www.nature.

com/scientificreports

OPEN Red‑tailed hawk algorithm

for numerical optimization
and real‑world problems
Seydali Ferahtia 1,2, Azeddine Houari 1, Hegazy Rezk 3, Ali Djerioui 2, Mohamed Machmoum 1,
Saad Motahhir 4* & Mourad Ait‑Ahmed 1

This study suggests a new nature-inspired metaheuristic optimization algorithm called the red-tailed
hawk algorithm (RTH). As a predator, the red-tailed hawk has a hunting strategy from detecting the
prey until the swoop stage. There are three stages during the hunting process. In the high soaring
stage, the red-tailed hawk explores the search space and determines the area with the prey location.
In the low soaring stage, the red-tailed moves inside the selected area around the prey to choose
the best position for the hunt. Then, the red-tailed swings and hits its target in the stooping and
swooping stages. The proposed algorithm mimics the prey-hunting method of the red-tailed hawk
for solving real-world optimization problems. The performance of the proposed RTH algorithm
has been evaluated on three classes of problems. The first class includes three specific kinds of
optimization problems: 22 standard benchmark functions, including unimodal, multimodal, and fixed-
dimensional multimodal functions, IEEE Congress on Evolutionary Computation 2020 (CEC2020),
and IEEE CEC2022. The proposed algorithm is compared with eight recent algorithms to confirm its
contribution to solving these problems. The considered algorithms are Farmland Fertility Optimizer
(FO), African Vultures Optimization Algorithm (AVOA), Mountain Gazelle Optimizer (MGO), Gorilla
Troops Optimizer (GTO), COOT algorithm, Hunger Games Search (HGS), Aquila Optimizer (AO), and
Harris Hawks optimization (HHO). The results are compared regarding the accuracy, robustness,
and convergence speed. The second class includes seven real-world engineering problems that will
be considered to investigate the RTH performance compared to other published results profoundly.
Finally, the proton exchange membrane fuel cell (PEMFC) extraction parameters will be performed
to evaluate the algorithm with a complex problem. The proposed algorithm will be compared with
several published papers to approve its performance. The ultimate results for each class confirm the
ability of the proposed RTH algorithm to provide higher performance for most cases. For the first
class, the RTH mostly got the optimal solutions for most functions with faster convergence speed.
The RTH provided better performance for the second and third classes when resolving the real word
engineering problems or extracting the PEMFC parameters.

Optimization algorithms have played a vital role in recent decades in assisting engineers and designers in find-
ing improved solutions for saving time, money, and e­ nergy1. Numerical optimization methods often employ
simple and widely used ­models2. Nevertheless, these algorithms require several gradient information to find
superior solutions circling a specific point in a small ­space3. Furthermore, they are sensitive to starting points,
mainly when the issues under consideration include multiple local solutions. Inappropriate initial point selection
makes searching for the global optimum solution complex and ­inconsistent4. In recent years, many complicated
optimization problems have evolved in various disciplines. These problems frequently involve several choices of
optimization variables, complex nonlinear constraints, and objective f­ unctions5. These algorithms have been used
for civil e­ ngineering6, electrical engineering, image p
­ rocessing7, medical and biological applications, and others.
As a result, existing numerical approaches cannot address these complicated problems promptly and precisely.
On the other hand, nature inspiration can provide concepts to develop artificial intelligence algorithms to solve
these complex problems.

1
Institut de Recherche en Énergie Électrique de Nantes Atlantique, IREENA, Nantes University, Saint‑Nazaire,
France. 2Laboratoire de Génie Electrique, Dept. of Electrical Engineering, University of M’sila, M’sila,
Algeria. 3College of Engineering at Wadi Addawaser, Prince Sattam Bin Abdulaziz University, Al‑Kharj, Saudi
Arabia. 4ENSA, University of Sidi Mohamed Ben Abdellah, Fez, Morocco. *email: [email protected]

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 1

Vol.:(0123456789)
 www.nature.com/scientificreports/

Recently, various nature-inspired metaheuristic algorithms (MAs) have been developed. These MAs mimic
the motions of live beings or natural events. These MAs have been elaborated and employed for solving many
optimization issues as competitive alternative ­solvers8. However, these MAs have a common drawback; they
frequently exhibit an extra sensitivity to adjusting user-defined parameters. Another disadvantage is that MAs
may not always reach the global optimum ­solution9. MAs are divided into two ­types8: single solution-based
and population-based. A single solution is processed during optimization in the single solution-based type. In
contrast, in the population-based type, solutions are developed in each optimization iteration. Population-based
MAs initiate the optimization process by producing a set of random individuals. Each of them represents a pos-
sible optimal solution. Iteratively, the population will be developed by substituting the actual population with a
recently created one based on certain stochastic ­operators10,11.
Regardless of how diverse these algorithms are, they always have one common characteristic: the searching
operations are divided into two phases: exploration and ­exploitation12. As a result, during the early stages of
the search process, a well-designed optimizer’s exploratory behaviors must have an enriched-enough random
character to distribute more random solutions ­effectively13. Hence, it enhanced diverse parts of the search space.
After the exploration phase, the exploitation phase is carried out. The optimizer accelerates the search process
in a narrow area instead of the whole search space by concentrating on the neighborhood of the best-obtained
solutions. An efficient optimizer must strike an acceptable and precise balance between the exploration and
exploitation phases. Otherwise, the risk of becoming locked in local optima (local solutions) and having imma-
ture convergence downsides grows. According to the No Free Lunc (NFL) theorem [36], all the proposed MAs
show a comparable average performance when resolving all possible optimization problems. In other words,
no algorithm can be considered a general best algorithm. As a result, the NFL theorem promotes creating and
developing more new efficient optimizers. Technically, each MA employs unique evolution mechanisms. Accord-
ing ­to14, several fundamental concepts can be used to characterize them. Among these concepts:

• Parallelism is utilized in population-based algorithms such as SA and HA. Different individuals are sent out
at once to complete a task, and the results are compared. According to the comparison, additional ideas are
utilized to assess the population evolution or to create new individuals.
• Acceptance is utilized in three cases: 1. Admit temporary solutions that weaken objective function due to
search space expansion, 2. Handling the cost function’s constraints. There are two ways to cope with the
constraints. The first way excludes any solutions if they match the start conditions. The second procedure is
used if any solution can be allocated a numerical value. In this case, all solutions must participate, and the
initial conditions may match infeasible solutions. 3. Adding restrictions to allowed solutions that improve
the optimal solution by, at minimum, the limiting level. When comparing values produced from previous
calculations, this strategy aids in avoiding numerical issues.
• Elitism: the best solution must be kept from one iteration to the next in repeating population-based algo-
rithms. The elitism principle is utilized to accomplish this by retaining the individual who outlined the best
solution and utilizing it as a reference for the following iteration or upgrading it if a better solution is located.
The notion of elitism may also be applied to many individuals, directing an élite group of them to the next
iteration.
• Selection is a probability-based method that generates new individuals from a pool of available ones. This
approach may incorporate weights into the probabilistic selection process, in which random individuals are
chosen to produce new ones.
• Decay or Reinforcement: Decay allows for more initial freedom, followed by incremental flexibility constraints.
This approach is based on a decremental updating factor (< 1) used at each iteration. Reinforcement is
employed similarly in some cases by applying an incremental factor higher (> 1).
• Immunity is gained through finding characteristics of solutions that lead to appropriate settings. Immunity
prefers solutions with characteristics similar to those attributes.
• Self-Adaptation is a process that allows the parameters of algorithms to be updated in response to the progress
of optimization.

A novel high-efficiency SA optimization algorithm is suggested in this study to compete with existing opti-
mizers. The basic concept of the suggested optimizer is based on the hunting skills of Red-tailed Hawks, one of
the most intelligent birds of prey. The benefits of evolutionary and swarm approaches have been included in the
RTH algorithm’s design to outperform the existing optimization algorithms. The proposed RTH contains three
stages: high soaring, low soaring, and stooping and swooping stages. The high soaring stage is quite similar to
evolutionary approaches’ search behavior. The RTH algorithm relays and gathers all search points, beginning
with the original position and progressing to the best position. The second and third phases imitate SA’s behavior
while moving to the best position, using the information of the preceding site for each point.
Despite the availability of metaheuristic algorithms, there is a continuous need for developing and creating
new algorithms that may provide better performance for such problems than the actual ones. The NFL theory
confirms this need for new optimizers. To this end, this paper suggests this novel algorithm. The main contribu-
tion of this paper is to propose a novel nature-inspired metaheuristic optimization algorithm for better solving
optimization issues. The proposed algorithm benefit from the unique hunting strategy of the red-tailed hawk
and applies it to tackle various optimization problems. The suggested RTH algorithm’s performance has been
examined on three optimization problem classes. The first class contains three types of optimization functions:

• Twenty-three standard benchmark functions: unimodal, multimodal, and fixed-dimensional multimodal

functions.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 2

Vol:.(1234567890)
www.nature.com/scientificreports/

• IEEE Congress on Evolutionary Computation 2020 (CEC2020) with 15 and 20 search space dimensions.
• IEEE Congress on Evolutionary Computation 2022 (CEC2022) with 10 and 20 search space dimensions.

The proposed algorithm is compared with eight recent algorithms to confirm its contribution to solving the
problems of this class. The considered algorithms in the comparison are Farmland fertility Optimizer (FO)15,
African Vultures Optimization Algorithm (AVOA)16, Mountain Gazelle Optimizer (MGO)17, Artificial Gorilla
Troops Optimizer (GTO)18, COOT a­ lgorithm19, Hunger Games Search (HGS)20, Aquila Optimizer (AO)21, and
Harris Hawks optimization (HHO)22. The findings’ accuracy (mean value), robustness (standard deviation), and
convergence speed are compared. The second class consists of seven real-world engineering problems that will
be thoroughly investigated compared to previously published solutions. These problems include the optimal
design of an I-shaped beam, a three-bar truss Design, design of a tubular column, a piston lever, a corrugated
design, and design tension/compression spring. The third class provides the parameters extraction for proton
exchange membrane fuel cell (PEMFC), which will be used to assess the suggested algorithm with a difficult
task. To validate its performance, the proposed algorithm will be compared to numerous published articles for
three types of PRMFC: BSC 500W, NedStack PS6, and SR-12 500W.
The rest of this paper is organized as follows: Section “Related works” presents a related works part that
shows and explains a set of related works. Section “Red-tailed hawk algorithm” reviews the hunting behavior of
red-tailed, gives the inspiration source, and describes each stage of the proposed RTH algorithm. This section
includes the mathematical model of the proposed algorithm. Section “Results and discussion” presents the out-
comes of RTH in solving the considered problems. This paper ends with a conclusion in Section “Engineering
optimization problems”.

Related works
In recent decades, there has been a rise in the assessment and application of metaheuristic algorithms to tackle
optimization problems. In the literature, the population-based MAs may be divided into four categories based on
their ­inspiration23,24: evolutionary (EA), Physics-based (PA), Human-based (HA), and swarm-based MAs (SA).
EAs imitate biological evolutionary processes, including recombination, mutation, and selection. The Genetic
Algorithm (GA)25,26, Differential Evolution (DE)27, Biogeography-Based Optimizer (BBO)28, and Mind evolu-
tionary algorithm optimization (MEDA)29 are the most well-known EAs. Physical phenomena-inspired algo-
rithms are based on physical laws such as gravity, magnetic force, etc. Gravitational Search Algorithm (GSA)30,
Gradient-based ­optimizer31, and Energy Valley Optimizer (EVO)32 are a few examples. Human-based MAs
imitate some human activities and behaviors. Socio Evolution and Learning Optimization (SELO)33, Social
Network Search (SNS)34, and Human Felicity Algorithm (HFA)35 are some examples of this category. Swarm-
based MAs mimic the social behaviors of animals or organisms living in swarms, communities, or p ­ acks36.
Particle Swarm Optimization (PSO)37, Salp Swarm Algorithm (SSA)38, and Jellyfish Search algorithm (JSA)39
are the most well-known MAs in this category. The swarm-based MAs have been getting more attention in
the last years due to the availability of inspiration sources and their efficacy in resolving various optimization
problems. New papers propose or review this category of optimization algorithms, such as Mountain Gazelle
Optimizer (MGO)17, Advances in Spotted Hyena ­Optimizer40, Advances in Tree Seed A ­ lgorithm41, Advances
in Sparrow Search A ­ lgorithm42, Butterfly Optimization Algorithm (BOA)43, Advanced Butterfly Optimization
Algorithm (ABOA)44, Modified Butterfly Optimization Algorithm with Lagrange Interpolation (MBOALI)45,
African Vultures Optimization Algorithm (AVOA)16, Whale optimization algorithm (WOA)46, Artificial Gorilla
Troops Optimizer (GTO)18, COOT a­ lgorithm19, Weibull Flight based Moth Flame Optimization (WF-MFO)47,
Hunger Games Search (HGS)20, Aquila Optimizer (AO)21 and so on. On the other hand, Quantum-inspired
metaheuristic algorithms developed by combining Quantum Computing (QC) principles into metaheuristic
algorithms are gaining more i­ nterest48. The performance of these algorithms is considerably enhanced by using
the QC for boost exploration and exploitation and faster convergence.

Red‑tailed hawk algorithm

This part discusses the proposed RTH algorithm. The inspiration source and hunting strategy are discussed in
the first subsection. Then, the mathematical model mimicking the red-tailed hawk’s behavior is presented, and
each stage of the algorithm is analyzed.

Inspiration and behavior during hunting. The red-tailed hawk (Buteo jamaicensis) is a bird of prey
that breeds over much of North America, from Alaska’s interior and northern Canada to Panama and the West
Indies. The red-tailed hawk lives in various environments and elevations, such as deserts, grasslands, forests,
agricultural fields, and cities. The red-tailed hawk is a predatory, carnivorous eater. Almost every little animal
they come upon may be seen as possible ­prey49. Small mammals, such as rodents, are their most common prey,
although they also eat birds, fish, reptiles, invertebrates, and amphibians. Prey varies greatly depending on geo-
graphical and seasonal availability; however, rodents comprise 85 percent of a hawk’s d ­ iet50.
The red-tailed hawk soars with its wings in a mild dihedral, flapping as little as possible to save energy. Unlike
other hawks, the red-tails can fly for long distances thanks to this feature. Because soaring is the most efficient
flying mode for these hawks, it is utilized more f­ requently51. It moves between 32 and 64 km/h (20–40 mph)
when soaring or flapping its wings. The large wings allow the red-tailed to reach 190 km/h (120 mph) when
­plunging52. Red-tailed hawks can fly fast and powerfully while repeatedly diving at perceived threats during nest
­protection53. As illustrated in Fig. 1, the red-tailed has three types of flying:

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 3

Vol.:(0123456789)
 www.nature.com/scientificreports/

a- High Soaring b- Low Soaring c- Stopping & Swooping

Figure 1.  Behavior of red-tailed hawk during hunting.

• High soaring (Fig. 1a): It flies highly with its wings in a mild dihedral, flapping as little as possible to save
energy to explore the selected area.
• Low soaring (Fig. 1b): after selecting the target position, the red-tailed fly with a low soaring in a spiral move-
ment around the prey. This movement allows it to detect the best location and time to hit the target.
• Stooping and swooping (Fig. 1c): after selecting the best location and moment in the previous step, the
red-tailed swooped its prey by stooping and raising its acceleration (from 32–64 to 190 km/h) in a curved
direction.

Mathematical model. The suggested red-tailed hawk (RTH) algorithm mimics the red-tailed hawk’s hunt-
ing behavior. The actions taken at each hunt stage are presented and modeled. This algorithm includes three
stages, high soaring, low soaring, and stooping and swooping.

1 High soaring: the red-tailed hawk will soar far into the sky, looking for the best location in terms of food
availability. Figure 2 illustrates the behavior of the red-tailed hawks during the high soaring stage, and Eq. (1)
represents the mathematical model of this stage :
X(t) = Xbest + (Xmean − X(t − 1)) · Levy(dim) · TF(t) (1)
where X(t) represents the red-tailed hawk position at the iteration t, Xbest is the best-obtained position, Xmean
is the positions’ mean, Levy represents the levy flight distribution function that can be calculated accord-
ing to Eq. (2), and TF(t) denotes the transition factor function that can be calculated according to Eq. (3).

Figure 2.  Behavior of red-tailed hawk during high soaring stage.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 4

Vol:.(1234567890)
www.nature.com/scientificreports/

µ·σ
Levy(dim) = s −1
|υ|β
 � � � 
 Ŵ(1 + β) · sin πβ
2  (2)
σ = � �
 1 − β/2

1 + β
�
Ŵ( 2) · β · 2
where s is a constant (0.01), dim is the problem dimension, β is a constant (1.5), u, and υ are random num-
bers [0 to 1].
  
TF(t) = 1 + sin 2.5 + t Tmax (3)

where Tmax represents the max number of iterations.

2 Low soaring: the hawk surrounds the prey by flying much lower to the ground in a spiral line. This stage is
illustrated in Fig. 3, and its model can be expressed as follows:
X(t) = Xbest + (x(t) + y(t)) · StepSize(t)
(4)
StepSize(t) = X(t) − Xmean
where x and y denote direction coordinates which can be calculated as follows
 � � 
t x(t)
� �
x(t) = R(t) · sin (θ(t))  R(t) = R0 · r − Tmax · rand  x(t) =
�
max |x(t)|
� �
y(t) (5)
y(t) = R(t) · cos (θ(t))
�
θ (t) = A · 1 − t Tmax · rand y(t) =
� � �
max �y(t)�
 

where R0 represents the initial value of the radius [0.5–3], A denotes the angel gain [5–15], rand is a random
gain [0–1], and r is a control gain [1, 2]. These parameters help the hawk fly around the prey with spiral
movements, as explained in Fig. 4.
3 Stooping and Swooping: In this stage, the hawk suddenly stoops and attacks the prey from the best-obtained
position in the low soaring stage. Figure 2 explains the behavior of the red-tailed hawks during this stage.
This stage can be modeled as follows:
X(t) = α(t) · Xbest + x(t) · StepSize1(t) + y(t) · StepSize2(t) (6)
where each step size can be calculated as follows
StepSize1(t) = X(t) − TF(t) · Xmean
(7)
StepSize2(t) = G(t) · X(t) − TF(t) · Xbest

X(t)

X(t+1)

Xbest
Figure 3.  Behavior of red-tailed hawk during low soaring stage.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 5

Vol.:(0123456789)
 www.nature.com/scientificreports/

0.5

y 0

-0.5

-1
0 1
5 0.5
10 0
15 -0.5
20 -1 x
iterations

Figure 4.  Evolution of the direction coordinates as a function of iterations.

where α and G are the acceleration and the gravity factors, respectively, they can be defined as follows:
 
α(t) = sin2 2.5 − t Tmax

  (8)
G(t) = 2 · 1 − t Tmax


where α represents the hawk’s acceleration that increases with the increase of t to enhance the convergence
speed, and G is the gravity effect that decreases to reduce the exploitation diversity when the hawk is much
near the prey. This phase is explained in Fig. 5.

X(t)

G(t)
X(t+1)

Xbest
Figure 5.  Behavior of red-tailed hawk during stooping and swooping stages.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 6

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 6 presents the evolution of the solutions of a 2-dimensional problem as a function of the iterations.
During the first iterations, the hawk explored the search space (exploration phase). It took significant steps to
detect the prey position. The length of these steps decreases as the hawk is near the prey position. The hawk
spent a small number of iterations in the exploration due to the hybrid evolutionary-swarm updating mecha-
nism. It avoided attempting on the local minim thanks to the high soaring stage based on the Levy function. The
low soaring enhances the convergence speed in the exploitation phase, where the stooping and the swooping
stages strengthen the RTH algorithm’s accuracy. Figure 7 presents the hunting behavior for 2-dimensional and
3-dimensional problems. The hawk explores the whole search space from these figures to avoid attempting local

Howk' position
-31.97033
Prey' position
-31.970332

-31.970334
60
-31.970336

40 -31.970338 70
-15.98644
-15.98642 80
-15.9864 90
20 -15.98638
-15.98636 100
dim2

-20

-40

0
-60 10
-60 20
-40 30
-20 40
50
0
60
20
70
40 80
60 90 iteration
dim1
80 100

Figure 6.  Agents’ evolution as a function of the iterations for a 2-dimensional problem.

Figure 7.  Agents’ evolution for 2-dimensional and 3-dimensional problems.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 7

Vol.:(0123456789)
 www.nature.com/scientificreports/

solutions. Due to its unique updating mechanism gets closer to the optimal solution without falling into local
ones with a fast convergence rate.
To better understand the characteristics of the proposed RTH and its basics, the RTH concepts are presented
in Table 1.

ALGORITHM 1:RTHPSEUDOCODE
1 Initialization: random generation within the search space.
2 While t<Tmax do
3 High soaring stage: for i=1:Npop do
4 Calculate Levy flight distribution Eq(2)
5 Calculate the transition factor TF Eq(3)
6 Update positions Eq(1)
7 end
8 Low soaring stage: for i=1:Npop do
9 Calculate direction coordinates Eq(5)
10 Update positions Eq(4)
11 end
12 Stooping and Swooping stage: for i=1:Npop do
13 Calculate the acceleration and the gravity factors Eq(8)
14 Calculate the step size Eq(7)
15 Update positions Eq(6)
16 end
17 end

Results and discussion

RTH’s performance was evaluated using several test functions. Three classes of test functions are used in this
study:

• Standard test ­functions54,

• CEC 2020 benchmark test f­ unctions55,
• CEC 2022 benchmark test f­ unctions56.

Standard test functions. Unimodal, multimodal, and fixed-dimension multimodal test functions are
used to evaluate the RTH p ­ erformance54. Unimodal test functions (F1-F7) are used to put an algorithm’s exploi-
tation abilities to the test, whereas multimodal test functions (F8-F13) are used to experiment with the algo-
rithm’s exploration performance. The fixed dimension test functions (TF14-TF23) demonstrate the algorithm
exploration ­capability57. These functions are provided in Table A1 (appendix). For all the tests, the population
size for each algorithm has been set at 30 with max iterations of 1000.
This study ran RTH 30 times and reported the findings, including the average and standard deviation values of
the best-so-far solutions obtained in each run to provide statistically significant results. The test is carried out for
six other algorithms to underline the effectiveness and superiority of RTH to that of different methods: Particle
Swarm Optimization (PSO)37 as the best-known optimization method, Salp Swarm Algorithm (SSA)38 as one of
the most used algorithms, Hunger Games Search (HGS)20, COOT a­ lgorithm19, Artificial Eco-system Optimizer
(AEO)4, and Aquila Optimizer (AO)21 as recently developed metaheuristics. Table 2 summarizes the parameter
setting for the RTH algorithm, where the parameters are set by the try-and-error method. The other algorithms
are based on their default parameters. Table 3 shows the results of RTH and the other used algorithms on these
standard test functions using the best, average, and standard deviation (StD) values.

Concept Projection on the algorithm

Classification according to inspiration type Swarm-based
Classification according to evolution type Population-based
Parallelism The algorithm uses several positions and their mean value
Acceptance Not used due to the low random behavior of the algorithm
Elitism Used at each phase (Xbest)
Selection The high soaring phase used this concept based on Levey’s distribution
Decay Used in α and G Eq. (8) to raise the exploration ability and converge toward the best solution
Reinforcement Used in TF Eq. (3) to enhance the exploitation ability and avoid attempting in local optima
Immunity Not used
Used in TF Eq. (3) and α and G Eq. (8), these parameters are updated according to the current
Self-adaptation
iteration

Table 1.  The RTH concepts.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 8

Vol:.(1234567890)
www.nature.com/scientificreports/

Parameter Range Cosen value

A [5–25] 15
R0 [0.1–2] 0.5
r [0.1–2] 1.5

Table 2.  The RTH paramters.

Unimodal functions results (F 01–07) because unimodal test functions have just one global optimum solution,
they may be used to assess an algorithm’s convergence speed (exploitation phase). From Table 3 results and the
curves illustrated in Fig. 8 (fnc 05 curves are not included due to the exact similarity in the results), the RTH
outperformed all algorithms in all test functions in terms of accuracy and convergence speed. These results
demonstrate RTH’s capabilities in quick exploitation, which can enable RTH in fast convergence. This capacity
is derived from the high soaring phase that uses the adaptive transition factor (TF) and Lévy distribution.
Multimodal functions results (F 08–13) Multimodal test functions include multiple local solutions. Their num-
ber grows exponentially as the number of search space dimensions grows (optimization variables). Having
more than one optimal is beneficial if the purpose is to test an algorithm’s exploration capabilities. As a result,
they are well-suited for assessing optimization algorithms’ local optima avoidance and exploratory behavior.
The obtained results in Table 3 and Fig. 9 demonstrated that RTH has superior exploration capabilities to
other methods for five of six functions in terms of finding the optimal solution with the lowest number of
iterations. These functions approve the ability of the RTH to handle complicated problems.
Fixed-dimensional multimodal function (F 14–23) similar to the multimodal test functions, these functions
include a lot of local optima with more complexity due to the minimization of the negative cost function.
Hence, they are suitable for evaluating the accuracy of the algorithm. As illustrated in Fig. 10 and reported
in Table 3, the proposed RTH has reached the global optimum for all the considered functions. However, its
robustness, represented by the standard deviation (StD), has been reduced to the increased complexity. RTH’s
exploration is due to its different optimization phases and the acceleration and gravity effects.

IEEE CEC 2020. The IEEE Congress on Evolutionary Computation (CEC) offers yearly benchmark func-
tions to test and evaluate optimization algorithms’ ability to identify optimal solutions (minimization problems).
This collection of functions is commonly recognized by the year as standard functions established by the IEEE
Congress on Evolutionary Computation, for example, CEC 2020 and CEC 2022. These test functions include
unimodal, basic multimodal, hybrid, and composition functions. First, the proposed RTH will be compared
with the considered algorithms using the ten functions of the IEEE CEC ­202055. The characteristics of CEC 2020
functions are presented in Table A2. Their 3D maps are shown in Fig. 11 for the unimodal function (F1), Fig. 12
for the basic functions (F2-F4), and Fig. 13 for the composition functions (F8–F10).
All the algorithms are run 30 times for each function. The results for these functions for D = 15 are presented
in Table 4 and. These include the best, the main, and the standard deviation (StD). The fitness evolutions during
the optimization process for each function are presented in Fig. 14.
The three last rows show the scores of each algorithm. It can be seen from the L1 row that the proposed
RTH can find the optimal solution for five functions from ten, followed by MGO by three functions, then HGS,
AVOA, and GTO by two functions, where the other algorithms achieved only one optimal solution. From the
penultimate row, the proposed RTG can accomplish the best mean results for five problems, followed by MGO
by four, then HGS by two. The last row presents the StD results that confirm the algorithm’s robustness. Based
on these results, the proposed RTH provided the best results in this criterion four times. The evolution of the
average cost function is presented in Fig. 13. From this figure, the proposed RTH has fast convergence speed
compared to the other algorithms. This can be explained by the unique exploitation and exploitation strategies
of the proposed RTH algorithm.
The functions of the CEC 2020 benchmark will be tested with increased search space dimension where D
is equal to 20. This allows us to investigate the performance with more complex cases. The results are provided
in Table 5.
It can be seen from the L1 row that the proposed RTH can find the optimal solution for six functions, followed
by MGO by three, then HGS, AVOA, and GTO by two functions. From the penultimate row, the proposed RTG
has got the best mean results for five problems, followed by MGO by three, then HGS by two. From the last row,
the StD results of the RTH is the bet by five times, which confirms the algorithm’s robustness. Based on these
results, the proposed RTH provided the best results in this criterion four times. The evolution of the average
cost function is presented in Fig. 15. Similar to the previous case, the proposed RTH has fast convergence speed
compared to the other algorithms.
ANOVA is the abbreviation for analysis of variance. ANOVA is frequently used to evaluate equality across
different means by comparing variance between groups with variation within groups. This test will be used to
assess the results between the optimizers for both D = 15 and D = 20 cases. The best outcome for each case is
marked in red.
For the unimodal function (function 01), the ANOVA ranking is presented in Fig. 16. From this figure; the
RTH generates the best solutions in terms of accuracy with the lowest variation range. On the other hand, the
FO and the HHO cannot find any solution.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 9

Vol.:(0123456789)
 www.nature.com/scientificreports/

Func Metric MGO FO COOT HGS AO HHO GTO AVAO RTH

Unimodal functions
Best 1.21 × ­10−254 1.46 × ­10−38 9.63 × ­10−4 0 0 1.95 × ­10−13 0 0 0
F1 Mean 6.50 × ­10 −240
1.93 × ­10 −35
4.22 × ­10 −3
0 7.71 × ­10 −200
9.75 × ­10−9 0 0 0
StD 0 5.11 × ­10−35 3.30 × ­10−3 0 0 1.97 × ­10−8 0 0 0
Best 5.84 × ­10−140 2.16 × ­10−21 7.95 × ­10−3 0 7.47 × ­10−155 4.08 × ­10−8 0 0 0
F2 Mean 1.20 × ­10−127 5.58 × ­10−20 1.69 × ­10−2 5.78 × ­10−191 2.73 × ­10−100 1.17 × ­10−5 0 3.36 × ­10−308 0
StD 6.45 × ­10−127 9.78 × ­10−20 4.79 × ­10−3 0 1.50 × ­10−99 1.94 × ­10−5 0 0 0
Best 1.96 × ­10−65 0 5.10 × ­10−16 0 2.3714 × ­10−135 4.16 × ­10−19 0 0 0
F3 Mean 1.01 × ­10−4 0 1.78 × ­10−12 6.6714 × ­10−129 8.07 × ­10−9 5.73 × ­10−11 0 0 0
StD 5.53 × ­10−4 0 3.54 × ­10−12 3.6541 × ­10−128 1.21 × ­10−9 1.29 × ­10−10 0 0 0
Best 1.25 × ­10−92 4.69 × ­10−15 0.0245 0 5.34 × ­10−156 2.20 × ­10−8 0 0 0
F4 Mean 1.95 × ­10−79 1.54 × ­10−13 0.0948 4.11 × ­10−144 2.23 × ­10−100 1.44 × ­10−5 0 1.4239E− 301 0
StD 1.06 × ­10−78 1.47 × ­10−13 0.0350 2.21 × ­10−143 1.22 × ­10−100 2.56 × ­10−5 0 0 0
Best 0 0 0 0 0 0 0 0 0
F5 Mean 0 0 0 0 0 0 0 0 0
StD 0 0 0 0 0 0 0 0 0
Best 0 0.7533 0.0001 1.13 × ­10−20 7.84 × ­10−9 2.15 × ­10−6 6.16 × ­10−33 8.26 × ­10−16 0
F6 Mean 1.77 × ­10−31 1.2063 0.0044 3.30 × ­10−15 2.26 × ­10−5 3.37 × ­10−4 4.98 × ­10−24 2.82 × ­10−14 2.67 × 10−33
StD 6.06 × ­10 −31
0.1781 0.0024 1.58 × ­10−14
5.50 × ­10 −5
6.70 × ­10 −4
2.72864 × ­10 −23
5.74 × ­10 −14
4.97 × 10−33
Best 7.63 × ­10 −6
2.13 × ­10 −4
1.14 × ­10 −3
1.59 × ­10−7
1.46 × ­10 −7
1.08 × ­10 −4
8.71 × ­10−7
6.19 × ­10 −6
2.38 × 10−6
F7 Mean 9.93 × ­10−5 1.98 × ­10−3 2.51 × ­10−3 6.30 × ­10−4 4.85 × ­10−5 7.74 × ­10−4 5.47 × ­10−5 7.78 × ­10−5 4.07 × 10−5
StD 8.49 × ­10−5 9.81 × ­10−4 1.25 × ­10−3 1.11 × ­10−4 6.61 × ­10−5 6.13 × ­10−4 4.92 × ­10−5 6.77 × ­10−5 4.54 × 10−5
Multimodal functions
Best − 4189.83 − 2703.44 − 3594.96 − 4189.83 − 4189.67 − 4189.83 − 4189.83 − 4189.83 − 3833.00
F8 Mean − 4189.83 − 2255.04 − 3059.81 − 4174.04 − 3878.30 − 4097.29 − 4189.83 − 4064.56 − 3003.14
StD 2.15 × 10−12 192.93 262.38 86.50 645.70 351.65 3.73 × ­10−12 265.98 268.73
Best − 20.64 − 74.41 − 88.01 − 90 − 90 − 90 − 90 − 90 − 90
F9 Mean − 0.69 − 59.95 − 81.66 − 90 − 90 − 90 − 90 − 90 − 90
StD 3.77 6.63 4.23 0 0 1.84 × 10–8 0 0 0
Best − 347.92 − 22003.75 − 22023.75 − 22023.75 − 22023.75 − 22023.75 − 22023.75 − 22023.75 − 22023.75
F10 Mean − 69.88 − 22003.75 − 22004.44 − 22023.75 − 22021.58 − 22007.29 − 22006.02 − 22022.47 − 22013.08
StD 88.94 7.40 × 10−12 48867365 1.64 × ­10−10 4.58 7.08 4.73 4.88 10.15
Best 0 0.25 0.06 0 0 0 0 0 0
F11 Mean 0 0.60 0.17 6.41E− 05 0.02 0 0 0 0
StD 0 0.10 0.06 0 3.46E− 12 0 0 0 0
Best 4.71 × 10−32 1.57 × ­10−11 4.27 × ­10−17 4.71 × 10−31 1.28 × ­10−9 1.03 × ­10−12 4.71 × 10−31 4.71 × 10−31 4.71 × 10−31
F12 Mean 4.71 × 10−32 6.49 × ­10−6 3.43 × ­10−11 4.71 × 10−31 1.50 × ­10−6 2.22 × ­10−7 4.71 × 10−31 1.07 × ­10−30 4.71 × 10−31
StD 1.67 × 10−47 1.33 × ­10−5 6.92 × ­10−11 8.91 × ­10−47 3.11 × ­10−6 4.48 × ­10−7 8.91 × ­10−47 3.29 × ­10−30 8.91 × ­10−47
Best 1.35 × 10 −32
8.59 × ­10−12
3.75 × ­10−15
1.35 × 10 −32
1.49 × ­10−11
1.31 × ­10−11
1.35 × 10 −32
1.35 × 10 −32
1.35 × 10−32
F13 Mean 1.35 × 10 −32
5.33 × ­10−6
6.46 × ­10−12
1.35 × 10 −32
7.04 × ­10−8
1.42 × ­10−7
1.35 × 10 −32
1.90 × ­10−30
1.35 × 10−32
StD 5.57 × 10 −48
1.10 × ­10−5
1.33 × ­10−11
5.57 × 10 −48
1.52 × ­10−7
2.57 × ­10−7
5.57 × 10 −48
5.76 × ­10−30
5.57 × 10−48
Fixed− dimensional multimodal function
Best 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
F14 Mean 1.00 1.33 1.00 1.98 2.43 1.53 1.00 3.00 1.00
StD 2.10 × 10−16 0.38 1.96E− 13 2.98 3.32 1.26 4.12 × 10−17 3.54 2.90
Best 3.075 × 10−4 7.166 × ­10−4 3.273 × ­10−4 3.07493 3.132 × ­10−4 3.097 × ­10−4 3.075 × 10−4 3.127 × ­10−4 3.075 × 10−4
F15 Mean 3.075 × 10−1 1.163 × ­10−3 6.392 × ­10−4 6.32275 4.366 × ­10−4 3.974 × ­10−4 3.685 × ­10−4 6.611 × ­10−4 3.762 × ­10−3
StD 7.53 × 10−12 2.73 × ­10−4 2.69 × ­10−4 2.36376 8.266 × ­10−5 7.429 × ­10−5 2.323 × ­10−4 4.906 × ­10−4 7.558 × ­10−2
Best − 1.032 − 1.032 − 1.032 − 1.032 − 1.032 − 1.032 − 1.032 − 1.032 − 1.032
F16 Mean − 1.032 − 1.032 − 1.032 − 1.032 − 1.032 − 1.032 − 1.032 − 1.032 − 1.032
StD 6.25 × 10−16 3.54 × ­10−5 3.29 × ­10−9 6.71 × ­10−16 1.57 × ­10−4 3.33 × ­10−6 6.71 × ­10−16 6.90 × ­10−3 6.65 × ­10−16
Best 0.398 0.398 0.398 0.398 0.398 0.398 0.398 0.398 0.398
F17 Mean 0.398 0.542 0.398 0.398 0.398 0.398 0.398 0.398 0.398
StD 0 0.725 2.16 × ­10−8 0 1.55 × ­10−4 9.02 × ­10−9 0 1.62 × ­10−9 0
Continued

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 10

Vol:.(1234567890)
www.nature.com/scientificreports/

Func Metric MGO FO COOT HGS AO HHO GTO AVAO RTH

Best 3 >3 3 3 >3 >3 3 >3 3
F18 Mean 3 >3 >3 3 >3 >3 3 12.9850033 3.9
StD 1.9710−15 4.26 × 10−4 1.22 × 10−8 3.92 × 10−15 2.27 × 10−2 2.144 × 10−5 8.96 × 10−16 21.94 4.93
Best − 3.863 − 3.854 − 3.863 − 3.863 − 3.863 − 3.863 − 3.863 − 3.863 − 3.863
F19 Mean − 3.863 − 3.787 − 3.863 − 3.863 − 3.856 − 3.818 − 3.863 − 3.825 − 3.863
StD 2.58 × ­10−15 0.263 3.20 × ­10−9 2.71 × ­10−15 2.84 × ­10−3 6.94 × ­10−2 2.682 × 10−15 0.142 2.696 × ­10−15
Best − 3.322 − 3.009 − 3.322 − 3.322 − 3.295 − 3.187 − 3.322 − 3.322 − 3.322
F20 Mean − 3.267 − 2.751 − 3.299 − 3.282 − 3.208 − 2.831 − 3.290 − 2.888 − 3.263
StD 0.06 0.39 0.05 0.07 0.08 0.33 0.05 0.62 0.06
Best − 10.153 − 5.69 − 10.153 − 10.153 − 10.153 − 5.055 − 10.153 − 10.153 − 10.153
F21 Mean − 10.153 − 3.221 − 9.157 − 9.983 − 10.149 − 5.045 − 10.153 − 10.153 − 7.184
StD 6.02 × ­10−15 1.278 2.58 0.93 0.01 0.02 6.68 × ­10−15 5.55 × ­10−4 2.68
Best − 10.403 − 6.306 − 10.403 − 10.403 − 10.403 − 5.088 − 10.403 − 10.403 − 10.403
F22 Mean − 10.403 − 2.917 − 9.541 − 10.403 − 10.399 − 5.068 − 10.403 − 10.224 − 6.073
StD 6.60 × 10−16 1.20 2.28 7.34 × ­10−15 0.01 0.05 9.33 × ­10−16 0.97 2.790
Best − 10.536 − 4.055 − 10.536 − 10.536 − 10.536 − 5.129 − 10.536 − 10.536 − 10.536
F23 Mean − 10.536 − 2.838 − 10.536 − 10.356 − 10.533 − 5.119 − 10.536 − 10.236 − 6.210
StD 1.32 × ­10−15 0.61 7.71 × ­10−6 0.99 0.01 0.01 9.33 × 10−16 1.166 2.81
Best 15 5 10 19 11 9 21 19 22
Results Mean 14 3 6 13 4 5 20 9 16
StD 10 3 2 7 3 2 14 7 10

Table 3.  standard test functions results. The best results are marked in bold.

fnc 01 fnc 02
0
10 0 10
Average Best So-far

Average Best So-far

10 -200 10 -200

100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

fnc 03 fnc 04
10 0 10 0
Average Best So-far

Average Best So-far

10 -200 10 -200

100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

fnc 06 fnc 07
10 0 10 0
Average Best So-far

Average Best So-far

10 -20

10 -40 10 -5
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

iterations iterations
HGS COOT MGO FO AO HHO GTO AVOA RTH

Figure 8.  Convergence curves of the unimodal functions.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 11

Vol.:(0123456789)
 www.nature.com/scientificreports/

fnc 08 fnc 09
-1000

Average Best So-far

Average Best So-far

-2000
100
-3000

-4000
0 200 400 600 800 1000 0 2 4 6 8 10

104 fnc 10 fnc 11

0
Average Best So-far

Average Best So-far

0
10
-0.5

-1

-1.5

-2
10-20
0 200 400 600 800 1000 0 200 400 600 800 1000

fnc 12 fnc 13
Average Best So-far

Average Best So-far

100 100

0 200 400 600 800 1000 0 200 400 600 800 1000
iterations iterations
HGS COOT MGO FO AO HHO GTO AVOA RTH
1 1

Figure 9.  Convergence curves of the multimodal functions.

For the basic functions (fnc 02–04), their results are presented in Fig. 17. For fnc 02 for both cases (D = 15
and D = 20), all the optimizers except for the F and the HHO got the optimal solution where the COOT provided
the lower variation range and the MGO include the set of the best result. The results of function 03 are similar
to those of function 02, with slight superiority to the HGS in precision and stability. For function 04, the RTH
provided identical results for both cases (D = 15 and D = 20). This confirms its ability to resolve this problem
effectively.
Hybrid functions (fnc 05–07) ANOVA results are presented in Fig. 18. The results of function 06 are not
included because of the exact similarity of the results. These figures confirm the outperformance of the proposed
RTH for the three functions compared to the other considered algorithms. The MGO and GTO algorithms
performed the second best after the RTH.
Composition functions (fnc 08–10) ANOVA results are presented in Fig. 19. For function 08; the MGO pro-
vided the best performance where the proposed RTH for the best result, but its variation range is higher due to
some ultimate results far from the best. The performance of the algorithms is much closer to each other except
for the FO and the HHO for function 09, with light superiority for the MGO. The RTH provides more stable
performance and higher accuracy for function 10, followed by the MGO.

IEEE CEC 2022. In this case, the proposed RTH will be compared with the considered algorithms using the 12
functions of the IEEE CEC ­202256. The characteristics of CEC 2022 functions are presented in Table A3. Their
3D maps are illustrated in Fig. 20 for the unimodal function (F1), Fig. 21 for the basic functions (F2-F5), and
Fig. 22 for the composition functions (F9-F12).
Similar to the CEC 2020 cases, all the algorithms are run 30 times for each function for D = 10 and D = 20.
The obtained results for these functions for D = 21 are presented in Table 6.
From the L1 row, the proposed RTH can find the optimal solution for eight from twelve functions, followed
by GTO by six functions, then the HGS by five functions, then AVOA by four. This can approve its ability to get
the optimal solution and escape from the local solutions. In comparison, FO, AO, and HHO cannot provide any
best solution. From the penultimate row, the MGO achieved the best mean results by six functions, followed by
the proposed RTH by five functions, then the HGS and GTO by two. This approves the accuracy of the proposed
RTH. FO, COOT, AO, HHO, and AVOA algorithms didn’t get any mean best results for all the tests. From the
last row, the proposed RTH provided the best results by five functions, followed by the proposed MGO by four
times. This can approve its robustness. Figure 23 presents the fitness evolution for each function of the CEC
2022 with D = 10. These figures support the findings of the CEC 2020, where the proposed RTH gives an excel-
lent convergence speed.
For D = 20 case, the results are provided in Table 7.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 12

Vol:.(1234567890)
www.nature.com/scientificreports/

Average Best So-far

Average Best So-far

fnc 14 fnc 15

102
100

100 10-5
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000
Max iterations Max iterations
Average Best So-far

Average Best So-far

fnc 16 fnc 17

2
1
0 100
-1
2 4 6 8 10 12 14 16 18 20 100 200 300 400 500 600 700 800 900 1000
Max iterations Max iterations
Average Best So-far

Average Best So-far

fnc 18 fnc 19
105
-2

-3

0 -4
10
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000
Max iterations Max iterations
Average Best So-far

Average Best So-far

fnc 20 fnc 21
0 0

-2 -5

-4 -10
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000
Max iterations Max iterations
Average Best So-far

Average Best So-far

fnc 22 fnc 23
0 0

-5 -5

-10 -10
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

Max iterations Max iterations

HGS COOT MGO FO AO HHO GTO AVOA RTH

Figure 10.  Convergence curves of the multimodal functions.

From the L1, the proposed RTH can find the optimal solution for six from twelve functions, followed by GTO
by only three functions, then the HGS by two functions. The increased complexity can explain the decreased
number of optimal solutions due to increased search space dimensions. These results approve the RTH’s ability
to get the optimal solution better and escape from the local solutions even in complicated problems. From L2,
the MGO achieved the best mean results by five functions, followed by the proposed RTH by four. This approves
the accuracy of the proposed RTH compared to the MGO algorithm. From L3, the proposed RTH provided the
best results by five functions, followed by the proposed MGO by three times. This can approve its robustness.
Compared to D = 10, the RTH can generally handle problems that include optimization variables better. Figure 24
presents the fitness evolution for each function of the CEC 2022 with D = 20. These figures approve the findings
of the CEC 2020 and CEC 2022 with D = 10, where the proposed RTH presents a faster convergence speed.
Figure 25 shows the ANOVA ranking results for unimodal functions for both cases. From this figure, the
RTH and the MGO generate the best solutions in terms of accuracy with the lowest variation range for D = 10.
However, when the search space dimensions increased, the performance of the MGO decreased where the RTH
performance was constant.
The basic functions (fnc 02–05) results are presented in Fig. 26. For fnc 02 for both cases (D = 15 and D = 20),
all the optimizers except for the FO and the HHO got the optimal solution where the MGO and RTH provided the
lower variation range for both cases. The results of fnc 03 show that the proposed RTH cannot resolve this prob-
lem effectively. The MGO and the HGS provide the best performance for this case. For fnc 04, all the optimizers
provided near performance with a slight advantage to the MGO. The same comments can be extracted for fnc 05.
Hybrid functions (fnc 06–08) ANOVA results are presented in Fig. 27. The results of the fnc 06 confirm the
outperformance of the proposed RTH. All the optimizers provide similar performance for the other two func-
tions with an advantage to the MGO, HGS, and RTH algorithms.
Composition functions (fnc 09–12) ANOVA results are presented in Fig. 28. For function 09, the MGO
and the RTH provided the best performance for both cases. For the other functions, the performance of the

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 13

Vol.:(0123456789)
 www.nature.com/scientificreports/

10 10

4.5

3.5

2.5

1.5

0.5

0
-100

-50

-80 -100
50 -40 -60
0 -20
40 20
100 80 60
100

Figure 11.  CEC 2020 F1 (unimodal) 3D map.

Figure 12.  CEC 2020 basic functions’ 3D map (F2–F4).

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 14

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 13.  CEC 2020 composition functions’ 3D map (F8–F10).

algorithms is much closer to each other, with light superiority for the FO and the HGS for fnc 10, the GTO and
the RTH for fnc 11, and the MGO and the COOT for fnc 12.
These mathematical test functions are used to elaborate the performance of the proposed RTH algorithm
compared to other new and robust algorithms. The statistical results are summarized in Fig. 29.
From the results obtained, the RTH algorithm can provide very competitive performance for solving vari-
ous optimization problems. The accuracy and stability of this proposed algorithm have been tested compared to
the prementioned algorithms. The convergence speed has been checked and confirmed. After mathematically
approving its performance, the proposed RTH algorithm will be tested with some published papers for several
engineering tests.

Engineering optimization problems

As mathematics and engineering real-world complex examples, the proposed RTH algorithm will be used to
solve seven real-world complex engineering problems.

(a) Optimal design of I-Shaped beam: this issue, which tries to reduce the vertical deflection of the beam,
is another common engineering optimization p ­ roblem34. It satisfies the cross-sectional area and stress
restrictions simultaneously under specific loads, as explained in Fig. 30. This optimization problem can be
expressed as follows:
5000
f (x) = 3
x3
 2
(x − 2x ) x x
 
x3 2 4
12 + 1 4 6 + 2bx4 x2 − 4 2
g1 (x) = 2x1 x3 + x3 (x2 − 2x4 ) ≤ 300
(9)
18x2 × 104 15x1 × 103
g2 (x) = 3
+ ≤ 56
(x2 − 2x4 )x32 + 2x3 x13
 2
x3 (x2 − 2x4 ) + 2x1 x3 4x4 + 3x2 (x2 − 2x4 )
10 ≤ x1 ≤ 50, 10 ≤ x2 ≤ 80, 0.9 ≤ x3 ≤ 5, 0.9 ≤ x4 ≤ 5
where the optimization variables are: the width of the flange (x1), the height of the section (x2), the thick-
ness of the web (x3), and the thickness of the flange (x4).

From Table 8, the proposed RTH obtains the best solutions for solving this optimization problem compared
to other published algorithms. In addition, the PFOA provided similar performance to the proposed RTH, but
the RTH achieved the stop criteria after 300 iterations, proving better convergence speed and accuracy.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 15

Vol.:(0123456789)
 www.nature.com/scientificreports/

MGO FO COOT HGS AO HHO GTO AVOA RTH

Unimodal function
F1 Best 154.539 8576.36 × ­106 41887 2181 4024949 4911.13 × ­106 104.345 102.058 100.592
Mean 4974.367 1610.29 × ­10 7
112335 1135972 11527453 1208.31 × ­107 5013.844 7186.27 4365.139
StD × ­103 4.779 2.867 × ­106 45.662 4.50 × ­103 8.05 × ­103 3.55 × ­106 5.93 8.32 4.00
Basic functions
F2 Best 1126.91 3620.76 1572.07 1412.69 2007.40 2996.31 1678.96 1925.81 1571.65
Mean 2020.35 3987.02 2334.50 1773.44 2667.60 3857.45 2423.14 2569.96 2401.16
StD 382.45 184.61 260.14 245.91 387.24 363.03 404.35 294.22 430.07
F3 Best 723.00 911.14 741.18 725.33 774.33 895.80 746.42 766.98 751.74
Mean 752.73 955.17 768.25 751.06 809.70 963.07 815.69 810.54 806.92
StD 19.811 19.78 14.66 17.71 21.97 31.88 34.12 32.44 33.36
F4 Best 1901.38 70470.76 1902.476 1901.19 1909.34 1958.76 1902.26 1902.13 1900.08
Mean 1903.72 294336.46 1906.02 1908.62 1918.19 91339.86 1908.18 1908.71 1903.69
StD 1 200755.19 1.98 7.30 9.88 82844.70 4.70 3.65 2.66
Hybrid functions
F5 Best 2492.57 755289 14050 13508.5 96570.2 1270090 1986.35 3909.53 2015.57
Mean 62543.12 2969214 196629 249879 1098805 11324500 2563.61 244561 2439.74
StD 57658 1323317 139143 325361 1047277 12488412 286.09 169180 249.82
F6 Best 2319.99 2319.99 2319.99 2319.99 2319.99 2319.99 2319.99 2319.99 2319.99
Mean 2319.99 2319.99 2319.99 2319.99 2319.99 2319.99 2319.99 2319.99 2319.99
StD × ­10−3 4.47 × ­10−8 3.96 × ­10−2 8.27 × ­10−6 1.39 × 10−9 6.48 × ­10−3 0.115 1.39 × 10−9 1.39 × 10−9 1.39 × 10−9
F7 Best 2269.802 164541.7 4580.516 5310.281 7155.679 852667.0 2136.214 3331.492 2161.322
Mean 38399.91 527133.1 66495.89 125533.1 191838.5 6232589 2581.630 94355.00 2597.401
StD 41656.35 241929.1 88216.10 119168.3 151067.9 3376108 208.5411 108669.8 239.2218
Composition functions
F8 Best 2300 3107.01 2302.31 2300.51 2308.68 2980.88 2300 2299.52 2300
Mean 2300.96 3811.25 2304.30 2578.74 2311.52 4326.10 2304.54 2388.05 2556.90
StD 0.653 381.30 1.44 445.75 0.81 963.94 6.87 338.93 678.38
F9 Best 2796.06 3059.19 2501.81 2500 2550.93 2858.15 2805.19 2801.05 2800
Mean 2808.72 3133.14 2819.69 2801.70 2846.57 3191.31 2839.22 2836.49 2831.28
StD 7.79 41.27 62.36 57.56 59.29 138.81 46.35 21.04 15.06
F10 Best 2900 3628.14 2901.42 2900 2909.24 3482.782 2900 2900 2900
Mean 2935.55 4107.85 3019.66 2973.60 3059.41 4035.529 2975.479 3021.413 2920.14
StD 84.64 316.40 97.67 100.56 84.6 309.486 102.096 102.84 61.44
L1 Best 3 1 1 2 1 1 2 2 5
L2 Mean 4 1 1 2 1 1 1 1 5
L3 StD 3 1 1 1 0 0 1 1 4

Table 4.  CEC 2020 benchmark test function results with D = 15. The best results are marked in bold.

(b) Three-bar truss design: In this example, a 3-bar planar truss is c­ onsidered34, as represented in Fig. 31. The
objective function and its constraints are presented as follows
 √ 
f (x) = 2 2x1 + x2 × l
√
2x1 + x2
g1 (x) = P √ 2 −σ ≤0
2x1 + 2x2 x1
x2
g2 (x) = P √ 2 −σ ≤0 (10)
2x1 + 2x2 x1
x2
g3 (x) = P √ −σ ≤0
2x2 + x1
l = 100cm, P = 2kN cm3 , σ = 2kN cm3 , 0 ≤ x1 , x2 ≤ 1
 

Table 9 provides the statistical results of these algorithms. It is clear that the best results are obtained using
the proposed RTH algorithm and the AHA optimizer. However, the required number of iterations is only 670,

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 16

Vol:.(1234567890)
www.nature.com/scientificreports/

fnc 01 fnc 02

Average Best So-far

Average Best So-far

1010 5000
4000

3000

2000
200 400 600 800 1000 200 400 600 800 1000
fnc 03 fnc 04
Average Best So-far

Average Best So-far

1600
1400
1200
105
1000

800
200 400 600 800 1000 200 400 600 800 1000
fnc 05 fnc 06
Average Best So-far

Average Best So-far

2330

105 2325

2320
200 400 600 800 1000 200 400 600 800 1000
fnc 07 fnc 08
Average Best So-far

Average Best So-far

6000
5000
4000
105
3000

200 400 600 800 1000 200 400 600 800 1000
fnc 09 fnc 10
Average Best So-far

Average Best So-far

3800
3600 8000
3400 6000
3200
3000 4000

200 400 600 800 1000 200 400 600 800 1000
Iterations Iterations
HGS COOT MGO FO AO HHO GTO AVOA RTH

Figure 14.  CEC 2020 fitness evolution for D = 15.

much lower than the other reported algorithms, including the AHA, which requires 15,000 iterations to achieve
the optimal value.

(c) Design of a tubular column: this problem illustrates the structure of a uniform column with a tubular
cross-section that can support a compressive load at the lowest possible c­ ost34, as explained in Fig. 32.
The two design variables in this issue are the mean diameter of the column (x1) and the thickness of the
tube (x2). The column is constructed from a material having a yield stress of y and an elasticity modulus
of E. The following equation represents this optimization problem:

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 17

Vol.:(0123456789)
 www.nature.com/scientificreports/

MGO FO COOT HGS AO HHO GTO AVOA RTH

Unimodal function
F1 Best 103.05 16806140527 224340.05 130.33 7001885.6 13513137381 100.25 100.49 100.20
Mean 2840.68 29506906789 637140.05 10976.98 39584617.2 21692845681 2554.76 3266.99 1715.68
StD × ­103 3312.12 4605945120 311729.08 16272.32 21104674.4 4650504699 2610.12 3863.87 1565.21
Basic functions
F2 Best 1472.78 4746.85 2112.23 1455.18 2653.59 4630.30 2276.34 1926.65 1966.20
Mean 2333.59 5471.61 3206.10 1905.89 3385.10 5383.45 3238.12 2838.21 3156.77
StD 438.85 259.64 452.76 220.67 474.10 357.14 463.24 482.15 596.00
F3 Best 743.17 982.24 776.23 735.22 803.50 931.26 798.78 785.60 796.21
Mean 783.98 1061.86 815.46 762.18 855.62 1028.41 859.93 836.36 844.61
StD 20.88 36.71 28.20 17.65 27.78 31.34 41.12 30.18 30.29
F4 Best 1902.61 98274.78 1905.98 1901.75 1915.55 9210.16 1905.96 1906.25 1900.93
Mean 1910 358004.03 1910.40 1903.68 1932.14 227855.65 1918.30 1914.45 1909.82
StD 9.30 246927.96 2.57 1.20 12.14 190896.04 9.38 6.26 4.61
Hybrid functions
F5 Best 13793.98 963554.04 71241.37 35642.41 180500.1 258875.09 4762.99 25666.73 2793.50
Mean 103507.19 3221412.17 384815.44 769776.7 640537.1 3585480.15 32023.92 480253.8 6392.68
StD 60422.30 1429844.08 239613.83 527020.1 339804.200 2798617.66 34478.79 413365.6 3120.73
F6 Best 1604.05 1604.05 1604.05 1604.05 1604.05 1604.05 1604.05 1604.05 1604.05
Mean 1604.05 1604.05 1604.05 1604.05 1604.05 1604.05 1604.05 1604.05 1604.05
StD × ­10−12 1.16 1.16 1.16 1.16 1.16 1.16 1.16 1.16 1.16
F7 Best 3544.55 271391.75 9867.50 34895.32 42748.08 318315.91 2621.89 19735.31 2443.60
Mean 65648.76 763153.83 144945.05 284910.30 530727.72 3573597.72 4921.82 226243.15 3648.84
StD 57276.90 423466.26 92024.16 253278.10 450075.13 3473899.96 1940.52 189687.04 947.33
Composition functions
F8 Best 2300 4347.34 2306.97 2300 2314.95 3880.60 2300 2300 2300
Mean 2300.70 5396.11 2309.54 3209.51 2321.03 6004.87 2418.29 2768.84 2671.51
StD 0.63 576.35 1.13 1247.79 4.81 892.123 639.17 1107.05 1004.51
F9 Best 2821.03 3125.32 2821.07 2850.66 2863.87 3081.44 2837.50 2860.88 2861.11
Mean 2853.78 3281.45 2883.40 2917.45 2927.31 3341.05 2905.20 2948.24 2918.42
StD 26.25 59.29 37.24 34.10 44.59 149.48 49.74 49.63 40.15
F10 Best 2913.93 4160.30 2911.42 2909.25 2941.17 3605.15 2913.10 2910.51 2901.27
Mean 2973.47 4890.83 2979.09 2931.18 3010.99 4473.18 2969.76 2982.71 2959.72
StD 33.02 539.25 31.04 28.29 23.33 643.61 35.28 34.40 31.12
L1 Best 3 1 1 2 1 1 2 2 6
L2 Mean 3 1 1 2 1 1 1 1 5
L3 StD 2 1 1 2 1 1 1 1 5

Table 5.  CEC 2020 benchmark test function results with D = 15. The best results are marked in bold.

f (x) = 9.8x1 x2 + 2x2

P
g1 (x) = −1≤0
πx2 x1 σy
8PL2
g2 (x) =  −1≤0
π 3 Ex2 x1 x12 + x22

(11)
2 x1
g3 (x) = − 1 ≤ 0, g4 (x) = − 1 ≤ 0,
x1 14
2 x2
g5 (x) = − 1 ≤ 0, g6 (x) = −1≤0
x2 8
σy = 500kgf cm2 , P = 0.85 × 106 kgf cm3 ,
 
2 ≤ x1 ≤ 14, 0.2 ≤ x2 ≤ 0.8

This problem has already been tackled using several algorithms. The best results of these algorithms, as well
as the suggested RTH, are shown in Table 10. Based on these results, the proposed RTH achieved the best results
with the lowest number of iterations (300).

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 18

Vol:.(1234567890)
www.nature.com/scientificreports/

fnc 01 fnc 02

Average Best So-far

Average Best So-far

1010 6000

4000

2000
200 400 600 800 1000 200 400 600 800 1000
fnc 03 fnc 04
Average Best So-far

Average Best So-far

2000

1500
105
1000

200 400 600 800 1000 200 400 600 800 1000
fnc 05 fnc 06
Average Best So-far

Average Best So-far

104

105

102
200 400 600 800 1000 200 400 600 800 1000
fnc 07 fnc 08
Average Best So-far

Average Best So-far 8000

6000

4000
105

200 400 600 800 1000 200 400 600 800 1000
fnc 09 fnc 10
Average Best So-far

Average Best So-far

3800 12000
3600 10000
8000
3400
6000
3200
3000 4000

200 400 600 800 1000 200 400 600 800 1000
Iterations Iterations
HGS COOT MGO FO AO HHO GTO AVOA RTH

Figure 15.  CEC 2020 fitness evolution for D = 20.

(d) Speed Reducer Design: speed reducer is an essential part of the gearbox in mechanical systems. It may be
used for a variety of ­purposes71. As shown in Fig. 33, the weight of the speed reducer must be reduced
under 11 restrictions in this optimization problem. Face width (x1), the module of teeth (x2), number of
teeth in the pinion (x3), length of the first shaft between bearings (x4), length of the second shaft between
bearings (x5), the diameter of first shafts (x6), and diameter of the second shafts (x7) are the seven variables
in this issue. This problem can be modeled as

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 19

Vol.:(0123456789)
 www.nature.com/scientificreports/

Figure 16.  CEC 2020 unimodal function (F1) ANOVA results.

Figure 17.  CEC 2020 basic functions ANOVA results.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 20

Vol:.(1234567890)
www.nature.com/scientificreports/

107 D = 15 106 D = 20
6 14
105 104 2700
2.5 3400 3000 12

5 3200
12
2600
2800 10
2
3000
10 2500
4 2800 8
1.5 2600
2400
8
fnc 05

2600
6
3 1 2400 2300
2400
6 4
2200 2200
2 0.5 2200
2000 4 2 2100

1 0 1800 2000
2 0 2000

0 0
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

106 106
14 14
3200 2600 12000 2400
12 12
3000 2500
10000
10 10 2300
2800 2400
8000
8 8
fnc 07

2600 2300 2200

6000
6 2200
6
2400
2100
4 2200 2100 4 4000

2 2000 2000 2 2000 2000

0 0
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

Figure 18.  CEC 2020 hybrid functions ANOVA results.

D = 15 D = 20
6000 2302.5 2308 2313 2335 8000 2312 2335
2301.5
2330
2307
2302 2312 2311
2330
5000 2325
2306 6000 2301
2310
2301.5 2311 2320
fnc 08

2305 2325
4000 2301 2310
2315 2309
2304 2300.5
2310 4000 2320
2308
2300.5 2309
3000 2303 2305

2302 2300 2300 2307 2315

2300 2308
2000
2000
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

3800
3400
3600

3200 3400
fnc 09

3200
3000
3000
2800
2800
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

5000 5000

4500 4500

4000 4000
fnc 10

3500 3500

3000 3000

2500 2500
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

Figure 19.  CEC 2020 composite functions ANOVA results.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 21

Vol.:(0123456789)
 www.nature.com/scientificreports/

10 8

10

2
100

50
0
100 0
80 60 40 20 0 -50
-20 -40 -60 -80 -100 -100

Figure 20.  CEC 2022 F1 (unimodal) 3D map.

Figure 21.  CEC 2022 basic functions’ 3D map (F2–F5).

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 22

Vol:.(1234567890)
www.nature.com/scientificreports/

Figure 22.  CEC 2022 composition functions’ 3D map (F9–F12).

f (x) = 0.7854x1 x22 3.3333x32 + 14.9334x3 − 43.0934

 

− 1.508x1 x62 + x72 + 7.4777 x63 + x73 + 0.7854 x4 x62 + x5 x72

     

27 397.5
g1 (x) = − 1 ≤ 0, g2 (x) = −1≤0
x1 x3 x22 x1 x3 x22
1.93 1.93x53
g3 (x) = 4
− 1 ≤ 0, g4 (x) = −1≤0
x2 x3 x6 x2 x3 x74
 2
745x4

6
x2 x3 + 16.9 × 10
g5 (x) = −1≤0
110x63 (12)
 2
745x5

6
x2 x3 + 157.5 × 10
g6 (x) = −1≤0
85x73
x2 x3 5x2 x1
g7 (x) = − 1 ≤ 0, g8 (x) = − 1 ≤ 0, g9 (x) = −1≤0
40 x1 12x2
1.5x6 + 1.9 1.1x7 + 1.9
g10 (x) = − 1 ≤ 0, g9 (x) = −1≤0
x4 x5
2.6 ≤ x1 ≤ 3.6, 0.7 ≤ x2 ≤ 0.8, x3 ∈ [17, 18, · · · 28], x4 ≥ 7.3
x5 ≤ 8.3, 2.9 ≤ x6 ≤ 3.9, 5 ≤ x7 ≤ 5.5

The results of RTH and other published algorithms are reported in Table 11. Among these algorithms, the
RTH has the lowest number of iterations, while its metrics in terms of best, mean, and StD are better than those
other algorithms.

(e) Piston lever: the essential goal of this problem is to identify the piston components (x1), (x2), (x3), and (x4)
by reducing the oil volume when the piston’s lever is raised from 0° to 45°34. This problem is presented in
Fig. 34, and it can be modeled as follows:

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 23

Vol.:(0123456789)
 www.nature.com/scientificreports/

MGO FO COOT HGS AO HHO GTO AVOA RTH

Unimodal function
Best 300 7075.29 300.23 300 439.34 5774.01 300 300 300
F1 Mean 300 12242.75 302.53 321.05 2127.41 9833.45 300 302.72 300
StD 4.09 × ­10−12 3256.05 2.71 115.26 1730.22 1056.31 1.93E-10 8.05 3.95 × 10−14
Basic functions
Best 400.001 519.53 400.002 400.39 400.189 502.241 400.002 400.013 400
F2 Mean 405.32 765.59 419.209 421.58 421.727 850.32 408.968 415.57 404.022
StD 3.77 153.57 28.72 29.89 36.835 321.15 17.82 25.613 3.25
Best 600 633.70 600.40 600 605.12 626.52 601.18 602.62 600.82
F3 Mean 600.17 643.93 603.26 600.67 618.32 648.29 607.01 617.85 611.97
StD 0.25 5.38 2.44 0.60 6.92 11.60 5.17 10.68 9.77
Best 802.99 839.46 807.34 813.93 810.07 831.72 808.96 808.96 810.95
F4 Mean 813.73 856.94 818.33 838.07 823.00 844.06 827.20 833.13 823.68
StD 7.25 6.92 5.92 14.68 8.21 9.92 7.41 11.01 8.90
Best 900 1170.61 900.01 900.91 903.56 1142.71 902.40 955.37 902.36
F5 Mean 904.88 1375.01 902.57 1141.03 993.59 1537.75 956.74 1186.98 1053.05
StD 9.07 124.92 4.16 252.20 81.42 205.56 66.27 196.57 174.49
Hybrid functions
Best 1846.46 430643.97 1917.56 1835.65 3197.23 3007.22 1805.15 1877.29 1802.74
F6 Mean 2122.37 8578353.07 3444.352 6035.25 10648.95 2588374.51 2008.43 3592.39 1840.81
StD 344.20 7161565.78 1736.32 2355.11 5945.94 8077679.39 707.50 1969.34 57.95
Best 2001 2066.31 2009.90 2000.01 2021.64 2039.65 2006.52 2020.59 2012.93
F7 Mean 2020.96 2094 2029.99 2016.77 2048.19 2106.54 2029.45 2038.85 2038.61
StD 6.29 11.28 9.91 7.44 24.12 30.63 11.62 20.07 25.97
Best 2214.70 2233.23 2208.56 2219.43 2208.16 2228.73 2206.81 2221.88 2211.59
F8 Mean 2226.69 2246.53 2224.47 2222.47 2227.78 2249.43 2221.81 2225.57 2221.34
StD 21.87 6.53 4.02 3.47 6.11 24.51 3.39 3.35 2.31
Composition functions
Best 2529.28 2635.33 2529.28 2529.28 2530.96 2634.54 2529.28 2529.28 2529.28
F9 Mean 2529.28 2680.84 2529.29 2531.32 2587.69 2730.09 2529.45 2534.18 2529.28
StD 0 21.64 0.03 11.14 32.55 45.59 0.59 26.83 0
Best 2500.22 2509.88 2500.33 2500.41 2500.71 2502.28 2500.31 2500.37 2500.10
F10 Mean 2531.19 2528.46 2546.63 2557.65 2588.60 2812.47 2527.34 2564.39 2547.56
StD 52.07 10.33 57.59 61.64 54.02 455.78 49.60 65.20 62.76
Best 2600 2822.82 2600.39 2600 2604.60 2787.33 2600 2600 2600
F11 Mean 2721.97 2946.83 2731.62 2805.30 2690.65 3284.55 2638.41 2719.58 2739.04
StD 123.03 81.79 183.31 170.69 93.38 406.47 88.85 138.41 147.85
Best 2861.41 2884.28 2859.00 2861.41 2862.75 2873.20 2861.44 2861.41 2860.18
F12 Mean 2864.73 2905.10 2864.85 2864.07 2867.99 2985.81 2864.64 2866.86 2866.71
StD 1.49 13.50 2.33 1.43 3.45 78.94 2.13 5.65 5.74
L1 6 0 1 5 0 0 3 4 8
L2 6 0 0 2 0 0 2 0 5
L3 4 2 0 1 0 0 0 0 5

Table 6.  CEC 2022 benchmark test function results for D = 10. The best results are marked in bold.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 24

Vol:.(1234567890)
www.nature.com/scientificreports/

fnc 01 fnc 02

Average Best

Average Best
105
103

100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000
fnc 03 fnc 04
700 920
Average Best

Average Best
680 900
660 880
640 860
620 840
820
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

fnc 05 fnc 06
1010
Average Best

Average Best
4000

2000
105

100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

fnc 07 fnc 08
2300 7000
Average Best

Average Best
6000
2200 5000
4000
2100 3000

100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

fnc 09 fnc 10
3200
Average Best

Average Best

4000
3000
3500
2800
3000
2600
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

fnc 11 fnc 12
Average Best

Average Best

5000
3200
4000
3000
3000

100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

Iterations Iterations
HGS COOT MGO FO AO HHO GTO AVOA RTH

Figure 23.  CEC 2022 fitness evolution for D = 10.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 25

Vol.:(0123456789)
 www.nature.com/scientificreports/

MGO FO COOT HGS AO HHO GTO AVOA RTH

Unimodal function
Best 300.02 30,224.81 2435.79 477.80 37,670.33 32,487.71 300.16 3008.42 300
F1 Mean 300.08 59,629.62 4778.73 2836.77 56,386.78 56,267.25 307.43 8776.13 300
StD 0.145 9986.83 1631.82 2289.21 13,929.99 20,389.83 16.521 4062.34 9.5 × 10−12
Basic functions
Best 444.90 1504.50 449.28 415.00 463.19 912.12 414.82 428.68 400
F2 Mean 454.95 2200.73 465.98 462.83 513.93 1750.55 458.46 464.59 439.81
StD 10.311 435.219 24.26 30.513 44.726 529.31 19.660 15.015 18.155
Best 600.24 662.56 607.77 600.14 620.86 653.72 606.71 623.472 612.84
F3 Mean 605.31 684.74 629.01 602.34 639.25 684.30 629.32 645.34 634.42
StD 4.190 7.393 9.322 4.048 7.358 12.220 10.058 9.661 11.915
Best 821.89 968.55 835.56 849.58 829.52 922.31 853.73 845.77 842.78
F4 Mean 850.46 999.82 865.50 900.93 871.90 954.79 882.85 890.06 875.42
StD 16.05 17.376 15.214 29.79 16.95 20.278 17.669 23.452 22.258
Best 930.94 3277.01 992.25 1238.17 1666.064 2849.926 1249.81 1788.12 1206.33
F5 Mean 1169.31 4948.35 1513.81 2462.87 2371.00 3673.03 1848.52 2378.20 1874.44
StD 194.436 859.562 376.653 636.69 375.57 471.658 343.371 304.611 324.049
Hybrid functions
Best 1882.54 90,365,085.95 2223.48 1947.71 25,861.92 57,613,430.85 1862.02 2022.830 1800.30
F6 Mean 4612.37 389,128,839.2 5328.58 13,826.44 155,894.16 867,615,884.2 5794.91 7461.50 4610.99
StD 3201.410 190,973,416.5 3227.276 9392.515 117,944.76 838,420,299.2 4755.08 5751.65 2683.249
Best 2027.05 2192.06 2052.78 2023.81 2070.94 2172.70 2035.89 2027.46 2069.44
F7 Mean 2060.66 2239.15 2110.38 2085.49 2133.50 2239.80 2117.42 2156.65 2142.50
StD 26.909 26.480 38.913 52.669 72.288 46.489 40.241 66.119 55.245
Best 2220.77 2286.92 2228.55 2221.19 2229.99 2238.83 2221.56 2228.03 2210.52
F8 Mean 2224.82 2418.70 2267.21 2253.21 2247.24 2422.96 2243.35 2238.87 2246.50
StD 6.695 70.141 54.040 47.844 31.615 148.88 43.417 9.448 26.753
Composition functions
Best 2480.782 2720.715 2480.884 2480.782 2494.789 2639.303 2480.781 2480.782 2480.781
F9 Mean 2480.797 2913.051 2485.666 2486.190 2565.406 3000.960 2480.814 2483.454 2480.781
StD 0.013 90.519 5.330 5.828 43.243 286.987 0.048 3.432 3.63 × 10 −12
Best 2500.472 2610.636 2500.668 2414.572 2500.670 5714.266 2500.754 2500.760 2500.745
F10 Mean 2878.745 2820.891 3040.516 2806.015 3065.910 6675.800 3726.354 3359.881 3745.876
StD 639.116 160.398 855.475 165.485 910.325 514.280 1165.580 767.330 726.295
Best 2900 5193.046 2619.646 2600.055 2876.708 5654.721 2900 2600.002 2900
F11 Mean 2930 6317.223 2941.369 2990.848 3186.476 7419.084 2965.112 2910.003 2916.667
StD 46.609 539.848 167.852 161.029 185.885 816.256 88.733 95.952 37.905
Best 2939.506 3158.671 2943.284 2940.273 2983.960 3213.355 2946.685 2940.228 2938.921
F12 Mean 2984.007 3266.632 2973.659 2967.016 3033.970 3559.546 3026.821 2981.706 2972.224
StD 57.231 55.531 24.066 29.855 40.245 254.008 59.346 47.495 30.449
L1 3 0 0 2 0 0 0 1 6
L2 5 0 0 1 0 0 0 1 4
L3 3 2 1 1 0 0 0 0 5

Table 7.  CEC 2022 benchmark test function results for D = 10. The best results are marked in bold.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 26

Vol:.(1234567890)
www.nature.com/scientificreports/

fnc 01 fnc 02
1010 104

Average Best

Average Best
105

100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000
fnc 03 fnc 04
Average Best

Average Best
720 1100
700
680 1000
660
640
620 900

100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000
fnc 05 fnc 06
1010
Average Best

Average Best
104

103
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000
fnc 07 fnc 08
5
2600 10
Average Best

Average Best
2400
104
2200
103
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000
fnc 09 fnc 10
4500
Average Best

Average Best
8000
4000
6000
3500
3000 4000
2500
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000
fnc 11 fnc 12
15000
Average Best

Average Best

10000 4000

5000
3000
100 200 300 400 500 600 700 800 900 1000 100 200 300 400 500 600 700 800 900 1000

Iterations Iterations
HGS COOT MGO FO AO HHO GTO AVOA RTH

Figure 24.  CEC 2022 fitness evolution for D = 10.

1 2
f (x) = πx (L2 − L1 )
4 3
g1 (x) = QL cos(θ) − RF ≤ 0
g2 (x) = Q(L − x4 ) − Mmax ≤ 0
g3 (x) = 1.2(L2 − L1 ) − L1 ≤ 0
x3
g4 (x) = − x2 ≤ 0
2 (13)
−x4 (x4 sin(θ) + x1 ) + x1 (x2 − x4 cos(θ))
R=
L1
 
2
F = πPx3 4, L1 = (x4 − x2 )2 + x12 , L2 = (x4 sin(θ) + x1 )2 + (x2 − x4 cos(θ))2


θ = 45◦ , Q = 10000lbs, L = 240in, 1.8 × 106 lbs.in, P = 1500psi

0.05 ≤ x1 ≤ 500, 0.05 ≤ x2 ≤ 120

According to the obtained results in Table 12, the SNS performs better than the proposed RTH algorithm
in terms of mean and StD. However, achieving the stop criteria took many iterations (5000), whereas the RTH
requires only 130 iterations.f) Corrugated design: This task seeks to reduce the weight of a corrugated bulkhead
in a chemical ­tanker66. The optimization variables are width (x1), depth (x2), length (x3), and plate thickness (x4).
Its mathematical model of this problem can be expressed as follows:

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 27

Vol.:(0123456789)
 www.nature.com/scientificreports/

104 D = 10
2.5

301 312 1000 301 335 301

2 900 330
310
300.5 300.5 325 300.5
800
308
1.5 320
700
300 306 300 315 300
600
1 310
304
500
299.5 299.5 305 299.5
302 400
0.5 300

299 300 300 299 295 299

0
MGO FO COOT HGS AO HHO GTO AVOA RTH

104 D = 20
12
300.8 400 301
10 300.7
380
300.6 300.5
8
300.5
360

6 300.4 300

300.3 340

4
300.2 299.5
320
300.1
2
300 300 299

0
MGO FO COOT HGS AO HHO GTO AVOA RTH

Figure 25.  CEC 2022 unimodal function (F1) ANOVA results.

D = 10 D = 20
2000 3000 480 460

470
1500 440
fnc 02

2000 460
1000 420
450
1000
500 440 400

MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

680
700
660
fnc 03

640 650

620

600 600
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

880
1000
860
950
fnc 04

840
900
820 850

800 800
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

2000 8000
940 920

930
6000
915
fnc 05

1500 920 910 4000

910 905
2000
1000 900 900
0
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

Figure 26.  CEC 2022 basic functions ANOVA results.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 28

Vol:.(1234567890)
www.nature.com/scientificreports/

107 D = 10 109 D = 20
4 104
4 3500 8000 8000 2200 15000 8000
3 2200
2100 3
3 3000 6000 6000 10000 6000 2100

fnc 06
2
2000 2
2 2000
2500 4000 4000 4000
5000 1
1900 1900
1 1
2000 2000 2000
2000 1800 0 0 1800
0 0
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

2200 2600

2150
2400
fnc 07

2100
2200
2050

2000 2000
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

2350 2800

2300 2600
fnc 08

2250 2400

2200 2200
MGO FO COOT HGS AO HHO GTO AVOA RTH MGO FO COOT HGS AO HHO GTO AVOA RTH

Figure 27.  CEC 2022 hybrid functions ANOVA results.

Figure 28.  CEC 2022 composite functions ANOVA results.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 29

Vol.:(0123456789)
 www.nature.com/scientificreports/

25

20

15

10

0
Best Mean StD Best Mean StD Best Mean StD Best Mean StD Best Mean StD
Standard CEC 2020 D = 15 CEC 2020 D = 20 CEC 2022 = 10 CEC 2022 = 20

MGO FO COOT HGS AO HHO GTO AVOA RTH

Figure 29.  Classes one test results.

Figure 30.  Cantilever beam schematic illustration.

MA Worst (× ­10−2) Mean (× ­10−2) Best (× ­10−2) StD(× ­10−5) Elapsed iterations

58
SOS NA 1.30884 1.30741 4.0 5000
CS59 1.35365 1.32165 1.30747 13.45 5000
AOS60 1.38140 1.31788 1.30741 15.55 100,000
SNS34 1.30764 1.30743 1.30741 4.31 × ­10−02 3600
PFOA61 1.307412 1.307412 1.307412 2.05 × ­10−4 24,000
AHA62 1.343036 1.340146 1.339957 7.91 × ­10−5 15,000
FA63 1.339960 1.339957 1.339956 2.51 × ­10−7 NA
RTH 1.307410 1.307410 1.307400 1.00 × ­10−8 300

Table 8.  Comparison of RTH results with the other algorithms for the I-shaped beam problem.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 30

Vol:.(1234567890)
www.nature.com/scientificreports/

H H

H x2

P P
Figure 31.  Three-bar truss schematic illustration.

MA Worst Mean Best StD Elapsed iterations

CS59 NA 264.066900 263.9715600 9.00 × ­10−05 15,000
GWO64 263.904218 263.897955 263.896006 1.61 × ­10−03 50,000
WCA​65 263.896201 263.895903 263.895843 8.71 × ­10−05 5250
PSO66 264.584903 263.957414 263.895843 1.37 × ­10−01 50,000
CGO67 263.896007 263.895851 263.895844 2.51 × ­10−05 100,000
SNS34 263.895856 263.895846 263.895843 3.31 × ­10−6 4800
PFOA61 263.895844 263.895843 263.895842 2.01 × ­10−6 18,000
RL-BA68 263.924700 263.900300 263.895840 6.06 × ­10−6 NA
AHA62 263.895843 263.895843 263.895843 1.09 × ­10−7 15,000
RTH 263.895843 263.895843 263.895843 5.78 × 10−14 670

Table 9.  Comparison of RTH results with the other used algorithms for the three-bar truss design problem.
The best results are marked in bold.

x1

Figure 32.  Tubular column schematic illustration.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 31

Vol.:(0123456789)
 www.nature.com/scientificreports/

MA Worst Mean Best StD Elapsed iterations

ISA69 26.532 26.531 26.531 1.70 × ­10−4 3000
CS59 26.53972 26.53504 26.53217 1.93 × ­10−3 15,000
AOS60 26.60831361 26.53161399 26.53137828 1.03 × ­10−3 100,000
SNS 34
26.48637095 26.48636249 26.48636147 2.22 × ­10−6 1250
PFOA61 26.48636148 26.48636150 26.48636152 2.00 × ­10−8 24,000
AOA70 27.16391 26.80510 26.53730 0.02761 3000
AOA-NM70 26.53169 26.53133 26.53132 5.36 × ­10−6 3000
RTH 26.48636147 26.48636147 26.48636147 7.23 × 10−15 300

Table 10.  RTH results and the other used algorithms for tubular column designing problem. The best results
are marked in bold.

x7

x5 x1 x2
x4
x3

x6

Figure 33.  Speed reducer design schematic illustration.

MA Worst Mean Best StD Elapsed iterations

CS59 3009.00000 3007.19970 3000.9810 4.96 250,000
ABC72 NA 2997.05841 2997.05841 0.20 30,000
WCA​65 2994.50558 2994.47439 2994.47107 7.40 × ­10−3 15,150
APSO73 4443.01764 3822.64062 3187.63047 3.66 × ­10−2 30,000
CGO67 2995.50493 2994.46540 2994.44365 0.11 100,000
SNS34 2994.47110 2994.47101 2994.47107 7.00 × ­10−6 3750
62
AHA 2994.47116 2994.47165 2994.47323 4.25 × ­10−4 30,000
ARSCA74 NA NA 2995.821 NA NA
RTH 2994.42400 2994.42400 2994.42400 4.63 × ­10−13 212

Table 11.  Comparison of RTH results with other algorithms for speed reducer designing problems.

5.885x4 (x1 + x3 )
f (x) = 
x1 + x32 − x22 

 
x3 
 
+ 8.94 x1 + x32 − x22  ≤ 0

g1 (x) = −x4 x2 0.4x1 +
6
 4
x3   3
 
g2 (x) = −x4 x22 0.2x1 + + 2.2 8.94(x1 + x32 − x22 ) ≤0 (14)
12
g3 (x) = −x4 + 0.0156x1 + 0.015 ≤ 0
g4 (x) = −x4 + 0.0156x3 + 0.015 ≤ 0
g5 (x) = −x4 + 0.015 ≤ 0
g6 (x) = x2 − x3 ≤ 0
0 ≤ x1 , x2 ≤ 100, x3 ≤ 100, 4 ≤ x4 ≤ 5,
The statistical results of this problem are provided in Table 13. Although AOA achieved the best results, these
results were only obtained after 40,000 iterations, unlike the proposed RTH for the stop value after 129 iterations,

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 32

Vol:.(1234567890)
www.nature.com/scientificreports/

x1

x2
Figure 34.  Piston lever design schematic illustration.

MA Worst Mean Best StD Elapsed iterations

CS59 168.5920 40.2319 8.4271 59.06 50,000
PSO66 294 166 122 51.7 50,000
ISA69 610.6 226.5 8.4 111.2 12,500
AOS60 167.6650 33.7413 8.4191 93.47 100,000
CGO67 167.4728 45.0487 8.4128 67.25 100,000
SNS34 167.4728 24.3190 8.4127 47.72 5000
SSA75 653.4973 276.9405 8.4220 121.42 25,000
MFO75 167.4727 91.1239 8.4126 80.27 25,000
MVO75 356.2368 138.4470 8.4289 138.51 25,000
EO75 167.4727 100.6675 8.4127 79.30 25,000
RTH 167.4727 30.8287 8.4127 71.54 130

Table 12.  Comparison of RTH results with the other used algorithms for the piston lever problem. The best
results are marked in bold.

MA Worst Mean Best StD Elapsed iterations

FA76 NA 10.23 7.21 1.95 12,000
LF-FA76 NA 8.83 6.95 1.26 12,000
LS-LF-FA76 NA 7.44 6.86 0.67 12,000
AD-IFA76 NA 7.21 6.84 0.58 12,000
AOS60 7.06694 7.06081 6.84296 6.49 × ­10−04 100,000
SNS34 6.84307 6.84298 6.84296 2.09 × ­10−05 3125
DMO 77
5.9617 5.031 5.002 0.996 40,000
RTH 6.84296 6.84296 6.84295 4.51 × 10−15 129

Table 13.  Comparison of RTH results with the other used algorithms for the corrugated bulkhead design
problem. The best results are marked in bold.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 33

Vol.:(0123456789)
 www.nature.com/scientificreports/

x1

Figure 35.  Speed reducer design schematic illustration.

MA Worst Mean Best StD Elapsed iterations

PSO66 0.071802 0.019555 0.012857 1.17 × ­10−2 20,000
CPSO79 0.012924 0.012730 0.012675 5.20 × ­10−5 200,000
HPSO80 0.012719 0.012707 0.012665 1.58 × ­10−5 81,000
WCA​65 0.012952 0.012746 0.012665 8.06 × ­10−5 11,750
MCEO81 0.013509 0.012720 0.012661 3.79 × ­10−5 2000
EO82 0.013997 0.013017 0.012666 3.91 × ­10−4 15,000
SNS34 0.012766 0.012685 0.012665 2.39 × 10−5 9000
SCMWOA83 NA 0.013400 0.012670 2.40 × ­10−4 2460
RL-BA68 0.012928 0.012745 0.012676 7.19 × ­10−4 NA
DDAO84 0.017320 0.0151829 0.012907 1.26 × ­10−3 NA
RTH 0.013185 0.012776 0.012665 1.41 × ­10−4 300

Table 14.  Comparison of RTH results with the other used algorithms for tension/compression spring design
problem. The best results are marked in bold.

which approves its solving speed.g) Design of tension/compression spring: tension/compression spring design
challenge aims to decrease the weight of a tension/compression spring, as detailed ­in78 and Fig. 35. Minimum
deflection, shear stress, surge frequency, outside diameter limitations, and design factors all play a role in this
problem. The optimization variables include the mean coil diameter (x1), the wire diameter (x2), and the number
of active coils (x3). This problem can be expressed as follows:

f (x) = (x3 + 2)x2 x12

x3 x23
g1 (x) = 1 − ≤0
71785x14
x x3 1
g2 (x) =  3 23 + −1≤0 (15)
12566 x2 x1 − x14 5108x12
140.45x1 x1 + x2
g3 (x) = 1 − ≤ 0, g4 (x) = ≤0
x3 x22 1.5
0.05 ≤ x1 ≤ 2, 0.05 ≤ x1 ≤ 2, 0.05 ≤ x1 ≤ 2,

Table 14 compares the RTH statistical results with other MAs. The RTH algorithm solves this problem in only
130 iterations, much less than the other considered algorithms with better results.

Proton Exchange membrane fuel cell parameters’ identification

Fuel cells are recent energy-generation devices that produce efficient power by directly using fuel (mainly hydro-
gen) and oxygen while emitting no pollutants or noise (water, heat, and electricity). Because of its benefits and
wide range of applications, proton exchange membrane fuel cells (PEMFCs) that employ polymers as elec-
trolytes are becoming more p ­ rominent85,86. Fuel cell modeling is a crucial stage that enables the investigation
and improvement of its p ­ erformance87. The PEMFC model is built from nonlinear differential equations that
explain internal chemical processes. These equations contain various empirical parameters that must be accu-
rately identified to improve model accuracy. As a result, identifying these parameters is essential for creating
an accurate model.

PEMFC model. The PEMFC output voltage can be provided as ­follows88

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 34

Vol:.(1234567890)
www.nature.com/scientificreports/

Vfc = ENernest − Vact − Vohm − Vcon (16)

where Vact is the activation voltage losses, Vohm is the ohmic voltage losses, Vcon represents the concentration
voltage losses, and ENernest denotes the thermodynamic potential voltage (the Nernst voltage). ENernest can be
calculated as follows:
 
ln(PO2 )
ENernest = 1.229 − 0.85 × 10−3 (T − 298.15) + 4.3085 × 10−5 × T ln(PH2 ) + (17)
2
where T represents the operating temperature, PH2 and PO2 are the partial pressures of the hydrogen and the
oxygen, respectively. They can be calculated as follows:
 
1
PH2 = 0.5 × Rha × PH2 O 
Rha ×PH2 O 1.635(i/ A)
− 1 (18)
× e T 1.334
Pa

 
1
PH2 = 0.5 × Rha × PH2 O 
Rha ×PH2 O 1.635(i/ A)
− 1 (19)
×e T 1.334
Pa

where Rha and Rhc represent the vapor humilities of both anode and cathode, Pa and Pc present the inlet pres-
sures of both anode and cathode (atm), A is the electrode surface (­ cm2), i is the current of the FC (A), and PH2O
represents the water vapor saturation pressure(atm).

NedStack PS6 BCS 500W SR-12 500W

N (number of cells) 65 32 32
A ­(cm2) 240 64 64
l (μm) 178 178 178
∗ (bar)
PH2 1.0 1.0 1
∗ (bar)
PO2 1.0 1.0 0.2095
T (K) 343 333 333
RHa 100%
RHc 100%

Table 15.  The characteristics of the considered PEMFCs.

Optimization variables ξ1 ξ2 ξ3 ξ4  RC (�) b (V)

Lower limit − 1.19969 0.001 3.6 × 10−5 −2.6 × 10−4 10 1 × 10−4 0.0136
Upper limit − 0.8532 0.005 9.8 × 10−5 −9.54 × 10−5 24 8 × 10−4 0.5

Table 16.  The upper and the lower limits of the empirical parameters.

Parameters SSA89 PO90 MPA90 IAEO91 MAEO92 ISSA93 EHBO94 HGSA95 NNA96 RTH
ξ1 − 1.010 − 1.200 − 0.986 − 0.810 − 0.856 − 1.098 − 1.200 − 1.11 − 1.060 − 1.029
ξ2 × ­10−3 3.220 4.042 2.609 5.17 2.73328 3.3352 3.310 3.753 3.744 3.100
ξ3 × ­10−5 5.450 9.800 3.600 8.790 6.634 5.903 4.200 9.710 9.690 6.487
ξ4. × ­10−5 − 1.420 − 1.929 − 1.929 − 1.900 − 1.928 − 1.928 − 1.930 − 1.935 − 19.302 − 1.936
 20.710 20.818 20.817 20.877 20.703 21.250 20.877 21.970 20.877 22.02
RC (�) × ­10−4 0.075 0.016 0.016 0.161 0.100 0.161 0.100 0.100 0. 100 0.100
b (V) × ­10−2 1.00 1 1 1.00 1.60 148.2 1.60 1.60 1.60 1.740
Best 1.219 1.156 1.156 1.16 1.157 1.16 1.170 1.169 1.1698 1.140
Worst 1.520 NA NA 1.17 NA 179.26 1.185 1.34 1.3670 1.140
Mean NA 1.88 1.16 1.16 2.9326 14.57 1.174 NA NA 1.140
StD 871 × ­10−2 NA NA 144 × ­10−3 NA NA 0.6 × ­10−4 3. 4 × ­10−4 5.64 × ­10−4 9.37 × 10−8

Table 17.  The extracted BSC 500W FC parameters. The best results are marked in bold.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 35

Vol.:(0123456789)
 www.nature.com/scientificreports/

SSA89 EO97 STSA98 IAEO91 EBHO94 mAEFA99 NNA96 RTH

ξ1 1.130 − 1.12171 − 0.853 − 1.1997 − 0.85396 − 1.149 − 0.8535 − 0.90568
ξ2 × ­10−5 3.460 3.77 2.840 3.4103 2.40 3.349 2.4316 3.44
ξ3 × ­10−5 4.590 7.81 6.790 3.60 3.60 3.60 3.7545 8.76
ξ4 × ­10−5 − 9.620 − 9.54 − 9.540 − 9.54 − 9.54 − 9.5 − 9.5400 −9
λ 12.910 16.60171 13.463 19.7903 13.465 13.097512 13.0802 17.80574
Rc × ­10−3 0.100 0.205 0.100 0.362 0.1 0.1 0.1000 8
b × ­10−2 6.000 0.0285 1.360 1.360 1.360 1.360 1.360 8.185
Best 2.181 2.40931 2.146 2.1459 2.14570 2.07974 2.14487 2.1058
Worst 2.251 3.02680 3.183 2.1459 2.14570 2.08019 2.1645 2.1058
Mean NA 2.61680 0.280 2.1459 2.14570 2.07987 NA 2.1058
StD 0.020 0.148 0.177 NA 5.69 × ­103 1.6 × ­10−4 5.848 × ­103 1.79 × 10−6

Table 18.  The extracted NedStack PS6 FC parameters. The best results are marked in bold.

Parameters PO90 LSHADE100 MPA90 VSDE101 MAEO92 ISSA93 TGA​102 RTH

ξ1 − 0.860 1.216 − 1.028 − 0.952 − 0.860 − 159 − 1.112 − 0.906
ξ2 × ­10−3 3.376 − 0.960 3.898 3.000 2.771 4.146 3.855 3.440
ξ3 × ­10−5 9.794 2.621 9.800 7.783 6.170 5.6443 4.370 8.760
ξ4. × ­10−5 − 0.954 3.60 − 0.954 − 2.00 − 0.954 − 2.2908 − 0.964 − 9.00
 23.00 − 9.54 23.00 20.29 22.99 13.78 23.00 17.806
RC (�) × ­10−4 6.723 0.154 6.723 1.00 6.707 1.00 2.19 8.00
b (V) × ­10−2 17.5 23.999 17.5 2.79 17.5 7.4 18.3 8.185
Best 1.057 1.216 1.057 1.0526 1.057 0.792 1. 104 0.5607
Worst NA 3.508 NA 1.1875 NA 1.793 5.504 0.5607
Mean 1.058 3.001 1.057 1.0834 6.43 1.46 2.064 0.5607
StD NA 0.12402 NA 0.1768 NA NA NA 1.79 × 10−6

Table 19.  The extracted SR-12 500W FC parameters. The best results are marked in bold.

30 600
Measurments
Estimated
28 Measurments
Estimated 500

26
400
Voltage (V)

24
Power (P)

300

22

200
20

100
18

16 0
0 5 10 15 20 25 30

Current (A)

Figure 36.  Voltage and power curves of BCS 500W using the proposed RTH algorithm.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 36

Vol:.(1234567890)
www.nature.com/scientificreports/

The activation voltage losses (Vact) can be obtained as

(20)
 
Vact = − ζ1 + ζ2 T + ζ3 T ln(CO∂ ) + ζ4 T ln(i)

where ξ1, ξ2, ξ3, ξ4 denote semi-empirical parameters; ­CO2 is the concentration of oxygen at the cathode’s surface
(mol.cm−3). It can be calculated as follows:
 
498
PO2
(21)
T
CO2 = e
5.08 × 106
The ohmic losses (Vohm) can be obtained as
Vohm = i(Rm + Rc ) (22)
where Rc is the resistance of the connectors, and Rm is the resistance of the membrane. Rm can be calculated as

65 9000
Measurments
Estimated
8000
Measurments
60 Estimated
7000

55 6000
Voltage (V)

Power (W)
5000
50
4000

45 3000

2000
40
1000

35 0
0 50 100 150 200

Current (A)

Figure 37.  Voltage and power curves of BCS 500W using the proposed RTH algorithm.

45 900
Measurments
Estimated
800
Measurments
Estimated
40
700

600
35
Voltage (V)

Power (W)

500

400
30
300

200
25

100

20 0
0 5 10 15 20 25 30 35
Current (A)

Figure 38.  Voltage and power curves of SR-12 500W using the proposed RTH algorithm.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 37

Vol.:(0123456789)
 www.nature.com/scientificreports/

 
l
Rm = ρm (23)
Am
where l and Am represent the membrane thickness (cm) and surface (­ cm2), respectively, and ρm represents the
membrane-specific resistivity (ohm × cm). ρm can be obtained as
 
   2.5
181.6 1 + 0.03i Am + 0.062T 303 i Am


ρm =    (24)
T−303
 − 0.634 − 3i Am e4.18 T

where λ is the membrane material’s water content.

Objective function. The sum square error (SSE) between the measured (Vdata) and model output data
(Vdata) will be used as an objective function. The objective function can be constructed as follows:
N
f (x) = (Vdata (k) − Vmodel (k, x))2 (25)
k=1

where N represents the data size, and x represents a vector containing seven unknown parameters.

(26)
 
x = ζ1 ζ2 ζ3 ζ4 Rc  b

The FC data are compared with those generated by the model, and the fitness value is calculated based on the
error between them. The model is developed in MATLAB script, and the data are loaded from the Excel sheet
file. The identification is an iterative process that updates the candidate solutions at each iteration by sending
them to the MATLAB script that includes the FC model and simulates it after that, and then generates the fitness
value. This process repeated until the last iteration.

Results. The suggested RTH algorithm will be used to extract the seven unknown characteristics of three
PEM fuel cells: the NedStack PS6, the BCS 500W, and the SR-12 500W. Table 1587 provided the accurate values of
the parameters, testing operating conditions, and measurement data for the tested PEMFC types. The upper and
lower limits of the empirical parameters are presented in Table 16. The results are compared to those published
for the same FC types.
Table 17 shows the comparative findings for the BCS 500 W type, Table 18 for the NedStack PS6, and Table 19
for the SR-12 PEM 500 W. The comparison findings are primarily based on sum square error (SSE).
Based on the findings reported in Table 17, the suggested method-based RTH performed the best of all the
reported techniques in this comparison to extract the parameters of the BCS 500W. The proposed RTH algo-
rithms obtained the minimal fitness function (SSE) by 1.14 × ­10−2. On the other hand, the mean fitness values
were similar to the best value. The RTH provides the lowest STD values compared to the different algorithms
(9.37 × ­10−8), indicating its robustness. Figure 36 depicts the experimental and estimated voltage and power
curves of BCS 500W using the suggested RTH algorithm. The estimated voltage and power curves match the
experimental curves. These curves demonstrate the proposed algorithm’s accuracy in deriving the best BCS
500W parameters.
Similar to the BDC 500W case, the suggested method-based RTH performed the best of all the cited methods
in this comparison to extract the parameters of the NedStack PS6. The proposed RTH best result is 2.058, similar
to the mean values. The RTH provides the lowest StD values compared to the different algorithms (1.79 × ­10−6),
approving its robustness. Figure 37 depicts the experimental and estimated voltage and power curves of NedStack
PS6 using the suggested RTH algorithm. The estimated voltage and power curves match the experimental curves.
Similar to the previous cases, the suggested method-based RTH performed the best of all the cited methods
in this comparison to extract the parameters of the SR 500W. The proposed RTH best result is 0.5607, similar to
the mean values. The RTH also provides the lowest StD value by 1.79 × ­10−6, approving its robustness. Figure 38
shwos the experimental and estimated voltage and power curves of SR 500W using the suggested RTH algorithm.
The obtained results for the standard test functions evaluated the performance of the proposed RTH algo-
rithm for both exploitation and exploitation phases compared to other MAs. From these results, the proposed
RTH has excellent exploitation and exploration abilities. In addition, its convergence speed has been approved.
Then, the proposed RTH efficiently has been evaluated for several real-world applications. Its performance has
been compared to several published algorithms for each problem. The achieved results confirm its ability to
handle these problems. Finally, a PEMFC parameters extraction has been performed using the proposed RTH
algorithm. The results have been compared to other well-known algorithms to approve its performance. The
results achieved were excellent. Hence, its performance has been elaborated and approved for various types of
optimization problems.

Conclusion and future works

This paper has proposed a novel metaheuristic optimization algorithm named the red-tailed hawk (RTH) algo-
rithm to solve various optimization tasks and problems. The proposed RTH is inspired by the red-tailed hawk’s
hinting behaviors of a predatory bird. A mathematical model has been developed to replicate the behavior of
red-tailed hawks. The main contribution of this paper is to propose a new optimizer that has high robustness and
fast convergence speed when solving various optimization problems. RTH’s performance was firstly evaluated

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 38

Vol:.(1234567890)
www.nature.com/scientificreports/

using three types of mathematical functions that express the nature of different optimization problems: twenty-
three standard benchmark test functions, IEEE Congress on Evolutionary Computation 2020 (CEC2020) with
15 and 20 search space dimensions, and CEC2022 with 10 and 20 search space dimensions. These functions
enable evaluating the exploitative ability, exploratory ability, and local optima avoidance of RTH. The results
are compared to other recent and robust optimizers, including Farmland Fertility Optimizer (FO), African
Vultures Optimization Algorithm (AVOA), Mountain Gazelle Optimizer (MGO), Gorilla Troops Optimizer
(GTO), COOT algorithm, Hunger Games Search (HGS), Aquila Optimizer (AO), and Harris Hawks optimization
(HHO). The results show that the proposed algorithm can provide the optimal solution for most of the considered
functions with fast convergence speed and good robustness. Then, the findings of the seven constrained engi-
neering design problems demonstrated that the RTH could show superior results to other published algorithms
in terms of precision, robustness, and convergence rate. To deeply investigate the performance of the proposed
RTH, the results of the proton exchange membrane fuel cell parameters extractions (PEMFC) have used the
proposed RTH algorithm compared to published ones. The ultimate results show the RTH’s ability to find better
parameters for the dynamic model of the PEMFC.
RTH’s performance is anticipated to be considerably improved by integrating more complicated processes and
combining effective operators and technics of other heuristics. Enhancing the proposed algorithm by including
other factors in the model, such as the wind effect and prey escaping, is possible. However, the bigger code size
of the RTH compared to other algorism like the PSO and the SSA can be a problem for its implementation. Bus,
this problem can be bypassed with the utilization of fast calculators.

Data availability
All data generated or analyzed during this study are included in this published article and its supplementary
information files.

Received: 13 March 2023; Accepted: 14 July 2023

References
1. Mahadeva, R., Kumar, M., Gupta, V., Manik, G. & Patole, S. P. Modified Whale Optimization Algorithm based ANN: a novel
predictive model for RO desalination plant. Sci. Rep. 13, 2901 (2023).
2. Zhang, C. & Ordóñez, R. Numerical optimization. Adv. Ind. Control https://​doi.​org/​10.​1007/​978-1-​4471-​2224-1_2 (2012).
3. Wu, G. Across neighborhood search for numerical optimization. Inf. Sci. 329, 597–618 (2016).
4. Zhao, W., Wang, L. & Zhang, Z. Artificial ecosystem-based optimization: A novel nature-inspired meta-heuristic algorithm.
Neural Comput. Appl. 32, 9383–9425 (2020).
5. Wang, L., Zhao, W., Tian, Y. & Pan, G. A bare bones bacterial foraging optimization algorithm. Cogn. Syst. Res. 52, 301–311
(2018).
6. Sharma, S., Saha, A. K. & Lohar, G. Optimization of weight and cost of cantilever retaining wall by a hybrid metaheuristic
algorithm. Eng. Comput. 38, 2897–2923 (2022).
7. Sharma, S., Saha, A. K., Majumder, A. & Nama, S. MPBOA: A novel hybrid butterfly optimization algorithm with symbiosis
organisms search for global optimization and image segmentation. Multimed. Tools Appl. 80, 12035–12076 (2021).
8. Talbi, E.-G. Metaheuristics From Design to Implementation (Wiley, 2009).
9. Dreo, J., Petrowsdki, A., Siarry, P., Taillard, E. & Chatterjee, A. Metaheuristics for Hard Optimization: Methods and Case Studies.
(2006).
10. Heidari, A. A., Abbaspour, R. A. & Jordehi, A. R. An efficient chaotic water cycle algorithm for optimization tasks. Neural
Comput. Appl. 28, 1–29 (2015).
11. Mafarja, M. et al. Evolutionary population dynamics and grasshopper optimization approaches for feature selection problems.
Knowledge-Based Syst. 145, 25–45 (2018).
12. Salcedo-Sanz, S. Modern meta-heuristics based on nonlinear physics processes: A review of models and design procedures.
Phys. Rep. 655, 1–70 (2016).
13. Hossein-Gandomi, A., Xin-She, Y. & Hossein-Alavi, A. Cuckoo search algorithm: A metaheuristic approach to solve structural
optimization problems. Eng. Comput. 29, 17–35 (2013).
14. Batrinu, F., Carpaneto, E. & Chicco, G. A unified scheme for testing alternative techniques for distribution system minimum
loss reconfiguration. in 2005 International Conference on Future Power Systems vol. 2005 6 pp. (IEEE, 2005).
15. Shayanfar, H. & Gharehchopogh, F. S. Farmland fertility: A new metaheuristic algorithm for solving continuous optimization
problems. Appl. Soft Comput. 71, 728–746 (2018).
16. Abdollahzadeh, B., Gharehchopogh, F. S. & Mirjalili, S. African vultures optimization algorithm: A new nature-inspired
metaheuristic algorithm for global optimization problems. Comput. Ind. Eng. 158, 107408 (2021).
17. Abdollahzadeh, B., Gharehchopogh, F. S., Khodadadi, N. & Mirjalili, S. Mountain Gazelle Optimizer: A new nature-inspired
metaheuristic algorithm for global optimization problems. Adv. Eng. Softw. 174, 103282 (2022).
18. Abdollahzadeh, B., Soleimanian Gharehchopogh, F. & Mirjalili, S. Artificial gorilla troops optimizer: A new nature-inspired
metaheuristic algorithm for global optimization problems. Int. J. Intell. Syst. 36, 5887–5958 (2021).
19. Naruei, I. & Keynia, F. A new optimization method based on COOT bird natural life model. Expert Syst. Appl. 183, 115352
(2021).
20. Yang, Y., Chen, H., Heidari, A. A. & Gandomi, A. H. Hunger games search: Visions, conception, implementation, deep analysis,
perspectives, and towards performance shifts. Expert Syst. Appl. 177, 114864 (2021).
21. Abualigah, L. et al. Aquila Optimizer: A novel meta-heuristic optimization algorithm. Comput. Ind. Eng. 157, 107250 (2021).
22. Heidari, A. A. et al. Harris hawks optimization: Algorithm and applications. Futur. Gener. Comput. Syst. 97, 849–872 (2019).
23. Mirjalili, S. & Lewis, A. The Whale optimization algorithm. Adv. Eng. Softw. 95, 51–67 (2016).
24. Faris, H. et al. An efficient binary Salp Swarm algorithm with crossover scheme for feature selection problems. Knowledge-Based
Syst. 154, 43–67 (2018).
25. Bian, F., Li, T. & Cong, P. Genetic programming. Fenxi Huaxue 26, 783–785 (1998).
26. Lapre, C. et al. Genetic algorithm optimization of broadband operation in a noise-like pulse fiber laser. Sci. Rep. 13, 1865 (2023).
27. Storn, R. & Price, K. Differential evolution: A simple and efficient heuristic for global optimization over continuous spaces. J.
Glob. Optim. 11, 341–359 (1997).

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 39

Vol.:(0123456789)
 www.nature.com/scientificreports/

28. Zheng, Y., Lu, X., Zhang, M. & Chen, S. Biogeography-based optimization. In Biogeography-Based Optimization: Algorithms
and Applications 27–49 (Springer, 2019). https://​doi.​org/​10.​1007/​978-​981-​13-​2586-1_2.
29. Bai, H., Cao, Q. & An, S. Mind evolutionary algorithm optimization in the prediction of satellite clock bias using the back
propagation neural network. Sci. Rep. 13, 2095 (2023).
30. Rashedi, E., Nezamabadi-pour, H. & Saryazdi, S. GSA: A gravitational search algorithm. Inf. Sci. 179, 2232–2248 (2009).
31. Ismaeel, A. A. K., Houssein, E. H., Oliva, D. & Said, M. Gradient-based optimizer for parameter extraction in photovoltaic
models. IEEE Access 9, 13403–13416 (2021).
32. Azizi, M., Aickelin, U., Khorshidi, A. & H. & Baghalzadeh Shishehgarkhaneh, M,. Energy valley optimizer: a novel metaheuristic
algorithm for global and engineering optimization. Sci. Rep. 13, 226 (2023).
33. Kumar, M., Kulkarni, A. J. & Satapathy, S. C. Socio evolution & learning optimization algorithm: A socio-inspired optimization
methodology. Futur. Gener. Comput. Syst. 81, 252–272 (2018).
34. Bayzidi, H., Talatahari, S., Saraee, M. & Lamarche, C.-P. Social network search for solving engineering optimization problems.
Comput. Intell. Neurosci. 2021, 1–32 (2021).
35. Verijkazemi, M. & Fazeli Veysari, E. A new optimization algorithm inspired by the quest for the evolution of human society:
Human felicity algorithm. Expert Syst. Appl. 193, 116468 (2022).
36. Baykasoğlu, A. & Ozsoydan, F. B. Evolutionary and population-based methods versus constructive search strategies in dynamic
combinatorial optimization. Inf. Sci. (Ny) 420, 159–183 (2017).
37. Kennedy, J. & Eberhart, R. Particle swarm optimization. in Proceedings of ICNN’95—International Conference on Neural Networks
vol. 4 1942–1948 (IEEE, 1995).
38. Mirjalili, S. et al. Salp Swarm algorithm: A bio-inspired optimizer for engineering design problems. Adv. Eng. Softw. 114, 163–191
(2017).
39. Chou, J.-S. & Molla, A. Recent advances in use of bio-inspired jellyfish search algorithm for solving optimization problems. Sci.
Rep. 12, 19157 (2022).
40. Ghafori, S. & Gharehchopogh, F. S. Advances in spotted hyena optimizer: A comprehensive survey. Arch. Comput. Methods Eng.
29, 1569–1590 (2022).
41. Gharehchopogh, F. S. Advances in tree seed algorithm: A comprehensive survey. Arch. Comput. Methods Eng. 29, 3281–3304
(2022).
42. Gharehchopogh, F. S., Namazi, M., Ebrahimi, L. & Abdollahzadeh, B. Advances in sparrow search algorithm: A Comprehensive
Survey. Arch. Comput. Methods Eng. 30, 427–455 (2023).
43. Chakraborty, P., Sharma, S. & Saha, A. K. Convergence analysis of butterfly optimization algorithm. Soft Comput. 27, 7245–7257
(2023).
44. Sharma, S., Khodadadi, N., Saha, A. K., Gharehchopogh, F. S. & Mirjalili, S. Non-dominated sorting advanced butterfly opti-
mization algorithm for multi-objective problems. J. Bionic Eng. 20, 819–843 (2023).
45. Sharma, S., Chakraborty, S., Saha, A. K., Nama, S. & Sahoo, S. K. mLBOA: A Modified Butterfly optimization algorithm with
lagrange interpolation for global optimization. J. Bionic Eng. 19, 1161–1176 (2022).
46. Chakraborty, S., Sharma, S., Saha, A. K. & Saha, A. A novel improved whale optimization algorithm to solve numerical optimiza-
tion and real-world applications. Artif. Intell. Rev. 55, 4605–4716 (2022).
47. Sahoo, S. K., Sharma, S. & Saha, A. K. A novel variant of moth flame optimizer for higher dimensional optimization problems.
J. Bionic Eng. https://​doi.​org/​10.​1007/​s42235-​023-​00357-7 (2023).
48. Gharehchopogh, F. S. Quantum-inspired metaheuristic algorithms: comprehensive survey and classification. Artif. Intell. Rev.
56, 5479–5543 (2023).
49. Preston, C. R. Wild Bird Guide: Red-Tailed Hawk (Wild Bird Guides) (Stackpole Books, 2000).
50. Pearlstine, E. V. & Thompson, D. B. Geographic variation in morphology of four species of migratory raptors. J. Raptor Res. 38,
334–342 (2004).
51. Ballam, J. M. The use of soaring by the red-tailed Hawk (Buteo jamaicensis). Auk 101, 519–524 (1984).
52. Day, L. The City Naturalist: Red Tailed Hawk. in The 79th Street Boat Basin Flora and Fauna Society (2007).
53. Knight, R. L., Andersen, D. E., Bechard, M. J. & Marr, N. V. Geographic variation in nest-defence behaviour of the Red-tailed
Hawk Buteo jamaicensis. Ibis 131, 22–26 (2008).
54. Yang, X.-S. Appendix A: Test Problems in Optimization. In Engineering Optimization 261–266 (Wiley, 2010). https://​doi.​org/​
10.​1002/​97804​70640​425.​app1.
55. Yue, CT and Price, KV and Suganthan, Ponnuthurai N and Liang, JJ and Ali, Mostafa Z and Qu, BY and Awad, Noor H and
Biswas, P. P. Problem definitions and evaluation criteria for the CEC 2020 special session and competition on single objective
bound constrained numerical optimization. Comput. Intell. Lab., Zhengzhou Univ., Zhengzhou, China, Tech. Rep 201911, (2019).
56. Suganthan, P. N. et al. Problem definitions and evaluation criteria for the CEC 2021 special session and competition on single
objective bound constrained numerical optimization. Technical Report, Nanyang Technological University, Singapore, May 2005
AND KanGAL Report 2005005, IIT Kanpur, India (2020).
57. Faramarzi, A., Heidarinejad, M., Mirjalili, S. & Gandomi, A. H. Marine Predators Algorithm: A nature-inspired metaheuristic.
Expert Syst. Appl. 152, 113377 (2020).
58. Cheng, M.-Y. & Prayogo, D. Symbiotic organisms search: A new metaheuristic optimization algorithm. Comput. Struct. 139,
98–112 (2014).
59. Kaveh, A. Advances in Metaheuristic Algorithms for Optimal Design of Structures (Springer, 2021). https://​doi.​org/​10.​1007/​
978-3-​030-​59392-6.
60. Azizi, M. Atomic orbital search: A novel metaheuristic algorithm. Appl. Math. Model. 93, 657–683 (2021).
61. Cao, S. et al. A novel meta-heuristic algorithm for numerical and engineering optimization problems: Piranha foraging opti-
mization algorithm (PFOA). IEEE Access https://​doi.​org/​10.​1109/​ACCESS.​2023.​32671​10 (2023).
62. Zhao, W., Wang, L. & Mirjalili, S. Artificial hummingbird algorithm: A new bio-inspired optimizer with its engineering applica-
tions. Comput. Methods Appl. Mech. Eng. 388, 114194 (2022).
63. Ghasemi, M. et al. A new firefly algorithm with improved global exploration and convergence with application to engineering
optimization. Decis. Anal. J. 5, 100125 (2022).
64. Mirjalili, S., Mirjalili, S. M. & Lewis, A. Grey Wolf Optimizer. Adv. Eng. Softw. 69, 46–61 (2014).
65. Eskandar, H., Sadollah, A., Bahreininejad, A. & Hamdi, M. Water cycle algorithm: A novel metaheuristic optimization method
for solving constrained engineering optimization problems. Comput. Struct. 110–111, 151–166 (2012).
66. Ravindran, A., Ragsdell, K. M. & Reklaitis, G. V. Engineering Optimization (Wiley, 2006). https://​doi.​org/​10.​1002/​97804​70117​
811.
67. Talatahari, S. & Azizi, M. Optimization of constrained mathematical and engineering design problems using chaos game opti-
mization. Comput. Ind. Eng. 145, 106560 (2020).
68. Meng, X. B., Li, H. X. & Gao, X. Z. An adaptive reinforcement learning-based bat algorithm for structural design problems. Int.
J. Bio-Inspired Comput. 14, 114–124 (2019).
69. Gandomi, A. H. & Roke, D. A. Engineering optimization using interior search algorithm. In 2014 IEEE Symposium on Swarm
Intelligence 1–7 (IEEE, 2014). https://​doi.​org/​10.​1109/​SIS.​2014.​70117​71.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 40

Vol:.(1234567890)
www.nature.com/scientificreports/

70. Yıldız, B. S. et al. A novel hybrid arithmetic optimization algorithm for solving constrained optimization problems. Knowledge-
Based Syst. 271, 110554 (2023).
71. Sattar, D. & Salim, R. A smart metaheuristic algorithm for solving engineering problems. Eng. Comput. 37, 2389–2417 (2021).
72. Akay, B. & Karaboga, D. Artificial bee colony algorithm for large-scale problems and engineering design optimization. J. Intell.
Manuf. 23, 1001–1014 (2012).
73. Ben Guedria, N. Improved accelerated PSO algorithm for mechanical engineering optimization problems. Appl. Soft Comput.
40, 455–467 (2016).
74. Yang, X. et al. An adaptive quadratic interpolation and rounding mechanism sine cosine algorithm with application to con-
strained engineering optimization problems. Expert Syst. Appl. 213, 119041 (2023).
75. Gupta, S. et al. Comparison of metaheuristic optimization algorithms for solving constrained mechanical design optimization
problems. Expert Syst. Appl. 183, 115351 (2021).
76. Wu, J. et al. An improved firefly algorithm for global continuous optimization problems. Expert Syst. Appl. 149, 113340 (2020).
77. Agushaka, J. O. et al. Improved Dwarf Mongoose optimization for constrained engineering design problems. J. Bionic Eng. 20,
1263–1295 (2023).
78. Arora, J. S. Introduction to Optimum Design (Elsevier, 2012). https://​doi.​org/​10.​1016/​C2009-0-​61700-1.
79. He, Q. & Wang, L. An effective co-evolutionary particle swarm optimization for constrained engineering design problems. Eng.
Appl. Artif. Intell. 20, 89–99 (2007).
80. He, Q. & Wang, L. A hybrid particle swarm optimization with a feasibility-based rule for constrained optimization. Appl. Math.
Comput. 186, 1407–1422 (2007).
81. MiarNaeimi, F., Azizyan, G. & Rashki, M. Multi-level cross entropy optimizer (MCEO): An evolutionary optimization algorithm
for engineering problems. Eng. Comput. 34, 719–739 (2018).
82. Faramarzi, A., Heidarinejad, M., Stephens, B. & Mirjalili, S. Equilibrium optimizer: A novel optimization algorithm. Knowledge-
Based Syst. 191, 105190 (2020).
83. El-Kenawy, E.-S.M. et al. Novel meta-heuristic algorithm for feature selection, unconstrained functions and engineering prob-
lems. IEEE Access 10, 40536–40555 (2022).
84. Ghafil, H. N. & Jármai, K. Dynamic differential annealed optimization: New metaheuristic optimization algorithm for engineer-
ing applications. Appl. Soft Comput. 93, 106392 (2020).
85. Olabi, A. G., Wilberforce, T. & Abdelkareem, M. A. Fuel cell application in the automotive industry and future perspective.
Energy 214, 118955 (2021).
86. Rezk, H. et al. Fuel cell as an effective energy storage in reverse osmosis desalination plant powered by photovoltaic system.
Energy 175, 423–433 (2019).
87. Rezk, H. et al. Optimal parameter estimation strategy of PEM fuel cell using gradient-based optimizer. Energy 239, 122096
(2022).
88. Cao, Y., Li, Y., Zhang, G., Jermsittiparsert, K. & Nasseri, M. An efficient terminal voltage control for PEMFC based on an
improved version of whale optimization algorithm. Energy Rep. 6, 530–542 (2020).
89. El-Fergany, A. A. Extracting optimal parameters of PEM fuel cells using Salp Swarm optimizer. Renew. Energy 119, 641–648
(2018).
90. Zaki Diab, A. A. et al. Fuel cell parameters estimation via marine predators and political optimizers. IEEE Access 8, 166998–
167018 (2020).
91. Rizk-Allah, R. M. & El-Fergany, A. A. Artificial ecosystem optimizer for parameters identification of proton exchange membrane
fuel cells model. Int. J. Hydrogen Energy 46, 37612–37627 (2021).
92. Menesy, S. et al. Effective parameter extraction of different polymer electrolyte membrane fuel cell stack models using a modified
artificial ecosystem optimization algorithm. IEEE Access 8, 31892–31909 (2020).
93. Sultan, H. M., Menesy, A. S., Kamel, S., Selim, A. & Jurado, F. Parameter identification of proton exchange membrane fuel cells
using an improved salp swarm algorithm. Energy Convers. Manag. 224, 113341 (2020).
94. Abdel-Basset, M., Mohamed, R., Elhoseny, M., Chakrabortty, R. K. & Ryan, M. J. An efficient heap-based optimization algorithm
for parameters identification of proton exchange membrane fuel cells model: Analysis and case studies. Int. J. Hydrogen Energy
46, 11908–11925 (2021).
95. Fahim, S. R. et al. Parameter identification of proton exchange membrane fuel cell based on hunger games search algorithm.
Energies 14, 5022 (2021).
96. Fawzi, M., El-Fergany, A. A. & Hasanien, H. M. Effective methodology based on neural network optimizer for extracting model
parameters of PEM fuel cells. Int. J. Energy Res. 43, 8136–8147 (2019).
97. Menesy, A. S., Sultan, H. M. & Kamel, S. Extracting Model Parameters of Proton Exchange Membrane Fuel Cell Using Equi-
librium Optimizer Algorithm. in 2020 International Youth Conference on Radio Electronics, Electrical and Power Engineering
(REEPE) 1–7 (IEEE, 2020). doi:https://​doi.​org/​10.​1109/​REEPE​49198.​2020.​90592​19.
98. Jiang, J., Xu, M., Meng, X. & Li, K. STSA: A sine Tree-Seed Algorithm for complex continuous optimization problems. Phys. A
Stat. Mech. its Appl. 537, 122802 (2020).
99. Houssein, E. H., Hashim, F. A., Ferahtia, S. & Rezk, H. An efficient modified artificial electric field algorithm for solving opti-
mization problems and parameter estimation of fuel cell. Int. J. Energy Res. 45, 20199–20218 (2021).
100. Fathy, A., Abdel Aleem, S. H. E. & Rezk, H. A novel approach for PEM fuel cell parameter estimation using LSHADE-EpSin
optimization algorithm. Int. J. Energy Res. 45, 6922–6942 (2021).
101. Fathy, A., Elaziz, M. A. & Alharbi, A. G. A novel approach based on hybrid vortex search algorithm and differential evolution
for identifying the optimal parameters of PEM fuel cell. Renew. Energy 146, 1833–1845 (2020).
102. Kamel, S., Jurado, F., Sultan, H. & Menesy, A. Tree growth algorithm for parameter identification of proton exchange membrane
fuel cell models. Int. J. Interact. Multimed. Artif. Intell. 6, 11 (2020).

Author contributions
The authors confirm their contribution to the paper as follows: study conception and design: S.F., A.D.; data col-
lection: R.H., S.M.; analysis and interpretation of results: A.H., M.M. M.A.; draft manuscript preparation: S.F.,
R.H. All authors reviewed the results and approved the final version of the manuscript.

Competing interests
The authors declare no competing interests.

Additional information
Supplementary Information The online version contains supplementary material available at https://​doi.​org/​
10.​1038/​s41598-​023-​38778-3.
Correspondence and requests for materials should be addressed to S.M.

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 41

Vol.:(0123456789)
 www.nature.com/scientificreports/

Reprints and permissions information is available at www.nature.com/reprints.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.

© The Author(s) 2023

Scientific Reports | (2023) 13:12950 | https://doi.org/10.1038/s41598-023-38778-3 42

Vol:.(1234567890)