International Journal of Pure and Applied Mathematics
Volume 118 No. 18 2018, 3813-3818
ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version)
url: http://www.ijpam.eu
Special Issue
MODIFIED BIOGEOGRAPHY BASED OPTIMIZATION ALGORITHM FOR
POWER DISPATCH USING MAX/MAX PRICE PENALTY FACTOR
Ramesh Ramachandaran1 and P. Avirajamanjula2
1
Research Scholar, Department of Electrical and Electronics Engineering, PRISTUniversity, India
2
Associate Professor, Department of Electrical and Electronics Engineering, PRISTUniversity, India
Abstract-This paper presents combined economic and
emission power dispatch (CEED) when the fuel cost
function is formulated as cubic function. Max/max price
penalty factor is considered in the multi-objective
function of (CEED). The fuel cost is presented with four
parameters (a, c, d, and e). Modified Biogeography
Based Optimization (MBO) approach is employed to
determine the optimal candidate solution. The MBO is
applied to minimize both the fuel cost and the emission
dispatch such as SO2, NOx, and CO2. In this study, in
order to evaluate the presented method, MBO is tested
on 6-unit system with cubic fuel cost and cubic emission
functions. The results obtained from MBO method are
compared with Lagrange method particle swarm
optimization and simulated annealing algorithms. The
results show that the MBO algorithm is better than the
others at solving such the problem of combined emission
and power dispatch.
Keywords: Generation allocation, economic dispatch,
max/max price penalty, modified biogeography based
optimization.
1. INTRODUCTION
The basic objective of economic dispatch (ED) of
electric power generation is to schedule the committed
generating unit outputs so as to meet the load demand at
minimum operating cost while satisfying all unit and system
equality and inequality constraints. The problem of ED is
posed to determine the production levels of generators to
ensure the balance between the production and the
consumption at the lowest cost [1–3]. This makes the ED
problem a large-scale highly non-linear constrained
optimization problem. In addition, the increasing public
awareness of the environmental protection and the passage
of the Clean Air Act Amendments of 1990 have forced the
utilities to modify their design or operational strategies to
reduce pollution and atmospheric emissions of the thermal
power plants.
The electrical energy supply system faces its main
problems, namely generator efficiency, transmission
efficiency, distribution system, or combination of these three
problems. Previous efforts to solve these problems were
concentrated on minimizing operational cost of fuel
consumption which has become the objective function and
other requirements as the constraints.
The rapid development of the art of digital
computing has controlled this problem by developing
several algorithms for determining the amount of energy
which can be generated by the central and transfer through
each link to meet the requirements of the consumers as
economically as possible, while respecting the constraints of
the system and all stations [4]. The other objective is the
reduction in greenhouse gas emissions, improving the
quality of the energy, a better system efficiency and high
reliability and availability [5].
3813
ijpam.eu
Different techniques have been reported in the
literature pertaining to environmental/economic dispatch
problem. In [6,7] the problem has been reduced to a single
objective problem by treating the emission as a constraint.
This formulation, however, has a severe difficulty in getting
the trade-off relations between cost and emission.
Alternatively, minimizing the emission has been handled as
another objective in addition to the cost. A linear
programming based optimization procedures in which the
objectives are considered one at a time was presented in
[8].Generally, the fuels used in thermal power plants (coal,
fuel oil, natural gas…) have bad environmental effects
because they produce carbon dioxide (CO2), sulfur dioxide
(SO2), and nitrous oxide (NOx) which are not only
greenhouse gases, but also toxic [9]. The ‗‗Air Act
Amendment‘‘ in 1990 (revised objectives for the period
2013–2020 after the amendment of 8 December in Doha)
which mandates the production of electric power industries
to SO2 emission reduction of 10 million tons/year and NO2
to 2 million tons/year [10] should compel producers to
consider these proportions in the commitment of units for
producing electrical energy and optimal allocation if they
want to reach these objectives. Several strategies such as
genetic algorithm (GA) [11], refined genetic algorithm
(RGA) [12], simple recursive approach [13], particle multi-
objective evolutionary algorithms (MOEA) [14] and swarm
optimization (PSO) [15] have been proposed to solve
various complex CEED problems.
A novel idea in view point of biogeography has
been projected by Dan Simon [16]. The biogeography based
optimization (BBO) strategy has been exploited for sensor
selection problem in aircraft engine health estimation [16].
The application of biogeography based optimization
algorithm is also extended to many power system problems
which includes power dispatch problem [17] and power
systems state estimation problem [18].This paper exploits
the modified biogeography based optimization (MBO)
algorithm to solves the combined economic and emission
dispatch problem in electrical power system. The use of the
MBOalgorithm to solve the CEED problem is therefore
justifiable by the method produces optimal solutions and
outperforms near optimal solver with respect to computation
time. The rest of the proposed research paper is sorted into
six sections. The CEED problem formulation is portrayed in
section 2. A BBO intelligent method is exhibited thoroughly
in section 3. Section 4 elucidates the philosophy of the MBO
strategy for solving CEED problem. Section 5 delineated the
Matlab simulation results and finally conclusion is devoted
in section 6.
2.COMBINED ECONOMIC AND EMISSION
DISPATCH PROBLEMS (CEED)
2.1 Economic model
The emission dispatch (ED) problemmay be
formulated as a nonlinear constrained problem [19]. The ED
International Journal of Pure and Applied Mathematics
problem assumes cubic cost function along with system
power demand and operational limit constraints. As
mentioned above, the objective of ED problems is to
minimize the fuel cost of committed generators (units)
subjected to operating constraints. Practically, the economic
power dispatch problem is usually formulated as,
Minimize
n (1)
 FCi Pi 
FCT 
i 1
Subject to
n (2)
P  P
i 1
i D  PL
Pi Min  Pi  Pi Max (3)
where, n is the number of units, FCTis the total fuel
cost, FCiand Piare the cost function and the real power
output of ith unit respectively, PDis the total demand, PLis the
transmission loss. Pi Min and Pi Max are the lower and upper
bounds of the ith unit respectively. The equality constraint,
(2) states that the total generated power should be balanced
by transmission losses and power consumption while (3)
denoting unit‘s operation constraints.
Traditionally, the fuel cost of a generator is usually
defined by a single cubic cost function,
FCi ( Pi )  ai Pi3  bi Pi 2  ci Pi  d i (4)
where, ai, bi,ciand diare cost coefficients of the ith
unit.
Conventionally, transmission loss is calculated
using B-matrix loss formula [2], i.e.,
PL  PT BP  PT B0  B00 (5)
where, P denotes the real power output of the
committed units in vector form, and B, B0 and B00 are loss
coefficients in matrix, vector and scalar respectively, which
are assumed to be constant, and reasonable accuracy can be
achieved when the actual operating conditions are close to
the base case where the B-coefficients were derived. In the
summary, the objective of economic power dispatch
optimization is to minimize FT subject to the constraints (2)
and (3).
2.2 Emission model
The toxic gases released by thermal units due to the
burning of fossil fuel sources such as SO2, NOx, and CO2
can individually contribute to the minimization of global
emissions through [20]:
E SO 2i ( Pi )  a SO 2i Pi3  bSO 2i Pi2  c SO 2i Pi  d SO 2i (6)
E NOxi ( Pi )  a NOxi Pi3  bNOxi Pi2  c NOxi Pi  d NOxi
ECO 2i ( Pi )  aCO 2i Pi  bCO 2i Pi  cCO 2i Pi  d CO 2i
3 2
In this work, we integrate the price penalty factor hi
(maximum fuel cost/maximum emission) for each gas
emission equation (6)–(8). So that, the problem can be
transformed into a function penalty, which is:
FTi ( Pi )  FCi ( Pi )  hSO 2 i E SO 2 i ( Pi ) (9)
 h NOxi E NOxi ( Pi )  hCO 2 i ECO 2 i ( Pi )
3814
Special Issue
where hSO2, hNOx, and hCO2 are price penalty factors
for SO2, NOx, and CO2, respectively, blending the emission
costs with the normal fuel costs.
h 
FCi ( PMaxi ) (10)
SO 2 i
E SO 2i ( PMaxi )
FCi ( PMaxi ) (11)
hNOxi 
E NOxi ( PMaxi )
FCi ( PMaxi ) (12)
hCO 2i 
ECO 2i ( PMaxi )
3. BIOGEOGRAPHY BASED OPTIMIZATION (BBO)
A novel optimization algorithm in view point of
biogeography has been first projected by Dan Simon [16].
Biogeography is defined as the investigation of dispersion of
biological species in ecosystem and it can analogous to the
mathematical optimization problems in the field of
engineering, medical science, economic science, sport,
business enterprise, town future planning and so on. The
species transmigrate from one island to another islands
having positive geographic environment. Every island is in
fact identified as habitat. The fitness of each habitat is
denoted as Habitat Suitability Index (HSI). Habitat
properties such as temperature, rain and land surface are
indicated as Suitability Index Variables (SIV).
The Biogeography Based Optimization (BBO)
algorithm is roused on fundamental basis of scientific
mechanisms of actual biogeography. The BBO is a
stochastically streamline metaheuristic algorithm spurred by
the relocation framework of biological communities. The
BBO technique usances a lexicon equivalent to the
terminologies used in biogeography whereas each habitat is
typified as a feasible solution of an optimization problem.
The variables of each solution in such problem are denoted
as SIV. The habitats are assessed based HSI which is
correspondent to the target fitness (objective) function of the
engineering problems. Generally in BBO technique, HSI are
reliant factors whereas SIV are the autonomous factors.
This nature inspired BBO technique can be devised
as follows.
1. The BBO begins by instating the algorithm
parameters such as number of SIVs, maximum
species count, stopping condition (Maximum
iteration or tolerance), maximum immigration and
emigration rates E and I, mutation probability and
modification probability.
2. Create initial islands arbitrarily.
3. Compute emigration rates (species leaving from the
islands) and the immigration rates (species landing
to the islands) for every island.
4. Based on immigration rates, pick the immigration
islands probabilistically. Choose emigration islands
based on emigration rates and fitness proportionate
selection.
5. Migrate arbitrarily chosen SIV from the selected
(7) island in the step 4.
6. Based on mutation probability, execute mutation of
(8)
every island probabilistically.
7. Compute the HSI (fitness value) of every individual
island.
8. If the convergence criteria are not achieved, go to
third step; otherwise, stop and go to the next step.
9. Output the global optimum solution and
corresponding HSI (fitness value).
The efficiency and performance of the naturally
inspired BBO technique was analyzed and contrasted with
seven other EA optimization algorithms [16]. Correlation of
International Journal of Pure and Applied Mathematics
the efficiency of BBO algorithm with other existing
techniques employing power system problem demonstrates
the huge superiority of BBO technique [17, 18]. The author
additionally gives an illustration of the utilization of BBO
algorithm in sensor selection problem of aircraft engine
[16].
3.1 Modified Biogeography Based Optimization (MBO)
The migration and mutation are two principle
operators in the BBO algorithm. Generally, BBO gives a
distinct solution space running from lower limit to upper
limit and with predefined granularity. This paper proposes a
novel migration operator called improved migration process,
keeping in mind that to enhance the fast convergences
efficiency and performance, and it give rise to a global
optimal solution/habitat to the engineering minimization or
maximization problems.
Mutation Operator
Mutation is a probabilistic operation that arbitrarily
adjusts the habitat‘s SIV in light of its earlier probability of
presence. Specifically, an arbitrarily created SIV substitutes
a chosen SIV in habitat HI based on mutation probability. In
spite of the fact that mutation is not the most imperative
operator, the chance of enhancement in habitat solution is
gotten by distracting the habitat‘s SIV in mutation process.
Generally, mutation probability is inversely proportional to
habitat probability and is characterized by

 PI 
m I  mmax 1   (12)

 Pmax 

wheremmax is maximum possible mutation
probability, Pmax = argmax(PI), I = 1,…….., n (n is number
of population) and PI is the habitat probability. The mutation
operation leans to increase diverseness between the habitats.
Improved Migration Operator
In actual biogeography, species migration process
is defined as deracination of species among various habitats
whereas in BBO algorithm, migration is a probabilistic
operation that modifies each habitat HI by distributing
properties among the habitats. In basic BBO work [16],
probability that the habitat HI is chosen as the immigrating
habitat is corresponding to its immigration rate λ I, and
probability that the habitat HJ is chosen as the emigrating
habitat is corresponding to its emigration rate. Migration can
be conveyed as:
H I (SIV)  H J (SIV) (13)
A single SIV is conceived as an optimization
problem variable and the arrangement of all conceivable
SIVs is the search space from which an optimum habitat
will be identified. (10) depicts that a habitat property of the
habitat HI is substituted by the property of another habitat
HJ. Each and every HI occupies in an island having its own
emigration rate μJ and immigration rate λI. A decent habitat
solution has generally low λ and high μ, where the
opposition is valid for a worst habitat solution. The
emigration and immigration rates are derived from the
fitness function of the habitat solution. They can be
conveyed as:

 KI 
 
KI 

 I  I 1  ,  I  E   (14)

 n 
 
 n 
 4. SIMULATION RESULTS AND ANALYSIS
where n is the number of habitat; K I is the fitness
rank of the Ith habitat (if KI = n, then solution is poor and if
KI = 1, then solution is good); E is emigration rate and I is
immigration rate. I and E are frequently set equivalent to 1
or marginally under 1. Note of that the migration rate (11) is
3815
Special Issue
linear equations, but typically it may be a more difficult
curve.
Haiping Ma and Dan Simon [16] presented the
mixed migration operator which is speculation of the basic
BBO migration operation. The mixed migration operator is
propelled by mixed crossover in GA. In mixed crossover,
rather than replicating a parent's gene to a child
chromosome, the off spring are acquired by joining parents‘
gene, so that the equation (10) can be rewritten by:
H I (SIV)  H I (SIV)  (1   ) H J (SIV) (15)
where   0,1
In [16]  is chosen as a constant. The optimal
simulation result pointed in the paper equivalent to α = 0.5.
This paper proposes a novel migration process known as
improved migration operator which can be expressed as
follows:
H I (SIV)   (H I )H I (SIV)  (1   (H I ))H J (SIV) (16)
KI
 (H I )  (17)
KI  KJ
It ought to be noticed that both expressions (12)
and (13) can give a new habitat in the problem search space,
whereas the expression (10) simply give re-development of
the new habitat variables from the choice of constant pool.
From the expression (9), the mutation operation simply
creates a new SIV but not individuals from the group of
initial habitat‘s SIVs. Henceforth, a technique utilizing (9)
may require high initial habitats or very huge mutation rate.
By substituting equation (13) in (14) to obtain the improved
version of migration operator and it is characterized as:
H I ( SIV ) 
K I H I (SIV)  K J H I (SIV)
K I  K J  (18)
Improved migration operation is an alluring BBO
modification from unique perspectives. From another
point of view, the local optimal solution can be lesser extent
to degrade due to migration process. The algorithm steps in
improved BBO method are briefly explained as follows:
Generate n random habitats H1, H2, H3…………..…Hn;
Calculate fitness value HSI for all habitats;
While the stooping criteria are not achieved do
Calculate emigration rate μ and immigration rate λ for every
habitat using fitness value HSI;
For each habitat (feasible point)
For each SIV (habitat properties)
Choose habitat HI with probability α λI;
If HI is chosen then
Choose HI with probability α μI;
If HJ is chosen then calculate new habitat using (15)
End If
End If
Choose habitat HI (SIV) based on mutation probability mI;
If HI (SIV) is chosen then
Substitute HI (SIV) with a new randomly generated SIV;
End If
Next for
Recalculate fitness value HSI;
Next for
End while
To verify the effectiveness and capability of the
proposed method, 6-unit system is used for analysis and the
results are reported in the next section.
The presented modified biogeography based
optimization has been tested on a 6-unit system with SO2,
NOx, and CO2 emission. The algorithm is implemented
inMATLAB simulation software. The cost coefficients &
International Journal of Pure and Applied Mathematics Special Issue

generation limits andemission coefficients are given Table 1 25.268, 17.301 and 22.431 MW for the 6 units, respectively.
and 2 respectively [21]. We can note that the all 6 units satisfy the generation limits
constraint (Pmin and Pmax). The emissions are increased when
Table 1 Cost coefficient and power limits of 6-unit system the demand is increased (from 3146.831 to 5335.805 kg for
ai bi ci di Pmin Pmax SO2 emission, from 2406.236 to 3811.178 kg for NOx
P1 0.001 0.092 14.5 -136 50 200 emission, and from 2564.572 to 4328.074 kg for CO2
P2 0.0004 0.025 22 -3.5 20 80 emission). From Table 4, we can see that the fuel costs
P3 0.0006 0.075 23 -81 15 50 obtained by modified biogeography based optimization
P4 0.0002 0.1 13.5 -14.5 10 50 algorithm are 2704.922 $, 3188.128 $, 3727.426$, and
P5 0.0013 0.12 11.5 -9.75 10 50 4315.633 $ for the 4 levels: 150, 175, 200, and 225 MW,
P6 0.0004 0.084 12.5 75.6 12 40 respectively. Similarly, the computational time for the
presented MBO method is lesser than existing methods
Table 3 presents the results of CEEDgeneration (3.375, 3.179, 3.291 and 2.751 sec for all 4 load levels). It is
output for each unit of the 6-unit system and computational concluded that presented modified biogeography based
time of the 4 levels of demand (150, 175, 200 and 221 MW). optimization algorithm is better than Lagrange [21], PSO
In the first load level, the power dispatch gives 50, 20, 15, [22]and SA [23]methods.

Table 2 Emission coefficientsof 6-unit system

SO2 emission NOx emission CO2 emission
ai bi ci di ai bi ci di ai bi ci di
P1 0.0005 0.15 17 -90 0.0012 0.052 18.5 -26 0.0015 0.092 14 -16
P2 0.0014 0.055 12 -30.5 0.0004 0.045 12 -35 0.0014 0.025 12.5 -93.5
P3 0.001 0.035 10 -80 0.0016 0.05 13 -15 0.0016 0.055 13.5 -85
P4 0.002 0.07 23.5 -34.5 0.0012 0.07 17.5 -74 0.0012 0.01 13.5 -24.5
P5 0.0013 0.12 21.5 -19.75 0.0003 0.04 8.5 -89 0.0023 0.04 21 -59
P6 0.0021 0.08 22.5 25.6 0.0014 0.024 15.5 -75 0.0014 0.08 22 -70

Table 3 Results of CEED for the 6-unit system

Method SA [23] MBO SA [23] MBO SA [23] MBO SA [23] MBO
Load (MW) 150 MW 175 MW 200 MW 225 MW
P1 (MW) 50.000 50.000 50.000 50.000 50.000 50.000 50.000 50000
P2 (MW) 20.178 20.000 30.030 23.770 32.904 31.028 38.900 37249.71
P3 (MW) 15.001 15.000 15.000 15.000 15.003 15.000 18.690 18050.65
P4 (MW) 24.108 25.268 28.596 32.646 36.574 38.825 42.059 44323.37
P5 (MW) 19.827 17.301 25.421 23.918 30.938 29.482 35.351 35376.29
P6 (MW) 20.886 22.431 25.953 29.666 34.581 35.666 40.000 40.000
Fuel cost ($/h) 2705.212 2704.922 3220.513 3188.128 3735.73 3727.426 4321.515 4315.633
ET of SO2
3138.446 3146.831 3763.479 38.59485 4553.973 4592.636 5287.306 5335.805
(kg/h)
ET of NOx
2379.351 2406.236 2789.92 28.54134 3285.648 3325.346 3781.192 3811.178
(kg/h)
ET of CO2
2568.946 2564.572 3094.688 31.29204 3714.333 3715.689 4324.3 4328.074
(kg/h)
FTOTAL
10261.49 10255.21 12280.04 12241.67 14421.3 14413.71 16790.69 16784.34
($/h)
CPU time
34.157 3.375 31.3873 3.179 39.4835 3.291 21.9241 2.751
(sec)

Table 4 Comparison of results of fuel cost ($/h)

Load (MW) Lagrange [21] PSO [22] SA [23] MBO
150 2,729 2,734 2,705 2,705
175 3,475 3,236 3,221 3,188
200 4,210 3,785 3,736 3,727
225 5,131 4,402 4,322 4,316

formulated with cubic cost and emissions functions.

5. CONCLUSION Modified biogeography based optimization algorithm is
The combined economic and emission load developed to determine the solution to the dispatch problem.
dispatch problem of both fuel cost and emission functions is The problem has been formulated as multi-objective

3816
International Journal of Pure and Applied Mathematics Special Issue

optimization problem with competing fuel cost and emission [15] Abido M (2009) Multi-objective particle swarm
impact objectives. The proposed approach has a diversity- optimization for environmental/economic dispatch problem.
preserving mechanism to find widely different Pareto- Electr Power Syst Res 79:1105–1113
optimal solutions.The 6-unit system multi-objective dispatch [16] Simon, D. (2008). "Biogeography-based
problem is solved for max/max price penalty factor and 4 optimization". IEEE Transactions on Evolutionary
load levels (150, 175, 200, and 225 MW). The conclusion Computation. 12: 702–713. doi:10.1109/tevc.2008.919004.
obtained is that modified biogeography based optimization [17] Bhattacharya, Chattopadhyay P (2010) Application
algorithm can give a better solution than the other methods of biogeography- based optimization for solving
such as SA, PSO and Lagrange technique. It can be multiobjective economic emission load dispatch problems.
concluded that themodified biogeography based Electr Power ComponSyst 38:340–365
optimization algorithm is capable of giving the optimal [18] K. Jamuna and K.S. Swarup, ―Biogeography based
solution to such optimization problem in the power systems. optimization for optimal meter placement for security
constrained state estimation‖, Swarm and Evolutionary
6. REFERENCES Computation, Elsevier Publishers, vol. 1, no. 2, June 2011,
[1] Narimani MR (2011) A new modified shuffle frog pp.89-96.
leaping algorithm for non-smooth economic dispatch. World [19] Sonmez Y (2013) Estimation of fuel cost curve
ApplSci J 12(6):803–814 parameters for thermal power plants using the ABC
[2] Park JB, Jeong YW, Sin JR, Lee KY (2010) An algorithm. Turkish J ElectrEngComputSci 21:1827–1841
Improved particle swarm optimization for non-convex [20] . Kim YS, Eoz IK, Park JH (1991) Economic
economic load dispatch problems. IEEE Trans Power Syst power dispatch for piecewise quadratic cost function using
25(1):156–166 neural network. In: IEE International conference on
[3] Kamboj VK, Bath SK, Dhillon JS (2015) A novel advances in power system control, operation and
hybrid DErandom search approach for unit commitment management, Hong Kong, Nov. 1991
problem. Neural ComputAppl, 1–23 [21] Krishnamurthy S, Tzoneva R (2012) Impact of
[4] Shaw B, Ghoshal SP, Mukherjee V (2011) Solution price penalty factors on the solution of the combined
of combined economic and emission dispatch problems economic emission dispatch problem using cubic criterion
using hybrid craziness- based PSO with differential functions, power and energy society general meeting, pp. 1–
evolution. In: IEEE symposium on differential evolution 9
(SDE), pp 1–8 [22] Ziane, Ismail, FaridBenhamida, and AmelGraa.
[5] Kumar RM, Bavisetti T, Kiran K (2012) "Simulated annealing algorithm for combined economic and
Optimization of combined economic and emission dispatch emission power dispatch using max/max price penalty
problem: a comparative study for 30 bus systems. J Electr factor." Neural Computing and Applications 28.1 (2017):
Electron Eng (IOSR-JEEE) 2:37–43 197-205.
[6] Brodesky SF, Hahn RW. Assessing the influence of [23] Pitono J, Soepriyanto A, Purnomo MH (2014)
power pools on emission constrained economic dispatch. Advance optimization of economic emission dispatch by
IEEE Trans Power Syst 1986;1(1):57–62. particle swarm optimization (PSO) using cubic criterion
[7] Granelli GP, Montagna M, Pasini GL, Marannino functions and various price penalty factors. KursorSci J
P. Emission constrained dynamic dispatch. Electr Power InfTechnol 7(3):153–164
Syst Res 1992;24:56–64.
[8] Farag A, Al-Baiyat S, Cheng TC. Economic load
dispatch multiobjective optimization procedures using linear
programming techniques. IEEE Trans Power Syst
1995;10(2):731–8.
[9] Basu M (2011) Economic environmental dispatch
using multiobjective differential evolution. Elsevier Appl
Soft Comput 11:2845–2853
[10] Manteaw ED, Odero NA (2012) Combined
economic and emission dispatch solution using ABC_PSO
hybrid algorithm with valve point loading effect. Int J Sci
Res Publi 2(12):1–9
[11] Xu JX, Chang CS, Wang XW. Constrained
multiobjective global optimization of longitudinal
interconnected power system by genetic algorithm. IEE
Proc—GenerTransmDistrib 1996;143(5):435–46.
[12] Sudhakarn M, Slochanal SMR, Sreeram R,
Chandrasekhar N (2004) Application of refined genetic
algorithm to combined economic emission dispatch. J Inst
Eng (India) 85:115–119
[13] Balamurugan R, Subramanian S (2008) A
Simplified Recursive Approach to combined economic
emission dispatch. Electr Power ComponSyst 36:17–27
[14] Abido M (2003) Environmental/economic power
dispatch using multiobjective evolutionary algorithms. IEEE
Trans Power Syst 18:1529–1537

3817
3818