasubramani, K.S. Swarup, Environmental/economic dispatch using multi-
objective harmony search algorithm, Electr. Power Syst. Res. 81 (2011)
1778–1785.
[66] S. Rajasomashekar, P. Aravindhababu, Biogeography based optimization techni-
que for best compromise solution of economic emission dispatch, Swarm Evol.
Comput. 7 (2012) 47–57.
[67] A. Ghasemi, A fuzzified multi objective interactive honey bee mating optimization
for environmental/economic power dispatch with valve point effect, Int. J. Electr.
Power Energy Syst. 49 (2013) 308–321.
[68] R. Zhang, J.Z. Zhou, L. Mo, S. Ouyang, X. Liao, Economic environmental dispatch
using an enhanced multi-objective cultural algorithm, Electr. Power Syst. Res. 99
(2013) 18–29.
[69] Z.G. Lu, T. Feng, X.P. Li, Low-carbon emission/economic power dispatch using
the multi- objective bacterial colony chemotaxis optimization algorithm consid-
ering carbon capture power plant, Int. J. Electr. Power Energy Syst. 53 (2013)
106–112.
[70] B.S. Rao, K. Vaisakh, Multi-objective adaptive clonal selection algorithm for
solving environmental/economic dispatch and OPF problems with load uncer-
tainty, Int. J. Electr. Power Energy Syst. 53 (2013) 390–408.
[71] M.D.A.C.E. Silva, C.E. Klein, V.C. Mariani, L.D.S. Coelho, Multi-objective scatter
search approach with new combination scheme applied to solve environmental/
economic dispatch problem, Energy 53 (2013) 14–21.
[72] P.K. Roy, S. Bhui, Multi-objective quasi-oppositional teaching learning based
optimization for economic emission load dispatch problem, Int. J. Electr. Power
Energy Syst. 53 (2013) 937–948.
[73] M.D. Mostafa, A.R. Nasrudin, Multi-objective backtracking search algorithm for
economic emission dispatch problem, Appl. Soft Comput. 40 (2016) 479–494.
[74] B.Y. Qu, J.J. Liang, Y.S. Zhu, Z.Y. Wang, P.N. Suganthan, Economic emission
dispatch problems with stochastic wind power using summation based multi-
objective evolutionary algorithm, Inform. Sci. 351 (2016) 48–66.
[75] M. Basu, Dynamic economic emission dispatch using evolutionary programming
and fuzzy satisfying method, Int. J. Emerg. Electr. Power Syst. 8 (2007) 1–15.
[76] M. Basu, Dynamic economic emission dispatch using non-dominated sorting
genetic algorithm-II, Int. J. Electr. Power Energy Syst. 30 (2008) 140–149.
[77] X. Xia, A.M. Elaiw, Optimal dynamic economic dispatch of generation: a review,
Electr. Power Syst. Res. 80 (2010) 975–986.
[78] G.C. Liao, A novel evolutionary algorithm for dynamic economic dispatch with
energy saving and emission reduction in power system integrated wind power,
Energy 36 (2011) 1018–1029.
[79] M. Basu, Particle swarm optimization based goal-attainment method for dynamic
economic emission dispatch, Elect. Power Compon. Syst. 34 (2006) 1015–1025.
[80] R. Arul, S. Velusami, G. Ravi, A new algorithm for combined dynamic economic
emission dispatch with security constraints, Energy 79 (2015) 496–511.
[81] C.X. Guo, J.P. Zhan, Q.H. Wu, Dynamic economic emission dispatch based on
group search optimizer with multiple producers, Electr. Power Syst. Res. 86
(2012) 8–16.
[82] K. Deb, S. Agrawal, A. Pratap, T. Meyarivan, A fast elitist non-dominated sorting
genetic algorithm for multi-objective optimization: NSGA-II, PPSN VI (2000)
849–858.
[83] K. Deb, Multi-Objective Optimization Using Evolutionary Algorithms, John Wiley
and Sons, USA, 2001.
[84] R.A. Abarghooee, J. Aghaei, Stochastic dynamic economic emission dispatch
considering wind power, IEEE Power Eng. Autom. Conf. 1 (2011) 158–161.
[85] B. Bahmani-Firouzi, E. Farjah, T. Niknam, Multi-objective stochastic dynamic
economic emission dispatch enhancement by fuzzy adaptive modified theta-
particle swarm optimization, J. Renew. Sustain. Energy 4 (2012) 133–137.
[86] B. Bahmani-Firouzi, E. Farjah, R. Azizipanah-Abarghooee, An efficient scenario-
based and fuzzy self-adaptive learning particle swarm optimization approach for
11
Swarm and Evolutionary Computation 38 (2018) 1–11
dynamic economic emission dispatch considering load and wind power uncer-
tainties, Energy 50 (2013) 232–244.
[87] T. Niknam, F. Golestaneh, M.S. Sadeghi, θ-Multiobjective teaching-learning-based
optimization for dynamic economic emission dispatch, IEEE Syst. J. 6 (2012)
341–352.
[88] X.W. Jiang, J.Z. Zhou, H. Wang, Y.C. Zhang, Dynamic environmental economic
dispatch using multi-objective differential evolution algorithm with expanded
double selection and adaptive random restart, Energy Syst. 49 (2013) 399–407.
[89] Y.S. Zhu, J. Wang, B.Y. Qu, P.N. Suganthan, Multi-objective dynamic economic
emission dispatch integrating wind farm, Power Syst. Technol. 39 (2015)
1315–1322.
[90] S. Mondal, A. Bhattacharya, S.H.N. Dey, Multi-objective economic emission load
dispatch solution using gravitational search algorithm and considering wind
power penetration, Int. J. Electr. Power Energy Syst. 44 (2013) 282–292.
[91] F. Yao, Y.D. Zhao, K. Meng, X. Zhao, H.C.I. Henbert, K.T. Wong, Quantum-
inspired particle swarm optimization for power system operations considering
wind power uncertainty and carbon tax in Australia, IEEE Trans. Ind. Inform. 8
(2012) 880–888.
[92] R.H. Liang, J.H. Liao, A fuzzy-optimization approach for generation scheduling
with wind and solar energy systems, IEEE Trans. Power Syst. 22 (2007)
1665–1674.
[93] H.T. Jadhav, R. Roy, Gbest guided artificial bee colony algorithm for environ-
mental/economic dispatch considering wind power, Expert Syst. Appl. 40 (2013)
6385–6399.
[94] J.C. Lee, W.M. Lin, G.C. Liao, T.P. Tsao, Quantum genetic algorithm for dynamic
economic dispatch with valve-point effects and including wind power system, Int.
J. Electr. Power Energy Syst. 33 (2011) 189–197.
[95] G.C. Liao, A novel evolutionary algorithm for dynamic economic dispatch with
energy saving and emission reduction in power system integrated wind power,
Energy 36 (2011) 1018–1029.
[96] C.H. Peng, H.J. Sun, J.F. Guo, G. Liu, Dynamic economic dispatch for wind-
thermal power system using a novel bi-population chaotic differential evolution
algorithm, Int. J. Electr. Power Energy Syst. 42 (2012) 119–126.
[97] A.R.A. Wafa, Optimization of economic/emission load dispatch for hybrid
generating systems using controlled elitist NSGA-II, Electr. Power Syst. Res. 105
(2013) 142–151.
[98] C.C. Kuo, Wind energy dispatch considering environmental and economic factors,
Renew. Energy 35 (2010) 2217–2227.
[99] B.Y. Qu, J.J. Liang, Y.S. Zhu, Z.Y. Wang, P.N. Suganthan, Economic emission
dispatch problems with stochastic wind power using summation based multi-
objective evolutionary algorithm, Inform. Sci. 351 (2016) 48–66.
[100] G. Ali, G. Milad, J.G. Mohammad, E. Mohammad, Modeling of wind/environ-
ment/economic dispatch in power system and solving via an online learning meta-
heuristic method, Appl. Soft Comput. 43 (2016) 454–468.
[101] J. Aghaei, T. Niknam, R. Azizipanah-Abarghooee, J.M. Arroyo, Scenario-based
dynamic economic emission dispatch considering load and wind power uncer-
tainties, Int. J. Electr. Power Energy Syst. 47 (2013) 351–367.
[102] B. Bahmani-Firouzi, E. Farjah, R. Azizipanah-Abarghooee, An efficient scenario-
based and fuzzy self-adaptive learning particle swarm optimization approach for
dynamic economic emission dispatch considering load and wind power uncer-
tainties, Energy 50 (2013) 232–244.
[103] R. Azizipanah-Abarghooee, T. Niknam, A. Roosta, A.R. Malepour, M. Zare,
Probabilistic multi-objective wind-thermal economic emission dispatch based on
point estimated method, Energy 37 (2012) 322–335.
[104] A.Y. Saber, G.K. Venayagamoorthy, Plug-in vehicles and renewable energy sources
for cost and emission reductions, IEEE Trans. Ind. Electron. 58 (2011)
1229–1238.
[105] A.Y. Saber, G.K. Venayagamoorthy, Resource scheduling under uncertainty in a
smart grid with renewables and plug-in vehicles, IEEE Syst. J. 6 (2012) 103–109.
[106] A. Gholami, J. Ansari, M. Jamei, Environmental/economic dispatch incorporating
renewable energy sources and plug-in vehicles, IET Gener. Transm. Dis. 8 (2014)
2183–2198.
[107] U.K. Debnath, I. Ahmad, D. Habibi, A.Y. Saber, Energy storage model with
gridable vehicles for economic load dispatch in the smart grid, Int. J. Electr.
Power Energy Syst. 64 (2015) 1017–1024.
[108] Z. Yang, K. Li, Q. Niu, A. Foley, A self-learning TLBO based dynamic economic/
environmental dispatch considering multiple plug-in electric vehicle loads, J.
Mod. Power Syst. Clean Energy 2 (2014) 298–307.
[109] H.L. Li, X.M. Bai, W. Tan, J.W. Dong, N. Li, Research on dynamic economic
dispatch based on smart grid, Power Syst. Technol. 37 (2013) 1547–1554.
[110] A.A. Moghaddam, A. Seifi, T. Niknam, Multi-operation management of a typical
micro-grids using particle swarm optimization: a comparative study, Renew. Sust.
Energy Rev. 16 (2012) 1268–1281.
[111] F.A. Mohamed, H.N. Koivo, Multi-objective optimization using mesh adaptive
direct search for power dispatch problem of micro-grid, Int. J. Electr. Power
Energy Syst. 42 (2012) 728–735.
[112] G.C. Liao, Solve environmental economic dispatch of smart micro-grid containing
distributed generation system – using chaotic quantum genetic algorithm, Int. J.
Electr. Power Energy Syst. 43 (2010) 779–787.
[113] L. Guo, W. Liu, J. Cai, B. Hong, C. Wang, A two-stage optimal planning and design
method for combined cooling, heat and power microgrid system, Energy Convers.
Manag. 74 (2013) 433–445.
[114] M. Motevasel, A.R. Seifi, T. Niknam, Multi-objective energy management of CHP
(combined heat and power)-based micro-grid, Energy 51 (2013) 123–136.