International Journal of Optimum design of linear and circular

Microwave and Wireless

Technologies antenna arrays using equilibrium
optimization algorithm
cambridge.org/mrf
Ali Durmus1 and Rifat Kurban2
1
Department of Electricity and Energy, Vocational College, Kayseri University, Kayseri, Turkey and 2Department of
Research Paper Computer Technologies, Vocational College, Kayseri University, Kayseri, Turkey

Cite this article: Durmus A, Kurban R (2021). Abstract

Optimum design of linear and circular antenna
arrays using equilibrium optimization In this paper, equilibrium optimization algorithm (EOA), which is a novel optimization algo-
algorithm. International Journal of Microwave rithm, is applied to synthesize symmetrical linear antenna array and non-uniform circular
and Wireless Technologies 13, 986–997. https:// antenna array (CAA). The main purpose of antenna array synthesis is to achieve a radiation
doi.org/10.1017/S1759078720001774
pattern with low maximum side lobe level (MSL) and narrow half-power beam width
Received: 16 July 2020 (HPBW) in far-field. The low MSL here is an important parameter to reduce interference
Revised: 18 December 2020 from other communication systems operating in the same frequency band. A narrow
Accepted: 18 December 2020 HPBW is needed to achieve high directionality in antenna radiation patterns. Entering the lit-
First published online: 26 January 2021
erature as a novel optimization technique, EOA optimally determined the amplitude and pos-
Key words: ition values of the array elements to obtain a radiation pattern with a low MSL and narrow
Antenna array synthesis; circular antenna HPBW. The EOA is inspired by models of the control volume mass balance used to predict
array; equilibrium optimization algorithm; equilibrium as well as dynamic states. To demonstrate the flexibility and performance of the
linear antenna array; maximum side lobe level proposed algorithm, 10-element, 16-element and 24-element linear arrays and eight-element,
Author for correspondence: 10-element and 12-element CAAs are synthesized. The MSL and HPBW values of radiation
Ali Durmus, pattern obtained with the EOA are very successful compared to the results of other optimiza-
E-mail: [email protected] tion methods in the literature.

Introduction
The great increase in wireless services such as cellular mobile communication, satellite, sonar,
and radar in recent years has made it necessary to use the existing frequency spectrum effi-
ciently. Since the desired radiation patterns cannot be obtained with a single antenna, antenna
arrays capable of effective and directed radiation are formed by combining multiple antennas.
Antenna arrays must be designed optimally and accurately in order to have features such as
high directivity and signal gain with an extended coverage. The positions, amplitudes, and
phase angles of the elements should be determined in order to increase the properties of
the radiation patterns of antenna arrays in the far-field. The region in the direction of max-
imum radiation is named as the main beam lobe. Side lobes other than the main lobe need
to be eliminated. However, in practical applications, they can only be suppressed as much
as possible [1]. The main purpose of creating a radiation model with a lower maximum
side lobe level (MSL) is to prevent interference from other communication systems operating
in the same frequency band. Half-power beam width (HPBW) is the angle at which the main
beam decreases 50% or 3 dB from its peak value. In addition, a radiation pattern with a narrow
HPBW shall be needed to provide the high directivity required to communicate with long dis-
tances. However, it is very difficult to design an antenna array with both narrower HPBW and
lower MSL values. Because while improving HPBW value, MSL value gets worse. Of course,
the opposite is also true, which means that it is very difficult to obtain narrower HPBW
while lowering the MSL value. Therefore, despite the increase in electromagnetic pollution,
it has become possible to design an antenna array with low MSL and narrow HPBW with
the help of metaheuristic algorithms [2].
Antenna arrays can take different names according to their geometric features such as
linear, circular, and elliptical. Linear antenna arrays (LAAs) have a structure where elements
are positioned along a linear line and are one of the most preferred geometries in the literature
due to their simplicity. In the last decade, several studies have been done on LAA synthesis, an
© The Author(s), 2021. Published by important electromagnetic problem. Plant growth simulation algorithm was used for the pat-
Cambridge University Press in association with tern nulling of LAA by amplitude control [3]. Dib proposed a symbiotic organisms search
the European Microwave Association (SOS) method for the design of linear array patterns with low MSL [4]. The biogeography-
based optimization (BBO) method is utilized to synthesize linear and elliptical antenna arrays
[5]. Khodier and Al-Aqeel proposed a particle swarm optimization (PSO) algorithm to adjust
the amplitude, phase and location of linear and circular array elements for optimum radiation
patterns [6]. Taguchi’s optimization method and self-adaptive differential evolution

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 International Journal of Microwave and Wireless Technologies 987

algorithms are implemented to suppress the MSL and perform equilibrium as well as dynamic states. EOA is an algorithm that
null steering for LAA by controlling different parameters of the functions on the volume of control according to the effect of
array elements [7]. Das et al. adopted moth flame optimization dynamic mass balance. In this algorithm, a mass balance equation
(MFO) to design linear and circular antenna array (CAA) for is used to measure the number of mass inputs. In order to find the
side lobe reduction [8]. Runner-root algorithm (RRA) is used to equilibrium of the system, the most suitable one among the
reduce MSL and null depths in LAAs [9]. Prabhakar and volumes generated is determined. The EOA algorithm has
Satyanarayana proposed a hybrid salp swarm whale optimization many advantages such as having an easy and flexible application
algorithm and experimented on typical benchmark functions and structure, differentiation of possible solutions (individuals) in the
cylindrical conformal antenna array synthesis [10]. Adaptive search space, creating a good balance between exploration and
flower pollination algorithm (AFPA) is used to optimize linear exploitation operators [24]. Extensive experiments show that
array elements in order to obtain minimum MSL [11]. EOA has superior performance in designing optimal LAA than
Almagboul et al. implemented atom search optimization algo- SOS, BBO, PSO, Taguchi, and MFO methods. Moreover, the
rithm to suppress MSL and steering the null in the radiation pat- results achieved by EOA for designing optimal CAA are signifi-
tern of LAAs [12]. cantly better than GA, PSO, SA, BBO, FA, SOA, CSO, Taguchi,
In addition, with the rapid developments in communication and MFO methods. The numerical results show that EOA has
systems, the optimum design of antenna arrays with complex better results than existing methods in terms of obtaining lower
geometries such as CAAs has become an important requirement MSL and narrower HPBW.
in the literature. Panduro et al. proposed a genetic algorithm (GA) The rest of this paper is organized as follows: EOA is briefly
to design non-uniform CAAs for side lobe reduction [13]. PSO is described in section “Equilibrium optimization algorithm”. In
used to synthesize non-uniform CAAs to obtain minimum MSL section “Design of antenna optimization problems”, the design
[14]. Rattan et al. proposed a simulated annealing (SA) method of LAA and CAA is examined. Numerical results obtained with
for the optimization of CAAs of isotropic radiators [15]. EOA are given in comparison with the results of other optimiza-
Khodier and Al-Aqeel implemented cuckoo search (CS) algo- tion algorithms in section “Numerical results”. Finally, section
rithm to minimize MSL with and without null steering in the “Conclusions” concludes the paper.
design of CAAs [16]. BBO is employed to determine the ampli-
tude and position of the CAA elements to provide a radiation pat-
tern with a reduced MSL [17]. Sharaqa and Dib applied firefly Equilibrium optimization algorithm
algorithm (FA) to design CCA and concentric CAA (CCAA) by Meta-heuristic algorithms are used to search the domain globally
adjusting the amplitude and positions of the elements to reduce to find optimal solutions and have many applications in various
MSL [18]. Parallel implementation of seeker optimization algo- fields [25]. EOA, proposed by Faramarzi et al., is inspired by the
rithm (SOA) is proposed to design CAA and CCAA with the models of volume regulation mass balance which are used to
low MSL at a fixed beam width [19]. Ram et al. proposed a cat approximate dynamic as well as equilibrium states [24]. Each par-
swarm optimization (CSO) for CAA and CCAA synthesis by ticle in EOA, referred as a possible solution, works as a search
adjusting current excitation weights and inter-element separations agent with its concentration or position. Such agents change
[20]. Taguchi method is applied to design CAA for reduced MSL their concentration correlated with best-so-far solutions, referred
[21]. Improved chicken swarm optimization is proposed to sup- as equilibrium candidates, arbitrarily to achieve the state of equilib-
press the MSL of LAA, CAA, and random antenna arrays [22]. rium or the optimal result. The phrase “generation rate” changes
Zheng et al. adopted improved invasive weed optimization with the behavior of exploration and exploitation. EOA differs from
random mutation and Lévy flight algorithm to design LAAs the comparison algorithms used in this paper with its unique elit-
and CAAs with lower MSL and narrower main lobe beam [23]. ism operator. In [24], it is statistically proved that EOA is better
Different number of array elements are used to synthesize LAA than GA, PSO; grey wolf optimizer, gravitational search algorithm,
and CAA in various studies. In order to compare the proposed and salp swarm algorithm; evolution strategy with covariance
method consistently with most of the existing studies in the litera- matrix adaptation (CMA-ES), success history-based parameter
ture, the number of elements of the antenna arrays is specifically adaptation differential evolution (SHADE), SHADE with linear
determined in our article. In this study, first, 10-element, population size reduction hybridized with semi-parameter adapta-
16-element, and 24-element LAA synthesis is performed. The tion of CMA-ES on numerous unimodal, multimodal, and com-
amplitudes of the antenna array elements are determined by position functions and engineering application problems.
using equilibrium optimization algorithm (EOA) [24]. EOA, a
very recent metaheuristic algorithm, is used for the first time in
LAA and CAA synthesis in this study. The results achieved Initialization phase
with EOA are compared with the outcomes of other algorithms
in the literature for LAA design. In addition, the position (inter- In EOA, an initial concentration population is created with uni-
element spacing) and the phase values of the array elements are form randomization, based on the number of particles and
λ/2 and zero, respectively. Second, the position and amplitude dimensions in equation 1. Generation of the initial population
values of the eight-element, 10-element and 12-element non- in EOA method is important because it affects the search space
uniform CAAs are determined using EOA. The MSL and of iterations.
HPBW values achieved with the proposed method EOA are com-
pared with the results of the well-known algorithm in the litera- Ciinitial = Cmin + randi (Cmax − Cmin ) i = 1, 2, . . . , n, (1)
ture for non-uniform CAA. The aim of this study is to obtain
radiation patterns with lower MSL and narrower HPBW for linear Ciinitial is the initial position vector of ith particle, Cmin and
and CAA design by using EOA. The EOA is inspired by the mod- Cmax are the dimension boundary limits, randi is a randomized
els of the control volume mass balance used to predict variable within the range of [0,1], and n is the population size

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 988 Ali Durmus and Rifat Kurban

of particles. Every population particle is evaluated sequentially,

and the fitness values are determined.

Equilibrium candidates
Equilibrium state is the global optimum of the problem and Ceq is
the final state of the convergence. In EOA, as given in equation 2,
there are four best-so-far particles and plus another calculated
particle, that is, the average of four particles is identified as equi-
librium candidates. Equation 2 is significant because the first four
candidates help EOA for better exploration ability while the aver-
aged one helps in exploitation:

eq,pool = {C
C
eq(1) , C

eq(2) , C

eq(3) , C

eq(4) , C

eq(ave) }. (2)

All the particles update their concentration by selecting ran-

domly one among the candidates.

Exponential term
Concentration updating is realized by the exponential term (F ) to
ensure the EOA has a reasonable balance between exploration and
exploitation by using equations 3 and 4.

F
= a1 sign(
r − 0.5)[e−lt − 1], (3)

(a2 (Iter/Max iter))

Iter
t = 1− , (4)
Max iter

where t is a function of iteration, a1 and a1 are constant values

used to manage the ability of exploration and exploitation, sign
is the direction, r and λ are random vectors, respectively.
Recommended values for a1 and a1 are equal to 2 and 1,
respectively.

Generation rate
Equations 5–7 calculate the generation rate term which provides Fig. 1. One-dimensional visualization of population modifying their positions.
the exact solution by improving the exploitation phase:

=G
G
0 F,

(5) where V is the unit. The significance of equation 8 can be
summed up as follows:




0 = GCP
C),

• The first term is a global optimum candidate, the second and
G (C eq − l (6)
the third terms are control combinations.
• Second term contributes more to the exploration by taking
 advantage of the direct difference between a sample particle
 0.5r1 r2 ≥ GP
GCP = , (7) and an equilibrium candidate.
0 r2 , GP
• Third term contributes more to exploitation from small varia-
tions in concentration.
where G
0 is the initial value, r2 and r1 are random numbers
between 0 and 1, GCP is the generation rate control parameter,
Figure 1 shows a 1D visualization of working mechanism of
GP is the generation probability. A good equilibrium is reached
the main equation terms. C1 − Ceq represents the second term
by using GP = 0.5, between exploration and exploitation.
of the final equation and Ceq − λC1 represents the third term of
At final, the updating rule of EOA is given in equation 8.
the updating rule of EOA.
EOA has a memory-saving technique to help every particle



=C
C
eq + (C
eq )F
+ G (1 − F),

−C
(8) keep track of its location in space with its fitness value. This tech-

V
l nique adds to the potential of exploitation; however, it can raise

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 International Journal of Microwave and Wireless Technologies 989

Fig. 2. Pseudo code of EOA.

the risk of being trapped into a local minimum. Detailed pseudo Circular antenna array
code of EOA is given in Fig. 2.
Figure 4 illustrates the structure of CAA with M isotropic
elements placed non-uniformly on a circle. The radius of the
Design of antenna optimization problems circle placed in the x-y plane is a, and all elements in this structure
exhibit similar radiation properties (isotropic source).
Linear antenna array The array pattern of CAA can be defined by its array factor.
The LAA with the 2M isotropic array element is positioned sym- The CAA’s array factor (AF) can be described as in equations 10–
metrically in two sections along the x-axis with M elements as 13 [1].
illustrated in Fig. 3. The distance between the elements in this

M
array is determined as λ/2. The all-phase excitation weights are AF(u) = In · e( j.ka. cos (u−wn )+an ) , (10)
set to zero in the antenna array. n=1
The antenna array factor, given in equation 9, was obtained by
assuming that the amplitude coefficients of the array elements are 
M
ka = di , (11)
symmetrical around the origin [1]. i=1


M
AF(u) = 2 In cos (kdn sin u + yn ), (9) 
n

n=1
wn = (2p · di )/ka, (12)
i=1
where In and υn are the amplitude and phase excitation weight of
nth element in the array, respectively. The wave number is k and an = −ka · cos(u0 − wn ), (13)
spacing between the array elements is d. θ is the angle of scanning. where αn and In are the phase and excitation amplitude of the nth
The M value here is half the total number of elements in the LAA. array element, respectively. di is the distance between the two

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 990 Ali Durmus and Rifat Kurban

Fig. 3. M-element linear antenna array.

antenna array elements. The number of waves is denoted by k.

The nth array element angular position in the x-y plane is denoted
by wn. The M value indicates the total number of elements of the Fig. 4. M-element circular antenna array.
CAA on the ring.

in dB. The function MHPBW can be formulated as in equation 17.


Cost function formulation Mo − MHPBWmax , Mo . MHPBWmax
MHPBW = , (17)
0, elsewhere
The most important element in antenna array design is the cre-
ation of antenna structures that communicate with long dis-
tances and are not affected by interference from other where MHPBWmax and Mo are the desired maximum HPBW and
communication systems operating in the same frequency band. the value of HPBW obtained by EOA, respectively.
In order to achieve this, it is necessary to design antenna arrays
to obtain radiation patterns with lower MSL and narrower
HPBW values. In addition, another important parameter in Numerical results
the application of the antenna design is the physical size of
In this study, the analysis of antenna arrays with two different
the antenna and this value is desired to be low in designs. The
geometries, linear and circular, is performed. In the first group
purpose of this article is to determine the position and ampli-
of examples, the amplitude values of the LAA elements with
tude values of the array elements by the EOA to obtain the radi-
10, 16, and 24 elements are optimally determined by EOA. In
ation diagram with the lowest MSL and the narrowest HPBW
the second group of examples, EOA is used to optimize the amp-
values in order to make optimum antenna design. In order to
litude and position values of the CAA having 8, 10, and 12 ele-
achieve the optimum antenna synthesis, it is important to design
ments. The aim of all simulation examples is to achieve the
an appropriate cost function to reduce MSL and HPBW. The
radiation patterns with narrower HPBW and lower MSL values.
desired radiation pattern can be obtained by using the cost func-
All the results in this paper are obtained using a computer with
tion given in equation 14.
16 GB RAM and 2.6 GHz i7 processor in MATLAB. In all simu-
Ccost = FMSL · MMSL + FHPBW · MHPBW , (14) lation examples, the population number of the EOA method is
fixed to 40.
where FMSL and FHPBW are the weight factors. MMSL and MHPBW
are the functions used to decrease HPBW and suppress the MSL
values, respectively. The FMSL function can be formulated as in
Optimum design of LAA
equation 15.
wn1 p In the first example, the amplitudes of the 10-element symmetric
FMSL = mMSL (w) · d w + mMSL (w) · d w, (15) LAA are determined by using the EOA method to perform the
−p wn2 optimum design. Figure 5 shows the radiation patterns obtained
with EOA and other optimization algorithms. The reduced MSL
where wn1 and wn2 are the two angles at the first nulls on each and HPBW values achieved by using EOA, SOS [4], BBO [5],
side of the main beam. μMSL(w) can be given in equation 16. PSO [6], Taguchi [7], MFO [8], and AFPA [11] are listed in
 Table 1.
10 (w) − MSLd , 10 (w) . MSLd The MSL and HPBW values of the array pattern achieved with
mMSL (w) = , (16)
0, elsewhere EOA are −27.01 dB and 12.49°, respectively. As can be clearly
seen from Table 1 and Fig. 5, the MSL value of the radiation pat-
where MSLd is the desired value of MSL. ε0(w) is the array factor tern with EOA is better than the results of other compared

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 International Journal of Microwave and Wireless Technologies 991

Fig. 5. Array pattern for 10-element LAA design.

Table 1. Comparative results in terms of MSL and HPBW values for 10-element LAA design

Parameter EOA SOS [4] BBO [5] PSO [6] Taguchi [7] MFO [8] AFPA [11]

MSL (dB) −27.01 −25.28 −25.21 −24.62 −24.87 −26.07 −25.33

HPBW (°) 12.49 12.24 12.24 12.24 12.24 12.60 12.22

Fig. 6. Array pattern for 16-element LAA design.

algorithms. The HPBW value of EOA is better than MFO and it is [11] are illustrated in Fig. 6. MSL and HPBW values of
very close to the HPBW value of other optimization algorithms. 16-element LAAs are given in Table 2 comparatively.
The number of elements of the LAA is determined as 16 in the It is apparent from Fig. 6 and Table 2, it can be seen that
second example. The radiation patterns achieved by using EOA, MSL values of radiation pattern obtained by using EOA are
SOS [4], BBO [5], PSO [6], Taguchi [7], MFO [8], and AFPA better than those of SOS [4], BBO [5], PSO [6], Taguchi [7],

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 992 Ali Durmus and Rifat Kurban

Table 2. Comparative results in terms of MSL and HPBW values for 16-element LAA design

Parameter EOA SOS [4] BBO [5] PSO [6] Taguchi [7] MFO [8] AFPA [11]

MSL (dB) −40.35 −33.39 −33.06 −30.63 −31.21 −40.25 −33.49

HPBW (°) 8.99 8.28 8.28 8.00 8.28 9.00 8.38

Fig. 7. Array pattern for 24-element LAA design.

Table 3. Comparative results in terms of MSL and HPBW values for 24-element LAA design

Parameter EOA SOS [4] BBO [5] PSO [6] Taguchi [7] RRA [9] AFPA [11]

MSL (dB) −41.21 −39.37 −37.14 −34.46 −35.02 −41.08 −37.19

HPBW (°) 6 6 6 6 6 6 6

Table 4. The amplitude values of LAA elements using EOA method

Number of elements Amplitude values [I1, I2, I3, …….IM]

10-Element LAA (Fig. 5) [1.0000 0.8907 0.7000 0.4737 0.3308]

16-Element LAA (Fig. 6) [1.0000 0.9347 0.8147 0.6586 0.4891 0.3283 0.1931 0.1102]
24-Element LAA (Fig. 7) [1.0000 0.9706 0.9137 0.8336 0.7360 0.6272 0.5145 0.4042 0.3019 0.2123 0.1381 0.1113]

MFO [8], and AFPA [11]. All side lobe levels of the radiation the HPBW is 6°. The results of the SOS [4], BBO [5], PSO [6],
pattern achieved with EOA are lower than −40.35 dB. The Taguchi [7], RRA [9], and AFPA [11] method show the MSL
HPBW value of EOA method is 8.99°. The results of HPBW values of −39.37, −37.14, −34.46, 35.02, −41.08, and −37.19 dB,
values achieved by SOS [4], BBO [5], PSO [6], Taguchi [7], respectively, whereas the HPBW values are 6° for all the cases.
MFO [8], and AFPA [11] are 8.28°, 8.28°, 8.00°, 8.28°, 9.00°, The amplitude values of elements optimally determined by
8.38°, respectively. using EOA for the synthesis of LAA with 10, 16, and 24 elements
The number of elements chosen is 24 for LAA synthesis in the are listed in Table 4.
third example. The array patterns achieved by using EOA, SOS Figure 8 shows the converge curve of the LAA with 10, 16, and
[4], BBO [5], PSO [6], Taguchi [7], RRA [9], and AFPA [11] 24 elements by using EOA. As shown in Fig. 8, 10-element LAA
are plotted in Fig. 7. It is shown in Table 3 that the side lobe levels problem converges earlier than those of others due to its low
obtained by the EOA method are suppressed to −41.21 dB and complexity.

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 International Journal of Microwave and Wireless Technologies 993

Fig. 8. Converge curve of LAA with 10, 16, 24 elements.

Fig. 9. Array pattern for 8-element CAA design.

Table 5. Comparative results in terms of MSL, di, and HPBW values for 8-element CAA design

Parameter EOA GA [13] PSO [14] SA [15] BBO [17] FA [18] SOA [19] CSO [20] Taguchi [21] MFO [8]

MSL (dB) −15.95 −9.81 −10.80 −11.99 −12.25 −12.99 −11.11 −15.76 −15.17 −15.78
HPBW (°) 31.50 32.22 31.68 26.64 18.00 32.76 31.68 32.76 36.72 32.76
di 5.78 4.40 4.49 5.87 9.07 4.65 4.40 5.79 4.47 5.80

Optimum design of circular antenna array

In the practical designs of circular arrays, the physical size
In the optimum design of circular arrays, suppression of MSL of the antenna (di) is an important parameter and this value
plays an important role in preventing undesired interference. is desired to be very small. In addition, it also needs a

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 994 Ali Durmus and Rifat Kurban

Fig. 10. Array pattern for 10-element CAA design.

Table 6. Comparative results in terms of MSL, di, and HPBW values for 10-element CAA design

GA PSO SA BBO SOA CSO Taguchi MFO CS

Parameter EOA [13] [14] [15] [17] FA [18] [19] [20] [21] [8] [16]

MSL (dB) −16.29 −10.85 −12.30 −12.99 −13.99 −13.30 −12.50 −14.92 −15.03 −15.81 −12.33
HPBW (°) 29.90 25.56 24.48 18.36 16.20 24.84 24.12 16.74 29.16 30.60 24.20
di 5.90 6.08 5.90 8.02 9.23 6.07 6.10 10.63 5.91 5.91 6.05

Fig. 11. Array pattern for 12-element CAA design.

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 International Journal of Microwave and Wireless Technologies 995

Table 7. Comparative results in terms of MSL, di, and HPBW values for 12-element CAA design

Parameter EOA GA [13] PSO [14] SA [15] BBO [17] FA [18] SOA [19] CSO [20] Taguchi [21] MFO [8]

MSL (dB) −18.29 −11.80 −13.67 −13.92 −14.27 −14.20 −14.11 −17.44 −17.68 −17.79
HPBW (°) 15.08 20.52 20.88 19.44 14.40 20.52 20.34 16.92 16.92 16.92
di 10.61 7.77 7.15 7.95 10.61 7.21 7.14 10.62 10.62 10.62

Table 8. The amplitude and position values of CAA elements using EOA method

Number of
elements Amplitude values [I1, I2, I3, …….IM] Position (inter element spacing) values [d1, d2, d3, …….dM]

8-Elements [0.9039 0.5090 1.0000 0.9550 0.4431 0.0992 0.8523 0.6986] [0.3730 1.3290 0.9222 0.5306 0.9059 0.2444 1.1164 0.3683]
(Fig. 9)
10-Elements [0.9938 0.3262 0.3043 0.4843 0.5275 1.0000 0.3166 0.3033 0.9554 [0.3487 0.8546 0.5515 0.9056 0.0000 0.5657 1.0000 0.3657 0.9864
(Fig. 10) 0.5301] 0.3218]
12-Elements [0.7338 0.5670 0.5926 0.4354 0.6999 1.0000 0.5514 0.6308 0.3313 [0.5533 1.1084 1.3470 1.4997 0.4652 0.5074 0.6138 0.9598 1.2281
(Fig. 11) 0.6236 0.3295 0.9636] 0.7298 1.0599 0.5416]

Fig. 12. Converge curve of CAA.

narrower HPBW value for the circular array to have higher The optimum design problem of the 10-element CAAs is
directionality. examined in the second example. The array patterns obtained
In the second group of examples, first, eight-element CAA with EOA and different meta-heuristic optimization algorithms
synthesis is performed. The radiation pattern obtained by using are plotted in Fig. 10. In Table 6, the HPBW, di, and MSL values
the EOA method is shown in Fig. 9. For comparison purposes, of array pattern achieved by EOA, GA [13], PSO [14], SA [15],
array patterns obtained by using GA [13], PSO [14], SA [15], BBO [17], FA [18], SOA [19], CSO [20], Taguchi [21], MFO
BBO [17], FA [18], SOA [19], CSO [20], Taguchi [21], and [8], and CS [16] are tabulated.
MFO [8] are also illustrated in Fig. 9. It is apparent from Fig. 9 As can be clearly seen from Table 6 and Fig. 10, MSL and di
and Table 5 that the results achieved by EOA are better than values achieved by the EOA technique are better than the result
those of GA, PSO, FA, SOA, CSO, Taguchi, and MFO in terms of other optimization algorithms. The HPBW value achieved by
of MSL and HPBW. Although the best values of HPBW are EOA is better than the values found using the MFO methods.
obtained by using SA and BBO, MSL values obtained by these The HPBW values of the compared algorithms are less than
algorithms are quite high for optimum design. The physical size the value of the proposed method, but if we look in terms of
of the antenna is the same in all algorithms except the value of MSL values, it is unacceptably high for an optimum antenna
BBO is quite large, which is undesirable for design. design.

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 996 Ali Durmus and Rifat Kurban

In the last example, the number of elements of the CAA is 5. Sharaqa A and Dib N (2014) Design of linear and elliptical antenna arrays
selected as 12. The array pattern of EOA is compared with using biogeography-based optimization. Arabian Journal for Science and
those of GA [13], PSO [14], SA [15], BBO [17], FA [18], SOA Engineering 39, 2929–2939.
6. Khodier M and Al-Aqeel M (2009) Linear and circular array optimization:
[19], CSO [20], Taguchi [21], and MFO [8] in Fig. 11. Table 7
a study using particle swarm intelligence. Progress in Electromagnetics
listed the HPBW, di, and MSL values of array patterns achieved
Research (PIER) B 15, 347–373.
by EOA and those of GA [13], PSO [14], SA [15], BBO [17], 7. Dib N, Goudos S and Muhsen H (2010) Application of Taguchi’s
FA [18], SOA [19], CSO [20], Taguchi [21], and MFO [8]. optimization method and self-adaptive differential evolution to the synthesis
From Table 7 and Fig. 11, it is shown that the HPBW and MSL of linear antenna arrays. Progress in Electromagnetics Research 102, 159–180.
results achieved by EOA are better than the other compared algo- 8. Das A, Mandal D, Ghoshal SP and Kar R (2019) Moth flame optimiza-
rithms. If we examine the physical dimensions of the antennas, tion based design of linear and circular antenna array for side lobe reduc-
the antenna obtained with EOA is smaller than the antenna tion. International Journal of Numerical Modelling: Electronic Networks,
designed with CSO, Taguchi, and MFO. Devices and Fields 32, e2486.
The position and amplitude values of the elements determined 9. Subhashini KR (2020) Runner-root algorithm to control sidelobe level
and null depths in linear antenna arrays. Arabian Journal for Science
by using EOA for the radiation patterns in Figs 9–11 are tabulated
and Engineering 45, 1513–1529.
in Table 8. Figure 12 shows the converge curve of the CAA with 8,
10. Prabhakar D and Satyanarayana M (2019) Side lobe pattern synthesis
10, and 12 elements by using EOA. As shown in Fig. 12, an eight- using hybrid SSWOA algorithm for conformal antenna array. Engineering
element CAA problem converges earlier than those of others due Science and Technology, an International Journal 22, 1169–1174.
to its low complexity. 11. Salgotra R, Singh U, Saha S and Nagar AK (2020) Improved flower pol-
The results described in Figs 5–7 and 9–11 illustrate that the lination algorithm for linear antenna design problems. In Soft Computing
proposed method EOA can accurately achieve the optimum radi- for Problem Solving, 79–89, Springer, Singapore.
ation patterns of linear and CAAs with lower MSL and equal or 12. Almagboul MA, Shu F, Qian Y, Zhou X, Wang J and Hu J (2019) Atom
narrow HPBW. The HPBW and MSL values of the array patterns search optimization algorithm based hybrid antenna array receive beam-
obtained with EOA are quite good. The physical size of CAA forming to control sidelobe level and steering the null. AEU-
International Journal of Electronics and Communications 111, 152854.
obtained by EOA is generally better than the other compared
13. Panduro M, Mendez A, Dominguez R and Romero G (2006) Design
optimization algorithms. It can be concluded that EOA can be
of non-uniform antenna arrays for side lobe reduction using the method
used as a good alternative to other optimization methods for of genetic algorithm. AEU International Journal of Electronics and
antenna array synthesis. Communications 60, 713–717.
14. Shihab M, Najjar Y and Khodier NDM (2008) Design of non-uniform
circular antenna arrays using particle swarm optimization. Journal of
Conclusions Electrical Engineering 59, 216–220.
15. Rattan M, Patterh MS and Sohi BS (2009) Optimization of circular
In this paper, the optimum design of linear and CAAs with a dif-
antenna arrays of isotropic radiators using simulated annealing.
ferent number of elements is carried out with a novel optimiza- International Journal of Microwave Wireless Technology 1, 441–446.
tion method known as EOA. The position and amplitude values 16. Khodier M (2020) Optimization of circular antenna arrays using the
of the antenna array elements are determined for the first time cuckoo search algorithm. International Journal of RF and Microwave
by EOA technique to obtain radiation patterns with the lowest Computer-Aided Engineering 30, e22247.
MSL and narrower HPBW values. Linear and CAA examples 17. Singh U and Kamal TS (2011) Design of non-uniform circular antenna
with different numbers of elements are examined and EOA results arrays using biogeography-based optimization. IET Microwaves,
are compared with the results of other optimization algorithms in Antennas and Propagation 5, 1365–1370.
the literature. The results obtained by using EOA are very com- 18. Sharaqa A and Dib N (2014) Circular antenna array synthesis using firefly
algorithm. International Journal of RF and Microwave Computer-Aided
petitive in reducing the MSL compared to other optimization
Engineering 24, 139–146.
algorithms such as SOS [4], RRA[9], GA [13], PSO [6, 14], SA
19. Guney K and Basbug S (2014) A parallel implementation of seeker opti-
[15], BBO [5, 17], FA [18], SOA [19], CSO [20], Taguchi mization algorithm for designing circular and concentric circular antenna
[7, 21], and MFO [8]. The radiation patterns achieved by using arrays. Applied Soft Computing 22, 287–296.
EOA are generally better than well-known optimization methods 20. Ram G, Mandal D, Kar R and Ghoshal SP (2015) Circular and concen-
compared in terms of HPBW and physical size. With the EOA tric circular antenna array synthesis using cat swarm optimization. IETE
method, antenna arrays with different geometric structures such Technical Review 32, 204–217.
as elliptical, concentric circular can be synthesized in future stud- 21. Babayigit B and Senyigit E (2017) Design optimization of circular antenna
ies. Furthermore, the width, length, and height of microstrip patch arrays using Taguchi method. Neural Computing & Applications 28, 1443–
antennas can also be determined with EOA effectively. 1452.
22. Sun G, Zhao X, Liang S, Liu Y, Zhou X and Zhang Y (2019) A modified
chicken swarm optimization algorithm for synthesizing linear, circular and
random antenna arrays. IEEE 90th Vehicular Technology Conference
References
(VTC2019-Fall), Hawaii, USA.
1. Balanis CA (1997) Antenna Theory: Analysis and Design, 2nd Edn. 23. Zheng T, Liu Y, Sun G, Zhang L, Liang S, Wang A and Zhou X (2020)
New York: Wiley. IWORMLF: improved invasive weed optimization with random mutation
2. Mailloux JR (2005) Phased Array Antenna Handbook. Boston: Artech and Lévy flight for beam pattern optimizations of linear and circular
House. antenna arrays. IEEE Access 8, 19460–19478.
3. Guney K, Durmus A and Basbug S (2009) A plant growth simulation 24. Faramarzi A, Heidarinejad M, Stephens B and Mirjalili S (2020)
algorithm for pattern nulling of linear antenna arrays by amplitude con- Equilibrium optimizer: a novel optimization algorithm. Knowledge-Based
trol. Progress in Electromagnetics Research 17, 69–84. Systems 191, 105190.
4. Dib N (2016) Design of linear antenna arrays with low side lobes level 25. Kurban T, Civicioglu P, Kurban R and Besdok E (2014) Comparison of
using symbiotic organisms search. Progress in Electromagnetics Research evolutionary and swarm based computational techniques for multilevel
(PIER) B 68, 55–71. color image thresholding. Applied Soft Computing 23, 128–143.

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press

 International Journal of Microwave and Wireless Technologies 997

Ali Durmus received the B.Sc., M.Sc., and Rifat Kurban received the B.Sc., M.Sc., and Ph.D.
Ph.D. degrees in Electrical and Electronics degrees in Computer Engineering, from Erciyes
Engineering, from Erciyes University, Kayseri University, Kayseri Turkey in 2004, 2006, and
Turkey in 2003, 2005, and 2016, respectively. 2012, respectively. At the Department of
He served as a lecturer at the Department of Computer Engineering, Erciyes University, he
Electricity and Energy, Erciyes University from served as a research assistant from 2005 to
2010 to 2018. Currently, he is an assistant pro- 2012, and an assistant professor from 2012 to
fessor at the Department of Electricity and 2019. He worked as a visiting post-doctoral
Energy, Kayseri University. His research inter- researcher at the Department of Electrical and
ests are smart grids, antennas, antenna arrays, meta-heuristic algorithms, Computer Engineering, University of Tennessee, Knoxville, USA from 2015
and computational electromagnetics. to 2016. Currently, he is an assistant professor at the Department of
Computer Technologies, Kayseri University. His research interests are image
fusion, meta-heuristic algorithms, and internet of things (IoT) applications
in water management.

https://doi.org/10.1017/S1759078720001774 Published online by Cambridge University Press