See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/325308641

A Solution to Non-convex/Convex and Dynamic Economic Load Dispatch

Problem Using Moth Flame Optimizer

Article · May 2018

DOI: 10.1007/s41403-018-0034-3

CITATIONS READS

2 228

4 authors, including:

Vikram Kumar Kamboj Manisha Sharma

DAV University National Institute of Technology, Hamirpur
21 PUBLICATIONS   258 CITATIONS    14 PUBLICATIONS   127 CITATIONS   

SEE PROFILE SEE PROFILE

S.K. Bath
Punjab Technical University
16 PUBLICATIONS   218 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

HGWO-PS View project

All content following this page was uploaded by Manisha Sharma on 29 August 2018.

The user has requested enhancement of the downloaded file.

INAE Letters
https://doi.org/10.1007/s41403-018-0034-3

ORIGINAL ARTICLE

A Solution to Non‑convex/Convex and Dynamic Economic Load

Dispatch Problem Using Moth Flame Optimizer
Ashutosh Bhadoria1 · Vikram Kumar Kamboj1 · Manisha Sharma2 · S. K. Bath3

Received: 16 October 2017 / Accepted: 13 March 2018

© Indian National Academy of Engineering 2018

Abstract
Moth flame optimization algorithm is novel nature inspired heuristic paradigm inspired by navigation method of moths in
nature and based on the concept that the moth eventually converges toward the light. This paper presents the application of
MFO algorithm for the solution of non-convex and dynamic economic load dispatch problem of electric power system. The
performance of MFO algorithm is tested for non-convex, convex and dynamic economic load dispatch problem of seven
IEEE benchmarks and the results are verified by a comparative study with lambda iteration method, particle swarm opti-
mization algorithm, genetic algorithm (GA), artificial bee colony, evolutionary programming (EP) and grey wolf optimizer
(GWO). Also, in the proposed research, the impact of renewable energy sources (i.e. wind and solar) has been taken into
consideration along with conventional thermal power generating units. Also, critical analysis has been made for percentage
cost saving with due consideration of solar and wind power units and it has been experimentally observed that the addition
of renewable energy sources to conventional thermal power system results in significant cost saving. Comparative results
show that the performance of Moth Flame Optimizer Algorithm is better than recently developed GWO algorithm and other
well known heuristics and meta-heuristics search algorithms.

Keywords  Economic load dispatch problem (ELDP) · Grey wolf optimizer (GWO) · Moth flame optimization (MFO)

Introduction network, as and when required, is known to be commit-

Recent power system networks consists of various power
generating units such as thermal, hydro, nuclear, solar and
wind etc. Also, the load demand varies during a day and
attains different peak values. Thus, it is required to decide
which generating unit to turn on and at what time it is needed
in the power system network and also the sequence in which
the units must be shut down keeping in mind the cost effec-
tiveness of turning on and shutting down of respective units.
The entire process of computing and making these deci-
sions is known as unit commitment (UC). The unit which is
decided or scheduled to be connected to the power system
* Vikram Kumar Kamboj
ashutoshbhadoria.mits@gmail.com
1
Department of Electrical Engineering, DAV University,
Jalandhar, India
2
Department of Electrical Engineering, National Institute
of Technology, Hamirpur, India
3
Department of Electrical Engineering, GZS Campus College
of Engineering and Technology, Bathinda, Punjab, India
ted unit. Unit commitment in power systems refers to the
problem of determining the on/off states of generating units
that minimize the operating cost for a given time horizon.
Electrical power plays a pivotal role in the modern world
to satisfy various needs. It is therefore very important that
the electrical power generated is transmitted and distributed
efficiently in order to satisfy the power requirement. Electri-
cal power is generated in several ways. The most significant
crisis in the planning and operation of electric power genera-
tion system is the effective scheduling of all generators in
a system to meet the required demand. The economic load
dispatch (ELD) problem is the most important optimization
problem in scheduling the generation among thermal gen-
erating units in power system.
Economic dispatch in electric power system refers to the
short-term discernment of the optimal generation output of
various electric utilities, to meet the system load demand,
at the minimum possible cost, subject to various system
and operating constraints viz.operational and transmission
constraints. The economic load dispatch problem (ELDP)
means that the electric utilities (i.e. generator’s) real and

123
Vol.:(0123456789)

reactive power are allowed to vary within certain limits so
as to meet a particular load demand within lowest fuel cost.
The ultimate aim of the ELD problem is to minimize the
operation cost of the power generation system, while sup-
plying the required power demanded. In addition to this,
the various operational constraints of the system should
also be satisfied. The problem of ELD is usually multi-
modal, discontinuous and highly nonlinear. Although the
cost curve of thermal generating units are generally mod-
elled as a smooth curve, the input–output characteristics are
nonlinear by nature because of valve-point loading effects,
Prohibited Operating Zones (POZ), ramp rate limits etc.
In recent years, various evolutionary, heuristic and meta-
heuristics optimization algorithms have been developed
simulating natural phenomena such as: genetic algorithm
(GA) (Walters and Sheble 1993), Hopfield neural network
algorithm (HNNA) (Park et  al. 2010), Chaotic particle
swarm optimization (CPSO) (Cai et al. 2012) New particle
swarm optimization algorithm (NPSO), (Selvakumar and
Thanushkodi 2007), (ARCGA) Adaptive real coded genetic
algorithm (Amjady and Nasiri-Rad 2010), Fuzzy adaptive
hybrid particle swarm optimization algorithm (Niknam
2010), Improved coordinated aggregation-based particle
swarm optimization (ICA-PSO) (Vlachogiannis and Lee
2009), Improved particle swarm optimization algorithm
for the non-convex economic dispatch problem (Park et al.
2010), GA-API genetic algorithm with ant colony approach
(Ciornei and Kyriakides 2012), Shuffled differential evolu-
tion (SDE) (Reddy and Vaisakh 2014), Bat algorithm (BA)
and Firefly algorithm (FA) (Moradi et al. 2016), Differen-
tial evolution ant colony optimization technique (DEIANT)
and conducted on IEEE 30 bus system with 6 generating
units. (Rahmat et al. 2014) Oppositional chemical reaction
optimization and 10 unit system including valve point effect
without transmission loss with load demand (2700 MW). 40
unit system with considering the transmission loss (Hazra
et al. 2015). Improved harmony search algorithm (SWOA)
(Karthigeyan et al. 2015), Firefly algorithm and 3 thermal
unit system considering transmission loss and 2 wind units
in which the transmission loss is neglected with different
load demand (50, 100, 125, 150, 200) (Saxena and Gan-
guli 2015), Quassi-Oppositional TLBO and 13 unit system
with load demand (1800 MW). 15unit system included valve
point loading effect, Ramp rate, limits generation limits, Pro-
hibited zone with load demand (2630 MW). 40 unit sys-
tem included valve point loading effect with load demand
(10,500 MW) (Maity 2016), Symbiotic Organisms Search
Algorithm (SOS) (Guvenc 2016), Genetic Algorithm (GA)
and 3 unit system with load demand of (150 MW). 6 unit
system with load demand of (700 MW) (Sharma and Moses
2016), Opposition-based krill herd algorithm (OKHA)
(Bulbul et  al. 2016), Improved teaching learning based
optimization (I-TLBO) algorithm (Mondal 2016), Symbiotic
123
INAE Letters
organism search algorithm (SOS) (Tiwari and Manjaree
2016), Wasp Swarm Algorithm (WSA) (Pinto et al. 2007),
Monkey Search Algorithm (MSA) (Mucherino and Seref
2007), Bee Collecting Pollen Algorithm (BCPA) (Lu and
Zhou 2008), Cuckoo Search Algorithm (CSA) (Yang and
Deb 2009), Dolphin Partner Optimization (DPO) (Shiqin
et al. 2009), Firefly Algorithm (Yang 2010), Krill Herd (KH)
algorithm (Gandomi and Alavi 2012), Fruit fly Optimization
Algorithm (FOA) (Pan 2012), Distributed BBO (Khokhar
et al. 2012). Out of these heuristics evolutionary search algo-
rithm, some of these are used to solve economic load dis-
patch problem (ELDP), combined economic load dispatch
problem(CELDP), dynamic economic dispatch problem
(DEDP) and combined economic emission dispatch (CEED)
and are reported in numerous literatures as: evolutionary
programming (Sinha et al. 2003), particle swarm optimiza-
tion (Gaing 2003), genetic algorithm (Gaing 2003; Chiang
2005), improved genetic algorithm (Devi and Krishna 2008),
Adaptive PSO and Chaotic PSO (Kumar et al. 2009), car-
dinal Priority ranking based Decision making (Singh and
Dhillon 2009), Gravitational Search Algorithm (Duman
et al. 2010; Swain et al. 2012; Güvenç et al. 2012), biogeog-
raphy based optimization (Bhattacharya and Chattopadhyay
2010a, b; Rajasomashekar and Aravindhababu 2012), intelli-
gent water drop algorithm (Rayapudi 2011), hybrid harmony
search algorithm (Pandi and Panigrahi 2011), Firefly Algo-
rithm (Yang et al. 2012), Cuckoo search algorithm (Adriane
2013; Thao and Thang 2014), biogeography based optimiza-
tion (Rajasomashekar and Aravindhababu 2012), differential
harmony Search (Wang and Li 2013), Hybrid particle swarm
optimization and gravitational search algorithm (Dubey et al.
2013), differential evolution (Ravi and Christober Asir Rajan
2013), modified ant colony optimization (Gopalakrishnan
and Krishnan 2013), modified harmony search (Secui et al.
2014), Hybrid GA-MGA (Kherfane et al. 2014), artificial
bee colony (Aydin et al. 2014). Although no optimization
algorithm can perform general enough to solve all optimiza-
tions problems, each optimization algorithm have their own
advantages and disadvantages. The brief description of cat-
egorized literature survey is summarized below in Table 1.
The limitations of some of these well known optimiza-
tion algorithms are described below. The major limitations
of the numerical techniques and dynamic programming
method are the size or dimensions of the problem, large
computational time and complexity in programming. The
mixed integer programming methods for solving the eco-
nomic load dispatch problem fails when the participation
of number of units increases because they require a large
memory and suffer from great computational delay. Gradient
Descent method is distracted for Non-Differentiable search
spaces. The Lagrangian Relaxation (LR) approach fails to
obtain solution feasibility and solution quality of problems
and becomes complex if the number of units are more. The
 Table 1  Literature survey of economic load dispatch problem
Author’s name
Walters and Sheble (1993)
Park et al. (2010)
Cai et al. (2012)
Selvakumar and Thanushkodi (2007) This papers introduces a new PSO (NPSO)
Amjady and Nasiri-Rad (2010)
Niknam (2010)
123
Technical description
This method emulates the optimization technique
found in nature this optimization algorithm
does not require the strict continuity of classical
techniques. But allows discontinuity and non-
linearity this technique uses the payoff informa-
tion to solve the ELD problems and uses any
type of cost curve with some adjustments
This method presents the Hopfield Neural Net-
work is use to solve the economic load dispatch
problem with non-convex piecewise quadratic
cost function with including the transmission
losses under satisfying the various constraints
This papers shows the ED problems in a power
system for the actual power system operation
consider the non-linear characteristics of a gen-
erator such as ramp rate limits and prohibited
operating zone for the econ0mic load dispatch
and its solution to the non-convex economic
dispatch problems under satisfying the vari-
ous constraints and PSO is use to solve the
non-linear and non-continuous optimization
problems of economic load dispatch
This papers shows the realistic operation of
economic dispatch should take both valve point
effects and multi fuels into account to solve the
ELD problems
The proposed method is use to solve the ED
problems with considering a the valve point
effect in which the FAPSO algorithm is the
main optimizer and nm is the local search tech-
nique and use to find out the solution for non-
linear, non-smooth and non-convex problems
Constraints
Power balance constraints. Generator capacity
constraints. Valve point effect constraints
Equality constraints (system power balance
constraints)
In-equality constraints (generation limit)
Power balance constraints. Generator capac-
ity constraints. Valve point effect constraints.
Ramp rate limit constraints. Generator
operation constraints (upper bound and lower
bound). Line flow constraints
Active power balance constraints. Genera-
tor ramp rate limits constraints. Prohibiting
operating zone constraints. Valve point loading
constraints
Power balance constraints. Generator capacity
constraints. Valve point effect constraints
Power balance constraints. Output generator
constraints
Algorithms and load demand
INAE Letters
Genetic algorithm (GA) and 3 unit system includ-
ing the valve point loading effect and transmis-
sion losses with load demand of (850 MW)
Hopfield neural network algorithm and 3 unit
system including the transmission losses with the
load demand of (850 MW)
Chaotic particle swarm optimization (CPSO) and
considering IEEE-26 buses system. 6 unit system
including transmission losses with load demand
(1263 MW). 15 unit system including transmis-
sion losses with load demand (2630 MW)
New particle swarm optimization algorithm
(NPSO) and 6 unit system including prohibited
operating zone with load demand (1263 MW).
40 unit System including valve point effect with
load demand (10,500 MW)
Adaptive real coded genetic algorithm (ARCGA)
and 40 unit system including valve point loading
effect with load demand (10,500 MW). 10 unit
system including valve point loading effect and
multiple fuel source option, non-smooth and
non- convex cost function with a load demand
(2700 MW)
Fuzzy adaptive hybrid particle swarm optimiza-
tion algorithm and 13 unit thermal generation
system including valve point loading effect
is used with the load demand (1800 MW and
2520 MW). 40 generating unit system including
valve point loading effect with a load demand of
(10,500 MW)

Table 1  (continued)
Author’s name
123
Vlachogiannis and Lee (2009)
Park et al. (2010)
Ciornei and Kyriakides (2012)
Reddy and Vaisakh (2014)
Moradi et al. (2016)
Technical description
This paper focuses on the performance evaluation Equality constraints (system power balance
of a newly introduced improved coordinated
aggregation-based particle swarm optimization
used for solving the economic load dispatch
problem under satisfying the various constraints
This paper proposes a PSO-based approach for
the non -convex Economic Dispatch problem
with satisfying the heavy constraints
This paper proposes a novel hybrid stochastic
method to solve the non-convex Economic
Dispatch problem. This method is a combina-
tion of Ant colony approach (API) and Genetic
algorithm (GA) to solve ED problems
This method uses the DE is an evolutionary
computation method for optimizing non-linear
and non-differential continues space functions.
SFLA is a newly developed meta-heuristic algo-
rithm for combinatorial optimization, having a
concept, fewer parameters, high performance to
solve large scale non-convex ELD problems
This method is applied for solving online generat- Power balance constraints. Capacity of power
ing unit ELD problems with some practical
constraints and the transmission loses are
included
Constraints
constraints). In-equality constraints (generation
limit) Valve point loading constraints
Equality constraints and in-equality constraints
(power balance constraints, generation limit
constraints ramp rate limit constraints prohib-
ited operating zone and valve point loading
constraints
Power balance constraints. Transmission con-
straints. Generation limit constraints. Ramp
limit constraints. Prohibited operating zone
constraints. Spinning reserve constraints
Active power balance constraints. Generator
capacity constraints
generation constraints
Algorithms and load demand
Improved coordinated aggregation-based particle
swarm optimization (ICA-PSO) and conduct-
ing on IEEE-30 bus system with 6 generat-
ing unit with load demand (283.4 MW), 13
unit system with load demand (1800 MW and
2520 MW without power loss and 2520 MW
with power loss), 15 unit system with load
demand (2650 MW without power loss and 2650,
2630 MW considering power loss) etc.
Improved particle swarm optimization algorithm
and 40 unit system including valve point load-
ing with load demand (10,500 MW), 15 unit
system considering POZ, ramp rate limits, and
transmission network losses with load demand
(2630 MW),10 unit including valve point loading
with load demand (2700 MW), 140 unit system
including ramp rate limits and load demand
(49,342 MW)
GA-API genetic algorithm with ant colony
approach and 3-unit system including valve point
loading. 6-unit system (obtained from IEEE-30
bus system) including valve point loading and
ramp rate limits with load demand (1263 MW).
15-unit system including valve point loading and
ramp rate limits with load demand (2630 MW).
40-unit system considering the valve point effect
and power loss with load demand (10,500 MW)
Shuffled differential evolution SDE and 13 unit
system including the transmission losses with
load demand (1800 MW and 2520 MW). 40 unit
system including valve point loading, neglecting
the transmission losses with the load demand
(10,500 MW). 140 unit system including valve
point loading, ramp rate limits, POZ, transmis-
sion loses are not taken into account with load
demand (49,342 MW)
Bat algorithm (BA) and Firefly algorithm (FA) and
3 unit system
INAE Letters
 Table 1  (continued)
Author’s name Technical description
Rahmat et al. (2014) This method is use to solving economic load
dispatch problem with valve point effect and
satisfying the other constraints like ramp rate
limits, prohibited zone
Hazra et al. (2015) This method is use to solve continues and
non-smooth cost function ELD problems with
valve point loading effect and multi-fuel option
including various constraints with and without
transmission loss
Karthigeyan et al. (2015) This method is use to solve the Economic Load
Dispatch problems with the inclusion of ramp
rate limits
Saxena and Ganguli (2015) This method is use to revolves the wind and solar Equality constraints (System power balance
power estimation through probability density
function and optimization of ELD problems the
transmission loses are neglected by renewable
sources and considering them only for thermal
units
Maity (2016) This method is use to solve the ELD problems
with valve point loading effect of thermal
power plant but without considering the trans- In-equality constraints (generation limit ramp rate included valve point loading effect, Ramp rate,
mission loses
Guvenc (2016) This method is applied for solving the classical Power balance constraints. Capacity of power
ELD problems and non-convex ELD problems
with some practical constraints and the trans-
mission loses are also included
Sharma and Moses (2016) This method is applied for solving non-linear
ELD problems with some practical constraints
and the transmission loses are also included
Bulbul et al. (2016) This method is applied for solving the non-
smooth and highly non-linear ELD problems
with some practical constraints and the trans-
mission loses are included
123
Constraints
Equality constraints (system power balance
constraints). In-equality constraints (generation
limit)
Power balance constraints. Generation limit
constraints
Ramp rate limit constraints. Valve point loading
constraints
constraints). In-equality constraints (generation
limit)
Equality constraints (system power balance
constraints).
limit prohibited zone)
generation constraints. Ramp rate limits
Power balance constraints. Power generation
constraints
Power balance constraints. Capacity of power
generation constraints. Ramp rate constraints.
Prohibited operating zone constraints. Spinning
reserve. Constraints. Valve point loading effect
Algorithms and load demand
INAE Letters
Differential evolution Immumed ant colony optimi-
zation technique (DEIANT) and conducted on
IEEE 30 bus system with 6 generating units
Oppositional chemical reaction optimization
and 10 unit system including valve point effect
without transmission loss with load demand
(2700 MW). 40 unit system with considering the
transmission loss
Improved harmony search algorithm and 20 unit
system is used including ramp rate limits and
valve point loading effect with load demand
(925, 1000, 1500, 2000, 2500 MW)
Firefly algorithm and 3 thermal unit system
considering transmission loss and 2 wind units
in which the transmission loss is neglected with
different load demand (50, 100, 125, 150, 200)
Quassi-Oppositional TLBO and 13 unit system
with load demand (1800 MW). 15unit system
limits generation limits, Prohibited zone with
load demand (2630 MW)
40 unit system included valve point loading effect
with load demand (10,500 MW)
Symbiotic Organisms Search Algorithm (SOS) and
38 unit system with load demand of (6000 MW).
3 unit system with load demand of (835 MW). 15
unit system with load demand of (2630 MW)
Genetic Algorithm (GA) and 3 unit system with
load demand of (150 MW). 6 unit system with
load demand of (700 MW)
Opposition-based krill herd algorithm (OKHA)
and 6 unit system including prohibited zone and
ramp rate limits with load demand (1263 MW).
10 unit system including valve point effect
and multi -fuel options with load demand
(2700 MW). 40 unit system including transmis-
sion loses and valve point. 140 unit system with
load demand (49,342 MW)
 INAE Letters

Branch and Bound (BB) method employs a linear function

Symbiotic organism search algorithm (SOS) which

demand (2520 MW). 15 unit system is including

generation limit, valve point effect and transmis-

to represent fuel cost, start-up cost and obtains a lower and

generation limit, valve point loading, ramp rate

without transmission loses but generation limit

is conducting on IEEE-30 bus system and con-

sion loses are neglected with load demand of
and valve point loading is included with load

limits and prohibited zone with load demand

(2630 MW). 40 unit system including power
tion (I-TLBO) algorithm and 13 unit system
Improved teaching learning based optimiza- upper bounds. The difficulty of this method is the exponen-
tial growth in the execution time for systems of a large prac-
tical size. An Expert System (ES) algorithm rectifies the
complexity in calculations and saving in computation time.
But it faces the problem if the new schedule is differing from

sidering a 6 generating units

Algorithms and load demand

schedule in database. The fuzzy theory method using fuzzy

set solves the forecasted load schedules error but it suffers
from complexity. The Hopfield neural network technique
(10,500 MW) considers more constraints but it may suffer from numeri-
cal convergence due to its training process. The Simulated
Annealing (SA) and Tabu Search (TS) are powerful, general-
purpose stochastic optimization technique, which can theo-
retically converge asymptotically to a global optimum solu-
constraints). In-equality constraints (generation

tion with probability one. But it takes much time to reach the
near-global minimum. Particle swarm optimization (PSO)
This method is powerful meta-heuristic optimiza- Power balance constraints. Generation limit
Equality constraints (system power balance

has simple concept, easy implementation, relative robustness

limit ramp rate limit prohibited zone)

to control parameters and computational efficiency (Mir-

jalili and Lewis 2014), although it has numerous advantages,
it get trapped in a local minimum, when handling heavily
constrained problems due to the limited local/global search-
ing capabilities (Dhillon and Kothari 2010; Mirjalili et al.
2014). Differential Evolution (DE) algorithm has the abil-
ity to find the true global minimum regardless of the initial
constraints

parameters values and requires few control parameters. It has

Constraints

parallel processing nature and fast convergence as compared

to conventional optimization algorithm. Although, it does
not always give an exact global optimum due to premature
convergence and may require tremendously high computa-
ing the condition that give the extreme value of
branch of mathematics concerned with obtain-
practical nonlinearities such as ramp rate limit
plants but without the transmission loses and

and prohibited operating zone are taken into

tion time because of a large number of fitness evaluations.

the function under the given circumstances
tion algorithm in which a optimization is a
with valve point loading effect of thermal
This method is use to solve ELD problems

The Biogeography Based Optimization (BBO) is an efficient

algorithm for Power System optimization, which does not
take unnecessary computational time and is good for exploit-
ing the solutions. The solutions obtained by BBO algorithm
does not die at the end of each generation like the other
optimization algorithm, but the convergence becomes slow
Technical description

for medium and large scale systems. Gravitational Search

algorithm has the advantages to explore better optimized
consideration

results, but due to the cumulative effect of the fitness func-

tion on mass, masses get heavier and heavier over the course
of iteration. This causes masses to remain in close proxim-
ity and neutralise the gravitational forces of each other in
later iterations, preventing them from rapidly exploiting the
optimum (Mirjalili and Lewis 2014). Therefore, increasing
effect of the cost function on mass, masses get greater over
Tiwari and Manjaree (2016)

the course of iteration and search process and convergence

becomes slow. To overcome the limitation of GSA, Mir-
Table 1  (continued)

jalili and Lewis (2014) proposed an Adaptive gbest-Guided

Mondal (2016)
Author’s name

Gravitational Search algorithm (AgGGSA), in which the

best mass is archived and utilised to accelerate the exploita-
tion phase, enriching the weakness of GSA. Grey wolf opti-
mizer (GWO) is a recently developed powerful evolutionary

123
INAE Letters

Fig. 1  a–c Convergence of moth towards light

algorithm proposed by Mirjalili et al. (2014) and has the satisfying all generating utilities constraints and various
ability to converge to a better quality near-optimal solu- operational constraints of the electric utilities. The economic
tion and possesses better convergence characteristics than load dispatch problem (ELDP) is a constrained optimization
other prevailing techniques reported in the recent literatures. problem and it can be mathematically expressed as follows
Also, GWO has a good balance between exploration and (Dhillon and Kothari 2010):
exploitation that result in high local optima avoidance, but ∑
U
the computation of GWO algorithm becomes slow, when min[FC(Pn )] = (𝛼n P2n + 𝛽n Pn + 𝛾n ) $∕Hour
applied to economic dispatch problem of medium and large n=1 (1)
scale power system. To overcome the drawbacks of recently
Subject to:
developed GWO algorithm, newly developed Moth Flame
Optimizer (MFO) algorithm developed by Mirjalili (2015)
(1) The energy balance equation:
is tested for the solution of non-convex and dynamic eco-
nomic load dispatch problem of electric power system in the
In conventional economic load dispatch problem, the
proposed research.
energy balance equation may be mathematically represented
as per Eq. (2a).
Economic Load Dispatch Problem ∑
U
Formulation Pn = PDemand + PLoss . (2a)
n=1
The scheduling of electric utilities along with the distribu-
tion of the generation power which must be planned to meet In the proposed research, in order to study the impact of
the load demand for a specific time period represents the wind and solar power, the energy balance equation may be
Unit Commitment Problem (UCP). Economic Load Dispatch modified as per Eqs. (2b) and (2c).
Problem (ELDP) refers the optimal generation schedule for
∑
U
the generation system to deliver the required load demand Pn = PDemand + PLoss − PWind , (2b)
plus transmission loss with the optimal generation fuel cost. n=1
Noteworthy economical benefits can be achieved by search-
ing a better solution to the Economic Load Dispatch Prob- ∑
U

lem (ELDP). The economic dispatch problem is defined so Pn = PDemand + PLoss − PSolar . (2c)
as to optimize the total operational cost of an electric power n=1

system while meeting the total load demand plus transmis- To study the impact of combined solar-wind-thermal sys-
sion losses within utilities generating limits (Dhillon and tem, the power balance constrained can be handled as per
Kothari 2010). The overall objective of Economic Load Dis- Eq. (2d).
patch Problem (ELDP) of electric power system is to plan
∑
U
the devoted (Committed) electric utilities outputs so as to Pn = PDemand + PLoss − PSolar − PWind .
congregate the load demand at optimal operating cost while n=1 (2d)

123
 INAE Letters

(2) The inequality constraints: The constrained economic load dispatch problem can be
converted to unconstrained ELD problem using penalty of
definite value, which can be mathematically expressed as:
Pnmin ≤ Pn ≤ Pmax
n
(n = 1, 2, 3, … , U). (3)
∑
U

where 𝛼n , 𝛽n and 𝛾n are cost coefficients. PDemand is load min[FC(Pn )] = Fn (Pn ) + 1000
demand. PLoss is power transmission loss. U is the number n=1
|(∑ ∑ ∑
)|
of generating units. Pn is real power generation and will act | U U U
|
∗ || Pn − PDemand − Bnm Pn Pm ||.
as decision variable. | n=1 |
| n=1 m=1 |
The most simple and approximate method of express-
(5a)
ing power transmission loss, PLoss as a function of generator
powers is through George’s Formula using B-coefficients In order to study the impact of wind power, the Eq. (5a)
and mathematically can be expressed as (Dhillon and will be mathematically modelled as per Eq. (5b).
Kothari 2010):
∑
U
min[FC(Pn )] = Fn (Pn ) + 1000
∑∑
U U
n=1
PLoss = Pgn Bnm Pgm MW, (4) |(∑ ∑ ∑
)|
| U U U
|
n=1 m=1 |
∗| Pn − PDemand − PWind − Bnm Pn Pm ||
| n=1 |
| n=1 m=1 |
where, Pgn and Pgm are the real power generations at the nth
(5b)
and mth buses, respectively.
Bnm is the loss coefficients which are constant under cer- For the consideration of impact of solar power, the
tain assumed conditions and U is the number of generating Eq. (5a) will be mathematically modelled as per Eq. (5c).
units.

Fig. 2  PSEUDO code and Flow chart of MFO algorithm

Table 2  Economic load dispatch for 3-generating units system (load demand = 150 MW)

Method Load P1 (MW) P2(MW) P3(MW) Fuel cost (Rs./h) Ploss (MW) No. of iteration Elapsed time
demand (s)
(MW)

Lambda iteration (dos Santos 150 33.4401 64.0974 55.1011 1599.9 2.66 250 NA
Coelho and Lee 2008)
PSO (dos Santos Coelho and Lee 150 33.0858 64.4545 54.8325 1598.79 2.37 250 NA
2008)
GWO (Kamboj et al. 2015) 150 30.4998 64.6208 54.8994 1597.4815 2.3444 250 4.761541
MFO 150 32.8101 64.595 54.9369 1597.482 2.342 250 2.252332

123
INAE Letters

load dispatch problem having (U-1) variables can be repre-

∑
U
min[FC(Pn )] = Fn (Pn ) + 1000 sented as:
n=1
|(∑ )| ∑
U
| U ∑U U
∑ | min[FC(Pn )] = (𝛼n P2n + 𝛽n Pn + 𝛾n ) + 1000
|
∗| Pn − PDemand − PSolar − Bnm Pn Pm ||.
| n=1 | n=1
| n=1 m=1 | | (∑ ∑ ∑
)|
(5c) | U U U
|
∗ ||abs Pn − PDemand − Bnm Pn Pm ||,
| |
The Eq.  (5) represent the unconstrained economic | n=1 n=1 m=1 |
load dispatch problem including penalty factor of (6a)
∑U ∑U
m=1 nm n m . The complete unconstrained economic
n=1
B PP

Table 3  Economic load dispatch for 3-generating units system (load demand = 850 MW)

Method Load Generation scheduling Fuel cost (Rs./h) Best cost Average cost Worst cost Iteration time (s)
demand
(MW) U1 U2 U3

Lambda iteration 850 382.258 127.419 340.323 8575.68 – – – –

GA 850 382.2552 127.4184 340.3202 8575.64 – – – –
PSO 850 394.5243 200 255.4756 8280.81 – – – –
ABC 850 300.266 149.733 400 8253.1 – – – –
MFO 850 300.266 149.733 400 8253.105 8253.1052 8253.1052 8253.1052 2.468

Table 4  Economic load dispatch for 3-generating units system (load demand = 1050 MW)

Method Load Generation scheduling Cost (Rs ./h) Best cost Average cost Worst cost Iteration time (s)
demand
(MW) U1 U2 U3

Lambda iteration 1050 487.5 162.5 400 10212.459 – – – –

GA 1050 487.498 162.499 400 10212.44 – – – –
PSO 1050 492.699 157.3 400 10123.73 – – – –
ABC 1050 492.6991 157.301 400 10123.73 – – – –
MFO 1050 492.69 157.3 400 10123.73 10123.73 10123.73 10123.73 2.4855

Convergence of MFO for 3-Generating Unit Test System[Load Demand=850 MW] Convergence of MFO for 3-Generating Units Test System[Load Demand=1050 MW]
3.927
10
MFO MFO
4.012
10
3.925
10
Fule Cost(Rs./Hour)--------->
Fuel Cost(Rs./hour)-------->

3.923 4.01
10 10

3.921
10
4.008
10

3.919
10

4.006
3.917
10
10
200 400 600 800 1000 1200 1400 1600 1800 2000 200 400 600 800 1000 1200 1400 1600 1800 2000
Iteration--------> Iteration---------->

(a) (b)

Fig. 3  a, b The convergence curve of test case-II for load demand of 850 and 1050 MW

123
 INAE Letters

Worst cost
U

2161.316
min[FC(Pn )] = (𝛼n P2n + 𝛽n Pn + 𝛾n ) + 1000
n=1
| (∑

–
–
–
–
–
–
| U
∗ ||abs Pn − PDemand − PWind − PSolar
|
| n=1

Average cost
)|
∑ ∑

2050.304
U U
|
− Bnm Pn Pm ||.
|
n=1 m=1 |

–
–
–
–
–
–
(6b)
The complete unconstrained economic load dispatch

2029.665
problem with valve point effect having (U-1) variables can

Best cost
be represented as:

–
–
–
–
–
–
∑
U
(
min[FC(Pn )] = 𝛼n P2n + 𝛽n Pn + 𝛾n
n=1
|( )|)

Cost (Rs./h)
+ || 𝛿n × sin(𝜀n × (Pmin | + 1000

2412.709
2412.538
2252.572

2030.673
2030.259
2029.665
− P |

2140.97
n n
| |
(U )
∑ ∑
U U
∑
∗ abs Pn − PDemand − Bnm Pn Pm .
n=1 n=1 m=1
(7)

209.815819
227.0275
125.4804
209.0692
209.8235
210.0079
Moth‑Flame Optimizer and Mathematical

272.042
Formulation

U5
Moth-Flame Optimization algorithm is a novel nature
inspired heuristic paradigm, proposed by Mirjalili (2015),

28.37844

73.11176
inspired by navigation method of moths in nature and based
28.38
U4

75

75
75
75
on the concept that the moth eventually converges toward the
light. The main brainwave of this optimization is direction-
finding manner of moth in nature called transverse orienta-

112.673503
tion. Moths fly in night always securing a definite angle with
147.5229

109.4146
113.7999
113.4005
Table 5  Economic load dispatch for 5-generating units (load demand = 730 MW)

147.535

respect to moon in nature called useful mechanism for trav-

175
U3

elling in straight line for extensive distances. However these

visualize insects are fascinated in fatal spiral path around
fake lights. As moth eventually converges towards light,
109.0092

101.5736
102.0669
102.9911
109.014

The steps of convergence of moth towards light are show

113.02
125

in Fig. 1a–c and mathematical model of MFO is depicted

U2

below:
The mathematical model of moth flame optimization
218.0184
229.5195
225.3845

229.5247
229.5196

algorithm proposed by Mirjalili (2015) can be explained

218.028

229.803

with the help of following steps:

U1

For economic load dispatch problem, we assume that the

optimal values of fuel cost are moths and the generation
Load demand

scheduling are the position of moths in the space. The set of

moths in a matrix can be represented as:
(MW)

730
730
730
730
730
730
730

⎡ m1,1 m1,2 … m1,d ⎤

⎢m m … m ⎥
M=⎢
2,1 2,2 2,d ⎥
Lambda iteration

⎢
, (8)
⋮ ⋮ ⋮ ⋮⎥
⎢ ⎥
⎣ mn,1 mn,2 ⋯ mn,d ⎦
Method

APSO

MFO
ABC
PSO
GA

EP

123
 INAE Letters

Table 6  Economic load dispatch for 6-generating units

Comparison of results for 6-generating units system
Load demand Methods P1 (MW) P2 (MW) P3 (MW) P4 (MW) P5 (MW) P6 (MW) Fuel cost (Rs./h) Ploss Iteration time (s)

600 MW Lambda iteration 23.7909 10.22 95.25 10.12309 202.967 181.34 32132.29 14.7988 –
PSO 23.8602 10 95.6394 100.7081 202.8315 181.1978 32094.72 14.2373 –
MFO 23.88 10 95.66 100.71 202.89 181.08 32091.51 14.2361 3.680353
700 MW Lambda iteration 28.29 10.0901 118.9873 118 230.2372 213.9068 36912.32 19.5114 –
PSO 28.29 10 118.9583 118.6747 230.763 212.7449 36912.22 19.43 –
MFO 28.31 10 118.9789 118.6792 230.822 212.6321 36908.44 19.43036 4.290743
800 MW Lambda iteration 32.9521 14.7126 141.5988 136.0345 258.1009 243.8011 41897.25 27.5 –
PSO 32.586 14.4839 141.5475 136.0435 257.6624 243.0073 41896.7 25.33 –
MFO 32.6137 14.5005 141.5637 136.0446 257.7167 242.8893 41892.38 25.3285 3.619414
900 MW Lambda iteration 36.9889 22.1821 163.01 153.2168 284.1482 273.0581 47045.32 32.6131 NA
PSO 36.848 21.0774 163.9304 153.263 284.1696 272.7301 47045.25 31.98 NA
MFO 36.87737 21.09383 163.9459 153.2269 284.2233 272.6181 47040.35 31.98542 3.605298
1000 MW Lambda iteration 40.3969 28.1002 187 171.2136 310.721 303.1006 52362.07 40.5323 NA
PSO 41.1657 27.7786 186.5604 170.5795 310.8297 302.568 52361.65 39.4821 NA
MFO 41.19 27.79 186.57 170.57 310.88 302.45 52355.77 39.479 3.571859

123
 INAE Letters

Table 7  Economic load Test system-V

dispatch for 13-generating units
[load demand = 2520 MW] ELD using MFO Comparison with others algorithms
Unit no. Power (MW) Method Cost (Rs ./h)

1 1166.877271 SA (Victorie and Jeyakumar 2004) 24970.91

2 303.8276937 GA (Victorie and Jeyakumar 2004) 24398.23
3 299.7904073 GA-SA (Victorie and Jeyakumar 2004) 24275.71
4 60 EP-SQP (Victorie and Jeyakumar 2004) 24266.44
5 109.8665501 PSO-SQP (Victorie and Jeyakumar 2004) 24261.05
6 60 UHGA (Dakuo et al. 2008) 24172.25
7 159.7331001 GA-MU (Chiang 2007) 24170.76
8 60.03842743 IGAMU (Chiang 2007) 24169.98
9 109.8665501 ACO (Pothiya et al. 2010) 241169.9
10 40 HGA (Dakuo et al. 2008) 24169.92
11 40 EDSA (Chen 2007) 24169.92
12 55 MFO (proposed method) 24164.97
13 55

where d is the number of variables (i.e. population or dimen- The moth flame optimization (MFO) algorithm is three-
sion) and n is the number of moths (i.e. generating units). tuple that approximates the global optimal of optimization
The corresponding fitness values (i.e. optimal cost) for problem and can be represented as:
all moths (i.e. generating units) can be stored in a array as
represented below:
MFO = (I, P, T), (12)
where I, P and T are three functions.
⎡ OM1 ⎤ The function I generates a random population of moths
⎢ OM ⎥ and corresponding fitness values and mathematically can be
⎢ 2⎥
represented as:
OM = ⎢ OM3 ⎥, (9)
⎢⋮ ⎥
⎢ OM ⎥
I ∶ 𝛷 = {M, OM} (13)
⎣ n⎦ The function P moves the moths around the search space.
It receive the matrix M and returns its updated one eventu-
ally and mathematical represented as:
where n is the number of moths.
Similar to moths, the set of flames in a matrix can be P∶M→M (14)
represented as: The function T returns true if the termination criterion is
satisfied and return false if the termination criterion is not
⎡ F1,1 F1,2 F1,3 … F1,d ⎤
⎢F F2,2 F2,3 … F2,d ⎥ satisfied and mathematically can be represented as:
F = ⎢ 2,1 ⎥ , (10)
⎢… … … … ⎥ T ∶ M → {True, False} (15)
⎣ Fn,1 … … Fn,d ⎦n×d
The function I is used generate initial generation
schedule and to calculate fuel cost value. The random
where d is the number of variables (i.e. population or dimen- generation used in function I can be implemented using
sion) and n is the number of moths (i.e. generating units). algorithm mentioned below: The framework of MFO algo-
The corresponding fitness values for all flames can be rithm with I, P and T can be updated as:
stored in a array as represented below:
⎡ OF1 ⎤
⎢ ⎥
⎢ OF2 ⎥
OF = ⎢ OF3 ⎥. (11)
⎢ ⎥
⎢ ⋮⎥
⎢ ⎥
⎣ OFn ⎦

123
INAE Letters

Table 8  Economic load dispatch for 40-generating units [load demand = 10500 MW]

ELD using MFO Comparison with other algorithms
Units Generation Method Best Average Worst Standard No. of
(MW) deviation evaluation

1 514.2342193 HGPSO (Ling et al. 2008) 124,797.13 126,855.7 NA 1160.91 NA

2 114 SPSO (Ling et al. 2008) 124,350.4 12,607.4 NA 1153.11 NA
3 97.39991254 PSO (Victoire and Jeyakumar 2004) 123,930.45 124,154.49 NA NA 10,000
4 179.7331001 CEP (Sinha et al. 2003) 123,488.29 124,793 126,902.89 NA NA
5 97 HGAPSO (Ling et al. 2008) 122,780 124,575.7 NA 906.04
6 140 FEP (Sinha et al. 2003) 122,679.71 124,119.37 127,249.59 NA NA
7 300 MFEP (Sinha et al. 2003) 122,647.57 123,489.74 12,4356.47 NA NA
8 300 IFPE (Sinha et al. 2003) 122,624.35 123,382 12,5740.63 NA NA
9 300 TM (Liu and Cai 2007) 122,477.78 123,078.21 12,4693.81 NA 4050
10 130 EP-SQP (Victoire and Jeyakumar 2004) 122,323.97 122,379.63 NA NA 10,000
11 94 MPSO (Park et al. 2005) 122,252.26 NA NA NA NA
12 94 ESO (Pereira-Neto et al. 2005) 122,122.16 122,558.45 12,3143.07 NA 75,000
13 125 HPSOM (Ling et al. 2008) 122,112.4 124,350.87 NA 978.75 NA
14 304.5195802 PSO-SQP (Victoire and Jeyakumar 2004) 122,094.67 122,245.25 NA NA 10,000
15 125 PSO-LRS (Selvakumar and Thanushkodi 122,035.79 122,558.45 12,3461.67 NA 20,000
2007)
16 304.5195802 IGA (Ling and Leung 2007) 121,915.3 122,811.41 12,3334 NA 100,000
17 489.2793703 HPSOWM (Ling et al. 2008) 121,915.3 122,844.4 NA 497.44 NA
18 500 IGAMU (Chiang 2007) 121,819.25 NA NA NA NA
19 511.2793703 HDE (Wang et al. 2007) 121,813.26 122,705.66 NA NA 100
20 511.2793703 DEC(2)-SQP(1) (Coelho and Mariani 2006) 121,741.97 122,295.12 12,2839.29 386.181 1800
21 523.2793703 PSO (Selvakumar and Thanushkodi 2008) 121,735.47 122,513.91 12,3467.4 NA 20,000
22 523.2793703 APSO (Selvakumar and Thanushkodi 2008) 121,704.73 122,221.36 12,2995.09 NA 20,000
23 523.2793703 NPSO-LRS (Selvakumar and Thanushkodi 121,664.43 122,209.31 12,2981.59 NA 20,000
2007)
24 523.2793703 APSO(2) (Selvakumar and Thanushkodi 2008) 121,663.52 122,153.67 12,2912.39 NA 20,000
25 523.2793703 SOHPSO (Chaturvedi et al. 2008) 121,501.14 121,853.57 12,2446.3 NA 62,500
26 523.2793703 BBO (Bhattacharya and Chattopadhyay 2010) 121,479.5 121,512.06 12,1688.66 NA 50,000
27 10 GA-PS-SQP (Alsumait et al. 2010) 121,458 122,039 NA NA 1000
28 10 BF (Panigrahi and Pandi 2008) 121,423.63 121,814.94 NA 124.876 10,000
29 10 PS (Al-Sumait et al. 2007) 121,415.14 122,332.65 12,5486.29 NA 1000
30 87.79990459 FA 121,415.05 121,416.57 12,1424.56 1.784 25,000
31 190 ST-HDE (Wang et al. 2007) 121,698.51 122,304.3 NA NA 100
32 190 MFO 120,846.37 121,514.08 12,2042.44 1000
33 190
34 200
35 200 Best, mean and worst fuel cost obtained by MFO algorithm for 40-units test system
36 200
37 110 Best cost ($/h) Mean cost ($/h) Worst cost($/h)
38 110 12,0846.37 12,1514.08 12,2042.44
39 110
40 511.2793703

123

The maximum and minimum generating capacity of gen-
erating units can be defined as:
Pmax = [ P max1 P max2 P max3 ⋯ P maxn−1 P maxn ] ,
(16)
where ­Pmaxi represents the maximum capacity of ith unit
Pmin = [ P min1 P min2 P min3 ⋯ P minn−1 P minn ] ,
(17)
where ­Pmini represents the maximum capacity of ith unit
After the initialization, the function P is iteratively run
until the function T returns true. The P function is the main
function that moves the moths around the search space. In
order to mathematically model the transverse orientation,
we update the position of each moth with respect to a flame
using the following equation:
Mi = S(Mi , Fj ), (18)
where, S indicates the spiral function, Fj indicates the jth
flame and Mi indicates the ith moth.
By selecting logarithmic spiral as the main mechanism
of moths, the logarithmic spiral for MFO algorithm can be
represented as:
S(Mi , Sj ) = Di .ebt . cos(2𝜋t) + Fj , (19)
where Di indicates the distance of ith moth for jth flame, b
is a constant for defining the shape of the logarithmic spiral
and t is a random number in [− 1,1]. Di of Eq. (19) can be
calculated as:
| |
Di = |Fj − Mi | (20)
| |
The position updating moths with respect to n different
locations in search space may degrade the exploitation of
the best promising solutions. To resolve this issue, following
mathematical mechanism is adopted (Fig. 2):
( )
N−1
flameNo. = round N − l ∗ (21)
T
Test Systems, Results and Discussion
In order to show the effectiveness of the MFO Algorithm
for Economic Load Dispatch Problem, six benchmark test
system of small and medium scale power systems having
123
INAE Letters
standard IEEE bus systems have been taken into considera-
tion. The performance of the proposed MFO algorithm is
tested in MATLAB 2013a (8.1.0.604) software on ­Intel®
core™ i-5-3470S CPU@ 2.90 GHz, 4.00 GB RAM system.
Test System‑I: 3‑Generating Unit System
Considering Transmission Losses
The first test system consists of 3-Generating units with a
load demand of 150 MW (dos Santos Coelho and Lee 2008).
Test data of 3-Generating Unit System are taken from (dos
Santos Coelho and Lee 2008), Loss Coefficients Matri-
ces are used to calculate the corresponding Transmission
losses. The algorithm is tested for 250 iterations and the
corresponding results are compared with lambda iteration
method (dos Santos Coelho and Lee 2008), particle swarm
optimization (PSO) (dos Santos Coelho and Lee 2008) and
grey wolf optimizer (GWO) (Kamboj et al. 2015). Table 2
shows that optimal fuel cost for 3-unit generating model for
150 MW load demand using GWO and MFO algorithm is
1597.4815 Rs/h, power loss using MFO is 2.3420 MW and
Iteration time for MFO algorithm is 2.252 s, which shows
the superiority of MFO algorithm over GWO and population
based PSO algorithm. MFO algorithm completely converges
in 58 iterations and takes Iteration time of 2.2523 s while
GWO algorithm takes 92 iterations for convergence and con-
verges times of 4.761541 s.
Test System‑II: 3‑Generating Unit System Without
Transmission Losses
The second test system also consisting of 3-Generating
Unit System (Bestha et al. 2014) is tested for two differ-
ent load demands of 850 and 1050 MW including transmis-
sion losses. The corresponding results are compared with
lambda iteration method (Bestha et al. 2014), genetic algo-
rithm (GA) (Bestha et al. 2014), particle swarm optimization
(PSO) (Bestha et al. 2014; dos Santos Coelho and Lee 2008),
artificial bee colony (ABC) (Bestha et al. 2014) and grey
wolf optimizer (GWO) (Kamboj et al. 2015). Tables 3 and
4 shows the comparison of results with different methodolo-
gies and it is found that optimal value of fuel cost obtained
by MFO algorithm is much less that lambda iteration, GA,
PSO, ABC and GWO. The convergence curve of test case-II
is shown in Fig. 3a, b.
INAE Letters

Convergence of MFO for 5-Generating Unit Test System[Load Demand=730 MW]

Table 9  Test data for 10-generating unit system
3.36
10 MFO
Pmin Pmax a b c
3.35
10
Fuel Cost(Rs./Hour)-------->

150 455 1000 16.19 0.00048

3.34
10
150 455 970 17.26 0.00031
10
3.33
20 130 700 16.6 0.002
3.32
20 130 680 16.5 0.00211
10
25 162 450 19.7 0.00398
3.31
10
20 80 370 22.26 0.00712
200 400 600 800 1000 1200 1400 1600 1800 2000
Iteration--------> 25 85 480 27.74 0.00079
(a) The convergence curve of test case-III for Load demand of 730 MW 10 55 660 25.92 0.00413
Convergence of MFO for 13-Unit Test System[Load Demand=1800 MW] 10 55 665 27.27 0.00222
10 55 670 27.79 0.00173
4.28
10 MFO
Fuel Cost($/Hour)------->

4.27
10
Valve point effect is taken into consideration, but transmis-
sion losses are neglected while calculating optimal fuel cost.
10
4.26 The results obtained by MFO algorithm are compared with
200 400 600 800 1000 1200 1400 1600 1800 2000
lambda iteration method (Bestha et al. 2014), genetic algo-
Iteration------>
rithm (GA) (Bestha et al. 2014), particle swarm optimiza-
(b) The convergence curve of test case-V for Load demand of 1800 MW tion (PSO) (Bestha et al. 2014), APSO (Bestha et al. 2014),
4.399
10
Convergence of MFO for 13- Generating Units Test System[Load Demand=2520 MW]
artificial bee colony(ABC) (Bestha et al. 2014), evolution-
ary programming (EP) (Bestha et al. 2014) and grey wolf
MFO

4.396

optimizer (GWO) (Kamboj et al. 2015). Table 5 shows the

10
Fuel Cost(Rs./Hour)------->

4.393
10
comparison of results with different methodologies and it is
10
4.39 found that optimal value of fuel cost obtained by MFO algo-
4.387
rithm is much less that lambda iteration, GA, PSO, APSO,
10
ABC, EP and GWO.
4.384
10
200 400 600 800 1000 1200 1400 1600 1800 2000
Iteration-------->
Test System‑IV: 6‑Generating Unit System Without
(c) The convergence curve of test case-V for Load demand of 2520 MW
Valve Point Effect
Convergence of MFO for 40-Generating Units Test System[Load Demand=10500 MW]

10
5.14 MFO The fourth test case consists of 6-Generating unit System
10
5.13
without valve point loading (dos Santos Coelho and Lee
2008). The results of 6-generating units systems are tested
Fuel Cost($/Hour)------->

5.12
10

10
5.11
for load demands of 600, 700, 800, 900 and 1000 MW and
10
5.1
are shown in Table 6 and effectiveness of MFO for 6-gener-
10
5.09
ating unit system is compared with lambda iteration method
200 400 600 800 1000 1200 1400 1600 1800 2000
(dos Santos Coelho and Lee 2008), particle swarm optimiza-
Iteration-------->
tion (PSO) (dos Santos Coelho and Lee 2008) and grey wolf
(d) The convergence curve of test case-VI for Load demand of 10500 MW optimizer (GWO) (Kamboj et al. 2015). Corresponding anal-
ysis of results (Table 6) shows that MFO algorithm yields
Fig. 4  a The convergence curve of test case-III for Load demand of better fuel cost and power loss as compared to Lambda-Iter-
730 MW. b The convergence curve of test case-V for Load demand of ation Method, Particle Swarm Optimization Algorithm and
1800 MW. c The convergence curve of test case-V for Load demand
of 2520  MW. d The convergence curve of test case-VI for Load
Grey Wolf Optimizer. Also, the convergence of algorithm is
demand of 10,500 MW much better than these algorithms.

Test System‑V: 13‑Generating Unit System

Test System‑III: 5‑Generating Unit System with Valve Point Effect
Considering Valve Point Effect
The fifth test case consists of 13-Generating unit System
The third test system consists of 5-Generating Unit System with valve point loading (Victorie and Jeyakumar 2004;
(Bestha et al. 2014) is tested for load demand of 730 MW. Dakuo et al. 2008; Chiang 2007; Pothiya et al. 2010; Dakuo

123
 INAE Letters

Table 10  24-h Load Demand pattern, solar and wind power for 10-generating unit system
Hour Load demand Solar power Wind power (MW) Hour Load demand Solar power (MW) Wind power (MW)
(MW) (MW) (MW)

1 700 0 73.082 13 1400 79.604 76.7905

2 750 0 65.2575 14 1300 74.16 69.6285
3 850 0 74.075 15 1200 62.64 60.192
4 950 0 79.6075 16 1050 45.92 64.165
5 1000 0 76.379 17 1000 24.96 70.7205
6 1100 0 68.7235 18 1100 1.28 78.115
7 1150 0 71.6965 19 1200 0 77.9825
8 1200 6.08 70.4125 20 1400 0 77.952
9 1300 33.44 75.4215 21 1300 0 76.6275
10 1400 52.08 69.336 22 1100 0 79.525
11 1450 65.84 68.889 23 900 0 77.869
12 1500 76.16 78.5315 24 800 0 78.4485

Table 11  Solution of Dynamic Economic Load Dispatch considering Thermal Generating Units

Hour P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 Hourly power Hourly fuel cost ($)
supplied (MW)

1 410 150 20 20 25 20 25 10 10 10 700 19,070.844

2 450.88803 150 20 29.111967 25 20 25 10 10 10 750 19,901.009
3 455 150 63.041347 81.958653 25 20 25 10 10 10 850 21,575.36
4 455 150 115 130 25 20 25 10 10 10 950 23,270.543
5 455 185 130 130 25 20 25 10 10 10 1000 24,134.628
6 455 285 130 130 25 20 25 10 10 10 1100 25,875.198
7 455 335 130 130 25 20 25 10 10 10 1150 26,747.808
8 455 385 130 130 25 20 25 10 10 10 1200 27,621.968
9 455 455 130 130 55 20 25 10 10 10 1300 29,448.948
10 455 448 130 130 162 20 25 10 10 10 1400 31,526.48
11 455 455 130 130 162 63 25 10 10 10 1450 32,631.851
12 455 455 130 130 162 80 25 10 43 10 1500 33,931.372
13 455 455 130 130 155 20 25 10 10 10 1400 31,502.528
14 455 455 130 130 55 20 25 10 10 10 1300 29,448.948
15 455 385 130 130 25 20 25 10 10 10 1200 27,621.968
16 455 235 130 130 25 20 25 10 10 10 1050 25,004.138
17 455 185 130 130 25 20 25 10 10 10 1000 24,134.628
18 455 285 130 130 25 20 25 10 10 10 1100 25,875.198
19 455 385 130 130 25 20 25 10 10 10 1200 27,621.968
20 455 455 130 130 155 20 25 10 10 10 1400 31,502.528
21 455 455 130 130 55 20 25 10 10 10 1300 29,448.948
22 455 285 130 130 25 20 25 10 10 10 1100 25,875.198
23 455 150 87.429587 107.57041 25 20 25 10 10 10 900 22,420.381
24 455 150 36.538034 58.461966 25 20 25 10 10 10 800 20,735.47
Total fuel cost ($) = 636,927.91

et al. 2008; Chen 2007). The results of 13-generating units PSO-SQP, UHGA, GA-MU, IGAMU, ACO and EDSA. Cor-
systems are tested for load demand of 2520 MW and are responding analysis of results (Table 7) shows that MFO
shown in Table 7 and effectiveness of MFO for 13-generat- algorithm yields better fuel cost and power loss as compared
ing unit system is compared with SA, GA, GA-SA, EP-SQP, to SA, GA, GA-SA, EP-SQP, PSO-SQP, UHGA, GA-MU,

123
INAE Letters

IGAMU, ACO and EDSA. Also, the convergence of algo- Test System‑VII: 10‑Generating Unit System
rithm is much better than these algorithms. Considering Impact of Solar and Wind Power

Test System‑VI: 40‑Generating Unit System
with Valve Point Effect
The sixth test case consists of 14-Generating unit System
with valve point loading (Ling et al. 2008; Victoire and Jeya-
kumar 2004; Sinha et al. 2003; Liu and Cai 2007; Park et al.
2005; Pereira-Neto et al. 2005; Selvakumar and Thanush-
kodi 2007; Ling and Leung 2007; Chiang 2007; Wang et al.
2007; Coelho and Mariani 2006; Selvakumar and Thanush-
kodi 2008). The results of 40-generating units systems are
tested for load demand of 10,500 MW and are shown in
Table 8 and effectiveness of MFO for 13-generating unit sys-
tem is compared with others heuristics, meta-heuristics and
evolutionary algorithms. Corresponding analysis of results
(Table 8) shows that MFO algorithm yields better fuel cost
and power loss as compared to others well known heuristics,
meta-heuristics and evolutionary algorithms. The conver-
gence curves for test system-III–VI are shown in Fig. 4a–d.
In order to consider the impact of renewable energy sources
(i.e. wind and solar power), the case of dynamic ecomic
load dispatch problem of realistic power system has been
taken into consideration, which consist of 10-thermal power
generating units, whose fuel cost coefficients, minimum
and maximum power ratings are given in Table 9. The load
demand pattern of 24-h has been taken into consideration
to study the dynamic economic load dispatch problem. The
available wind power and solar power pattern for 24-h time
horizon is shown in Table 10. In order to study the impact
of available wind and solar power, four different test studies
are taken into consideration. In first case study, the dynamic
economic load dispatch problem is solved for 24-h time
period and corresponding dispatch of thermal generating
units is depicted in Table 11. The second case study consist
of dispatch of thermal generating units with due considera-
tion of available wind power and their corresponding solu-
tions are shown in Table 12. The third case study consist of

Table 12  Solution of dynamic economic load dispatch considering impact of wind power

Hour P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 Hourly power Hourly fuel cost ($)
supplied (MW)

1 336.918 150 20 20 25 20 25 10 10 10 626.918 17,861.445

2 394.7425 150 20 20 25 20 25 10 10 10 684.7425 18,817.932
3 455 150 25.013104 45.911896 25 20 25 10 10 10 775.925 20,332.897
4 455 150 72.31489 93.07761 25 20 25 10 10 10 870.3925 21,919.381
5 455 150 100.76969 117.85131 25 20 25 10 10 10 923.621 22,821.372
6 455 216.2765 130 130 25 20 25 10 10 10 1031.2765 24,678.351
7 455 263.3035 130 130 25 20 25 10 10 10 1078.3035 25,497.029
8 455 314.5875 130 130 25 20 25 10 10 10 1129.5875 26,391.378
9 455 414.31946 125.25904 130 25 20 25 10 10 10 1224.5785 28,054.167
10 455 455 130 130 85.664 20 25 10 10 10 1330.664 30,070.196
11 455 455 130 130 136.111 20 25 10 10 10 1381.111 31,108.529
12 455 455 130 130 162 34.4685 25 10 10 10 1421.4685 31,976.94
13 455 455 130 130 78.2095 20 25 10 10 10 1323.2095 29,918.48
14 455 415.3715 130 130 25 20 25 10 10 10 1230.3715 28,153.716
15 455 324.808 130 130 25 20 25 10 10 10 1139.808 26,569.809
16 455 170.835 130 130 25 20 25 10 10 10 985.835 23,888.578
17 455 150 98.729885 125.54962 25 20 25 10 10 10 929.2795 22,917.673
18 455 206.885 130 130 25 20 25 10 10 10 1021.885 24,515.022
19 455 307.0175 130 130 25 20 25 10 10 10 1122.0175 26,259.261
20 455 455 130 130 77.048 20 25 10 10 10 1322.048 29,894.881
21 455 408.3725 130 130 25 20 25 10 10 10 1223.3725 28,031.126
22 455 205.475 130 130 25 20 25 10 10 10 1020.475 24,490.505
23 455 150 45.35227 71.77873 25 20 25 10 10 10 822.131 21,106.616
24 427.71185 150 20 23.839654 25 20 25 10 10 10 721.5515 19,428.43
Total fuel cost ($) = 604,703.71

123
 INAE Letters

Table 13  Solution of dynamic economic load dispatch considering impact of solar power

Hour P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 Hourly power Hourly fuel cost ($)
supplied (MW)

1 410 150 20 20 25 20 25 10 10 10 700 19,070.844

2 450.88802 150 20 29.111976 25 20 25 10 10 10 750 19,901.009
3 455 150 59.805278 85.194722 25 20 25 10 10 10 850 21,575.383
4 455 150 121.95194 123.04806 25 20 25 10 10 10 950 23,270.821
5 455 185 130 130 25 20 25 10 10 10 1000 24,134.628
6 455 285 130 130 25 20 25 10 10 10 1100 25,875.198
7 455 335 130 130 25 20 25 10 10 10 1150 26,747.808
8 455 378.92 130 130 25 20 25 10 10 10 1193.92 27,515.587
9 455 455 126.56 130 25 20 25 10 10 10 1266.56 28,789.527
10 455 455 130 130 102.92 20 25 10 10 10 1347.92 30,423.091
11 455 455 130 107.16 162 20 25 10 10 10 1384.16 31,260.97
12 455 455 130 130 162 36.84 25 10 10 10 1423.84 32,030.933
13 455 455 130 130 75.396 20 25 10 10 10 1320.396 29,861.334
14 455 410.84 130 130 25 20 25 10 10 10 1225.84 28,074.341
15 455 322.36 130 130 25 20 25 10 10 10 1137.36 26,527.066
16 455 189.08 130 130 25 20 25 10 10 10 1004.08 24,205.522
17 455 160.04 130 130 25 20 25 10 10 10 975.04 23701.149
18 455 283.72 130 130 25 20 25 10 10 10 1098.72 25852.88
19 455 385 130 130 25 20 25 10 10 10 1200 27621.968
20 455 455 130 130 155 20 25 10 10 10 1400 31502.528
21 455 455 130 130 55 20 25 10 10 10 1300 29448.948
22 455 285 130 130 25 20 25 10 10 10 1100 25875.198
23 455 150 84.624093 110.37591 25 20 25 10 10 10 900 22420.425
24 455 150 36.183911 58.816089 25 20 25 10 10 10 800 20735.47
Total fuel cost ($) = 626422.63

dispatch of thermal generating units with due consideration Conclusions

of available solar power and their corresponding solutions
are shown in Table 13. In the fourth case study, the impact
of available solar and wind power has been verified on avail-
able thermal generating units and the corresponding dispatch
of thermal generating units has been reported in Table 14.
The Comparative analysis for hourly fuel cost of thermal
generating units with due consideration of renewable energy
sources are shown in Fig. Figure 5a–d. The Percentage cost
saving for dynamic economic load dispatch problem, while
considering impact of wind and solar power is depicted in
Fig. 6b. It has been practically found that addition of solar
and wind power to conventional thermal generating unit’s
results in 7.15% fuel cost saving for 24-h, which results in
significant annual cost saving.
In this research paper, application of MFO algorithm is
presented for the solution of non-convex and dynamic eco-
nomic load dispatch problem of electric power system. Per-
formance of MFO algorithm is tested for small and medium
scale power plants. The effectiveness of proposed MFO
algorithm is tested with the standard IEEE bus system con-
sisting of 3, 5, 6, 10, 13 and 40-generating units model con-
sidering transmission losses (Power Loss) and valve point
effect.
The results obtained show that MFO have been suc-
cessfully implemented to solve different ELD problems
moreover, MFO is able to provide very spirited results in
terms of minimizing total fuel cost and lower transmission

123
INAE Letters

Table 14  Solution of dynamic economic load dispatch considering impact of wind and solar power
Hour P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 Hourly power Hourly fuel cost ($)
supplied (MW)

1 336.918 150 20 20 25 20 25 10 10 10 626.918 17,861.445

2 394.7425 150 20 20 25 20 25 10 10 10 684.7425 18,817.932
3 455 150 24.336572 46.588428 25 20 25 10 10 10 775.925 20,332.895
4 455 150 66.859658 98.532842 25 20 25 10 10 10 870.3925 21,919.522
5 455 150 99.009815 119.61119 25 20 25 10 10 10 923.621 22,821.375
6 455 216.2765 130 130 25 20 25 10 10 10 1031.2765 24,678.351
7 455 263.3035 130 130 25 20 25 10 10 10 1078.3035 25,497.029
8 455 308.5075 130 130 25 20 25 10 10 10 1123.5075 26,285.263
9 455 308.5075 130 130 25 20 25 10 10 10 1123.5075 26,285.263
10 455 376.1385 130 130 25 20 25 10 10 10 1191.1385 27,466.928
11 455 455 130 130 33.584 20 25 10 10 10 1278.584 29,019.502
12 455 455 130 130 100.3085 20 25 10 10 10 1345.3085 30,369.532
13 455 428.6055 130 130 25 20 25 10 10 10 1243.6055 28,385.597
14 455 341.2115 130 130 25 20 25 10 10 10 1156.2115 26,856.321
15 455 262.168 130 130 25 20 25 10 10 10 1077.168 25,477.245
16 455 150 105.78541 129.12959 25 20 25 10 10 10 939.915 23,098.674
17 455 150 93.824152 105.49535 25 20 25 10 10 10 904.3195 22,493.677
18 455 205.605 130 130 25 20 25 10 10 10 1020.605 24,492.765
19 455 307.0175 130 130 25 20 25 10 10 10 1122.0175 26,259.261
20 455 455 130 130 77.048 20 25 10 10 10 1322.048 29,894.881
21 455 408.3725 130 130 25 20 25 10 10 10 1223.3725 28,031.126
22 455 205.475 130 130 25 20 25 10 10 10 1020.475 24,490.505
23 455 150 50.798694 66.332306 25 20 25 10 10 10 822.131 21,106.621
24 427.71184 150 20 23.839658 25 20 25 10 10 10 721.5515 19,428.43
Total fuel cost ($)= 591,370.14

loss. Also, convergence of MFO is very fast as compared
to Lambda Iteration Method, Particle Swarm Optimization
(PSO) algorithm, Genetic algorithm (GA), APSO, Artificial
Bee Colony (ABC), ACO, SA, SA-GA, BBO, HDE, APSO,
IGA and grey wolf optimizer (GWO) for small scale power
systems and much better than various well known heuristics
and meta-heuristics algorithms for medium scale systems.
Also, It has been observed that the MFO has the abil-
ity to converge to a better quality near-optimal solution
and possesses better convergence characteristics than other
widespread techniques reported in the recent literatures. It is
also clear from the results obtained by different trials show
that the MFO shows a good balance between exploration
and exploitation that result in high local optima avoidance.
Also, it has been experimentally observed that the addition
of renewable energy sources to conventional thermal power
system results in significant cost saving.
Thus, this algorithm may become very promising for
solving some more complex power system optimizations
problems such as: Economic Load Dispatch for quadratic
and cubical cost function, Single and Multi-objective Eco-
nomic Load Dispatch including valve point effect, Economic
Load Dispatch incorporating wind Power, Economic Load
Dispatch incorporating Solar Power, Hydro-Thermal and
Wind-Thermal Scheduling of electric power system. Ther-
mal Scheduling incorporating Smart Grids, Hydro-Thermal
Scheduling incorporating Smart Grids, Single and Multi
Objective Unit Commitment Problem formulation, Multi-
Objective and Multi-Area Unit Commitment Problem.

123
 INAE Letters

Fig. 5  a Comparative hourly fuel cost of thermal generating units parative hourly fuel cost of thermal generating units considering
considering impact of wind power. b Comparative hourly fuel cost of impact of wind and solar power. d Comparative hourly fuel cost of
thermal generating units considering impact of solar power. c Com- thermal generating units considering impact of wind & solar power

Future Scope

Recently developed algorithms like Multi Verse Optimizer

(MVO) (Mirjalili et al. 2015), Dragonfly Algorithm (DA)
(Mirjalili 2015), and Ions Motion Optimization algorithm
(IMO) developed by Professor Seyedali Mirjalili can be
applied for the solution of Non-Convex Economic Load
Dispatch Problem for improved performance and better
convergence.

References
Adriane BS (2013) Cuckoo search for solving economic dispatch load
problem. Intell Control Autom 4:385–390
Alsumait JS, Sykulski JK, Al-Othman AK (2010) A hybrid GA–PS–
SQP method to solve power system valve-point economic dispatch
problems. Appl Energy 87:1773–1781
Al-Sumait JS, Al-Othmam AK, Sykulski JK (2007) Application of
pattern search method to power system valve-point economic load
dispatch. Electr Power Energy Syst 29(10):720–730
Amjady N, Nasiri-Rad H (2010) Solution of nonconvex and nonsmooth
economic dispatch by a new adaptive real coded genetic algo-
rithm. Expert Syst Appl 37(7):5239–5245
Aydin D, Ozyon S, Yaşar Celal, Liao Tianjun (2014) Artificial bee
colony algorithm with dynamic population size to combined eco-
nomic and emission dispatch problem. Int J Electr Power Energy
Fig. 6  a Overall fuel cost of thermal generating units considering impact Syst 54:144–153
of wind and solar power. b Percentage cost saving of for dynamic eco-
nomic load dispatch problem considering impact of wind and solar power

123
INAE Letters
Bestha M, Reddy KH, Hemakeshavulu O (2014) Economic load dis-
patch with valve-point result employing a binary bat formula. Int
J Electr Comput Eng IJECE 4(1):101–107
Bhattacharya A, Chattopadhyay PK (2010a) Solving complex eco-
nomic load dispatch problems using biogeography-based optimi-
zation. Expert Syst Appl 37(5):3605–3615
Bhattacharya A, Chattopadhyay PK (2010b) Application of biogeog-
raphy-based optimization for solving multi-objective economic
emission load dispatch problems. Electr Power Compon Syst
38(3):340–365
Bhattacharya A, Chattopadhyay PK (2010c) Solving complex eco-
nomic load dispatch problems using biogeography-based optimi-
zation. Expert Syst Appl 37:3605–3615
Bulbul SMA et al (2016) Opposition-based krill herd algorithm applied
to economic load dispatch problem. Ain Shams Eng J. https​://doi.
org/10.1016/j.asej.2016.02.003
Cai J et al (2012) A hybrid CPSO–SQP method for economic dis-
patch considering the valve-point effects. Energy Convers Manag
53(1):175–181
Chaturvedi KT, Pandit M, Srivastava L (2008) Self-organizing hier-
archical particle swarm optimization for non-convex economic
dispatch. IEEE Trans Power Syst 23(3):1079–1087
Chen CL (2007) Non-convex economic dispatch: a direct search
approach. Energy Convers Manag 48:219–225
Chiang C-L (2005) Improved genetic algorithm for power economic
dispatch of units with valve-point effects and multiple fuels. Power
Syst IEEE Trans 20(4):1690–1699
Chiang CL (2007a) Genetic-based algorithm for power economic load
dispatch. IET Gener Transm Distrib 1(2):261–269
Chiang CL (2007b) Genetic-based algorithm for power economic load
dispatch. IEE Proc Gener Transm Distrib 1(2):261–269
Ciornei I, Kyriakides E (2012) A GA-API solution for the economic
dispatch of generation in power system operation. IEEE Trans
Power Syst 27(1):233–242
Coelho LDS, Mariani VC (2006) Combining of chaotic differential
evolution and quadratic programming for economic dispatch
optimization with valve-point effect. IEEE Trans Power Syst
21(2):989–996
Dakuo H, Fuli W, Zhizhong M (2008a) Hybrid genetic algorithm for
economic dispatch with valve point effect. Electr Power Syst Res
78:626–633
Dakuo H, Fuli W, Zhizhong M (2008b) A hybrid genetic algorithm
approach based on differential evolution for economic dispatch
with valve-point effect. Electr Power Energy Syst 30:31–38
Devi AL, Krishna OV (2008) Combined economic and emission dis-
patch using evolutionary algorithms—a case study. ARPN J Eng
Appl Sci 3(6):28–35
Dhillon JS, Kothari DP (2010) Power system optimization, 2nd edn.
Prentice Hall India, New Delhi
dos Santos Coelho L, Lee CS (2008) Solving economic load dispatch
problems in power systems using chaotic and Gaussian particle
swarm optimization approaches. Int J Electr Power Energy Syst
30(5):297–307
Dubey HM, Manjaree P, Panigrahi BK, Udgir M (2013) Economic load
dispatch by hybrid swarm intelligence based gravitational search
algorithm. Int J Intell Syst Appl (Ijisa) 5(8):21–32
Duman S, Güvenç U, Yörükeren N (2010) Gravitational search algo-
rithm for economic dispatch with valve-point effects. Int Rev
Electr Eng 5(6):2890–2895
Gaing Z-L (2003) Particle swarm optimization to solving the economic
dispatch considering the generator constraints. Power Syst IEEE
Trans 18(3):1187–1195
Gandomi AH, Alavi AH (2012) Krill Herd: a new bio-inspired opti-
mization algorithm. Communications in nonlinear science and
numerical simulation
Gopalakrishnan R, Krishnan A (2013) An efficient technique to solve
combined economic and emission dispatch problem using modi-
fied Ant colony optimization. Sadhana 38(4):545–556
Guvenc U et al (2016) Application of symbiotic organisms search algo-
rithm to solve various economic load dispatch problems. Inno-
vations in intelligent systems and applications (INISTA), 2016
International Symposium on IEEE, 2016
Güvenç U, Sönmez Y, Duman S, Yörükeren N (2012) Combined eco-
nomic and emission dispatch solution using gravitational search
algorithm. Sci Iran 19(6):1754–1762
Hazra S, Roy PK, Sinha A (2015) An efficient evolutionary algorithm
applied to economic load dispatch problem. Computer, Commu-
nication, Control and Information Technology (C3IT), 2015 Third
International Conference on IEEE, 2015
Kamboj VK, Bath SK, Dhillon JS (2015) Solution of non-convex eco-
nomic load dispatch problem using grey wolf optimizer. Neural
Comput Appl. https​://doi.org/10.1007/s0052​1-015-1934-8
Karthigeyan P et al (2015) Comparison of harmony search algorithm,
improved harmony search algorithm with biogeography based
optimization algorithm for solving constrained economic load
dispatch problems. Proc Technol 21:611–618
Kherfane RL, Younes M, Kherfane N, Khodja F (2014) Solving eco-
nomic dispatch problem using hybrid GA-MGA. Energy Proc
50:937–944
Khokhar B, Parmar KPS, Dahiya S (2012) Application of biogeog-
raphy-based optimization for economic dispatch problems. Int J
Comput Appl (0975–888) 47(13):25–30
Kumar KS, Rajaram R, Tamilselvan V, Shanmugasundaram V, Naveen
S, Jayabarathi IM (2009) Economic dispatch with valve point
effect using various PSO Techniques. Power 2(6):130–135
Ling SH, Leung FHF (2007) An improved genetic algorithm with
average-bound crossover and wavelet mutation operation. Soft
Comput 11(1):7–31
Ling SH, Iu HHC, Chan KY, Lam HK, Yeung BCW, Leung FH (2008)
Hybrid particle swarm optimization with wavelet mutation and
its industrial applications. IEEE Trans Syst Man Cybern Part B
Cybern 38(3):743–763
Liu D, Cai Y (2007) Taguchi method for solving the economic dispatch
problem with non-smooth cost functions. IET Gener Transm Dis-
trib 1(5):793–803
Lu X, Zhou Y (2008) A novel global convergence algorithm: bee col-
lecting pollen algorithm. In: Advanced intelligent computing
theories and applications. With aspects of artificial intelligence.
Springer, pp 518–525
Maity D et al (2016) Implementation of quassi-oppositional TLBO
technique on economic load dispatch problem considering various
generator constraints. Electrical Energy Systems (ICEES), 2016
3rd International Conference on IEEE, 2016
Mirjalili S (2015a) Moth-flame optimization algorithm: a novel nature
inspired heuristic paradigm. Knowl Based Syst. https​://doi.
org/10.1016/j.knosy​s.2015.07.006
Mirjalili Seyedali (2015b) Dragonfly algorithm: a new meta-heuristic
optimization technique for solving single-objective, discrete,
and multi-objective problems. Neural Comput Appl. https​://doi.
org/10.1007/s0052​1-015-1920-1
Mirjalili Seyedali, Lewis Andrew (2014) Adaptive gbest-
guided gravitational search algorithm. Neural Comput Appl
25(7–8):1569–1584
Mirjalili S, Mirjalili SM, Lewis A (2014) Grey wolf optimizer. Adv
Eng Softw 69:46–61
Mirjalili S, Mirjalili SM, Hatamlou A (2015) Multi-verse optimizer: a
nature-inspired algorithm for global optimization. Neural Comput
Appl. https​://doi.org/10.1007/s0052​1-015-1870-7
Mondal A et al (2016) Solving of economic load dispatch problem
with generator constraints using ITLBO technique. In: Electrical,
123

electronics and computer science (SCEECS), 2016 IEEE Stu-
dents’ Conference on IEEE, 2016
Moradi M, Badri A, Ghandehari R (2016) Non-convex constrained
economic dispatch with valve point loading effect using a grey
wolf optimizer algorithm. Thermal Power Plants (CTPP), 2016
6th Conference on IEEE, 2016
Mucherino A, Seref O (2007) Monkey search: a novel metaheuristic
search for global optimization. In: AIP conference proceedings,
p 162
Niknam T (2010) A new fuzzy adaptive hybrid particle swarm opti-
mization algorithm for non-linear, non-smooth and non-convex
economic dispatch problem. Appl Energy 87(1):327–339
Pan W-T (2012) A new fruit fly optimization algorithm: taking the
financial distress model as an example. Knowl Based Syst
26:69–74
Pandi VR, Panigrahi BK (2011) Dynamic economic load dispatch
using hybrid swarm intelligence based harmony search algorithm.
Expert Syst Appl 38(7):8509–8514
Panigrahi BK, Pandi VR (2008) Bacterial foraging optimization:
Nelder–Mead hybrid algorithm for economic load dispatch. IET
Gener Transm Distrib 2(4):556–565
Park JB, Lee KS, Shin JR, Lee KY (2005) A particle swarm optimiza-
tion for economic dispatch with nonsmooth cost functions. IEEE
Trans Power Syst 20(1):34–42
Park J-B et al (2010) An improved particle swarm optimization for
nonconvex economic dispatch problems. IEEE Trans Power Syst
25(1):156–166
Pereira-Neto A, Unsihuay C, Saavedra OR (2005) Efficient evolu-
tionary strategy optimization procedure to solve the nonconvex
economic dispatch problem with generator constraints. IEE Proc
Gener Transm Distrib 152(5):653–660
Pinto PC, Runkler TA, Sousa JMC (2007) Wasp swarm algorithm for
dynamic MAX-SAT problems. In: International conference on
adaptive and natural computing algorithms, Springer, Berlin, Hei-
delberg, pp 350–357
Pothiya S, Ngamroo I, Kongprawechnon W (2010) Ant colony opti-
misation for economic dispatch problem with non-smooth cost
functions. Electr Power Energy Syst 32:478–487
Rahmat NA, Musirin I, Abidin AF (2014) Differential evolution immu-
nized ant colony optimization (DEIANT) technique in solving
weighted economic load dispatch problem. Asian Bull Eng Sci
Technol 1(1):17–26
Rajasomashekar S, Aravindhababu P (2012) Biogeography based opti-
mization technique for best compromise solution of economic
emission dispatch. Swarm Evol Comput 7:47–57
Ravi CN, Christober Asir Rajan C (2013) Differential evolution tech-
nique to solve combined economic emission dispatch. In: 3rd
international conference on electronics, biomedical engineering
and its applications (ICEBEA’2013), pp. 26–27
Rayapudi SR (2011) An intelligent water drop algorithm for solv-
ing economic load dispatch problem. Int J Electr Electron Eng
5(2):43–49
Reddy AS, Vaisakh K (2014) Shuffled differential evolution for large
scale economic dispatch. Electr Power Syst Res 96:237–245
Saxena N, Ganguli S (2015) Solar and wind power estimation and
economic load dispatch using firefly algorithm. Proc Comput Sci
70:688–700
Secui DC, Bendea G, Cristina H (2014) A modified harmony search
algorithm for the economic dispatch problem. Stud Inf Control
23(2):143–152
123
View publication stats
INAE Letters
Selvakumar AI, Thanushkodi K (2007) A new particle swarm optimi-
zation solution to nonconvex economic dispatch problems. IEEE
Trans Power Syst 22(1):42–51
Selvakumar AI, Thanushkodi K (2008) Anti-predatory particle swarm
optimization: solution to non-convex economic dispatch prob-
lems. Electr Power Syst Res 78:2–10
Sharma U, Moses B (2016) Analysis and optimization of economic
load dispatch using soft computing techniques. In: Electrical,
electronics, and optimization techniques (ICEEOT), International
Conference on IEEE, 2016
Shiqin Y, Jianjun J, Guangxing Y (2009) A dolphin partner optimiza-
tion. In: intelligent systems, 2009. GCIS’09. WRI Global Con-
gress, pp. 124–128
Singh L, Dhillon JS (2009) Cardinal priority ranking based decision
making for economic-emission dispatch problem. Int J Eng Sci
Technol 1(1):272–282
Sinha N, Chakrabarti V, Chattopadhyay PK (2003a) Evolutionary pro-
gramming techniques for economic load dispatch. Evol Comput
IEEE Trans 7(1):83–94
Sinha N, Chakrabati R, Chattopadhyay PK (2003b) Evolutionary pro-
gramming techniques for economic load dispatch. IEEE Trans
Evol Comput 7(1):83–94
Swain RK, Sahu NC, Hota PK (2012) Gravitational search algorithm
for optimal economic dispatch. Proc Technol 6:411–419
Thao NTP, Thang NT (2014) Environmental economic load dispatch
with quadratic fuel cost function using cuckoo search algorithm.
Int J u-and e-Serv Sci Technol 7(2):199–210
Tiwari A, Manjaree P (2016) Bid based economic load dispatch using
symbiotic organisms search algorithm. In: Engineering and tech-
nology (ICETECH), 2016 IEEE International Conference on
IEEE, 2016
Victoire TAA, Jeyakumar AE (2004) Hybrid PSO-SQP for eco-
nomic dispatch with valve-point effect. Electr Power Syst Res
71(1):51–59
Victorie TAA, Jeyakumar AE (2004) Hybrid PSO-SQP for economic
dispatch with valve-point effect. Electr Power Syst Res 71:51–59
Vlachogiannis JG, Lee KY (2009) Economic load dispatch—a compar-
ative study on heuristic optimization techniques with an improved
coordinated aggregation-based PSO. IEEE Trans Power Syst
24(2):991–1001
Walters DC, Sheble GB (1993) Genetic algorithm solution of eco-
nomic dispatch with valve point loading. IEEE Trans Power Syst
8(3):1325–1332
Wang L, Li L-p (2013) An effective differential harmony search algo-
rithm for the solving non-convex economic load dispatch prob-
lems. Int J Electr Power Energy Syst 44(1):832–843
Wang S-K, Chiou J-P, Liu C-W (2007) Nonsmooth/non-convex eco-
nomic dispatch by a novel hybrid differential evolution algorithm.
IET Gener Transm Distrib 1(5):793–803
Yang X-S (2010) Firefly algorithm, stochastic test functions and design
optimisation. Int J Bio Inspired Comput 2:78–84
Yang X-S, Deb S (2009) Cuckoo search via Lévy flights. In: Nature and
biologically inspired computing. NaBIC 2009. World Congress,
pp 210–214
Yang X-S, Hosseini SSS, Gandomi AH (2012) Firefly algorithm for
solving non-convex economic dispatch problems with valve load-
ing effect. Applied Soft Computing 12(3):1180–1186