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

net/publication/352692877

Comparison Investigation into Power System Optimization and Constraint-

Based Generator Load Scheduling Using Metaheuristic Algorithms

Article in ECTI Transactions on Electrical Engineering Electronics and Communications · June 2021
DOI: 10.37936/ecti-eec.2021192.222310

CITATIONS READS

3 40

6 authors, including:

Dogga Raveendhra Dr Devineni Gireesh Kumar

North Carolina State University B V Raju Institute of Technology
51 PUBLICATIONS 356 CITATIONS 51 PUBLICATIONS 185 CITATIONS

SEE PROFILE SEE PROFILE

Srividya Palakaluri
Gokaraju Rangaraju Institute of Engineering & Technology
17 PUBLICATIONS 36 CITATIONS

SEE PROFILE

All content following this page was uploaded by Dr Devineni Gireesh Kumar on 22 June 2023.

The user has requested enhancement of the downloaded file.

200 ECTI TRANSACTIONS ON ELECTRICAL ENG., ELECTRONICS, AND COMMUNICATIONS VOL.19, NO.2 JUNE 2021

Comparison Investigation into Power System

Optimization and Constraint-Based Generator
Load Scheduling Using Metaheuristic
Algorithms

D. S. Naga Malleswara Rao1† , Dogga Raveendhra1 , Devineni Gireesh Kumar2 ,

Bharat Kumar Narukullapati3 ,
Davu Srinivasa Rao1 , and Srividya Devi Palakaluri1 , Non-members

ABSTRACT Teaching Learning-Based Optimization Algorithm

In this paper, a novel flower pollination algorithm
(FPA) is implemented to solve the problem of 1. INTRODUCTION
combined economic emission dispatch (CEED) in the
Nowadays, the electrical energy market is becom-
power system. The FPA is a new metaheuristic
ing increasingly combative. For sustainability in the
optimization technique, which takes a biological
present era, optimal power generation is required
approach to flower pollination. The FPA mimics the
to minimize the cost of fuel and emission dispatch.
characteristics of flower pollination according to the
The foremost objective of economic load dispatch
survival of the fittest concept. CEED represents a
(ELD) is to reduce the overall generation expenses
combination of the emission and economic dispatch
while addressing equality and inequality constraints.
functions, formulated into a single function using the
Under the existing conditions, power engineers face
penalty factor. In this paper, the effect of valve point
the major problem of generator rescheduling. In
loading in the power system network is considered
recent decades, numerous techniques have been
to obtain minimum fuel cost, minimum emissions,
implemented to decipher the problem of economic
and optimum power generation. The performance
load dispatch. Fossil fuel usage in electrical energy
of the proposed algorithm is evaluated using two
generation results in the release of harmful gases like
test systems, namely 10 and 14 generating units by
carbon dioxide (CO2 ), nitrogen oxides, and sulfur
contemplating the valve point loading effect as well
oxide which affect both the living and non-living in
as transmission loss. The results of the 10 and 14
many ways. It also affects climate change and causes
system units are compared with a learning-based op-
damage to materials and visibility. Hence, reducing
timization technique to demonstrate the effectiveness
atmospheric pollution presents a major challenge for
of the FPA. The findings reveal that the proposed
the electrical energy industry. Previously, the main
FPA gives better performance than other algorithms
purpose of economical dispatch has been confined to
with minimum fuel cost and emissions.
minimizing the overall cost of the power generation
network while rescheduling the already committed
Keywords: Valve Point Loading Effect, Economic
outputs.
Dispatch, Emission Dispatch, Combined Economic
Emission Dispatch, Flower Pollination Algorithm, Nevertheless, achieving optimal economic load
dispatch does not lead to minimum emissions and
Manuscript received on October 21, 2019 ; revised on March
the ability to meet the environmental regulations
29, 2021 ; accepted on March 31, 2021. This paper was enforced in recent decades; emission control has
recommended by Associate Editor Kaan Kerdchuen. become one of the operational objectives. This has
1 The authors are with the Department of Electrical &
Electronics Engineering, Gokaraju Rangaraju Institute of
resulted in the revelation of a new model to address
Engineering & Technology, Hyderabad, India. the economic emission dispatch (EED) problem [1–3].
2 The author is with the Department of Electrical &
Many researchers have addressed the EED issue and
Electronics Engineering, B. V. Raju Institute of Technology,
Narsapur, Medak, India.
attempted to establish a single objective function
3 The author is with the Department of Electrical & by linearly combining the other EED objectives.
Electronics Engineering, Vignan’s Foundation for Science, To determine the exchange between environmental
Technology and Research (Deemed to be University), Guntur, cost and fuel cost, the weights must be varied.
India.
† Corresponding author: [email protected] Some researchers have carried out simultaneous
©2021 Author(s). This work is licensed under a Creative optimization of multiple objectives to address the
Commons Attribution-NonCommercial-NoDerivs 4.0 License. EED problem using evolutionary algorithms [4, 5].
To view a copy of this license visit: https://creativecommons.
org/licenses/by-nc-nd/4.0/. Dieu et al. implemented a hybrid Hopfield neural
Digital Object Identifier 10.37936/ecti-eec.2021192.222310 network to solve the non-convex cost function based
COMPARISON INVESTIGATION INTO POWER SYSTEM OPTIMIZATION USING METAHEURISTIC ALGORITHMS 201

on optimal power dispatch [6]. Later, the economic

dispatch problem was solved using augmented La- N
X
grange Hopfield networks on prohibited operating F = Fi (Pi ) (1)
zones. These techniques are designed based upon the i=1

piecewise quadratic cost function and linear quadratic

programming [7]. Wang and Singh proposed a model 2. 2 Real power balancing constraint
to solve the problem of the environmental economic The equality constraint must be satisfied to
emission dispatch effect by employing an improved balance the power. The total power generated must
particle swarm optimization technique which is both be sufficient to satisfy demand and compensate line
deterministic and stochastic [8]. Pourakbari-Kasmaei loss in the system [15].
and Rashidi-Nejad implemented a state-of-the-art
effortless hybrid method (EHM) to provide an N
X
economical dispatch solution for the power network Pi − (PD + PL ) = 0 (2)
[9]. Pandian and Thanushkodi designed a futur- i=1
istic technique to solve the power system dispatch
where PD is the load demand and PL is the
problem based upon the cubic cost function and
transmission line loss.
considering transmission loss using multiagent-based
hybrid particles for swarm optimization [10]. This Here, the transmission loss is considered to achieve
method also resolves the randomness issue, with economic load dispatch. The B-coefficient method
a unique solution for the variable tuning problems (mostly used in the power industry) is used to
present in traditional power system optimization. calculate transmission loss. The power network line
Lohokare et al. implemented a novel aBBOmDE loss formula is expressed as follows [16].
method to resolve the economic dispatch problem
using both convex and non-convex cost functions [11]. N X
X N N
X
A novel reinforcement learning technique has also PL = Pi Bij Pj + Pi B0i + B00 (3)
been implemented for solving power system dispatch i=1 j=1 i=1
problems [12].
Alternatively, another approach is available to where Bij , B0i , and B00 are constants called loss
handle both emission and fuel cost contempora- coefficients. In this study, the B-coefficient method is
neously as participating objectives. In the last used as the generator capacity constraint. The power
decade, with the evolution of multi-objective search output of every generation unit should lie between
techniques, many researchers are exhibiting interest the minimum and maximum permitted power. The
in this approach. To solve the EED problem, following equation shows the inequality constraint.
metaheuristic methods are employed, specifically the
flower pollination algorithm (FPA) and teaching
learning-based optimization (TLBO) [13]. Pi,min ≤ Pi ≤ Pi,max for i = 1, 2, . . . , N (4)

2. PROBLEM FORMULATION 2. 3 Fuel cost function without the valve point

In this article, the load dispatch problem of the effect
power network can be solved by optimum generation The heat rate curve is also called the input
allocation while addressing equality and inequality and output curve of the thermal plant, represented
constraints and the generator power balance of the in Btu/hr. Multiplying the fuel cost with the
network. The cost of generation can be minimized heat rate curve gives the cost curve. Opening
through simplified quadratic cost functions, subject the inlet valves of the steam turbine increases
to various network constraints. The following the output of fossil fuel power plants. The loss
constraints and objective functions are considered for incurred by throttling is reduced when the gates are
resolving the ELD optimization problem. fully open, while increasing when gates are slightly
open. This indicates that the incremental heat rate
2. 1 Basic economic load dispatch objective characteristically rises discontinuously.
function
However, for the purpose of dispatch, the fuel cost
1) Consider the electrical power system network N function can be approximated as a quadratic function
has a number of generation units. of the active power outputs in the generating unit
2) Each and every unit is loaded with Pi MW [17].
so that it satisfies both load demand PD MW and
transmission loss PL MW. Fi (Pi ) = ai p2i + bi pi + ci (5)
3) Let Fi (Pi ) be the fuel input output of the ith
unit cost function. The units are selected to minimize where ai , bi , and ci are the fuel cost coefficients of
the total fuel cost F [14]. the ith generating unit.
202 ECTI TRANSACTIONS ON ELECTRICAL ENG., ELECTRONICS, AND COMMUNICATIONS VOL.19, NO.2 JUNE 2021

2. 4 Fuel cost function with the valve point

effect
Valve point loading is required to obtain an
accurate fuel cost curve for each generating unit.
Therefore, the sinusoidal function is associated with
the quadratic function, as given in Eq. (6). The
cost function of the steam-generating units and valve
point loadings are represented in [18].

Fi (Pi ) = ai + bi Pi + ci Pi2 + ei sin fi Pimin − Pi

(6)
where ei and fi are coefficients of the generation unit
with the valve point loading effect.

2. 5 Emission objective
Eq. (7) represents the emission function developed
by combining SOx and NOx . The total pollutants of
emission are represented by the combined quadratic
and exponential functions as [19]

N
X
αj + βj Pj + γj Pj2 + ξj exp (λj Pj )


FE (P ) =
j=1
(7)
where FE represents the total emissions, αj , βj , γj ,
ξj , and λj are the j th generator emission coefficients.

3. APPLYING THE FLOWER POLLINA-

TION ALGORITHM TO THE ECO-
NOMIC EMISSION DISPATCH (EED)
PROBLEM
Yang et al. [13] proposed the FPA inspired by
the pollination of flowering plants. The FPA has
certain advantages such as robustness and providing a
better quality solution compared to other methods. It
consists of one main parameter p (switch probability)
to facilitate easy and rapid implementation of the
algorithm to achieve the optimum solution. The
intrinsic property of the FPA algorithm is its ability
to find a space solution with high accuracy and
precision. Therefore, the FPA can be used along with
DE to address the problem of emission dispatch. The
main reason for implementing this technique is that it Fig. 1: Flower pollination algorithm (FPA) flow
has numerous advantages compared to other similar chart.
algorithms.
As shown in Fig. 1, implementation of the
algorithm first requires the amount of cross and
self-pollination to be determined. The next step
involves initializing a specified population number The FPA continues by initiating a new population,
(N ), with each containing a group of variables, depending on whether the random number is greater
optimized by applying the objective function. This or less than parameter p, which represents the
technique consists of a main indexed term such as self- or cross-pollination of flowers. The random
the flower constancy of every population to determine numbers range between 0 and 1. In the case of
the strength of the variables as well as minimizing global pollination, the agents pollinating these flowers
the objective function. Depending on the flower should move in random step sizes from one flower
constancy, a good sample can be found among the to another. This moment is a representation of levy
queued populations [20]. flight distribution, as given by Eq. (8) [20].
COMPARISON INVESTIGATION INTO POWER SYSTEM OPTIMIZATION USING METAHEURISTIC ALGORITHMS 203

Table 1: Simulation results of TLBO for the Table 2: Simulation results of the FPA for the
ten-unit system with the valve point loading effect for ten-unit system with the valve point loading effect for
different objectives where PD = 2000 MW. different objectives where PD = 2000 MW.
Number TLBO algorithm Number TLBO algorithm
of units Economic Emission of units Economic Emission
CEED CEED
Dispatch Dispatch Dispatch Dispatch
1 55.0000 54.9998 55.0000 1 55.0000 55.0000 55.0000
2 80.0000 79.9998 80.0000 2 80.0000 80.0000 80.0000
3 106.8984 80.9279 85.3015 3 106.9383 81.1338 85.2966
4 100.6541 82.8710 83.8637 4 100.5761 81.3645 83.8629
5 81.7543 159.9998 139.5767 5 81.5041 160.0000 139.5769
6 82.7328 239.9998 159.2863 6 83.0204 240.0000 159.2850
7 300.0000 295.6908 300.0000 7 300.0000 294.4847 300.0000
8 340.0000 299.2810 315.4372 8 340.0000 297.2678 315.4370
9 470.0000 395.1220 431.6587 9 470.0000 396.7602 431.6844
10 470.0000 392.6056 433.9235 10 470.0000 395.5843 433.9048
Min. cost Min. cost
111 497.6500 116 430.5135 113 282.4474 111 497.6301 116 412.4344 113 282.4213
($/h) ($/h)
Emission Emission
4572.3307 3932.4779 4129.0021 4572.1866 3932.2432 4129.0020
(t/h) (t/h)
Power loss Power loss
87.0897 81.4975 84.0476 87.0388 81.5952 84.0477
(MW) (MW)

4. SIMULATION RESULTS AND DISCUS-

  SION
πλ
λΓ (λ) sin In this paper, the proposed algorithms, FPA,
2 1
L∼ · , (s  s0 > 0) (8) and TLBO are tested on the EED problem with
λ s 1+λ
the valve point loading effect and loss. The
efficacy and practicality of the proposed methods
The newly generated population is given by are demonstrated on 10 generating units and 14-unit
systems. The proposed algorithms are implemented
xt+1 = xti + γL (λ) g∗ − xti using MATLAB 9.5 R2018b running on an Intel®


i (9)
Core™ i5 processor, 2.26 GHz, 8 GB RAM PC.
where xti is the ith generating solution vector and g∗ Initially, the fuel cost and emission objectives are
is the local best solution found in the ith generation advanced individually by taking the weighting factor
of the solution, and the parameter function γ is a W as 1 and 0 in the fitness function. The algorithm
factor of scaling to decide the step size taken by proposed to address the problem and bi-objectives are
the agents. L(λ) represents the entire levy flight treated at the same time as the competing objectives.
distribution function. Self-pollination occurs among
the neighbors of the current population without 4. 1 Case study 1: 10-unit test system
involving the current minima, so parameter  is
chosen to represent the step size from a unique In this test system, 10 thermal units with the
distribution. The expression for self-pollination is valve point loading effect are studied. The total load
mathematically represented in Eq. (10) [20]. demand is taken as 2000 MW [14]. The simulation
results using the proposed methods, TLBO, and
FPA with dissimilar objective functions are given in
xt+1 = xti +  xtj − xtk


i (10) Tables 1 and 2, respectively. For this test system,
20 independent trials are conducted, each containing
where xtj and xtk are the flowering characteristics of 300 iterations. Based on the data obtained, the
different flowers in the same population. comparisons of the ten thermal units tested by
The new population has the same flower constancy dissimilar methods are presented in Table 3. The
as previously expressed. If the new population is convergence characteristics of the TLBO algorithm
better than the previous, it is modernized in the and FPA for the cost objective, emission objective,
previous place or rejected. This generation and and fitness of the ten-unit system are compared in
comparison process continues up to N counts, and Figs. 2, 3, and 4.
the best among them is declared to be the current The proposed TLBO algorithm and FPA are initi-
global best and the optimum solution. ated using different objective functions and tabulated
204 ECTI TRANSACTIONS ON ELECTRICAL ENG., ELECTRONICS, AND COMMUNICATIONS VOL.19, NO.2 JUNE 2021

Table 3: Comparison of simulation results for the ten-unit system with the valve point loading effect for FPA
and TLBO.
Cost objective Emission objective CEED
TLBO FPA TLBO FPA TLBO FPA
Cost ($/h) 111 497.6500 111 497.6301 116 430.5135 116 412.4344 113 282.4474 113 282.4213
Emission (t/h) 4572.3307 4572.1866 3932.4779 3932.2432 4129.0021 4129.0020
Power loss (MW) 87.0397 87.0388 81.4975 81.5952 84.0476 84.0477

Fig. 2: Comparison of convergence characteristics

for the EED problem in the ten-unit system with the
valve point loading effect for the cost objective.

Fig. 3: Comparison of the convergence characteris-

in Tables 1 and 2, respectively. The comparative tics for the EED problem in the ten-unit system with
simulation results for the suggested methods are the valve point loading effect for fitness.
shown in Table 3. As can be observed from Table 3,
in the case of the cost objective, the cost obtained
by TLBO is 111 497.6500 $/h with emissions and
power loss of 4572.3307 t/h (tons per hour) and
87.0897 MW, respectively. The cost, emissions, and
power loss obtained by the FPA are 111 497.6301 $/h,
4572.1866 t/h, and 87.0388 MW, respectively.
For emission dispatch, the cost procured by TLBO
is 116 430.5135 $/h with emissions of 3932.4779 t/h
and power loss of 81.4975 MW. The cost, emissions,
and power loss obtained by FPA are 116 412.4344 $/h,
3932.2432 t/h, and 81.5952 MW, respectively. In
the case of combined economic emission dis-
patch (CEED), the cost obtained by TLBO is
113 282.4474 $/h with emissions and power loss of
4129.0021 t/h and 84.0476 MW, respectively. The
cost, emissions, and power loss obtained by FPA are
113 282.4213 $/h, 4129.0020 t/h, and 84.0477 MW,
respectively. The results indicate that the FPA offers
a better performance in comparison to TLBO. Fig. 4: Comparison of the convergence character-
Fig. 2 presents a comparison of the convergence istics for the EED problem in the ten-unit system
characteristics of the EED problem of the ten-unit with the valve point loading effect for the emissions
system, including the valve point loading effect in objective.
relation to the cost objective at a power demand of
2000 MW. As can be observed in Fig. 3, there are
two plots: one for TLBO and the other for the FPA,
COMPARISON INVESTIGATION INTO POWER SYSTEM OPTIMIZATION USING METAHEURISTIC ALGORITHMS 205

Table 4: Simulation results for TLBO in the 14-unit Table 5: Simulation results for FPA in the 14-unit
system with the valve point loading effect for different system with the valve point loading effect for different
objectives where PD = 2500 MW. objectives where PD = 2500 MW.
Number TLBO algorithm Number FPA
of units Economic Emission of units Economic Emission
CEED CEED
Dispatch Dispatch Dispatch Dispatch
1 419.2794 286.6902 329.5196 1 419.2794 286.6830 329.5196
2 374.3995 153.5419 224.7998 2 374.5743 153.4521 224.7998
3 130.0000 130.0000 130.0000 3 130.0000 130.0000 126.2424
4 130.0000 130.0000 119.9748 4 120.1869 130.0000 119.7331
5 299.5997 235.1755 249.7331 5 299.5997 235.4270 249.7331
6 184.8666 460.0000 384.3328 6 234.7337 460.0000 384.3327
7 234.7331 298.0071 234.7331 7 184.8667 297.6711 284.5997
8 159.7331 235.8714 259.4662 8 159.7332 236.0786 209.5996
9 162.0000 162.0000 162.0000 9 162.0000 162.0000 162.0000
10 153.2968 160.0000 160.0000 10 160.0000 160.0000 160.0000
11 80.0000 80.0000 80.0000 11 80.0000 80.0000 80.0000
12 80.0000 80.0000 80.0000 12 80.0000 80.0000 80.0000
13 85.0000 85.0000 85.0000 13 85.0000 85.0000 85.0000
14 52.3999 55.0000 53.9391 14 55.0000 55.0000 52.3999
Min. cost Min. cost
12 542.4661 14 214.6900 12 846.5707 12 437.7031 14 210.0114 12 816.6164
($/h) ($/h)
Emission Emission
8767.1429 5994.1748 6396.5984 8716.7541 5994.1740 6373.8274
(t/h) (t/h)
Power loss Power loss
45.3080 51.2861 53.4984 44.9739 51.3119 47.9600
(MW) (MW)

shown colored red and blue, respectively. The total

cost obtained by TLBO is around 112 600 $/h, around
112,500 $/h for the FPA at the first iteration. Finally,
the global minimum calculated by TLBO is at the
15th iteration and at the 20th iteration by the FPA.
Fig. 4 represents the comparative convergence
characteristics for the EED problem in the ten-unit
system, including the valve point loading effect for
the emissions objective at 2000 MW demand. As can
be observed in Fig. 5, there are two plots: one for
the FPA and the other for TLBO, shown colored
blue and red, respectively. The emissions obtained by
TLBO are above 3990 t/h, while for the FPA they are
also above 3990 t/h at the first iteration. Finally, the
global minimum is calculated by TLBO at the 270th
iteration and at the 140th iteration by the FPA.

4. 2 Case study 2: 14-unit test system

This case study involves a system with 14 thermal
units, including the valve point loading effect. The
total load demand is taken as 2500 MW [22, 23]. Fig. 5: Comparison of the convergence characteris-
The dispatch results using the proposed TLBO and tics for the EED problem in the 14-unit system with
FPA are presented in Tables 4 and 5, respectively. the valve point loading effect for the cost objective.
This test system consists of 20 separate trials, each
containing 500 iterations. Comparisons of the 14
thermal unit tests with various objects using different As can be observed from Table 6, in the cost
methods are shown in Table 6. The comparative objective case, the total cost obtained by TLBO is
TLBO and FPA convergence characteristics for the 12 542.4661 $/h with emissions and power loss being
cost objective, emissions objective, and fitness of the 8767.1429 t/h and 45.308 MW, respectively. The
14-unit system are presented in Figs. 5, 6, and 7. cost, emissions, and power loss obtained by the FPA
206 ECTI TRANSACTIONS ON ELECTRICAL ENG., ELECTRONICS, AND COMMUNICATIONS VOL.19, NO.2 JUNE 2021

Table 6: Comparison of simulation results for the 14-unit system with the valve point loading effect for FPA
and TLBO.
Cost objective Emission objective CEED
TLBO FPA TLBO FPA TLBO FPA
Cost ($/h) 12 542.4661 12 542.4661 12 542.4661 12 542.4661 12 542.4661 12 542.4661
Emission (t/h) 8767.1429 8716.7541 5994.1748 5994.1740 6396.5984 6373.8274
Power loss (MW) 45.3080 44.9739 51.2861 51.3119 53.4984 47.9600

Fig. 7: Comparison of the convergence character-

istics for the EED problem in the 14-unit system
Fig. 6: Comparison of the convergence characteris-
with the valve point loading effect for the emissions
tics for the EED problem in the 14-unit system with
objective.
the valve point loading effect for fitness.

are 12 437.7031 $/h, 8716.7541 t/h, and 44.9739 MW, the 245th iteration by the FPA.
respectively. The cost obtained for emission dis- Fig. 7 presents the comparative convergence
patch by TLBO is 14 214.69 $/h with emissions of characteristics of the EED problem in the 14-unit
5994.1748 t/h with a power loss of 51.2861 MW. The system, including the valve point loading effect
cost, emissions, and power loss obtained by the FPA for the emissions objective at a load demand of
are 14 210.0114 $/h, 5994.1740 t/h, and 51.3119 MW, 2500 MW. The emissions calculated by TLBO are
respectively. In the case of combined economic above 6350 t/h, and by the FPA above 6400 t/h at
emission dispatch (CEED), the cost obtained by the first iteration. Finally, the global minimum
TLBO is 12 846.5707 $/h with emissions and power calculated by TLBO is at the 25th iteration and at
loss of 6396.5984 t/h and 53.4984 MW, respectively. 50th iteration by the FPA. The results reveal that
The cost, emissions, and power loss obtained by the FPA offers a better performance in comparison
the FPA are 12 816.6164 $/h, 6373.8274 t/h, and to TLBO.
47.96 MW, respectively. The results reveal that the
FPA offers a better performance in comparison to 5. CONCLUSION
TLBO.
This paper applies the FPA TLBO optimization
Fig. 6 presents the comparative convergence char- methods to the valve point loading effect and
acteristics of the EED problem in the 14-unit system, combined economic emissions dispatch problem, re-
including the valve point loading effect for the cost spectively. This suggested FPA provides the best
objective at a load demand of 2500 MW. As can be global solution in systems of 10 and 14 units. The
observed in Fig. 5, there are two plots: one for TLBO algorithm gives a better optimal solution than that
and is the other for the FPA, colored red and blue, obtained from other different complex algorithms.
respectively. The total cost calculated by TLBO is The results demonstrate that the proposed approach
around 12 910 $/h and around 12 890 $/h by the FPA is more efficient for multi-objective optimization.
at the first iteration. Finally, the global minimum Therefore, it can be concluded that the solution for
calculated by TLBO is at the 395th iteration and at economic emissions dispatch provided by the FPA
COMPARISON INVESTIGATION INTO POWER SYSTEM OPTIMIZATION USING METAHEURISTIC ALGORITHMS 207

is the best compromise among the optimal solutions evolutionary programming based efficient parti-
of the two optimizing techniques. It can help the cle swarm optimization for economic dispatch
power system to adjust the level of generation to problem with valve-point loading,” European
improve the economy of the utility while causing less Journal of Scientific Research, vol. 52, no. 3,
damage to the environment. Therefore, the FPA pp. 385–397, 2011.
based solution is proposed as being more robust and [11] M. R. Lohokare, B. K. Panigrahi, S. S. Pattnaik,
reliable compared to other techniques and can be S. Devi, and A. Mohapatra, “Neighborhood
recommended for use in resolving other optimization Search-Driven Accelerated Biogeography-Based
problems encountered in power system networks. Optimization for Optimal Load Dispatch,” IEEE
Transactions on Systems, Man, and Cybernetics,
REFERENCES Part C (Applications and Reviews), vol. 42,
no. 5, pp. 641–652, 2012.
[1] M.-T. Tsai, H.-J. Gow, and W.-M. Lin, “A
[12] E. A. Jasmin, T. P. Imthias Ahamed, and
novel stochastic search method for the solution
V. P. Jagathy Raj, “Reinforcement Learning
of economic dispatch problems with non-convex
approaches to Economic Dispatch problem,”
fuel cost functions,” International Journal of
International Journal of Electrical Power &
Electrical Power & Energy Systems, vol. 33,
Energy Systems, vol. 33, no. 4, pp. 836–845,
no. 4, pp. 1070–1076, 2011.
2011.
[2] K. T. Chaturvedi, M. Pandit, and L. Srivastava,
“Self-organizing hierarchical particle swarm op- [13] X.-S. Yang, M. Karamanoglu, and X. He,
timization for nonconvex economic dispatch,” “Flower pollination algorithm: A novel approach
IEEE Transactions on Power Systems, vol. 23, for multiobjective optimization,” Engineering
no. 3, pp. 1079–1087, 2008. Optimization, vol. 46, no. 9, pp. 1222–1237,
[3] R. V. Rao, V. J. Savsani, and D. P. Vakharia, 2013.
“Teaching–learning-based optimization: A novel [14] A. Chatterjee, S. P. Ghoshal, and V. Mukherjee,
method for constrained mechanical design op- “Solution of combined economic and emission
timization problems,” Computer-Aided Design, dispatch problems of power systems by an
vol. 43, no. 3, pp. 303–315, 2011. opposition-based harmony search algorithm,”
[4] M. Basu, “Economic environmental dispatch International Journal of Electrical Power &
using multi-objective differential evolution,” Ap- Energy Systems, vol. 39, no. 1, pp. 9–20, 2012.
plied Soft Computing, vol. 11, no. 2, pp. 2845– [15] I. Ciornei and E. Kyriakides, “A GA-API So-
2853, 2011. lution for the Economic Dispatch of Generation
[5] J. Cai, Q. Li, L. Li, H. Peng, and Y. Yang, in Power System Operation” IEEE Transactions
“A hybrid CPSO–SQP method for economic on Power Systems, vol. 27, no. 1, pp. 233–242,
dispatch considering the valve-point effects,” 2012.
Energy Conversion and Management, vol. 53, [16] M. S. Javadi, A. Esmaeel Nezhad, and S.
no. 1, pp. 175–181, 2012. Sabramooz, “Economic heat and power dis-
[6] V. N. Dieu, W. Ongsakul, and J. Polprasert, patch in modern power system harmony search
“The augmented Lagrange Hopfield network for algorithm versus analytical solution,” Scientia
economic dispatch with multiple fuel options,” Iranica, vol. 19, no. 6, pp. 1820–1828, 2012.
Mathematical and Computer Modelling, vol. 57, [17] S. M. Sajjadi, A. Sadeghi Yazdankhah, and
no. 1–2, pp. 30–39, 2013. F. Ferdowsi, “A new gumption approach for
[7] V. N. Dieu and P. Schegner, “Augmented La- economic dispatch problem with losses effect
grange Hopfield network initialized by quadratic based on valve-point active power,” Electric
programming for economic dispatch with piece- Power Systems Research, vol. 92, pp. 81–86,
wise quadratic cost functions and prohibited 2012.
zones,” Applied Soft Computing, vol. 13, no. 1, [18] D. Rao and N. Kumar, “Comparisional Investi-
pp. 292–301, 2013. gation of Load Dispatch Solutions with TLBO,”
[8] L. Wang and C. Singh, “Stochastic economic International Journal of Electrical and Com-
emission load dispatch through a modified puter Engineering, vol. 7, no. 6, pp. 3246–3253,
particle swarm optimization algorithm,” Elec- 2017.
tric Power Systems Research, vol. 78, no. 8, [19] M. Pradhan, P. K. Roy, and T. Pal, “Grey wolf
pp. 1466–1476, 2008. optimization applied to economic load dispatch
[9] M. Pourakbari-Kasmaei and M. Rashidi-Nejad, problems,” International Journal of Electrical
“An effortless hybrid method to solve economic Power & Energy Systems, vol. 83, pp. 325–334,
load dispatch problem in power systems,” Energy 2016.
Conversion and Management, vol. 52, no. 8–9, [20] H. M. Dubey, M. Pandit, and B. K. Pan-
pp. 2854–2860, 2011. igrahi, “Hybrid flower pollination algorithm
[10] S. M. V. Pandian and K. Thanushkodi, “An with time-varying fuzzy selection mechanism
 208 ECTI TRANSACTIONS ON ELECTRICAL ENG., ELECTRONICS, AND COMMUNICATIONS VOL.19, NO.2 JUNE 2021

for wind integrated multi-objective dynamic Devineni Gireesh Kumar received

his B.Tech degree in Electrical & Elec-
economic dispatch,” Renewable Energy, vol. 83, tronics Engineering from Jawaharlal
pp. 188–202, 2015. Nehru Technological University, Anan-
[21] H. M. Dubey, M. Pandit, and B. K. Panigrahi, tapur, India in 2009, M.Tech degree in
Power Electronics & Drives from SRM
“A Biologically Inspired Modified Flower Polli- University, India in 2011. He is pursu-
nation Algorithm for Solving Economic Dispatch ing his Ph.D. in Electrical Engineering
Problems in Modern Power Systems,” Cognitive at Lovely Professional University since
2018. He is currently working as
Computation, vol. 7, no. 5, pp. 594–608, 2015. Assistant Professor in Department of
[22] N. Sinha, R. Chakrabarti, and P. K. Chattopad- Electrical & Electronics Engineering, B. V. Raju Institute of
hyay, “Evolutionary programming techniques for Technology, Narsapur, Medak, Telangana, India. His research
interests include solar PV inverters, multilevel converters,
economic load dispatch,” IEEE Transactions hybrid electric vehicles, and motor drives.
on Evolutionary Computation, vol. 7, no. 1,
pp. 83–94, 2003.
[23] A. Mahor, V. Prasad, and S. Rangnekar,
“Economic dispatch using particle swarm opti- Bharat Kumar Narukullapati re-
mization: A review,” Renewable and Sustainable ceived his B.Tech degree from Jawa-
Energy Reviews, vol. 13, no. 8, pp. 2134–2141, harlal Nehru Technological University,
Hyderabad, India and M.Tech degree
2009. from Jawaharlal Nehru Technological
University, Kakinada, India in Electrical
and Electronics Engineering. He is cur-
rently with the Department of Electrical
and Electronics Engineering, Vignan’s
D. S. Naga Malleswara Rao received Foundation for Science, Technology and
his B.Tech from Jawaharlal Nehru Tech- Research (Deemed to be University),
nological University, Hyderabad, India, Vadlamudi, Guntur, India. His research interests include power
M.Tech from Jawaharlal Nehru Tech- and industrial drives, multilevel inverters, distributed energy
nological University, Kakinada, India systems, and smart nano grids.
in Electrical & Electronics Engineering,
and Ph.D. from National Institute of
Technology, Jamshedpur, India. Cur-
rently, he is working as Associate Profes-
sor in Electrical & Electronics Engineer- Davu Srinivasa Rao received his
ing Department, Gokaraju Rangaraju B.Tech degree in Electrical & Elec-
Institute of Engineering & Technology, Hyderabad, India. His tronics Engineering from Jawaharlal
research interest includes power system economic operation Nehru Technological University, Hyder-
and control and soft computing techniques in electrical abad, India in 2007, M.Tech degree
engineering field. in Power Electronics from Jawaharlal
Nehru Technological University, Kak-
inada, India in 2012. He is pursu-
ing his Ph.D. in Electrical Engineering
at National Institute of Technology,
Andhra Pradesh, India. He is currently
Dogga Raveendhra received his B.Tech
working as Assistant Professor in Department of Electrical
degree in Electrical and Electronics En-
& Electronics Engineering, Gokaraju Rangaraju Institute of
gineering from Jawaharlal Nehru Tech-
Technology, Hyderabad, Telangana, India. His research
nological University, Kakinada, India,
interests include solar PV inverters, multilevel converters, and
in 2009 and the M.Tech and Ph.D.
motor drives.
degrees in Electric Drives and Power
Electronics, Electrical Engineering De-
partment from Indian Institute of Tech-
nology (IIT), Roorkee, India, in 2012
and 2019, respectively. He is currently
working as an Associate Professor in the Srividya Devi Palakaluri received
Department of Electrical Engineering, Gokaraju Rangaraju her Ph.D. in Power Systems from
Institute of Engineering and Technology, Hyderabad, India. Andhra University, India. She had
He was associated with renewable energy research center, overall experience of 14 years in teach-
American University of Sharjah, United Arab Emirates during ing. She had published more than
2019–20. During his Ph.D. program, he was also associated 17 research papers in various national
with FREEDM System Center, NC State University, USA journals and conferences and attended
as a visiting researcher. He was recipient of prestigious more the 35 workshops conducted by
of Most Promising Innovation Award from IKMC in 2019, National Mission on Education through
Best Innovation award from Zunik Energies in 2019 and Information and Communication Tech-
Excellence in Doctoral Research Awards from IIT Roorkee in nology (NMEICT), Indian Institute of
2019. From 2012 to 2014, he was with Electrical Engineering Technology Bombay (IITB), Indian Institute of Technology
Department, Graphic Era University, Dehradun, India as an Kanpur (IITK), etc. She is a Certified LabVIEW Associate
Assistant Professor. He is the author of 3 book chapters, more Developer (CLAD) certified by National Instruments. Her
than 23 research papers, and three published inventions. He areas of interests are power systems, network analysis, and
is recipient of prestigious Bhaskara Advanced Solar Energy fuzzy systems, etc.
(BASE) Fellowship/Internship in 2016. He has guided 7
M.Tech Thesis and 12 B.Tech Projects. His research interests
include power converters, renewable energy systems, control
techniques, FPGA controlled power converters, and electric
drives.

View publication stats