International Journal of Power Electronics and Drive Systems (IJPEDS)

Vol. 15, No. 2, June 2024, pp. 1266~1275

ISSN: 2088-8694, DOI: 10.11591/ijpeds.v15.i2.pp1266-1275  1266

Multi-objective economic load dispatch using hybrid NSGA-II

and PVDE techniques

Mothala Chandrashekhar, Pradyumna Kumar Dhal

Department of Electrical and Electronics Engineering, Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology,
Chennai, India

Article Info ABSTRACT

Article history: Over decades, numerous methods have been used to optimize objective
functions. Where cost and emissions clash. The improved non-dominated
Received Jul 24, 2023 sorting genetic algorithm (NSGA-II) employs elitism to discover the
Revised Nov 22, 2023 optimum value and speed convergence in multi-objective optimization
Accepted Dec 12, 2023 problems. Population variant differential evolution algorithm alters
differential evolution (DE). The main distinction between DE and population
variant differential evolution algorithm (PVDE) is population replenishment.
Keywords: NSGA-II and PVDE are combined in the suggested hybrid approach. The
hybrid technique solves multi-objective optimization problems efficiently by
Economic load dispatch combining two or more methods. The hybrid technique solves multi-
Emission objective optimization problems well. This optimization problem pits cost vs
Fuel cost pollution. The hybrid approach exposes half the population to the NSGA-II
Non-dominated sorting genetic algorithm and half to the PVDE algorithm. In optimization problems with
algorithm opposing aims, such as minimizing costs and emissions, a hybrid technique
Population variant differential is utilized to find the optimal solution. Elitist diversity-preserving strategies
evolution algorithm avoid optimization issues becoming converging too soon. A 10-generator
IEEE 39 bus test system was validated using this method. The hybrid NSGA-
II and PVDE methodology achieves global optimal solutions with more
durability, simplicity, and optimization performance than existing methods.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Pradyumna Kumar Dhal
Department of Electrical and Electronics Engineering
Vel Tech Rangarajan Dr Sagunthala R&D Institute of Science and Technology
Avadi, Chennai-600061, India
Email: pradyumna.dhal@rediffmail.com

1. INTRODUCTION
A systematic planning and economic operation of generating units for electric power generation are
always important in the electrical industry. In an extensive integrated power system network, the utilization
of fossil fuels over a decade demands the optimal operation of thermal units. A little saving related to the
operation of a generating system leads to a notable reduction in fuel consumption quantities and cost [1]. In
generating systems, the problem associated with economic dispatch is bringing off the minimum operation
cost. Over recent years, emphasis on diminishing operating costs has become an issue. Load forecasting,
security analysis, unit commitment, and economic load dispatch are vital issues in the modern power system.
One of the possible ways is the gentle and efficient operation of generating units economically.
The main objective of the economic load dispatch is the economic operation of thermal units with
minimum fuel cost while maintaining the system constraints [2]. Economic load dispatch is related to the
online process of the minute-to-minute requirement of load allocation among the connected thermal units
with an objective of total cost minimization [3]. Thermal units should generate for reasonable optimization

Journal homepage: http://ijpeds.iaescore.com

Int J Pow Elec & Dri Syst ISSN:2088-8694  1267

problems. Optimization concerns include premature convergence, keeping to the local optimal value, and
convergence rate [4].
In former days, economic load dispatch optimization was a single-objective optimization issue that
minimized overall cost [5]. Electricity industries generate electricity and pollute. Many nations have rules to
minimize hazardous fumes from power plants and protect the environment. This measure reduces power
sector NO2 by two million tons/year and SO2 by 10 million tons/year. Therefore, several studies include
emission as additional target, creating multi-objective environmental and economic load dispatch [6]–[8].
Navaneetha et al. [9] Used a strong linear programming tool to reach global optima for rescheduling
power dispatch under overloaded conditions utilizing the part reduction and third simplex techniques.
A genetic algorithm (GA), a meta-heuristic related to bio-inspired evolutionary algorithms, was proposed by
Xu et al. [10]. The authors optimize multi-objectively using this method. In an interconnected system, cost,
emission dispatch, transient stability, small perturbation stability, and power exchange conflict. This method
is used in three areas. GA efficiently addresses the issue, the authors discovered. Abido [11] recommend
population-based evolutionary algorithms like the new strength Pareto evolutionary algorithm.
A connectivity-based clustering technique with a diversity mechanism controls the Pareto set without
compromising trade-off features to maintain non-dominated solution persistence.
A detailed examination of non-dominated sorting genetic algorithm (NSGA), strength Pareto
evolutionary algorithm (SPEA), and niched Pareto genetic algorithms was performed [12]. SPEA predicts
non-dominated solutions from the existing population and stores them in a repository to evaluate the cost-
emission trade-off curve. SPEA takes longer to optimize. Palanichamy and Babu [13] used a single identical
objective function to optimize. This optimization used analytical solution to describe cost and emission
characteristics as total generation. The recommended technique reduces computing time in economic,
emission, and cost function integration. Agrawal et al. [14] created updated repository particles and used
fuzzy decision-making to fix stochastic approach defects. To improve the suggested fuzzy clustering with
particle swarm algorithm (FCPSO) elite particles of the repository, self-adaptive mutation, and other proven
methodologies are used on the IEEE 30 bus six generator system to create a good compromise solution of a
diverse Pareto.
Sivasubramani and Swarup [15] introduced a new harmony search method (HSA) to restore Pareto
front uniformity. Fast non-dominated sorting and ranking are used to expand HSA on 30 and 118 IEEE bus
test systems, resulting in a more diverse Pareto front than NSGA-based methods II. Bayon et al. [16]
improved the economic emission dispatch issue analytical solution. Niknam and Mojarrad [17] propose a
modified adaptive θ-PSO with a new mutation operator to address drawbacks such as stagnant optimal
solutions.
Jubril et al. [18] used semi-definite programming, a crucial mathematical programme. Cost and
transmission losses were two goals in a multi-objective optimization problem. They made them single-
objective convex by lowering vector objective to scalar objective. Diversity pertains to nonlinear weighted
selection. Multi-area power systems by Pandit et al. [19] reduce dynamic crowding distance. This technique
optimizes multi-objective system cost, emission, and dependability. Tie line restriction is added by multi-
objective multi-area economic dispatch. Fuzzy selection priorities decision-maker significance.
El-sobky and Abo-elnaga [20] solved the multi-objective optimization problem with security, power
balancing, and min-max power generation restrictions using trust region globalization. Weighting turns
multi-objective problems into unitary problems and applies them to 30 bus six generator test network to
compare outcomes with previous approaches. Economic and stability levels are competing goals in the multi-
objective optimization issue, according to Vempalle and Dhal [21]. Load dynamics on load side composition
affect transient stability. Risk-based criteria calculate system transient stability probabilistically. A multi-
objective programming paradigm is used to get a suitable Pareto set. The planned one on ten machines 39 bus
system stabilizes load variance.
For multi-objective to single-objective conversion, Chopra et al. [22] used the cost penalty factor.
The nature-inspired grey wolf optimization technique performed well on three test systems and yielded
comparable results. Rafi et al. [23] created a novel technique for difficult power systems with strong non-
linearity. Used population variant differential evolution algorithm (PVDE) to reduce population inactivity.
Price penalty factor mono-objectivizes bi-objective.
Vijay et al. [24] introduced the multi-objective economic load dispatch issue with valve point
loading impact and gearbox losses limitations. Amorim and Rocha [25] developed objective optimization
issues to reduce pollution, fuel cost, wind generation, and gearbox power losses within system constraints.
NSGA-III addressed several objective optimization issues. The recommended dominance relation method
employs reference points. NSGA-III was tested on IEEE 30 to demonstrate its potential. Sakthivel et al. [26]
established multi-area economic environmental dispatch (MAEED) to reduce fuel cost and pollutant
emissions with tie-line, valve point loading, multi-fuel, and power balancing restrictions. MAEED should use
multi-objective squirrel search. Elitist depositary mechanism with crowding distance and dominance theory
Multi-objective economic load dispatch using hybrid NSGA-II and PVDE … (Mothala Chandrashekhar)
1268  ISSN:2088-8694

sustains diversity with non-dominant solutions. The issue is tested on 40-, 10-, and 140-unit Korean power
grids. Multi-objective squirrel search algorithm (MOSSA) beats existing approaches. For combined
economic emission dispatch, Hassan et al. [27] proposed chaotic artificial ecosystem-based optimization
(CAEO). The CAEO algorithm simulates nicely. Tahir et al. [28] Showed combined economic emission
dispatch (CEED) without and with renewable energy. The price penalty made CEED one goal. We
investigate multi-objective optimization. Flower pollination algorithm (FPA) cuts costs and pollutants. FPA
tests 11–15-unit systems with and without renewable energy. FPA surpassed previous literature methods.
Heuristic approaches like GA, neural networks (NN), simulated annealing (SA), particle swarm
optimization (PSO), Ant Colony, and artificial bee colony (ABC) and its variants have improved economic
and emission dispatch issue analysis and combination of both and produced excellent and acceptable
solutions. Power systems are modernizing daily under grid and rule control. Quality of solutions, emission
control, convergence, decreased losses, treatment of all limitations, and global optimal solutions for
combined emission and economic dispatch must be improved continuously. The literature study suggests that
the article will concentrate on novel algorithms and how they might solve combined emission and economic
dispatch concerns. This study addresses CEED valve-point impact using NSGA-II and PVDE. The reference
Pareto-front is derived using real coded genetic algorithm’s (RCGA) weighted sum. The remaining sections
are: i) the CEED issue formulation is shown in section 2; ii) in section 3, we discuss how to apply NSGA-II,
PVDE, and a hybrid NSGA-II-PVDE to the CEED problem; iii) section 4 presents the simulation results of
many test scenarios; and iv) section 5 concludes.

2. PROBLEM FORMULATION
The power business needs well-planned economic generating unit operation. In an interconnected
power system with ‘n’ units, economic load dispatch is linked to scheduling thermal unit active power
production to reduce operating costs using nonlinear cost functions. Cost and thermal unit emissions must be
reduced in scheduling difficulties for environmental reasons. Thus, the issue is a multi-objective economic
load dispatch optimization problem (MOELD) with two competing goals of minimizing cost and emission by
fulfilling equality and inequality constraints.

2.1. Mathematical formulation

The objective function and associated constraints of the multi-objective economic load dispatch
problem are given by (1).

𝑀𝑖𝑛 ∑𝑛𝑥=1[𝐶𝐹𝑥 (𝑃𝑔𝑥 ) + 𝐸𝐹𝑥 (𝑃𝑔𝑥 )] (1)

Where 𝐶𝐹𝑥 : represents the cost function of thermal unit ‘x’; 𝐸𝐹𝑥 : indicates the emission function of thermal
unit ‘x’; 𝑃𝑔𝑥 : is the real power generation of thermal unit ‘x’; 𝑥’: is a number of thermal units varying from 1
to n. The mathematical expression of a cost function 𝐶𝐹𝑥 (𝑃𝑔𝑥 ) represented in quadratic form as (2).

𝐶𝐹𝑥 (𝑃𝑔𝑥 ) = 𝑎𝑥 + 𝑏𝑥 𝑃𝑔𝑥 + 𝑐𝑥 𝑃𝑔𝑥2 (2)

Where 𝑎𝑥 , 𝑏𝑥 , 𝑐𝑥 represents the cost coefficients of thermal unit ‘𝑥’. The mathematical expression of an
emission function 𝐸𝐹𝑥 (𝑃𝑔𝑥 )represented in quadratic form as (3).

𝐸𝐹𝑥 (𝑃𝑔𝑥 ) = 𝑥 + 𝛽𝑥 𝑃𝑔𝑥 + 𝛾𝑥 𝑃𝑔𝑥2 (3)

Where 𝑥 , 𝛽𝑥 , 𝛾𝑥 represents the emission coefficients of thermal unit ‘𝑥’. The inequality and equality
constraints of the problem are as (4).

𝑃𝑔𝑥𝑚𝑖𝑛 ≤ 𝑃𝑔𝑥 ≤ 𝑃𝑔𝑥𝑚𝑎𝑥 (4)

Where 𝑃𝑔𝑥𝑚𝑖𝑛 , 𝑃𝑔𝑥𝑚𝑎𝑥 is minimum and maximum real power generation limits of ‘𝑥’ unit. Equality constraint
relates to the sum of total generation and total demand [29]–[31].

∑𝑛𝑥=1 𝑃𝑔𝑥 = 𝑃𝐷 (5)

Where PD is the total demand; 𝑃𝑔𝑥 is the power generation of thermal unit ‘x’; ‘n’ represents the total number
of thermal units

Int J Pow Elec & Dri Syst, Vol. 15, No. 2, June 2024: 1266-1275
Int J Pow Elec & Dri Syst ISSN:2088-8694  1269

3. METHODOLOGY
Maintenance of power balance is essential between the variation of load demand and power
generation from thermal units. Linear programming, nonlinear programming, quadratic programming, the
gradient technique, the Newton Raphson method, and the lambda iteration method, were all used to find an
answer to the aforementioned optimization issue. Evolutionary algorithms like the non-dominated sorting
genetic algorithm-II (NSGA-II) and PVDE are more efficient in obtaining an optimal global solution for the
MOELD.

3.1. Non-dominated sorting genetic algorithm – II (NSGA-II)

NSGA-II, an upgraded version of the non-dominated sorting genetic algorithm (NSGA), uses elitism
to find the best value and accelerate convergence in multi-objective optimization problems. Figure 1 shows
the NSGA-II flowchart for MOELD cost and emission calculations. The NSGA-II implementation phases
are:
a) To begin, random populations are created and their fitness values are assessed. These parent populations
are chosen by ranking and crowding distance. Each population's rank is decided via non-dominated
sorting [32].
b) Crowding distance reveals population proximity to neighbours. Crowding distance is assessed for each
population to identify parent populations with the same rank.
c) NSGA-II operations, including binary crossover and polynomial mutation, produce offspring from
selected parent populations.
d) Iterations continue until the maximum number is achieved. Get the optimum cost and emission values
from the optimization problem.

Figure 1. Flowchart for MOELD using NSGA-II

3.2. Population variant differential evolution algorithm (PVDE)

DE-modified PVDE. Population refreshment separates DE/PVDE. The DE population is fixed but
regenerated in PVDE using interquartile range. Figure 2 shows the PVDE approach for IEEE 39 bus system
MOELD flowchart. PVDE procedures are:
a) Set the number of deciding factors to 1, power limits, scaling factor vector (tsf), crossover probability
vector (tcp), and refreshment factor β.
b) From tsf and tcp evaluate maximum, minimum, and median values [23].
Multi-objective economic load dispatch using hybrid NSGA-II and PVDE … (Mothala Chandrashekhar)
1270  ISSN:2088-8694

c) Initialize the random parent populations which have to satisfy both equality and inequality constraints,
show in (3)-(5).
d) For the refreshment of the parent populations, Varmin and Var are evaluated.
e) Varmin is the vector difference of power limits i.e. maximum and minimum limits multiplied by 0.01.
f) Var is evaluated on parent populations based on the interquartile range (IQR).
g) IQR is defined as the difference between the upper quartile and lower quartile range.
h) If Var is less than Varmin new generation limits are evaluated. A new population is generated with the new
generation of limits which is termed a refreshed population.
i) Determine the fitness values of refreshed parent populations and find the best population i.e. which is
having minimum fitness value.
j) Find scaling factor (SF), and crossover probability (CP) which are useful for the performance of mutation
and crossover.
k) Perform mutation, crossover, and selection process to generate offspring population.
l) If the fitness of the offspring population is less than the fitness of the parent population, the offspring
population is considered for the next generation. Otherwise, the parent population is considered for the
next generation.
m) Update cp and sf as well as the current best population. Process repeats until the iteration count reaches
the maximum iteration count.
n) Finally, a set of optimal solutions are obtained called Pareto optimal solutions evaluate an optimal
solution.

Figure 2. Flowchart for MOELD using PVDE

3.3. Multi-objective optimization problems using hybrid method

Section 2's issue formulation in (1)-(3) must maintain equality and inequality restrictions in (4) and
(5). This optimization issue was addressed hybrid. NSGA-II and PVDE are combined in the suggested hybrid
approach. The hybrid technique solves multi-objective optimization problems efficiently by combining two
or more methods. The hybrid technique solves multi-objective optimization issues well. In this optimization
challenge, cost, and emission clash. The hybrid technique executes the NSGA-II algorithm on half the
population and the PVDE algorithm on the other [33]. The hybrid technique (NSGA-II and PVDE)
implementation plan is shown in Figure 3. Hybrid technique process:

Int J Pow Elec & Dri Syst, Vol. 15, No. 2, June 2024: 1266-1275
Int J Pow Elec & Dri Syst ISSN:2088-8694  1271

a) Before starting the hybrid approach, the decision variables, goals, emission and cost coefficients,
maximum iterations, and system demand are initialised.
b) The output of each generator is represented as a decision variable and presented as Pg= [Pg1, Pg2, Pg3
…] where Pg is the decision vector and Pg1, Pg2, Pg3…….. are the decision variables.
c) Random population is generated, which has to satisfy in (4) and (5).
d) Initial half of the population is considered for NSGA-II and fitness is evaluated for each parent population
using (2) and (3).
e) Apply the non-dominated sorting based on the ranking and crowding distance, and the process continues
until all ranking fronts are obtained.
f) The corresponding genetic operators i.e. selection, crossover, and mutation are implemented to generate
half of the offspring population [33].
g) The remaining parent population includes PVDE. Initially, population is analysed using interquartile
range.
h) Appraisal of scaling factor and crossover probability supports the operation during the production of
offspring populations.
i) Best population is evaluated from the half-refreshed population. The sequence of mutation, crossover, and
selection is carried out to generate half the offspring populations.
j) The two half offspring populations are combined to form the population of original size.
k) The parent population is amalgamated with the offspring population which leads to enhancement of
population size, which is twice in number.
l) Based on the best fitness values the initial size of a population is selected and remaining discarded.
m) The newly produced population serves as the parent population for the next iteration, and the iteration
process continues until the maximum number of iterations is reached.

Figure 3. Flowchart of a hybrid method

4. SIMULATION RESULTS
In this we are focusing on the economic load dispatch (ELD) and combined economic and emission
dispatch for various test systems including different buses. Standard coefficients would be provided for each
system under test. The system considered would be provided with the full loads that need to be generated.
The range of power that can be generated with each generator of the system is also provided. These values
Multi-objective economic load dispatch using hybrid NSGA-II and PVDE … (Mothala Chandrashekhar)
1272  ISSN:2088-8694

were fed as input to optimization models of hybrid NSGA-II and PVDE method. This NSGA-II-PVDE is
validated IEEE 39 bus test network with ten generators. The results of the entire algorithm provide total cost
of the system (fuel cost, emission cost, and combined total cost) and the output (load) needed to be generated
from each generator system.

4.1. IEEE 39 bus system using NSGA-II

The single-line diagram of the IEEE 39 bus system is shown in Figure 4 and its parameter values are
given in Table 1. Cost and emission values of IEEE 39 bus system in MOELD using NSGA-II for 700 MW,
850 MW, 1000 MW, 1300 MW, and 1500 MW given in Table 2. The Pareto front for maximum load
demand of 1500 MW is shown in Figure 5.

Table 1. Parameter values employed to obtain a global optimal solution for IEEE 39 bus system
in MOELD using NSGA-II
Parameter Value Parameter Value
Crossover rate 0.8 Population size 80
Mutation rate 0.2 number of iterations 100

Table 2. Cost and emission values of IEEE 39 bus system in MOELD using NSGA-II for 700 MW, 850 MW,
1000 MW, 1300 MW, and 1500 MW
Load(MW) 700 850 1000 1300 1500
P1 400.52 440.46 451.29 444.98 454.96
P2 156 159.32 248.51 436.49 451.96
P3 23.47 85.94 93.56 125.50 129.89
P4 25 25 106.61 95.49 131.11
P5 20 63.24 25 121.40 160.90
P6 10 10 10 10 66.23
P7 10 10 10 10 25.59
P8 10 10 10 10 32.60
P9 25 25 25 25 20.47
P10 20 20 20 20 27.45
Cost (Rs) 14,79,315 16,73,628 18,74,080 22,94,360 26,38,125
Emission (lb) 1,994.99 2,206.74 2,407.15 3,024.34 3,139.86

4.2. IEEE 39 bus system using PVDE

The same initial parameter values are considered in PVDE, which was considered in NSGA-II case
(i). Cost and emission values of IEEE 39 bus system in MOELD using PVDE for 700 MW, 850 MW, 1000
MW, 1300 MW, and 1500 MW is given in Table 3. In a comparison of IEEE 39 bus system with NSGA-II,
cost value increased and emission values are decreased using PVDE. Figure 6 shows the Pareto front
obtained for IEEE 39 bus system for a load of 850 MW using PVDE.

Figure 4. Single line diagram of IEEE 39 bus system

Int J Pow Elec & Dri Syst, Vol. 15, No. 2, June 2024: 1266-1275
Int J Pow Elec & Dri Syst ISSN:2088-8694  1273

Table 3. Cost and emission values of IEEE 39 bus system in MOELD using PVDE for 700 MW, 850 MW,
1000 MW, 1300 MW, and 1500 MW
Load (MW) 700 850 1000 1300 1500
P1 360.18 360.62 440.24 450.72 458.14
P2 152.79 164.40 283.32 414.98 452.14
P3 40.12 98.77 47.43 131 129.98
P4 46.90 126.20 128.23 128.99 129.99
P5 10 10 10 21.14 54.89
P6 10 10 10 10.35 54.26
P7 10 10 10 10 16.50
P8 25 25 25 82.15 141.34
P9 20 20 20 25.25 28.94
P10 25 25 25 25 34.05
Cost (Rs) 14,80,967 16,77,759 18,77,406 23,01,943 26,58,865
Emission (lb) 1,888.32 2,029.47 2,389.87 2,945.30 3,091.92

Figure 5. Pareto front of IEEE 39 bus system using Figure 6. Pareto front of IEEE 39 bus system using
NSGA-II PVDE

4.3. MOELD using hybrid method

The parameters used for this hybrid method are as in Table 4, population parameter is considered
80, the number of iterations is 100, crossover value 0.90, and mutation value 0.01. The results obtained from
this method for MOELD are given in Table 4, which illustrates the cost and emission values of IEEE 39 bus
system in MOELD using hybrid method for 700 MW, 850 MW, 1000 MW, 1300 MW, and 1500 MW. The
cost and emission values for each load demand of 24 hours with one hour time horizon are given in Table 4.
The optimal value of cost is Rs 26,45,344 and Rs 14,81,924 for loads 1500 MW (highest) 700 MW (lowest)
which is illustrated in Table 4. Table 5 shows the comparative analysis of cost and emission values of IEEE
39 bus system in MOELD.

Table 4. Cost and emission values of IEEE 39 in MOELD using HM for 700 MW, 850 MW, 1000 MW,
1300 MW, and 1500 MW
Load(MW) 700 850 1000 1300 1500
P1 410 450.57 445.11 452.64 453.93
P2 150 150 253.75 394.27 442.46
P3 20 57.99 124 128.50 130
P4 20 91.42 77.12 129.99 130
P5 25 25 25 104.49 162
P6 20 20 20 20.5 69.82
P7 25 25 25 25 26.5
P8 10 10 10 24.57 39.43
P9 10 10 10 10 24.87
P10 10 10 10 10 20.96
Cost (Rs) 14,81,924 16,76,787 18,85,833 23,07,059 26,45,344
Emission (lb) 2,009.13 2,206.05 2,435.11 2,901.30 3,109.97

Multi-objective economic load dispatch using hybrid NSGA-II and PVDE … (Mothala Chandrashekhar)
1274  ISSN:2088-8694

Table 5. Comparative analysis of cost and emission values of IEEE 39 bus system in MOELD
Method NSGA-II PVDE Hybrid method (NSGA-II-PVDE)
Load (MW) Cost (Rs) Emission (lb) Cost (Rs) Emission (lb) Cost (Rs) Emission (lb)
700 14,79,315 1,994.99 14,80,967 1,888.32 14,81,924 2,009.13
850 16,73,628 2,206.74 16,77,759 2,029.47 16,76,787 2,206.05
1000 18,74,080 2,407.15 18,77,406 2,389.87 18,85,833 2,435.11
1300 22,94,360 3,024.34 23,01,943 2,945.30 23,07,059 2,901.30
1500 26,38,125 3,139.86 26,58,865 3,091.92 26,45,344 3,109.97

5. CONCLUSION
The power sector must design and operate low-emission producing units economically to fulfill
demand. MOELD and multi-objective unit commitment (MOUC) are emphasized in power systems to save
cost and emission by preserving system limits. Many methods exist for tackling optimization issues with
competing goals. We first solve the MOELD with two competing objectives—cost and emission—using
NSGA-II and PVDE. Both optimization methods are used to two IEEE 39 bus and 17-unit test systems with
24 load demands in one hour. NSGA-II has cheap cost and PVDE has low emissions, according to
simulations. Hybrid approaches tackle multi-objective optimization problems better. Hybrid technique
combines NSGA-II and PVDE to find optimum MOELD and MOUC solutions. Results suggest that MOUC
objective values are better than MOELD optimization problem. NSGA-II, PVDE, Hybrid, and modified
hybrid methods are compared using metrics. Metric analysis shows that the hybrid approach outperforms the
modified hybrid method. That indicates the Hybrid technique reduces cost and emissions better than NSGA-
II, PVDE, and modified hybrid for all three optimization issues. The NSGA-II-PVDE algorithm also
outperformed or equaled other meta-heuristic optimization methods. Formulating any multi-objective
function in optimum power flow using HBSA makes it easy to include power system emergent challenges
like the environmentally friendly EED without complicating the solution approach.

REFERENCES
[1] C. E. Lin, S. T. Chen, and C.-L. Huang, “A direct Newton-Raphson economic dispatch,” IEEE Transactions on Power Systems,
vol. 7, no. 3, pp. 1149–1154, 1992, doi: 10.1109/59.207328.
[2] A. S. Reddy and K. Vaisakh, “Shuffled differential evolution for economic dispatch with valve point loading effects,”
International Journal of Electrical Power & Energy Systems, vol. 46, pp. 342–352, Mar. 2013, doi: 10.1016/j.ijepes.2012.10.012.
[3] D. R. Bobo, D. M. Mauzy, and F. J. Trefny, “Economic generation dispatch with responsive spinning reserve constraints,” IEEE
Transactions on Power Systems, vol. 9, no. 1, pp. 555–559, 1994, doi: 10.1109/59.317566.
[4] C. Jassadapakorn and P. Chongstitvatana, “Self-adaptation mechanism to control the diversity of the population in genetic
algorithm,” International Journal of Computer Science and Information Technology, vol. 3, no. 4, pp. 111–127, Aug. 2011, doi:
10.5121/ijcsit.2011.3409.
[5] Ching-Tzong Su and Chien-Tung Lin, “New approach with a Hopfield modeling framework to economic dispatch,” IEEE
Transactions on Power Systems, vol. 15, no. 2, pp. 541–545, May 2000, doi: 10.1109/59.867138.
[6] O. P. Akkas, E. Cam, I. Eke, and Y. Arikan, “New optimization algorithms for application to environmental economic load
dispatch in power systems,” Electrica, vol. 18, no. 2, pp. 133–142, Jul. 2018, doi: 10.5152/iujeee.2018.1825.
[7] M. Basu, “Economic environmental dispatch using multi-objective differential evolution,” Applied Soft Computing, vol. 11, no. 2,
pp. 2845–2853, Mar. 2011, doi: 10.1016/j.asoc.2010.11.014.
[8] A. Zakariazadeh, S. Jadid, and P. Siano, “Economic-environmental energy and reserve scheduling of smart distribution systems:
A multiobjective mathematical programming approach,” Energy Conversion and Management, vol. 78, pp. 151–164, Feb. 2014,
doi: 10.1016/j.enconman.2013.10.051.
[9] C. Navaneetha, G. V. N. Kumar, and V. Rafi, “Reactive power sharing in islanded microgrid by using voltage recovery operation
of droop control strategy,” in 2021 International Conference on Intelligent Technologies (CONIT), Jun. 2021, pp. 1–7. doi:
10.1109/CONIT51480.2021.9498380.
[10] J. X. Xu, C. S. Chang, and X. W. Wang, “Constrained multiobjective global optimisation of longitudinal interconnected power
system by genetic algorithm,” IEE Proceedings - Generation, Transmission and Distribution, vol. 143, no. 5, pp. 435–446, 1996,
doi: 10.1049/ip-gtd:19960418.
[11] M. A. Abido, “Environmental/economic power dispatch using multiobjective evolutionary algorithms,” IEEE Transactions on
Power Systems, vol. 18, no. 4, pp. 1529–1537, Nov. 2003, doi: 10.1109/TPWRS.2003.818693.
[12] M. A. Abido, “Multiobjective evolutionary algorithms for electric power dispatch problem,” IEEE Transactions on Evolutionary
Computation, vol. 10, no. 3, pp. 315–329, Jun. 2006, doi: 10.1109/TEVC.2005.857073.
[13] C. Palanichamy and N. S. Babu, “Analytical solution for combined economic and emissions dispatch,” Electric Power Systems
Research, vol. 78, no. 7, pp. 1129–1137, Jul. 2008, doi: 10.1016/j.epsr.2007.09.005.
[14] S. Agrawal, B. K. Panigrahi, and M. K. Tiwari, “Multiobjective particle swarm algorithm with fuzzy clustering for electrical power
dispatch,” IEEE Transactions on Evolutionary Computation, vol. 12, no. 5, pp. 529–541, Oct. 2008, doi: 10.1109/TEVC.2007.913121.
[15] S. Sivasubramani and K. S. Swarup, “Environmental/economic dispatch using multi-objective harmony search algorithm,”
Electric Power Systems Research, vol. 81, no. 9, pp. 1778–1785, Sep. 2011, doi: 10.1016/j.epsr.2011.04.007.
[16] L. Bayon, J. M. Grau, M. M. Ruiz, and P. M. Suarez, “The exact solution of the environmental/economic dispatch problem,”
IEEE Transactions on Power Systems, vol. 27, no. 2, pp. 723–731, May 2012, doi: 10.1109/TPWRS.2011.2179952.
[17] T. Niknam and H. Doagou-Mojarrad, “Multiobjective economic/emission dispatch by multiobjective thetas-particle swarm
optimisation,” IET Generation, Transmission & Distribution, vol. 6, no. 5, pp. 363–377, May 2012, doi: 10.1049/iet-gtd.2011.0698.

Int J Pow Elec & Dri Syst, Vol. 15, No. 2, June 2024: 1266-1275
Int J Pow Elec & Dri Syst ISSN:2088-8694  1275

[18] A. M. Jubril, O. A. Komolafe, and K. O. Alawode, “Solving multi-objective economic dispatch problem via semidefinite programming,”
IEEE Transactions on Power Systems, vol. 28, no. 3, pp. 2056–2064, Aug. 2013, doi: 10.1109/TPWRS.2013.2245688.
[19] M. Pandit, L. Srivastava, and M. Sharma, “Environmental economic dispatch in multi-area power system employing improved differential
evolution with fuzzy selection,” Applied Soft Computing, vol. 28, pp. 498–510, Mar. 2015, doi: 10.1016/j.asoc.2014.12.027.
[20] B. El-sobky and Y. Abo-elnaga, “Multi-objective economic emission load dispatch problem with trust-region strategy,” Electric
Power Systems Research, vol. 108, pp. 254–259, Mar. 2014, doi: 10.1016/j.epsr.2013.11.003.
[21] R. Vempalle and P. K. Dhal, “Optimal analysis of time varying load radial distribution system with photovoltaic and wind
generating system using novel hybrid optimization technique,” Renewable Energy Focus, vol. 41, pp. 246–257, Jun. 2022, doi:
10.1016/j.ref.2022.03.004.
[22] N. Chopra, G. Kumar, and S. Mehta, “Multi-objective economic emission load dispatch using grey wolf optimization,” International
Journal of Advanced Engineering Research and Science, vol. 3, no. 11, pp. 26–33, 2016, doi: 10.22161/ijaers/3.11.5.
[23] V. Rafi, P. K. Dhal, M. Rajesh, D. R. Srinivasan, M. Chandrashekhar, and N. M. Reddy, “Optimal placement of time-varying
distributed generators by using crow search and black widow - Hybrid optimization,” Measurement: Sensors, vol. 30, p. 100900,
Dec. 2023, doi: 10.1016/j.measen.2023.100900.
[24] R. Vijay, J. S. Dhillon, and P. Mathuria, “Economic emission load dispatch using fuzzy decision based whale algorithm,” in 2019
8th International Conference on Power Systems (ICPS), Dec. 2019, pp. 1–6. doi: 10.1109/ICPS48983.2019.9067662.
[25] E. A. Amorim and C. Rocha, “Optimization of wind-thermal economic-emission dispatch problem using NSGA-III,” IEEE Latin
America Transactions, vol. 18, no. 09, pp. 1555–1562, Sep. 2020, doi: 10.1109/TLA.2020.9381797.
[26] V. P. Sakthivel, H. H. Goh, S. Srikrishna, P. D. Sathya, and S. K. Abdul Rahim, “Multi-objective squirrel search algorithm for
multi-area economic environmental dispatch with multiple fuels and valve point effects,” IEEE Access, vol. 9, pp. 3988–4007,
2021, doi: 10.1109/ACCESS.2020.3046257.
[27] M. H. Hassan, S. Kamel, S. Q. Salih, T. Khurshaid, and M. Ebeed, “Developing chaotic artificial ecosystem-based optimization algorithm
for combined economic emission dispatch,” IEEE Access, vol. 9, pp. 51146–51165, 2021, doi: 10.1109/ACCESS.2021.3066914.
[28] M. W. Tahir, G. Abbas, M. A. Ilyas, N. Ullah, and S. M. Muyeen, “Economic emission and energy scheduling for renewable rich
network using bio-inspired optimization,” IEEE Access, vol. 10, pp. 79713–79729, 2022, doi: 10.1109/ACCESS.2022.3194012.
[29] V. Kumar, J. Singh, Y. Singh, and S. Sood, “Optimal economic load dispatch using genetic algorithms,” International Scholarly
and Scientific Research & Innovation, vol. 9, no. 4, pp. 463–470, 2015, doi: https://doi.org/10.5281/zenodo.1107031.
[30] V. Rafi, S. Nadendla, V. B. Nayak, K. V. N. S. Reddy, G. U. Kumar, and A. Maniteja, “Minimization of losses in 119 bus radial
distribution network using PSO algorithm,” in 2023 7th International Conference on Trends in Electronics and Informatics
(ICOEI), Apr. 2023, pp. 927–932. doi: 10.1109/ICOEI56765.2023.10125873.
[31] P. K. Dhal, V. Rafi, R. J. P. Narayana, K. Poojitha, V. Abhigna, and M. Swetha, “Loss minimization by reorganization in
distribution network using WOA technique,” in 2022 6th International Conference on Trends in Electronics and Informatics
(ICOEI), Apr. 2022, pp. 721–726. doi: 10.1109/ICOEI53556.2022.9776809.
[32] K. Deb, A. Pratap, S. Agarwal, and T. Meyarivan, “A fast and elitist multiobjective genetic algorithm: NSGA-II,” IEEE
Transactions on Evolutionary Computation, vol. 6, no. 2, pp. 182–197, Apr. 2002, doi: 10.1109/4235.996017.
[33] M. Rajkumar, K. Mahadevan, S. Kannan, and S. Baskar, “NSGA-II technique for multi-objective generation dispatch of thermal
generators with nonsmooth fuel cost functions,” Journal of Electrical Engineering and Technology, vol. 9, no. 2, pp. 423–432,
Mar. 2014, doi: 10.5370/JEET.2014.9.2.423.

BIOGRAPHIES OF AUTHORS

Mothala Chandrashekhar became a research scholar in the Department of Electrical

and Electronics Engineering at Veltech Rangarajan Dr. Sagunthala R&D Institute of Science &
Technology, Avadi, Chennai.He was completed M.Tech. in 2012from Auroras Engineering
College, Jntu University Hyderabad in the discipline of power electronics. He is interested in the
area of power electronics, artificial intelligence, electrical machines, and neural network and fuzzy
logic system. He can be contacted at email: chandrashekhar207@gmail.com.

Pradyumna Kumar Dhal received his M.E. degree power systems from Thiagarajar
College of Engineering under Madurai Kamaraj University, Madurai. He received his Ph.D. degree
in power systems from Sathyabama University, Chennai. He is currently working as professor in
Electrical & Electronics Engineering Department at Vel Tech Rangarajan Dr. Sagunthala R&D
Institute of Science and Technology, Chennai. He published technical papers in International &
National Journals and Conferences. His area of interest is power stability, power quality, and
optimization techniques in power system. He is member in ISTE & IEEE. He can be contacted at
email: pradyumna.dhal@rediffmail.com.

Multi-objective economic load dispatch using hybrid NSGA-II and PVDE … (Mothala Chandrashekhar)