Archives of Computational Methods in Engineering (2023) 30:427–455

https://doi.org/10.1007/s11831-022-09804-w

SURVEY ARTICLE

Advances in Sparrow Search Algorithm: A Comprehensive Survey

Farhad Soleimanian Gharehchopogh1 · Mohammad Namazi2 · Laya Ebrahimi1 · Benyamin Abdollahzadeh1

Received: 4 June 2022 / Accepted: 2 August 2022 / Published online: 22 August 2022
© The Author(s) under exclusive licence to International Center for Numerical Methods in Engineering (CIMNE) 2022

Abstract
Mathematical programming and meta-heuristics are two types of optimization methods. Meta-heuristic algorithms can iden-
tify optimal/near-optimal solutions by mimicking natural behaviours or occurrences and provide benefits such as simplicity
of execution, a few parameters, avoidance of local optimization, and flexibility. Many meta-heuristic algorithms have been
introduced to solve optimization issues, each of which has advantages and disadvantages. Studies and research on presented
meta-heuristic algorithms in prestigious journals showed they had good performance in solving hybrid, improved and
mutated problems. This paper reviews the sparrow search algorithm (SSA), one of the new and robust algorithms for solving
optimization problems. This paper covers all the SSA literature on variants, improvement, hybridization, and optimization.
According to studies, the use of SSA in the mentioned areas has been equal to 32%, 36%, 4%, and 28%, respectively. The
highest percentage belongs to Improved, which has been analyzed by three subsections: Meat-Heuristics, artificial neural
networks, and Deep Learning.

1 Introduction there is no guarantee of finding a universal solution. Many

researchers have used meta-heuristic strategies to find the
The necessity for meta-heuristic algorithms has grown optimal solution to achieve the global optimal [4–6].
in recent decades as the complexity of diverse issues has Meta-heuristic algorithms have solved many optimization
increased. Previously, academics employed mathematical problems, most of which can solve high-dimensional optimi-
strategies to handle local optimization's deterministic and zation problems well. Large-scale global optimization issues
difficult-to-trap optimization issues. Because the search are widespread in scientific research and engineering appli-
space in actual optimization issues increases exponentially cations and have attracted much attention in recent years.
and the problem perspective shifts in a multidimensional The high-dimensional optimization problem is expressed
fashion, standard optimization methods frequently gener- as a two-dimensional d minimization problem according to
ate less-than-optimal solutions [1–3]. These techniques are Eq. (1). Where X = [x1 , x2 , … , xD ] represents the decision
inefficient in solving real optimization problems, which has vector, and F(X) is the fitness function.
increased interest in metaheuristic algorithms in the last two
decades. Due to intrinsic complexity constraints and many
min F(X), X = [x1 , x2 , … , xD ] (1)
design variables such as nonlinear and convex, most real- As the number of dimensions increases, many
world optimization issues, such as text processing, commu- metaheuristic methods become "dimension traps", mean-
nity detection, feature selection, optimization issues, setting ing that performance decreases rapidly and is easily opti-
machine learning parameters, etc., require meta-heuristic mized locally as the number of dimensions increases.
algorithms. Therefore, solving these optimization problems Therefore, hybridization operators must strike a balance
is complicated due to many local minimums. In addition, between exploitation and exploration in the optimization
process, and search efficiency must be improved. In gen-
eral, solutions with better fitness have higher growth per-
* Farhad Soleimanian Gharehchopogh
[email protected] formance, while solutions with poor fitness can maintain
population diversity and strengthen their exploration abil-
1
Department of Computer Engineering, Urmia Branch, ity. As a result, to strike a balance between exploration and
Islamic Azad University, Urmia, Iran exploitation, the population must be separated into two
2
Department of Computer Engineering, Maybod groups: the main population and the sub-population, using
Branch. Islamic Azad University, Maybod, Iran

13
Vol.:(0123456789)
428 F. S. Gharehchopogh et al.

a multinational concept. People in the main population can • Investigations of SSA methods from the aspects of
boost exploitation capacity and solution accuracy, whereas Hybridization, Improved, Variants of SSA, and optimi-
individuals in the subpopulation can assure population zation issues.
variety and exploration ability. This method ensures the • Improved SSA analysis by different methods
ability to exploit while avoiding the local optimum. Dur- • Analysis of SSA performance in solving diverse prob-
ing the exploration phase, the subpopulation is adjusted by lems based on convergence rate, exploration, and exploi-
directing individuals in the exploitation population, which tation factors.
can increase exploration efficacy [7]. • Focus on outlook works in line with the SSA algorithm
In meta-heuristic algorithms, graceful exploration
means the ability to search optimally globally. The entire The general structure of this paper is as follows: The SSA
population is allowed to explore the whole solution space algorithm and its operators will be explained in Sect. 2. In
to find a promising area. In contrast, good exploitation Sect. 3, SSA approaches will be divided into four categories:
demonstrates good local search capability. The population hybridization, improvement, SSA variations, and optimiza-
can use the valuable points to refine the search for a more tion concerns. In Sect. 4, we'll talk about discussions and
accurate optimal solution. The balance of exploration and comparisons; in Sect. 5, we'll wrap things up and look for-
exploitation is essential in improving optimization perfor- ward to future projects.
mance. Excessive attention to exploration leads to wasting
evolution in search of some parts of the solution space, and
as a result, the convergence rate is reduced.
On the other hand, paying more attention to exploita- 2 SSA: Sparrow Search Algorithm
tion risks losing diversity in the early stages of evolution.
As a result, the initial population will likely be trapped in SSA Algorithm [17] is a new nature-inspired algorithm
the local optimal. Therefore, achieving the right balance inspired by the behaviour of sparrows in 2020. Many animals
between exploration and operation in meta-heuristic algo- search for cuisine and avoid predators with their swarming
rithms improves performance in solving complex optimi- intelligence in the wild. The population of sparrows is no
zation problems [8–13]. exception. They are separated into two categories depend-
Engineering optimization issues have been researched ing on their fitness, determined by each sparrow's unique
and solved using a variety of methodologies. Meanwhile, posture. The person who has a better fit belongs to the pro-
meta-heuristic algorithms have performed well. Unlike tra- ducers. The remaining sparrows are explorers. In the whole
ditional optimization methods, Meta-heuristic algorithms population of sparrows, different people have different eating
do not require gradient information and can avoid local behaviors. In addition, several sparrows are responsible for
optimization. As a result, they can be used to solve engi- avoiding predators during the forage search process among
neering optimization challenges. the population. To cope with the dangers, they choose to
They can find an optimal solution regardless of the fly farther or closer to other sparrows. In short, the sparrow
physical nature of the problem. Most of them are inspired colony can search for more low-risk cuisine by constantly
by physical or natural phenomena. Examples include updating its position.
Farmland Fertility Algorithm [14], African Vultures Opti- For simplicity, sparrows' behaviour and related laws are
mization Algorithm (AVOA) [15], Starling murmuration described below.
optimizer [16], Sparrow Search Algorıthm [17], and Arti-
ficial Gorilla Troops Optimizer [18]. • Producers often have a lot of energy reserves and offer
SSA is a population-based meta-heuristic algorithm regions or forage search routes to all explorers. They are
developed by Xue and Shen in 2020 to solve continuous in charge of locating cuisine-rich locations. Individual
optimization problems [17]. The evaluation of the SSA fitness determines how much energy is stored in the body.
algorithm is performed with 19 known mathematical • When a sparrow detects a predator, individuals start
functions. This algorithm has demonstrated its ability to chirping with warning signals. When the alarm value
address computational complexity and solution conver- exceeds the safety threshold, manufacturers must direct
gence difficulties. The SSA algorithm outperforms the all explorers to a safe area.
GWO, Particle Swarm Optimization(PSO), and GSA • Each sparrow can be a producer if it seeks a better cuisine
algorithms in performance. The main contributions of this fountainhead, but the ratio of producers to explorers in
paper are as follows: the whole population is unchanged.
• Producers are sparrows with more vigour. Several hun-
• SSA algorithm analysis based on schematic and pseudo- gry probes are more inclined to fly to other locations in
code search of cuisine to replenish their energy levels.

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 429

• Explorers are looking for a producer who can provide the in which each element is 1. If it is ­R2 ≥ ST, some sparrows
best cuisine in search cuisine. Meanwhile, some explor- have discovered the hunter, and all the sparrows must fly
ers may constantly monitor producers and compete for quickly to other safe areas. When R ­ 2 < ST means no hunter
cuisine over prey. is around, the manufacturer enters the extensive search
• The group sparrows move quickly to a safe area to find a mode.
better position if they are aware of the danger, while the In the case of explorers, Rules 4 and 5 must be fol-
sparrows in the middle of the group walk randomly to get lowed. As previously said, some explorers keep tabs on
closer to others. most manufacturers. They leave their current place to
compete for cuisine when they learn that a producer has
SSA is suggested by imitating the search behaviour of the discovered delicious cuisine. If they win, they can eat right
sparrow and anti-hunting group. This algorithm has fewer away; otherwise, Rule 5 will apply. Position updating for
parameters, a more robust search capacity, and faster perfor- explorers is defined according to Eq. (5). In Eq. (5) xp
mance. The main stages of SSA can be explained as follows: does whether the manufacturer occupy the optimal posi-
Step 1 Create and initialize the solution. At this stage, tion. Xworst represents the worst place in the world right
the population size, maximum number of replicates, pro- now. A means a 1 × d matrix that is randomly assigned 1
ducer ratio (PD), and the ratio of sparrows in intensive care
−1
or -1 to each element inside A+ = AT (AAT ) . If i > 2n This
(PV) are all determined. The initial position of the sparrow indicates that the i probe with a worse fit value is more
population is shown in Eq. (2). They are produced randomly. likely to go hungry.
{ ( X t −X t )
⎡ x1,1 x1,2 ⋯ ⋯ x1,d ⎤
Q × exp worst i,j
if > n2
⎢x x2,2 ⋯ ⋯ x2,d ⎥ t+1
Xi,j = i2 (5)
X = ⎢ 2,1 (2)
⋮ ⎥⎥ XPt+1 + |Xi,j − XPt+1 | × A+ × L otherwise
t
⎢ ⋮ ⋮ ⋮ ⋮
⎣ xn,1 xn,2 ⋯ ⋯ xn,d ⎦
Step 3 After updating the position of the whole popula-
The number of sparrows in Eq. (2) is n, and the dimension tion, several sparrows are selected as scouts (exploration)
of choice variables is d. Each person's suitability for the fol- responsible for identification and warning. They usually
lowing procedure is determined using Eq. (3). In Eq. (3), n make up 10 to 20% of the total population. Updating their
defines the number of sparrows, and the value of each row position is defined according to Rule 6 according to Eq. (6).
in FX represents each person's fit. { t
Xbest + 𝛽 × |Xi,j
t t
− Xbest | f > fg
t+1 ( ) i
⎡ f [x1,1 x1,2 ⋯ ⋯ x1,d ] ⎤ Xi,j = t
|Xi,j −Xworst |
t t
(6)
Xi,j + K × f −f +𝜀 fi = fg
⎢ f [x x2,2 ⋯ ⋯ x2,d ] ⎥ ( i w)
FX = ⎢ 2,1 (3)
⎢ ⋮ ⋮ ⋮ ⋮ ⋮ ⎥⎥ In Eq. (6) Xbest is the current global optimal position. ε
⎣ f [xn,1 xn,2 ⋯ ⋯ xn,d ] ⎦ is a small constant to avoid a zero-division error. β acts as
a control parameter for step size and the normal distribu-
Step 2 In the SSA, producers with higher fitness val-
tion of random numbers with mean value 0 and variance 1.
ues are given preference over those who produce cuisine.
fg and fw are the current best and worst overall suitability
Because producers are in charge of finding cuisine and
values, respectively. K is a random number in the range
directing the entire population's movement, producers can
1 and − 1. fi is the current value of the sparrow. fi= fg
search for cuisine in a broader range compared to the posi-
indicates that sparrows in the middle of the population are
tion of the explorers. According to steps (1) and (2), during
aware of the hazard and should approach the rest. If fi> fg
each iteration, the manufacturers update their status with
then the person is at the edge of the group. Xbest indicates
Eq. (4).
the central location of the population and is safe around it.
{ ( )
X t
× exp −i
if R2 < ST K represents the direction in which the person moves and
t+1
Xi,j = i,j
t
𝛼×itermax (4) the step size's control factor.
Xi,j + Q × L if R2 ≥ ST Step 4 Each person's current position is compared to
the last repetition. The update is done if the new position
In Eq. (4), itermax is a constant with the highest number
is better than the previous one and saves the best position.
of iterations. t is the current iteration, and j = 1, 2, … , d ,
The survival of some sparrows may improve after the last
t
Xi,j represent the next value of jth sparrow in the iteration
two steps.
of t. α is a random number between 0 and 1. R2 (alert Step 5 If the number of repetitions is less than the maxi-
value) is a number in the range of 0 to 1, and ST (safe mum number, move on to step 2. Otherwise, the algorithm
threshold) is a value of 0.5 to 1.0. Q is a random number stops, and the best solution is obtained.
based on a normal distribution. L represents a 1 × d matrix

13
430 F. S. Gharehchopogh et al.

Input:
G: maximum number of iterations
PD: the number of producers
SD: the number of agents (sparrows) alerted to the risk
R2: the alarm value
n: the number of agents
Create a population of n agents and describe its essential parameters.
Output: , .
01: While (t < G)
02: Rank the fitness values to determine the current best and worst individual.
03: R2 =
04: For in (1, PD)
05: Using Eq. (4), upgrade the agent’s position;
06: Out of For
07: For in ((PD + 1), n)
08: Using Eq. (5), upgrade the agent’s position;
09: Out of For
10: For l in (1, SD)
11: Using Eq. (6), upgrade the agent’s position;
12: Out of For
13: Get the current fresh position;
14: if the fresh position is better than formerly, upgrade it;
15: t= t + 1
16: End While
17: Return , .

Fig. 1  Pseudo-code of SSA algorithm [17]

Figure 1 shows the pseudocode of the SSA algorithm.

The primary implementation of meta-heuristic algo-
rithms usually has a sequential approach. Sparrow posi-
tions are typically recorded in a n × d matrix, where n is Fig. 2  Flowchart of SSA algorithm [17]
the number of sparrows and d is the number of dimensions
in the search space. Using For-loops, the members of this Papers are collected based on the title, keywords and
matrix are changed one by one based on their value in abstract. Each paper has been thoroughly reviewed in
the previous iteration and some random sample numbers. terms of text and type of algorithm. Finally, the papers
For search agents or various dimensions, all matrix com- belonging to the SSA algorithm were grouped. Figure 5
ponents are simultaneously updated. Figure 2 shows the shows the search steps and the number of papers in dif-
flowchart of the SSA algorithm. ferent steps.
Since 2020, various researches have been conducted
to solve optimization problems with SSA. All the papers
worked by SSA are downloaded to calculate the number
of SSA papers. Then a grouping based on the percentage
of papers in various journals and the number of publica-
tions of SSA papers per year is reviewed. Figure 3 shows
the rate of papers published by SSA in multiple publica-
tions. The majority of publications belonged to IEEE jour-
nals (39%), followed by Elsevier (28%), Springer (11%),
Hindawi (9%), Others (7%), and Tandfonline and Wiley
(3%). Figure 3 shows that the highest percentage of papers
published belong to the IEEE. At first, we downloaded
all the papers belonging to SSA. Then the papers were
categorized based on different publishers. We used the
Google search engine and reliable sites such as Springer,
Elsevier etc., for searching. We also used other databases
in the field of indexing papers.
Figure 4 illustrates the number of SSA papers printed
per year. The number of SSA papers printed in 2020 is
7. As shown in Fig. 4, the use of SSA has increased over
time. Fig. 3  Percentage of papers published with SSA in various journals

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 431

of SSA, the Binary subcategory is used, and optimization

issues are used to solve diverse optimization issues to find
the best answer.

3.1 Hybridization

3.1.1 Meta‑Heuristics

According to the performed classifications, the SSA algo-

rithm in the field of meta-heuristics is combined with PSO,
water wave optimization (WWO), sine cosine algorithm
(SCA), firefly algorithm (FA), differential equation (DE),
whale optimization algorithm (WOA), topographical global
optimization (TGO), and BSS algorithms.
Fig. 4  Number of SSA papers published per year
PESSA [19], a hybrid approach based on PSO and an
enhanced SSA (ESSA), has been presented. The ESSA
strengthens the producer's random jump to ensure global
search ability, each scrounger continues to learn from the
producers' optimal experience, and the difference between
the best and worst individual will be imposed on the sparrow
with the optimal position when it detects danger to speed up
the search process. Ten fundamental functions validate PES-
SA's performance, and the experimental findings reveal that
PESSA outperforms the other twelve methods. Finally, the
suggested PESSA is tested in four situations, two of which
are 2D settings and 3D environments. The findings revealed
that the PESSA could obtain a more viable and effective
route than the other models.
The WWO-SSA [20] was designed to combine the ben-
efits of the WWO and SSA algorithms while avoiding their
limitations. WWO and SSA have been integrated to achieve
good performance by continually modifying the parameters
to increase WWO's capabilities in development and explo-
ration. Using CEC2017's benchmark features, the hybrid
algorithm's performance is compared to WWO and SSA's
original methods. WWOSSA is more efficient, according to
the findings of the experiments.
A hybrid SSA-PSO [21] model has been developed to
speed up convergence before individual SSA updates. In
addition, a novel fitness function based on maximum likeli-
hood parameter estimate was created and utilized for param-
eter initialization. The optimization performance of this
algorithm was superior to that of a single method, with more
Fig. 5  Review of papers belongs to the SSA algorithm
incredible convergence speed and more stable, accurate out-
puts, according to the findings of five sets of actual datasets.
3 Methods of SSA Furthermore, it efficiently handled the difficulties of slug-
gish convergence speed and low solution accuracy with the
Figure 6 shows the taxonomy of SSA methods. Classifica- help of the new fitness function. The experimental findings
tion is based on Hybridization, Improved, Variants of SSA, revealed that the hybrid SSA-PSO could acquire a superior
and Optimization issues. In hybridization, the combina- solution, convergence speed, and stability in software defect
tion of SSA with other algorithms is used. Improved uses estimation and prediction than a single SSA and PSO.
various subcategories to improve solutions. In Variants The SCA was initially created to increase the global
search capacity of the SSA algorithm because it has the

13
432 F. S. Gharehchopogh et al.

Fig. 6  Classification of SSA

methods

qualities of attaining high search and avoiding local optimi- efforts to develop a green future. As a result, getting the best
zation. Additionally, the labour collaboration structure of results is critical. A novel multiple-objective optimization
the sparrow in the SSA algorithm is redefined to improve approach called SSA-DE [24] is used to get the best SES
the algorithm's convergence ability. Finally, the enhanced level. SSA-purpose DE is to determine the best value for
cooperative SSA based on the sine cosine algorithm (SCA- system resource parameters.
CSSA) is developed [22] based on the new labour coopera- In Wireless Sensor Networks, a hybrid SSA with DE is
tion structure and SCA algorithm. The SCA-CSSA approach designed to alleviate the energy efficiency issue by cluster
is used to adjust the weight of AdaBoost-S4VM and the head selection [25]. The proposed approach combined the
critical parameters of S4VM to improve the precision of the SSA's high-level search efficiency with DE's lively poten-
AdaBoost-S4VM model for semi-supervised lung CT clas- tial, extending node lifespan. The hybrid model performs
sification. The suggested AdaBoost-ISSA-S4VM model was well in the number of alive nodes, throughput, and residual
compared against several hybrids and popular approaches on energy. Compared to comparable algorithms, the Improved
CEC2017 tasks and 12 benchmark tasks, including unimodal SSA employing the DE model to find the best potential
and multimodal tasks, to see how effective it was. cluster head demonstrated residual power and throughput
The SSA is used to improve the starting weights and development.
thresholds of the BP-ANN, addressing the problem that Due to sluggish convergence speed, low accuracy, and
the BP neural network is sensitive to beginning weights optimum local distance, the WOA algorithm is merged with
and thresholds. The firefly algorithm (FA) technique with the SSA and golden sine leading strategy (SGSWOA) [26].
FASSA [23] is presented to alleviate the weakness of SSA The producer's position update rule in the SSA is integrated
that it is easy to slip into the local optimum. Finally, China's into the encircling prey stage of WOA to extend the algo-
big battery manufacturing firm is chosen for the empirical rithm's search space and escape from the local optimum.
study. Comparative tests are conducted on the FASSA-BP, Then, when used with the golden Sine leading technique, it
BP, SSA-BP, and PSO-BP regarding the accuracy, stability, may balance exploration and development capabilities while
and other factors. The FASSA-BP model was shown to be improving the WOA algorithm's performance. Finally, the
more accurate in the study. experimental findings showed that the SGSWOA method
Because buildings play a significant role in energy effi- has superior convergence accuracy, convergence speed,
ciency, it is critical to implement sustainable energy source and resilience after optimizing 16 benchmark functions and
(SES) systems globally, especially given the rising inter- applying it to actual engineering optimization situations.
est in near-zero energy structures. Because of their signifi- A new greedy genetic SSA (GGSC-SSA) based on the
cant influence on energy usage and pollution, SES must be SCA method has been suggested [27]. The greedy method
entirely used in buildings to promote renewable energy and is first implemented to initialize the population and boost

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 433

its diversity. Second, GA operators are utilized to balance structural hyperparameters. The TGO-SSA technique out-
global search and local development capacities to update the performs the standard scenario prediction model to improve
population. Finally, adaptive weight is added to the routine neural network model accuracy and reduce training loss.
upgrade to improve the algorithm's flexibility and maxi- Python was used to test the suggested technique. The find-
mize the solution quality, and the SCA approach is used to ings revealed that the approach could perform better at situ-
update the scroungers. On TSP datasets, the GGSC-SSA is ation prediction.
also tested against the genetic algorithm (GA), simulated The Improved Bear Smell Search (IBSS) and SSA are
annealing (SA), PSO, grey wolf optimization (GWO), ant presented [31]. The BSS is included in the proposed work
colony optimization (ACO), and the artificial fish algorithm through crossover and mutation functions, thus the designa-
(AFA). The results showed that the GGSC-SSA significantly tion IBSS-SSA. The grid created the multi-objective func-
enhanced the solution precision, optimization speed, and tion with reactive power variations dependent on the avail-
perseverance. able resource power. The SSA procedure ensures that online
For Unmanned Aerial Vehicle Path Planning, a unique control signals are detected utilizing a parallel implemen-
SPSA is suggested [28]. PSO improves the discoverer posi- tation against active and reactive power variations. Under
tion updating rule to improve the search along the start–end power flow changes, the control technique based on the sug-
line. When impediments are encountered, adaptive variable gested methodology enhances the power controller's control
speed escape search is employed to increase path search parameters. The proposed approach, based on fluctuations
efficiency. Adaptive oscillation optimization increases path in the resource and load side characteristics, is used to regu-
smoothing and lowers path fluctuations. Finally, reduc- late the power flow management of the smart-grid system.
ing the nodes and smoothing procedure increases the path The proposed model manages energy resources to meet the
smoothness, making it more acceptable for path planning grid's power needs, including renewable energy and energy
in the actual world. It's also been proven that the SPSA has storage devices.
a faster convergence time and uses less energy than other Figure 7 shows the advantages of hybridization SSA with
algorithms. different algorithms. SSA combines SCA, PSO, DE, and
SSA and TGO [29] have been proposed for network BSS algorithms to solve optimization issues.
security situations to improve the accuracy and perfor- In the hybrid SSA-PSO model, the SSA algorithm uses
mance of the scenario prediction model [30]. TGO-SSA is the gbest agent to update the agents' position. By choos-
used to optimize neural network scenario prediction model ing the best optimal points, the SSA algorithm ignores the

Fig. 7  Advantages of hybridi-

zation SSA with different
algorithms

13
434 F. S. Gharehchopogh et al.

problem of premature convergence in the PSO algorithm.

The SSA-PSO model prevents premature convergence by
enhancing group search capabilities. In the optimization pro-
cess, the SSA-PSO model reduces premature convergence by
determining the optimal value for the parameters.

3.2 Improved

All SSA restrictions are slow convergence, optimum local

entrapment in certain circumstances, and exploration and
exploitation phases that can't cope with the extensive dimen-
sions. Due to the limitations of SSA, numerous approaches
have been developed in recent years to enhance SSA, includ-
ing chaos, Gaussian, Lévy flight, OBL, Random Walk, Strat-
egy mechanism, artificial neural networks (ANNs), and
Deep Learning. Improved strategies for boosting population
variety and speeding convergence by hybridization operators
have shown to be beneficial.

3.2.1 Chaotic
Fig. 8  The most critical chaotic targets in SSA
Turbulent systems are distinguished by their unpredictabil-
ity, periodicity, and parameter sensitivity. Turbulent map-
ping can be used to produce rough numbers between 0 and A crack segmentation approach based on adaptive T-dis-
1 instead of pseudo-random number generators to optimize tribution is utilized to enhance the selection of clustering
the parameters of metaheuristic algorithms. Experiments centres and make the cracks more accurate and comprehen-
have demonstrated that using chaotic sequences for ini- sive segmentation [52]. The simulation demonstrates that
tialization influences the entire algorithm process and that incorporating an adaptive t-distribution mutation mechanism
chaotic sequences produce better results than quasi-random into SSA improves its ability to resist getting caught in local
numbers. Chaotic mapping improves the variety of the spar- optimization. Second, the improved TSSA is utilized as the
row population's starting state. Chaotic mapping prevents K-means algorithm's starting clustering point. After verifica-
premature convergence and increases global optimization tion, the suggested technique considerably enhanced compli-
accuracy and convergence. Figure 8 shows the most critical cated crack pictures' segmentation accuracy and fitness. The
chaotic targets in SSA. Gaussian function's objective is to keep you from slipping
Table 1 shows the improvement of SSA by the chaotic into the local optimum.
method. Items such as the advantages and disadvantages of A multi-objective scheduling model based on an ISSA
chaotic SSA have been analyzed. has been developed to increase the efficiency of Micro-grid
combined heat and power (MCHP) [53]. The objective and
3.2.2 Gaussian economic advantage objectives are initially defined to mini-
mize total operating costs and environmental pollutant dis-
A modified SSA termed CASSA [51] was used for an charge. Then, a dynamic adaptive weight is used instead of
uncrewed aerial vehicle (UAV). The route planning chal- a single weighting technique, and the SSA is combined to
lenge is changed into a multi-dimensional task optimiza- create a multi-objective optimum timetable model for com-
tion issue once the 3D task space model and UAV route prehensive energy. Finally, multiple optimization scenarios
planning cost functions are established. Second, the cha- are constructed to test a typical day operation's suggested
otic approach is used to broaden the algorithm's popula- scheduling optimization model. The results demonstrated
tion, while an adaptive inertia weight is used to balance the that the multi-objective configuration outperformed the
algorithm's convergence rate and exploration capabilities. single-objective setup to meet power grid dispatching cri-
Finally, the Cauchy-Gaussian mutation technique improves teria. The proposed optimization approach improved the
the algorithm's capacity to overcome stagnation. According economic and environmental advantages of the integrated
to simulation data, the CASSA-generated routes are superior energy system. As a result, the Cauchy-Gaussian mutation
to the SSA, PSO, ABC, and WOA. method chooses the current best fitness person for mutation,

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 435

Table 1  Improvement of SSA by Chaotic method

Refs Models Objective Advantages Disadvantages

[32] Improved chaotic SSA (ICSSA) Using an enhanced chaotic SSA, Population diversity* Slow convergence rate*
identify parameters of robot
manipulators with unknown
payloads
Strong global searchability* High execution time*
Good convergence*
[33] CSSA Fault diagnosis Good convergence* High execution time*
[34] IHSSA-ICMIC Optimization of engineering Faster convergence* Achieve a solution in the final
problems iterations*
Strong global searchability*
[35] Logistic chaotic-SSA The chaotic time series prediction Balance between exploration and High execution time*
method exploitation*
Good convergence*
[36] Sin chaotic-SSA Scheduling strategy of regional Good convergence* Achieve a solution in the final
integrated energy model iterations*
Prevent useless search*
Global optimization capability*
[37] CM-SSA Energy optimization in microgrid Population diversity* Slow convergence rate*
Good convergence*
Prevent useless search*
[38] Chaotic sine mapping-SSA Predicting and optimizing net- Population diversity* Achieve a solution in the final
work weights iterations*
Strong global searchability*
Good convergence*
Prevent useless search*
[39] CM-SSA Optimal dispatch strategy of Faster convergence* High execution time*
microgrid energy storage
Strong global searchability*
Short running time*
[40] CSSA Position recognition problem Balance between exploration and Achieve a solution in the final
exploitation* iterations*
Global optimization capability*
[41] CMSSA Global optimization Strong global searchability* High iterations*
Good convergence*
Prevent useless search*
[42] CSSA Dynamic path planning Balance between exploration and High execution time*
exploitation*
Good convergence*
Prevent useless search*
[43] CSSA Solve continuous optimiza- Short running time* Achieve a solution in the final
tion problems and continuous iterations*
dimensions
Balance between exploration and
exploitation*
Global optimization capability*
[44] NCSSA TSP problem Prevent useless search* Slow convergence rate*
Update the situation without get-
ting lost*
Faster convergence*
[45] CSSA Solve continuous optimiza- Balance between exploration and High iterations*
tion problems and continuous exploitation*
dimensions
[46] CSSA-SCN Solve continuous optimization Strong global searchability* High execution time*

13
436 F. S. Gharehchopogh et al.

Table 1  (continued)
Refs Models Objective Advantages Disadvantages

Short running time*

Balance between exploration and
exploitation*
[47] CSSA Solve optimization problems and Update the situation without get- Achieve a solution in the final
strengthen the antenna ting lost* iterations*
Population diversity*
[48] CWTSSA Solve continuous optimiza- Update the situation without get- Slow convergence rate*
tion problems and continuous ting lost*
dimensions
Population diversity*
Strong global searchability*
[49] CMISSA Solve continuous optimization Balance between exploration and Slow convergence rate*
exploitation*
[50] CSSA Solve continuous optimization Update the situation without get- High iterations*
ting lost*

The asterisk (*) indicates the number of items

compares the positions before and after the mutation, and SSA to leap out of local optimums, and Powell local search
chooses the better position to move on to the next iteration. is used to improve the method's convergence. Finally, the
ISSA-SVM is an ISSA that solves the problem of SVM simulation results indicated that the revised algorithm's
hyperparameter selection and constructs the mid-long term positioning error is considerably decreased compared to the
load prediction model [54]. A novel dynamic adaptive original DV-Hop method, and positioning accuracy is effec-
t-distribution mutation improves the ISSA. The ISSA offers tively enhanced in irregular regions.
greater convergence precision, stability, and speed than Bernoulli's chaotic mapping, LF, mutation, crossover,
the SSA, evidenced by a comparison test using six bench- competition, and enhanced SSA are designed to obtain an
mark functions. The simulation results demonstrated that ideal energy configuration [56]. The results of several test
the ISSA-SVM successfully enhanced prediction accuracy functions and assessment indicators illustrate the superiority
compared to the original SVM, BP neural network, multiple of the upgraded SSA. The suggested method's usefulness is
linear regression, and other methods. The Gaussian function demonstrated by solving and assessing the best configura-
has always detracted from local optimism and population tion of an energy management model. The case study results
variety. reveal that an active distribution network with a multi-micro-
grid system offers significant economic and environmental
3.2.3 Levy Flight (LF) benefits under many scenarios. The proposed solutions are
critical for a multi-microgrid dynamic distribution network's
The LF technique improves the multi-objective SSA's capac- effective operation and environmental conservation.
ity to jump out of the local optimum. The simulation results The LF operation is added to the original SSA's pro-
confirmed the upgraded multi-objective SSA's efficacy. The ducer's search process to improve the algorithm's ability
wind–solar-diesel–storage micro-cost grids and loss were to hop out of the local optimum and optimize performance
reduced using a multi-objective function. The correct step [57]. The opposition-based learning (OBL) technique pro-
size for an LF is crucial since a big step size might make duces better SSA solutions, which helps to speed up the
individuals meander around the local optimal amount and be algorithm's convergence speed. On the one hand, numeri-
unable to find the ideal answer. An extremely tiny step size, cal tests based on traditional benchmark functions assess
on the other hand, may render the ideal value unachievable the LOSSA's performance. On the other hand, the Support
and result in merely a local optimization. On the other hand, Vector Machine (SVM) hyper-parameter optimization task is
LF's search approach in SSA is fully random. used to verify LOSSA's capacity to tackle actual situations.
A DV-Hop method is suggested to optimize using the The LOSSA is feasible and successfully handles machine
improved SSA (ISSA) [55]. The maximum hop distance learning algorithms' hyper-parameter optimisation problem.
error is utilized to adjust the hop distance from the unknown This study proposes an improved SSA based on LOSSA.
node to each anchor node to decrease the estimated distance The LOSSA's overall performance is good, according to the
error. Second, LF is used to improve the capacity of the experimental data. The LOSSA surpassed the SSA and other

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 437

natural heuristic algorithms in search accuracy, convergence it incorporates the Metropolis criteria into the simulated
speed, and stability. annealing (SA) process, allowing the system to accept new
Reverse learning and the LF random step are used in the solutions by determining whether to take them, allowing the
SPISSA [58]. In the initialization stage, this model increased algorithm to escape the local optimum and improve global
the variety and quality of sparrows, and in the subsequent search capacity. OBL is utilized to construct a reverse solu-
iteration stage, it improved the global search capabilities. tion to create a new sparrow population, increase population
Finally, the technique is integrated with the Spark distrib- variety, and enhance population quality [62]. Simultane-
uted computing framework to account for network incur- ously, the OBL can be utilized to direct the algorithm away
sion traffic's high-dimensional and large-scale aspects. The from the local optimum. Finally, the ISSA's performance is
population is estimated in parallel in the Spark framework tested using 2D grid maps of various specifications created
based on the data partition. Experiments have shown that on the MATLAB platform. The simulation results demon-
SPISSA can identify the best subset from the public data strate that the ISSA outperforms the SSA, PSO, and other
set. At the same time, the algorithm's computation time has standard intelligence algorithms in optimal performance and
been significantly lowered. can successfully jump out of the local optimum.
In the WSN, DV-Hop is a frequently utilized position- An improved SSA (ISSA) is used to suggest a distributed
ing method [59]. An ISSA is provided based on DV-Hop maximum power point tracking (DMPPT) [63]. First, the
wireless sensor network positioning technology. There are population was initialized using the centre of gravity reverse
two critical factors to DV advancement. The double com- learning technique, resulting in a superior spatial solution
munication radius method modifies the minimum hop distribution. Second, the learning coefficient was added to
count between nodes to reduce the estimated distance error; the discoverer's position update section to increase the algo-
second, instead of using the least-squares method. This rithm's global search capability. Simultaneously, the muta-
improved algorithm employs SSA to evaluate nodes' posi- tion operator was employed to improve the joiner's position
tions; simultaneously, SSA employs the Lévy flight strategy update and prevent the algorithm from sliding into the local
to improve performance further. Finally, simulation is used extreme value. The initial sparrow population is generated
to assess the method's placement accuracy, and the results using random initialization in the conventional SSA. The
show that ISSA-LF is superior. quality of the starting population influences the end con-
vergence accuracy for the intelligent algorithm of popula-
3.2.4 Opposition‑Based Learning (OBL) tion iteration. The initial population of the SSA is gener-
ated using centroid OBL (COBL), which assures the initial
DV-Hop is a widely used placement strategy in the WSN population's homogeneity and variety while also improving
[60]. Based on DV-Hop wireless sensor network positioning its fitness. The model's findings revealed that the ISSA could
technology, an ISSA is supplied. DV progress is dependent track the maximum power point (MPP) more precisely and
on two variables. The double communication radius method fast than the perturbation observation technique (P&O) and
modifies the minimum hop count between vertices to reduce the PSO and has superior steady-state performance.
the estimated distance error; second, instead of using the The ROSSA model [64] is based on SSA paired with
least-squares method. This enhanced algorithm utilizes SSA Random OBL (ROBL) and a linear decreasing approach.
to calculate node positions; simultaneously, SSA uses the The path planning challenge for mobile robots may theo-
Lévy flight strategy to improve performance even further. retically turn into an optimization problem that intelligent
Finally, simulation is used to evaluate the method's accu- optimization systems can handle. An SSA-based optimiza-
racy in terms of placement, and the findings demonstrate tion technique is presented in light of this assumption. The
that ISSA-LF is superior. A new defect diagnostic tech- declining linear approach balances the algorithm's capac-
nique based on LightGBM optimized by the elite opposite ity to search worldwide and exploit locally by altering the
SSA(EOSSA) [61] is proposed. The change in data distribu- algorithm parameters. ROBL increases the variety of the
tion is frequently ignored by dimension reduction methods population and improves the algorithm's exploration capa-
based on Euclidean distance. bilities. Trials demonstrate the ROSSA's superiority with
In the case of mobile robot route planning, population three conventional algorithms for 11 benchmark test func-
variety is insufficient in later rounds, making it simple to tions and comparative studies with PSO and SSA on the path
settle into a locally optimum solution. Using the sparrow planning issue.
search method, an enhanced SSA is presented to address A learning SSA (LSSA) [65] is introduced in the dis-
these issues in mobile robot path planning. First, the algo- coverer stage. The random reverse learning technique
rithm optimizes the initial population of sparrows using promotes population variety and flexibility in the search
OBL, which increases the quality of the initial solution and process. An upgraded sine and cosine guiding mechanism
improves the system's local search capabilities. Second, is added at the follower stage to make the discoverer's

13
438 F. S. Gharehchopogh et al.

search approach more thorough. Finally, we suggest a dif- 3.2.6 Strategy Mechanism

ferential-based local search. The method updates the best
solution acquired each time to avoid omitting high-quality A strategy mechanism combines local and global search by
solutions throughout the search process. In 12 benchmark strengthening the update of individuals in the community
functions, LSSA was compared against CSSA, ISSA, SSA, to search more in the problem space. This method makes
BSO, GWO, and PSO to ensure the method was feasi- it possible to move towards a global optimizer faster, even
ble. According to the simulation results, LSSA has a high for algorithms with heavy computational fitness functions.
degree of universality. Finally, robot route planning is used The strategy mechanism helps to change the control param-
to verify LSSA's practicability, and LSSA has high path eters in different optimization stages, or even for various
planning quality and reliability. optimization problems, the parameters are adjusted accord-
ing to the ongoing search feedback. This approach selects
an adaptive parameter and thus balances exploration and
3.2.5 Random Walk exploitation (Table 2).

An enhanced SSA employing the random walk approach 3.2.7 Artificial Neural Networks
(RWSSA) [66] is presented to maximize the distribution
and signal coverage of 5G base stations in open-pit mines. ANNs are used to predict and classify various issues such as
RWSSA is compared to SSA, MS-ALO, and PSO. In com- time series, price estimation, weather forecasting, estimat-
parison to other models, the convergence speed and accuracy ing the accuracy of industrial devices, etc. If the structure of
are good. Finally, the RWSSA is superior to previous algo- ANNs is improved, then their efficiency and accuracy will
rithms in numerous ways, making it more appropriate for increase. In this section, the combination of SSA with ANNs
5G base station distribution optimization in open-pit mines. is examined. The SSA algorithm is used to optimize radial-
In the 5G base station deployment optimization challenge, basis function (RBF), extreme learning machine (ELM),
RWSSA achieved superior performance and application. generalized regression neural network (GRNN), and Elman
An RW technique is presented to enhance the SSA [67]. networks. The ELM network is a feed neural network used
The SSA, the grey wolf optimization algorithm, and the for statistical classification, regression analysis, clustering,
WOA are all contrasted and studied. It is confirmed that the approximate spars, comparison, and training. The limitation
ISSA has a quick convergence speed and good optimization of the ELM network is that initial weights and thresholds
precision after a benchmark test function experiment. The are determined using traditional trial-and-error or network
RW strategy improvement sparrow search method's exact search methods. Traditional methods do not work well with
solution is employed as the starting neuron connection inaccuracy, so SSA has been used to solve ELM problems.
weight and threshold information of the BP neural network, Each hidden node has a return edge of its connection in the
which is then used to categorize dangerous URLs and fur- Elman network. Figure 9 shows the SSA schema on ANNs.
ther evaluate the updated technique's viability. The findings Table 3 shows the advantages and disadvantages of SSA
demonstrated that optimising the BP network by optimising with ANNs.
the RW strategy optimization SSA might increase hazardous Recently, neural networks have been widely used to
URL classification accuracy. predict various issues in artificial intelligence. In ANNs,
For optimal model parameter identification of proton a parameter called weight needs to be updated frequently
exchange membrane fuel cell (PEMFC) stacks, a new opti- to avoid significant errors. Since the learning process of
mization approach termed Adaptive SSA (ASSA) [68] is an ANN is strongly related to optimizing a target func-
suggested. The sparrows' locations in the solution space tion, choosing an optimization algorithm is a crucial step
are scattered randomly. An RW technique is used when no in designing the structure of an ANN. Figure 10 shows the
nearby sparrows surround the present individual. The ASSA steps of SSA-BP synthesis. The purpose of SSA is to opti-
is used to minimize the sum of squared error (SSE) between mize the weight of the BP network.
the empirical and estimated stack voltages in the PEMFC
stack by optimally selecting the parameters in the PEMFC 3.2.8 Deep Learning
stack. The approach is used in three case studies: Horizon
H-12, Ballard Mark V, and NedStack PS6 under various Deep learning (DL) has demonstrated advanced perfor-
operating circumstances, yielding SEE values of 0.82, 5.14, mance on various issues. Hyper-parameter settings are
and 0.097, respectively. The ASSA significance is demon- vital in deep learning performance and machine learning
strated by comparing the algorithm's outputs to CGOA, models. Deep neural networks have one or more hidden
GRA, and simple SSA. According to the final data, the pro- layers between the input and output layers. They typically
posed ASSA is the most efficient compared to the others. apply nonlinear transformations or activation functions

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 439

Table 2  Improving SSA with strategic methods

Refs Models Strategy Results Global convergence Exploration vs exploita- Complexity
tion

[69] EEMD-Tent-SSA-LS- Tent chaotic mapping, Wind power prediction Moderate Tuning dependent Moderate
SVM t-distribution
[70] Mixed Strategy SSA Non-linear adjustment, Increase positioning Slow Moderate High
(MSSSA) random distribution accuracy
[71] ISSA-gradient boosting gradient boosting Optimization issues Tuning dependent Tuning dependent Moderate
regression tree tech- regression tree, in the engineering
nique (ISSA-GBRT) t-distribution industry
[72] ISSA Neighborhood search Path planning approach Slow Less diverse solutions Moderate
strategy for mobile robots
[73] ESSA-DELM Trigonometric substitu- Optimization of engi- Fast Good Moderate
tion strategy and neering and continu-
Cauchy mutation ous issues
[74] ISSA Mutation, random Optimal reactive Slow Less diverse solutions High
distribution power dispatch and
distributed generation
placement
[75] ISSA Position updating Solve engineering and Fast Tuning dependent Moderate
strategy dynamic problems
[76] LLSSA Inverse learning Optimization of engi- Slow Good Low
strategy, spiral search neering and continu-
strategy ous issues
[77] ISSA Iterative local search Optimization of engi- Fast Tuning dependent High
strategy, a greedy neering and continu-
strategy ous issues
[78] IMSSA Position updating Sequential quadratic Tuning dependent Moderate Moderate
strategy programming for solv-
ing the cost minimi-
zation
[79] SSACBR t-distribution mutation Prediction of statistical Tuning dependent Good Low
operator, memetic data
algorithm, Case-based
reasoning
[80] EMSSA Hazard-aware transfer- Continuous optimiza- Slow Less diverse solutions Moderate
ring strategy, dynamic tion problems
evolutionary strategy,
uniformity-diversi-
fication orientation
strategy
[81] adaptive spiral flying Variable spiral search Optimization of engi- Tuning dependent Tuning dependent Moderate
SSA (ASFSSA) strategy neering and continu-
ous issues
[82] ISSA Position updating Optimization of engi- Moderate Tuning dependent Low
strategy neering and continu-
ous issues
[83] ISSA Mutation, random Water quality prediction Slow Tuning dependent Moderate
distribution

(logistics, tanh, or ReLU). In DL, hyperparameters include of meta-parameters can be done using meta-heuristic algo-
the number of layers, the neurons in each layer, the acti- rithms. GRU, LSTM, and CNN are the most important
vation function, the learning rate, the deletion rate, and deep learning networks. Table 4 shows the combination
the batch size. There is no optimal general configura- of SSA with deep learning algorithms.
tion for hyperparameter optimization. These parameters Figure 11 shows the percentage of Improved SSA based
can be optimized manually but are time-consuming and on different methods. The Chaotic and Strategy models'
require specialized knowledge. Automated optimization percentage is higher than Gaussian, LF, OBL, and Random

13
440 F. S. Gharehchopogh et al.

technique based on a multi-objective SSA (MOSSA) [122]

has been developed (ADN). Distributed generation and
time-varying loads can aid in sustainable development and
energy conservation. As a result, this study investigates the
ADN integrated optimization problem while considering
distributed generation and time-varying demand to improve
ADN power quality, economics, and energy savings. The
supremacy of the proposed MOOSSA for the multi-objec-
tive, multi-constraint, non-linear, high-dimensional ADN
integrated optimization problem is first proven. Second, the
ADN mathematical model for integrated optimization is cre-
ated. The MOOSSA reduced power loss and node voltage
fluctuation by 75.76% and 70.06%.
Fig. 9  SSA schema on ANNs
MOSSA is used to efficiently manage the functioning of
a microgrid (MG) [123]. This paper presents two optimiza-
Walk models. The percentage of using the Strategy Mecha- tion problems. The first is a single-objective issue that tries
nism model with SSA is lower than Chaotic. to reduce the overall operating cost or the total emission
from the system. The second issue is a multi-objective prob-
3.3 Variants of SSA lem that simultaneously includes total operational costs and
emissions. There is a new version of SSA available. To man-
3.3.1 Binary age the energy of the MG optimally, photovoltaic modules
(PV), wind turbines (WT), fuel cells (FC), micro-turbines
A Discrete SSA (DSSA) [121] with a global perturbation (MT), batteries (BSS), and the grid are all integrated into
technique has been developed to solve the TSP. The roulette- the MG. The suggested method is statistically tested using
wheel selection generates the population's first solution. The Friedman and Kruskal–Wallis ANOVA tests in non-para-
order-based decoding approach is then added to finish the metric analysis. In addressing the single objective issue, the
sparrow position update. The global perturbation technique recommended SSA achieved cost and emission depreciation
is used with Gaussian mutation and swap operator to bal- of 1.44% and 54.76%, respectively, compared to Krill Herd
ance exploration and exploitation capacity. Finally, the 2-opt (KH). In the multi-objective problem, the proposed MOSSA
local search enhances the solution's quality. These tactics saved 42.78% operating expenses and 0.118% emissions
improve the quality of the solution and speed up the conver- compared to ALO. The critical findings indicated SSA's
gence process. Experiments were conducted using 34 TSP resiliency in regulating the created MG's functioning.
benchmark datasets. A MOSSA-based wireless sensor clustering and routing
Furthermore, statistical tests confirm the significant dif- protocol model (MUSHROOM) has been presented [124].
ferences between the DSSA and other current approaches. A fitness function based on maximum neighbour node dis-
According to the results, the suggested strategy is more tance, the average distance to BS, and energy ratio are shown
competitive and resilient in solving the TSP. A novel DSSA to conduct clustering. With the excess energy of the next-
algorithm has been augmented with the genetic operator and hop node, sink distance, and node degree, MUSHROOM
local search to the robot route. Comparing the DSSA to other applied the fitness function to the routing process. The sug-
approaches (Hybrid FA, PSO, Adaptive ABC, etc.) revealed gested model has undergone extensive testing to guaran-
that the enhanced algorithm outperformed the others in the tee that it has achieved maximum energy competence and
examples examined. network longevity when compared to other techniques with
varied numbers of nodes.
3.3.2 Multi‑objective Optimization MOSSA [125] is a suggested multi-objective variation of
the SSA. MO-SSA performed well compared to other well-
Single or many designs are commonly involved in engineer- known optimization methods in tests (NSGA-II, NSGA-III,
ing optimization problems. Multi-objective optimization, as and MO-ALO). The MOSSA outperformed most baseline
opposed to single-objective optimization, which tries to dis- algorithms on numerous performance parameters.
cover the best solution to a given issue using an objective
function, includes optimizing two or multi-objective func- 3.4 Optimization Problems
tions and offering optimum solutions.
To the dynamic reconfiguration integrated optimization The term "optimization" refers to selecting the best option
model of an active distribution network, a novel solution from a set of alternatives. Engineers search for ideal

13
 Table 3  Combination of SSA with ANNs
Refs Models Objective Advantages Disadvantages

[84] SSA-RBF Predicting the temperature of the sensors Find the optimal value for RBF parameters Non-optimal updates of individual
ISSA-RBF
Reduce data training time Achieve a solution in the final iterations
Increase detection accuracy
Error reduction
[85] SSA-ELM The SSA-ELM model predicts the uniaxial Increase forecast accuracy Non-optimal updates of individual
compressive strength (UCS) of the cemented
paste backfill (CPB) under different condi-
tions
Discover the optimal value for ELM param- High execution time
eters
Settings for the number of layers and the
number of nodes
[86] Firefly Algorithm SSA (FASSA-GRNN) Prediction of industrial and laboratory materi- Enhance SSA search capability using FA Slow convergence rate
als
Advances in Sparrow Search Algorithm: A Comprehensive Survey

Determining the optimal weight for GRNN

Reduce the amount of output error
[87] SSA-ENN The SSA-ENN strategy can improve road Reduce data training time Achieve a solution in the final iterations
capacity and traffic stability
Increase detection accuracy
Error reduction
[88] SSA-BP The proposed SSA-BP algorithm can charac- Find the optimal value for RBF parameters* Slow convergence rate
terize the critical deformation dimensions
(height, length, tilt angle) within the mean
relative error of 10%
Reduce data training time
Increase detection accuracy
Error reduction
[89] FA-SSA-BPNN Optimization of sensor features and model Reduce data training time Achieve a solution in the final iterations
parameters
Improve accuracy in data training
[90] ICEEMD-SSA-BPNN Predicting the price of carbon and industrial Settings for the number of layers and the High execution time
materials number of nodes
Improve the internal structure of the network
Increase detection accuracy
Improve accuracy in data training
[91] WMF-SSA-MLELM Short-term multistep wind speed forecasting Reduce data training time Achieve a solution in the final iterations
Improve accuracy in data training

13
441
 Table 3  (continued)
442

Refs Models Objective Advantages Disadvantages

13
Find the optimal value for network parameters
[92] SSA-BP predicting possible threats based on com- Improve the internal structure of the network High execution time
mander mood (PTP-CE)
Improve accuracy in data training
Reduce data training time
[93] CMSSA-Elman Short-term PV Power Forecasting Based on Settings for the number of layers and the Achieve a solution in the final iterations
Time-Phased and Error Correction number of nodes
Error reduction
[94] SSA-BP Optimization of the BP Neural Network Algo- Settings for the number of layers and the Slow convergence rate
rithm with SSA for the Processing of Coal number of nodes
Mine Water Source Data
Reduce data training time
Find the optimal value for network parameters
[95] SSA-ELM Predicting air pollution Increase detection accuracy Non-optimal updates of individual
Improve the internal structure of the network
Find the optimal value for network parameters
[96] SSA-BP Based on the SSA-BP Neural Network, an Reduce data training time High execution time
assessment algorithm for network security is
developed
Improve accuracy in data training
[97] Tent Cauchy SSA (TCSSA-BP) Regression prediction of material grinding Settings for the number of layers and the Achieve a solution in the final iterations
particle size number of nodes
Reduce data training time
[98] SSA-KELM Intelligent Fault Diagnosis Error reduction High execution time
Improve accuracy in data training
[99] SSA-DBN Predictability and accuracy of diagnosis Find the optimal value for network parameters Slow convergence rate
Settings for the number of layers and the
number of nodes
Error reduction
Improve accuracy in data training
[100] SSA-BP Forecasting hydropower generation Improve accuracy in data training Slow convergence rate
Improve the internal structure of the network
Error reduction
[101] SSA-KELM From water quality assessment to environmen- Settings for the number of layers and the High iterations
tal water quality management number of nodes
Improve the internal structure of the network
[102] SSA-BP neural network Prediction of industrial and laboratory materi- Increase detection accuracy Non-optimal updates of individual
als
F. S. Gharehchopogh et al.
 Advances in Sparrow Search Algorithm: A Comprehensive Survey 443

Non-optimal updates of individual

Slow convergence rate

Slow convergence rate

Find the optimal value for network parameters High iterations
Disadvantages

Find the optimal value for network parameters

Improve accuracy in data training

Improve accuracy in data training

Improve accuracy in data training

Classification of flame images

Increase detection accuracy

Good optimization ability

Reduce data training time

Reduce data training time

Error reduction

Error reduction

Error reduction
Advantages

Predicting the boiling point temperature of

Fig. 10  SSA-BP combination flowchart [90]

Wind and solar power forecasting

parameters to save time or expense while improving their

designs' operational efficiency, effectiveness, excellence,
Blood glucose prediction

or lifespan. Traditional approaches such as regular linear

Fire flame recognition

programming are acceptable for fundamental optimization

problems, and some employ gradient information to identify
working fluid

the best solution. Still, real-world engineering optimization

problems are often nonlinear, indistinguishable, complex,
Objective

and multifaceted. Solving these problems using classical

optimization methods is relatively difficult. Hence, meta-
heuristic algorithms that do not require gradient information
are needed to solve them. The optimization section uses the
[106] ISSA-FSCN(Fast stochastic configuration

SSA algorithm to solve forecasting, error detection, energy

management, complex optimization, clustering, scheduling,
and object detection in engineering disciplines. The algo-
rithm has proven its effectiveness in most issues. Table 5
provides an overview of SSA in optimization.
Figure 12 shows the percentage of SSA application in dif-
ferent areas of optimization. As shown in Fig. 12, the highest
rate of SSA utilization in Complex Optimization belongs
Table 3  (continued)

[105] SSA-KELM

to optimization in continuous and discrete problems. Fault

network)
[103] SSA-BP

[104] SSA-BP
Models

Diagnosis is equal to 15%, which belongs to industrial and

advanced issues. Clustering and Object Recognition are
equal to 8%. The lowest percentage belongs to Location
Refs

Optimization, and 5% is obtained.

13
444 F. S. Gharehchopogh et al.

Table 4  Hybridization of SSA with deep learning algorithms

Refs Models Objective Advantages Disadvantages

[107] SSA-BI-GRU​ Bidirectional GRU (Bi-GRU) and Improve accuracy in data training High iterations
time-series production forecasting
approach based on the integration
of (SSA)
Reduce data training time
Error reduction
[108] LSTM-SSA Short-term wind speed forecasting Find the optimal value for network Problem of Overfitting with an
parameters increasing number of iterations
Increase detection accuracy
Improve accuracy in data training
Increase detection accuracy
[109] SCGRU-HSSA Recognition of a linear source Improve accuracy in data training Reduction of performance of middle
contamination neurons by increasing repetitions
Settings for the number of layers
and the number of nodes
[110] SSA-CNN COVID-19 diagnosis and categori- Error reduction Non-optimal updates of individual
zation based on chest CT scans
High execution time
[111] VMD-ISSA-GRU​ Short-Term Photovoltaic Power Increase detection accuracy Reduction of performance of middle
Forecasting neurons by increasing repetitions
Find the optimal value for network
parameters
Error reduction
Improve the internal structure of
the network
[112] CEEMDAN-SSA-GRU​ Wind power prediction Find the optimal value for network Problem of Overfitting with an
parameters increasing number of iterations
Reduce data training time
Error reduction
[113] BSSA-CNN Optimal brain tumour diagnosis Improve the internal structure of High iterations
based on deep learning the network
Reduce data training time
[114] IMEFD-ODCNN-SSA Design fall detection systems for Error reduction Reduce network speed in detecting
smart homecare samples
Improve the internal structure of
the network
[115] TA-SSALSTM Electric vehicle load forecast Improve accuracy in data training Reduce network speed in detecting
samples
Settings for the number of layers
and the number of nodes
[116] ESSA-CNN Optimal brain tumour detection Find the optimal value for network Reduction of performance of middle
parameters neurons by increasing repetitions
Reduce data training time
Error reduction
[117] SWT-ISSA-LSTM Water quality prediction Error reduction High execution time
Improve the internal structure of
the network
[118] LSTM-SSSA Accurate ultra-short-term wind Increase detection accuracy Reduce network speed in detecting
speed prediction samples
Find the optimal value for network
parameters
Error reduction
Improve the internal structure of
the network

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 445

Table 4  (continued)
Refs Models Objective Advantages Disadvantages

[119] SSA-LSTM Residential high-power load Find the optimal value for network High execution time
prediction parameters
Reduce data training time
Error reduction
[120] ISSA-DELM Accurate damage degree prediction Find the optimal value for network High iterations
parameters
Error reduction

which leads to minor improvements in modelling. Studies

have shown that the search process is better with methods
such as Chaotic and OBL.
The explorers update manufacturers' search positions to
direct the next search. In each iteration, the solutions pro-
duced in previous generations can be used as a historical
experience to guide explorers' investigations. The histori-
cal background of the manufacturers not only shows the
explorers to the optimal points but also prevents the algo-
rithm from being placed in the local optimal. Better use of
manufacturers' experience leads to increased performance in
convergence and optimization. In addition, since the Crawler
Update Strategy may not apply to a variety of problems,
Vanguard agents improve the effect of the current strategy,
which is used as feedback information to set optimization
Fig. 11  Percentage diagram of improved SSA based on different strategies. Figure 13 shows the percentage of SSA methods
methods based on four different areas.
Table 6 shows the general advantages and disadvantages
of the SSA algorithm. The SSA algorithm also suffers from
4 Discussion the problem of operational incompatibility in complex
issues, and sometimes the accuracy of the solution is often
Global exploration and local exploitation are critical com- unsatisfactory at the specified time required.
ponents of the SSA algorithm's search process. An equal Various meta-heuristic algorithms have been successfully
balance of exploration and exploitation must be struck to developed in recent decades to solve optimization problems.
obtain a beneficial performance. In the early phases of the The SSA algorithm uses a variety of strategies to balance
search process, search agents should always be deployed global exploration and local exploitation. In particular, the
across the search space. They must, however, converge on opposite learning strategy increases the search scope in the
the best candidate in the following stages of the search decision space. The gaussian approach is used to improve
process. Agent During the global search phase, it changes the performance of elite solutions. At the same time, random
its location by learning from the best international agent search is used to diversify people in the community. Stud-
to expedite convergence and increase exploitation ability. ies have shown that improved SSA has a more vital ability
On the other hand, the individual at the local stage renews to find better solutions than SSA. Hence, it is an effective
his position by concurrently learning from present and ran- evolutionary optimizer with a robust search capability and
dom elements to promote population variety. Studies have convergence rate for global optimization problems. The SSA
shown that exchanging the global search section and the algorithm has been improved by including hybrid factors to
local search process uses a random distribution, sometimes update positions to increase the utilization rate. In addition,
losing the best agent in the search space. SSA uses algorithms such as PSO and DE to increase the
Random search makes it possible to explore more sig- ability to exploit and explore the approach.
nificant parts of the search space than the local approxi- The main disadvantage of SSA is the fast or sometimes
mations created in gradient-based optimization. However, slow coverage of the problem search space, and a proper
random search often leads to a large number of analyzes, balance between the search steps is not established. This
requires modified methods by creating an appropriate

13
446 F. S. Gharehchopogh et al.

Table 5  An overview of SSA in the field of optimization

Refs Models Application Advantages Disadvantages

[126] SSA-DBN Prediction Strong global searchability Non-optimal updates of individual

Update the situation without Slow convergence rate
getting lost
[127] Multipoint optimal minimum Fault Diagnosis Prevent useless search Achieve a solution in the final
entropy deconvolution adjusted iterations
(MOMEDA-SSA)
Better solution than other exist- High iterations
ing techniques
[128] SSA Prediction Update the situation without Non-optimal updates of individual
getting lost
Good convergence
Short running time
Update the situation without
getting lost
[129] SSA Energy management system Group and intelligent search Achieve a solution in the final
towards the optimal solution iterations
Balance between exploration and High iterations
exploitation
Better solution than other exist-
ing techniques
[130] SSA Complex optimization Strong global searchability High execution time
Update the situation without Achieve a solution in the final
getting lost iterations
[131] SSA Optimal scheduling Prevent useless search Non-optimal updates of individual
Better solution than other exist-
ing techniques
[132] SSA Object recognition Balance between exploration and Non-optimal updates of individual
exploitation
Update the situation without Slow convergence rate
getting lost
Population diversity
[133] SSA Complex optimization Global optimization capability Achieve a solution in the final
iterations
Balance between exploration and High iterations
exploitation
High quality of solution and
computation efficiency
[134] ISSACPM (control parameteri- Complex optimization Group and smart search towards Slow convergence rate
zation method (CPM)) the optimal solution
Balance between exploration and
exploitation
Better solution than other exist-
ing techniques
[135] IVMD-MSE-SSA-ELM Prediction Prevent useless search Non-optimal updates of individual
Better solution than other exist-
ing techniques
[136] SSA Energy management system Balance between exploration and Non-optimal updates of individual
exploitation
Update the situation without Slow convergence rate
getting lost
Population diversity
[137] SSA Complex optimization Global optimization capability Achieve a solution in the final
iterations

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 447

Table 5  (continued)
Refs Models Application Advantages Disadvantages

Balance between exploration and High iterations

exploitation
High quality of solution and
computation efficiency
[138] SSA Object recognition Prevent useless search Non-optimal updates of individual
Better solution than other exist-
ing techniques
[139] ISSA Energy management system Faster convergence Non-optimal updates of individual
Global optimization capability Slow convergence rate
[140] SSA Energy management system Strong global searchability High execution time
Update the situation without Achieve a solution in the final
getting lost iterations
[141] ISSA Complex optimization Prevent useless search Non-optimal updates of individual
Better solution than other exist-
ing techniques
[142] SSA Clustering Group and intelligent search Achieve a solution in the final
towards the optimal solution iterations
Balance between exploration and High iterations
exploitation
Better solution than other exist-
ing techniques
[143] ISSA Object recognition Group and intelligent search Non-optimal updates of individual
towards the optimal solution
Balance between exploration and
exploitation
Better solution than other exist-
ing techniques
[144] ISSA Location optimization Prevent useless search Non-optimal updates of individual
Better solution than other exist- Slow convergence rate
ing techniques
[145] LEACH-Wireless Gateway Rota- Clustering Global optimization capability High execution time
tion (WGR)-SSA
Balance between exploration and
exploitation
High quality of solution and
computation efficiency
[146] SSA-based Resource Manage- Optimal scheduling Strong global searchability Slow convergence rate
ment (SSARM)
Update the situation without
getting lost
[147] ISSA Complex optimization Group and intelligent search Achieve a solution in the final
towards the optimal solution iterations
Balance between exploration and High iterations
exploitation
Better solution than other exist-
ing techniques
[148] Active Disturbance Rejection Complex optimization Prevent useless search Slow convergence rate
Control (LADRC-SSA)
Better solution than other exist-
ing techniques
[149] SSA Prediction Faster convergence Non-optimal updates of individual
Global optimization capability
[150] SSA-PID Complex optimization Balance between exploration and Achieve a solution in the final
exploitation iterations

13
448 F. S. Gharehchopogh et al.

Table 5  (continued)
Refs Models Application Advantages Disadvantages

Update the situation without High iterations

getting lost
Population diversity
[151] SSA-XG-Boost Prediction Global optimization capability Non-optimal updates of individual
Balance between exploration and
exploitation
High quality of solution and
computation efficiency
[152] ISSA Fault diagnosis Balance between exploration and Achieve a solution in the final
exploitation iterations
Update the situation without High iterations
getting lost
Population diversity
[153] SSA Optimal scheduling Prevent useless search Non-optimal updates of individual
Better solution than other exist-
ing techniques
[154] SSA-SVM Fault diagnosis Update the situation without High execution time
getting lost
Good convergence
Short running time
Update the situation without
getting lost
[155] SSA Fault diagnosis Group and intelligent search Achieve a solution in the final
towards the optimal solution iterations
Balance between exploration and High iterations
exploitation
Better solution than other exist-
ing techniques
[156] SSA Optimal scheduling Prevent useless search High execution time
Better solution than other exist-
ing techniques
[157] SSA Fault diagnosis Balance between exploration and Non-optimal updates of individual
exploitation
Update the situation without
getting lost
Population diversity
[158] SSA Optimal scheduling Group and intelligent search High execution time
towards the optimal solution
Balance between exploration and
exploitation
Better solution than other exist-
ing techniques
[159] SSA Clustering Faster convergence Non-optimal updates of individual
Global optimization capability
[160] SSA Complex optimization Prevent useless search Slow convergence rate
Better solution than other exist-
ing techniques
[161] SSA-LA (SSA Based on Locali- Location optimization Strong global searchability Slow convergence rate
zation Algorithm)
Update the situation without
getting lost
[162] SSA Energy management system Faster convergence Non-optimal updates of individual
Global optimization capability

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 449

Table 5  (continued)
Refs Models Application Advantages Disadvantages

[163] SSAE-SSA-SVM Fault diagnosis Update the situation without High execution time
getting lost
Good convergence
Short running time
Update the situation without
getting lost
[164] SSA Complex optimization Faster convergence Non-optimal updates of individual
Global optimization capability
[165] SSA Threshold image segmentation Prevent useless search High execution time
Better solution than other exist-
ing techniques
[166] SSA Wireless sensor network cov- Nationwide coverage of the Non-optimal updates of individual
erage optimization network
Good convergence

operator for the SSA equations. OBL is one of the suitable

methods for the weaknesses of SSA. OBL-SSA aims to deal
with the drawbacks such as local search area confinement,
premature convergence, and balancing of the search process.
There are two main reasons for the weakness of SSA in high-
dimensional spaces. The first is the poor convergence rate.
The ability to search the high-dimensional target space is
insufficient. The second is diversity preservation because
preserving diversity in a high-dimensional target space with
a vast search space is challenging. Hence, effective mainte-
nance of diversity in SSA is necessary to deal with multi-
objective problems.

Fig. 12  Percentage of SSA application in different areas of optimiza- 5 Conclusions and Future Works
tion

Optimization problems have attracted the scientific commu-

nity to various meta-heuristic algorithms. As the complexity
of the problems increases, the need for new metaheuristic
algorithms has become very acute. Optimization aims to
find the best solution to a problem of all possible values to
​​
maximize or minimize output. Many researchers use practi-
cal meta-heuristic algorithms to find the optimal solution to
obtain the global optimal. Because the search space grows
exponentially in real optimization problems and the prob-
lem perspective becomes multidimensional, meta-heuristic
algorithms are viable for generating optimal solutions. This
paper examines the SSA algorithm from hybridization,
Improved, Variants, and Optimization aspects. The SSA
algorithm can run on most optimization problems due to
its ease of implementation and rapid increase in the spread
of agents in the problem space. Studies have shown that
SSA uses the scout search concept, making it possible to
Fig. 13  Percentage of SSA methods based on four different areas track population characteristics in the optimization process.

13
450 F. S. Gharehchopogh et al.

Table 6  Advantages and Factors

disadvantages of the SSA
algorithm Advantages ✓ A few parameters and simple implementation
✓ Excellent performance for optimization problems
✓ High-quality of solutions
✓ Good convergence properties and low generation costs
✓ The principle of balance between exploitation and exploration
✓ Short computational time
✓ Prevent the premature convergence
✓ Getting quality results effectively in less computational time
✓ Diversity of the population
✓ The balance between local seek and global seek
✓ SSA is highly competitive in finding optimal values
Disadvantages ✓ Incomplete exploitation in the solution of complex problems
✓ Incoherence in the local and global seek
✓ Increase of iterations with increasing the size of the issues

WOA, FA, SCA, and DE may suffer from premature con- for non-decomposition large-scale global optimization. Expert
vergence, stagnation, and sensitivity to the formulation. In Syst Appl 198:116895
9. Gharehchopogh FS (2022) Advances in tree seed algorithm: a
addition, WOA and FA contain more internal parameters comprehensive survey. Arch Comput Methods Eng 29:3281
than SSA, which, if not adjusted correctly, can reduce the 10. Banaie-Dezfouli M, Nadimi-Shahraki MH, Beheshti Z (2021)
efficiency of the exploratory value in the optimization pro- R-GWO: representative-based grey wolf optimizer for solving
cess. The results showed that the domain of Improved had engineering problems. Appl Soft Comput 106:107328
11. Goldanloo MJ, Gharehchopogh FS (2022) A hybrid OBL-based
the best performance. Chaotic, Gaussian Mutation, Levy firefly algorithm with symbiotic organisms search algorithm
flight, OBL, Random Walk, and Strategy mechanism meth- for solving continuous optimization problems. J Supercomput
ods have improved SSA. Orienting future work to complex 78(3):3998–4031
and improved algorithms to solve the optimization problem 12. Samadi Bonab M, Ghaffari A, Soleimanian Gharehchopogh F,
Alemi P (2020) A wrapper-based feature selection for improving
will be complicated. performance of intrusion detection systems. Int J Commun Syst
33(12):e4434
13. Nadimi-Shahraki MH, Taghian S, Mirjalili S, Faris H (2020)
References MTDE: an effective multi-trial vector-based differential evolution
algorithm and its applications for engineering design problems.
Appl Soft Comput 97:106761
1. Gharehchopogh FS (2022) An improved tunicate swarm algo- 14. Shayanfar H, Gharehchopogh FS (2018) Farmland fertility: a
rithm with best-random mutation strategy for global optimization new metaheuristic algorithm for solving continuous optimization
problems. J Bion Eng 19:1177 problems. Appl Soft Comput 71:728–746
2. Zamani H, Nadimi-Shahraki MH, Gandomi AH (2021) QANA: 15. Abdollahzadeh B, Gharehchopogh FS, Mirjalili S (2021) Afri-
quantum-based avian navigation optimizer algorithm. Eng Appl can vultures optimization algorithm: a new nature-inspired
Artif Intell 104:104314 metaheuristic algorithm for global optimization problems. Com-
3. Gharehchopogh FS, Gholizadeh H (2019) A comprehensive sur- put Ind Eng 158:107408
vey: whale optimization algorithm and its applications. Swarm 16. Zamani H, Nadimi-Shahraki MH, Gandomi AH (2022) Starling
Evol Comput 48:1–24 murmuration optimizer: a novel bio-inspired algorithm for global
4. Nadimi-Shahraki MH, Taghian S, Mirjalili S, Zamani H, Bah- and engineering optimization. Comput Methods Appl Mech Eng
reininejad A (2022) GGWO: Gaze cues learning-based grey wolf 392:114616
optimizer and its applications for solving engineering problems. 17. Xue J, Shen B (2020) A novel swarm intelligence optimiza-
J Comput Sci 61:101636 tion approach: sparrow search algorithm. Syst Sci Control Eng
5. Gharehchopogh FS, Shayanfar H, Gholizadeh H (2020) A com- 8(1):22–34
prehensive survey on symbiotic organisms search algorithms. 18. Abdollahzadeh B, Soleimanian Gharehchopogh F, Mirjalili S
Artif Intell Rev 53(3):2265–2312 (2021) Artificial gorilla troops optimizer: a new nature-inspired
6. Ghafori S, Gharehchopogh FS (2021) Advances in spotted hyena metaheuristic algorithm for global optimization problems. Int J
optimizer: a comprehensive survey. Arch Comput Methods Eng Intell Syst 36(10):5887–5958
29:1569 19. Wang Z (2022) A parallel particle swarm optimization and
7. Nadimi-Shahraki MH, Fatahi A, Zamani H, Mirjalili S, Oliva D enhanced sparrow search algorithm for unmanned aerial vehicle
(2022) Hybridizing of whale and moth-flame optimization algo- path planning. Res Sq. https://​doi.​org/​10.​21203/​rs.3.​rs-​13755​15/​
rithms to solve diverse scales of optimal power flow problem. v1
Electronics 11(5):831 20. Li H, Zhang B, Li J, Zheng T, Yang H (2021) using sparrow
8. Nadimi-Shahraki MH, Zamani H (2022) DMDE: diversity- search hunting mechanism to improve water wave algorithm.
maintained multi-trial vector differential evolution algorithm

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 451

In: 2021 IEEE International Conference on Progress in Infor- 39. Zheng Y, Liu F (2021) Optimal dispatch strategy of microgrid
matics and Computing (PIC) energy storage based on improved sparrow search algorithm. In:
21. Yang L, Li Z, Wang D, Miao H, Wang Z (2021) Software 2021 40th Chinese Control Conference (CCC)
defects prediction based on hybrid particle swarm optimization 40. Li H, Su J, Liu W, Zhang Y, Zhou X (2021) Indoor positioning
and sparrow search algorithm. IEEE Access 9:60865–60879 model based on support vector regression optimized by the spar-
22. Zhang J, Xia K, He Z, Yin Z, Wang S (2021) Semi-supervised row search algorithm. In: 2021 11th IEEE International Confer-
ensemble classifier with improved sparrow search algorithm ence on Intelligent Data Acquisition and Advanced Computing
and its application in pulmonary nodule detection. Math Probl Systems: Technology and Applications (IDAACS)
Eng 2021:6622935 41. Ma B, Lu P, Zhang L, Liu Y, Zhou Q, Chen Y, Qi Q, Hu Y (2021)
23. Shi L, Ding X, Li M, Liu Y (2021) Research on the capabil- Enhanced sparrow search algorithm with mutation strategy for
ity maturity evaluation of intelligent manufacturing based on global optimization. IEEE Access 9:159218–159261
firefly algorithm, sparrow search algorithm, and BP neural 42. Yang Y, Liu J, Wang Q, Yang S (2021) Dynamic path planning
network. Complexity 2021:5554215 for AGV based on Tent chaotic sparrow search algorithm. In:
24. Liu B, Rodriguez D (2021) Renewable energy systems opti- 2021 International Conference on Information Control, Electrical
mization by a new multi-objective optimization technique: a Engineering and Rail Transit (ICEERT)
residential building. J Build Eng 35:102094 43. Chengtian O, Yujia L, Donglin Z (2021) An adaptive chaotic
25. Kathiroli P, Selvadurai K (2021) Energy efficient cluster head sparrow search optimization algorithm. In: 2021 IEEE 2nd
selection using improved Sparrow Search Algorithm in Wire- International Conference on Big Data, Artificial Intelligence and
less Sensor Networks. J King Saud Univ Comput Inf Sci. Internet of Things Engineering (ICBAIE)
https://​doi.​org/​10.​1016/j.​jksuci.​2021.​08.​031 44. Zheng L, Cui J (2021) A novel chaos sparrow search algorithm
26. Huang S, Huang H (2021) A novel whale optimization algo- for TSP problem. In: 2021 17th International Conference on
rithm with sparrow algorithm and golden sine leading strategy. Computational Intelligence and Security (CIS)
In: 2021 5th Asian Conference on Artificial Intelligence Tech- 45. Jianhua L, Zhiheng W (2021) A hybrid sparrow search algorithm
nology (ACAIT) based on constructing similarity. IEEE Access 9:117581–117595
27. Wu C, Fu X, Pei J, Dong Z (2021) A novel sparrow search 46. Zhang C, Ding S (2021) A stochastic configuration network
algorithm for the traveling salesman problem. IEEE Access based on chaotic sparrow search algorithm. Knowl-Based Syst
9:153456–153471 220:106924
28. Yu W, Liu J, Zhou J (2021) A novel sparrow particle swarm 47. Liang Q, Chen B, Wu H, Ma C, Li S (2021) A novel modi-
algorithm (SPSA) for unmanned aerial vehicle path planning. fied sparrow search algorithm with application in side lobe level
Sci Program 2021:5158304 reduction of linear antenna array. Wirel Commun Mob Comput
29. Henderson N, Rego M, Sacco WR Jr (2015) A new look at 2021:9915420
the topographical global optimization method and its applica- 48. Yang X, Liu J, Liu Y, Xu P, Yu L, Zhu L, Chen H, Deng W
tion to the phase stability analysis of mixtures. Chem Eng Sci (2021) A novel adaptive sparrow search algorithm based on cha-
127:151 otic mapping and T-distribution mutation. Appl Sci 11(23):11192
30. Hu C, Liu G, Li M (2021) A network security situation prediction 49. Tang Y, Li C, Li S, Cao B, Chen C (2021) A fusion crosso-
method based on SA-SSA. In: 2021 14th International Sympo- ver mutation sparrow search algorithm. Math Probl Eng
sium on Computational Intelligence and Design (ISCID) 2021:9952606
31. Kumaravel S, Ponnusamy V (2020) An efficient hybrid technique 50. Song W, Liu S, Wang X, Wu W (2020) An Improved Sparrow
for power flow management in smart grid with renewable energy Search Algorithm. In: 2020 IEEE Intl Conf on Parallel & Distrib-
resources. Energy Sourc Part A Recov Utiliz Environ Effects. uted Processing with Applications, Big Data & Cloud Comput-
https://​doi.​org/​10.​1080/​15567​036.​2020.​18552​74 ing, Sustainable Computing & Communications, Social Comput-
32. Li X, Gu J, Sun X, Li J, Tang S (2022) Parameter identification ing & Networking (ISPA/BDCloud/SocialCom/SustainCom)
of robot manipulators with unknown payloads using an improved 51. Liu G, Shu C, Liang Z, Peng B, Cheng L (2021) A modified
chaotic sparrow search algorithm. Appl Intell 52:10341 sparrow search algorithm with application in 3d route planning
33. He D, Liu C, Jin Z, Ma R, Chen Y, Shan S (2022) Fault diagnosis for UAV. Sensors (Basel) 21(4):1224
of flywheel bearing based on parameter optimization variational 52. Wang Z, Liu Y, Song R (2021) A crack image segmentation
mode decomposition energy entropy and deep learning. Energy algorithm based on adaptive T-distribution. In: 2021 China Auto-
239:122108 mation Congress (CAC)
34. Wang Z, Huang X, Zhu D (2022) A Multistrategy-integrated 53. Xia L, Gao J, Liao Q, Yue X, Xiang H, Wu X, Tan X (2021)
learning sparrow search algorithm and optimization of engineer- Coordinated dispatch of combined heat and power microgrid
ing problems. Comput Intell Neurosci 2022:2475460 based on the improved sparrow search algorithm. In: 2021 11th
35. Wang W, Liu F, Wang W, Cheng M (2021) The chaotic time International Conference on Power and Energy Systems (ICPES)
series prediction method based on sparrow search algorithm opti- 54. Li J, Lei Y, Yang S (2022) Mid-long term load forecasting model
mization. In: 2021 2nd International Conference on Intelligent based on support vector machine optimized by improved sparrow
Computing and Human-Computer Interaction (ICHCI) search algorithm. Energy Rep 8:491–497
36. Ji F, Song Z, Yang S, Bai X, Yuan X (2021) Scheduling strategy 55. Jiang Z, Hu W, Qin H (2021) WSN node localization based on
of regional integrated energy system based on improved sparrow improved sparrow search algorithm optimization. In: Interna-
search algorithm. In: 2021 China Automation Congress (CAC) tional Conference on Sensors and Instruments (ICSI 2021). SPIE,
37. Wang P, Zhang Y, Yang H (2021) Research on economic opti- pp 1–8
mization of microgrid cluster based on chaos sparrow search 56. Wenzhi S, Zhang H, Tseng M-L, Weipeng Z, Xinyang L (2022)
algorithm. Comput Intell Neurosci 2021:5556780 Hierarchical energy optimization management of active distri-
38. Xu Y (2021) Optimization of BP artificial neural network regres- bution network with multi-microgrid system. J Ind Prod Eng
sion prediction model based on Improved Sparrow search algo- 39(3):210–229
rithm with Sine chaotic mapping and its application. In: 2021 57. Chen DZ, Zhao JD, Huang P, Deng X, Lu T (2021) An improved
IEEE International Conference on Emergency Science and Infor- sparrow search algorithm based on levy flight and opposition-
mation Technology (ICESIT) based learning. Assembly Autom 41(6):697–713

13
452 F. S. Gharehchopogh et al.

58. Chen H, Ma X, Huang S (2021) A feature selection method for 76. Ouyang C, Zhu D, Qiu Y (2021) Lens learning sparrow search
intrusion detection based on parallel sparrow search algorithm. algorithm. Math Probl Eng 2021:9935090
In: 2021 16th International Conference on Computer Science 77. Yan S, Yang P, Zhu D, Zheng W, Wu F (2021) Improved sparrow
& Education (ICCSE) search algorithm based on iterative local search. Comput Intell
59. Lei Y, De G, Fei L (2020) Improved sparrow search algorithm Neurosci 2021:6860503
based DV-Hop localization in WSN. In: 2020 Chinese Automa- 78. Tian H, Wang K, Yu B, Song C, Jermsittiparsert K (2021) Hybrid
tion Congress (CAC) improved Sparrow Search Algorithm and sequential quadratic
60. Jia J, Yuan S, Shi Y, Wen J, Pang X, Zeng J (2022) Improved programming for solving the cost minimization of a hybrid pho-
sparrow search algorithm optimization deep extreme learning tovoltaic, diesel generator, and battery energy storage system.
machine for lithium-ion battery state-of-health prediction. iSci- Energy Sourc Part A Recov Utiliz Environ Effects. https://​doi.​
ence 25(4):103988 org/​10.​1080/​15567​036.​2021.​19051​11
61. Fang Q, Shen B, Xue J (2022) A new elite opposite sparrow 79. Yan A, Guo Y (2021) Feature weight optimization method based
search algorithm-based optimized LightGBM approach for on t-memetic algorithm. In 2021 40th Chinese Control Confer-
fault diagnosis. J Ambient Intell Humaniz Comput. https://​ ence (CCC)
doi.​org/​10.​1007/​s12652-​022-​03703-5 80. Ma J, Hao Z, Sun W (2022) Enhancing sparrow search algorithm
62. Li F, Lin Y, Zou L, Zhong L (2021) Improved sparrow search via multi-strategies for continuous optimization problems. Inf
algorithm applied to path planning of mobile robot. In: 2021 Process Manage 59(2):102854
International Conference on Computer Information Science 81. Ouyang C, Qiu Y, Zhu D (2021) Adaptive spiral flying sparrow
and Artificial Intelligence (CISAI) search algorithm. Sci Program 2021:6505253
63. Yuan J, Zhao Z, Liu Y, He B, Wang L, Xie B, Gao Y (2021) 82. Liang Q, Chen B, Wu H, Han M (2021) A novel modified spar-
DMPPT control of photovoltaic microgrid based on improved row search algorithm based on adaptive weight and improved
sparrow search algorithm. IEEE Access 9:16623–16629 boundary constraints. In: 2021 IEEE 6th International Confer-
64. Zhang G, Zhang E (2021) A random opposition-based sparrow ence on Computer and Communication Systems (ICCCS)
search algorithm for&nbsp;Path Planning Problem. In: Artifi- 83. Song C, Yao L, Hua C, Ni Q (2021) A water quality prediction
cial Intelligence: First CAAI International Conference, CICAI model based on variational mode decomposition and the least
2021, Hangzhou, China, June 5–6, 2021, Proceedings, Part II. squares support vector machine optimized by the sparrow search
Springer, Hangzhou, pp 408–418 algorithm (VMD-SSA-LSSVM) of the Yangtze River, China.
65. Ouyang C, Zhu D, Wang F (2021) A learning sparrow search Environ Monit Assess 193(6):363
algorithm. Comput Intell Neurosci 2021:3946958 84. Jiang S, Wang D, Xing X (2022) Temperature compensation of
66. Chang Z, Gu Q, Lu C, Zhang Y, Ruan S, Jiang S (2021) 5G an Eddy-current displacement sensor using an improved sparrow
private network deployment optimization based on RWSSA in search algorithm method. J Sens 2022:6021182
open-pit mine. IEEE Trans Ind Inform 18:5466 85. Hu Y, Li K, Zhang B, Han B (2022) Strength investigation of the
67. Ma Y, Guan Q, Guo F, Zhang G (2021) Malicious URL clas- cemented paste backfill in alpine regions using lab experiments
sification model based on improved sparrow search algorithm. and machine learning. Constr Build Mater 323:126583
In: 2021 IEEE 11th International Conference on Electronics 86. Chen Y, Duan W, Yang Y, Liu Z, Zhang Y, Liu J, Li S (2022)
Information and Emergency Communication (ICEIEC) Rapid in measurements of brown tide algae cell concentrations
68. Zhu Y, Yousefi N (2021) Optimal parameter identification of using fluorescence spectrometry and generalized regression neu-
PEMFC stacks using Adaptive Sparrow Search Algorithm. Int ral network. Spectrochim Acta Part A Mol Biomol Spectrosc
J Hydrog Energy 46(14):9541–9552 272:120967
69. Li Z, Luo X, Liu M, Cao X, Du S, Sun H (2022) Wind power 87. Ding H, Pan H, Bai H, Zheng X, Chen J, Zhang W (2022) Driv-
prediction based on EEMD-Tent-SSA-LS-SVM. Energy Rep ing strategy of connected and autonomous vehicles based on
8:3234–3243 multiple preceding vehicles state estimation in mixed vehicular
70. Man Y, Guangwu C, Zongshou W (2021) TDOA positioning traffic. Physica A 596:127154
method based on mixed strategy sparrow search algorithm. In 88. Xin J, Chen J, Li C, Lu R-k, Li X, Wang C, Zhu H, He R (2022)
2021 CAA Symposium on Fault Detection, Supervision, and Deformation characterization of oil and gas pipeline by ACM
Safety for Technical Processes (SAFEPROCESS) technique based on SSA-BP neural network model. Measurement
71. Yang H, Liu X, Song K (2022) A novel gradient boosting 189:110654
regression tree technique optimized by improved sparrow 89. Jiang H, Wang J, Mao W, Chen Q (2022) Determination of afla-
search algorithm for predicting TBM penetration rate. Arab J toxin B1 in wheat based on colourimetric sensor array technol-
Geosci 15(6):461 ogy: optimization of sensor features and model parameters to
72. Zhang Z, He R, Yang K (2022) A bioinspired path planning improve the model generalization performance. Microchem J
approach for mobile robots based on improved sparrow search 175:107173
algorithm. Adv Manuf 10(1):114–130 90. Ji Z, Niu D, Li M, Li W, Sun L, Zhu Y (2022) A three-stage
73. Quan L, Li A, Cui G, Xie S (2021) Using enhanced sparrow framework for vertical carbon price interval forecast based
search algorithm-deep extreme learning machine model to on decomposition–integration method. Appl Soft Comput
forecast end-point phosphorus content of BOF. Ind Manuf Eng 116:108204
2021(1):1–22 91. Zhang H, Peng Z, Tang J, Dong M, Wang K, Li W (2022) A
74. Chen Q, Wang W, Wang H (2021) Optimal reactive power dis- multi-layer extreme learning machine refined by sparrow
patch and distributed generation placement based on a hybrid search algorithm and weighted mean filter for short-term multi-
co-evolution algorithm and bi-level programming. Int Trans step wind speed forecasting. Sustain Energy Technol Assess
Electr Energy Syst 31(12):e13246 50:101698
75. Liu Q, Zhang Y, Li M, Zhang Z, Cao N, Shang J (2021) Multi- 92. Wang X, Liu J, Hou T, Pan C (2021) The SSA-BP-based potential
UAV path planning based on fusion of sparrow search algo- threat prediction for aerial target considering commander emo-
rithm and improved bioinspired neural network. IEEE Access tion. Defence Technol. https://​doi.​org/​10.​1016/j.​dt.​2021.​05.​017
9:124670–124681 93. Chen J, Zhang N, Liu G, Li J, Qiao Y (2021) Short-term PV
power forecasting based on time-phased and error correction.

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 453

In: 2021 6th International Conference on Power and Renewable 110. Baghdadi NA, Malki A, Abdelaliem SF, Magdy Balaha H, Bad-
Energy (ICPRE) awy M, Elhosseini M (2022) An automated diagnosis and clas-
94. Yan P, Shang S, Zhang C, Yin N, Zhang X, Yang G, Zhang Z, sification of COVID-19 from chest CT images using a transfer
Sun Q (2021) Research on the processing of coal mine water learning-based convolutional neural network. Comput Biol
source data by optimizing BP neural network algorithm with Med 144:105383
sparrow search algorithm. IEEE Access 9:108718–108730 111. Jia P, Zhang H, Liu X, Gong X (2021) Short-term photovol-
95. Liu B, Ye S (2021) Research on seasonal PM 2.5 predication in taic power forecasting based on VMD and ISSA-GRU. IEEE
hangzhou city based on SSA-ELM Model. In: 2021 7th Annual Access 9:105939–105950
International Conference on Network and Information Systems 112. Chen G, Shan J, Li DY, Wang C, Li C, Zhou Z, Wang X, Li
for Computers (ICNISC) Z, Hao JJ (2019) Research on wind power prediction method
96. Zhang R, Pan Z, Yin Y (2021) Research on assessment algorithm based on convolutional neural network and genetic algorithm.
for network security situation based on SSA-BP neural network. In: 2019 IEEE Innovative Smart Grid Technologies - Asia
In: 2021 7th International Symposium on System and Software (ISGT Asia)
Reliability (ISSSR) 113. Liu T, Yuan Z, Wu L, Badami B (2021) Optimal brain tumor
97. Zhang S, Zhang J, Wang Z, Li Q (2021) Regression prediction diagnosis based on deep learning and balanced sparrow search
of material grinding particle size based on improved sparrow algorithm. Int J Imaging Syst Technol 31(4):1921–1935
search algorithm to optimize BP neural network. In: 2021 2nd 114. Vaiyapuri T, Lydia EL, Sikkandar MY, Díaz VG, Pustokhina
International Symposium on Computer Engineering and Intel- IV, Pustokhin DA (2021) Internet of things and deep learning
ligent Communications (ISCEIC) enabled elderly fall detection model for smart homecare. IEEE
98. Jin C, Chen L, Ma H, Xue X, Mei S (2021) Intelligent fault diag- Access 9:113879–113888
nosis approach for PV arrays based on sparrow search algorithm 115. Chen Z, Peng D, Li J, Wang D, Zhao H, Xu C (2021) Electric
optimized Kernel extreme learning machine. In: 2021 IEEE 5th vehicle load forecast based on TA-SSA-LSTM. In: 2021 IEEE
Conference on Energy Internet and Energy System Integration 3rd International Conference on Power Data Science (ICPDS)
(EI2) 116. Liu T, Yuan Z, Wu L, Badami B (2021) An optimal brain tumor
99. Xie S, Li L (2021) Improvement and application of deep belief detection by convolutional neural network and enhanced sparrow
network based on sparrow search algorithm. In: 2021 IEEE Inter- search algorithm. Proc Inst Mech Eng H 235(4):459–469
national Conference on Advances in Electrical Engineering and 117. Song C, Yao L, Hua C, Ni Q (2021) A novel hybrid model for
Computer Applications (AEECA) water quality prediction based on synchrosqueezed wavelet trans-
100. Wang H, Wu X, Gholinia F (2021) Forecasting hydropower gen- form technique and improved long short-term memory. J Hydrol
eration by GFDL-CM3 climate model and hybrid hydrological- 603:126879
Elman neural network model based on Improved Sparrow Search 118. Tian Z, Chen H (2021) A novel decomposition-ensemble pre-
Algorithm (ISSA). Concurr Comput Pract Exp 33(24):e6476 diction model for ultra-short-term wind speed. Energy Convers
101. Song C, Yao L, Hua C, Ni Q (2021) Comprehensive water quality Manage 248:114775
evaluation based on kernel extreme learning machine optimized 119. Ma Y, Tang Y, Li B, Qi B (2020) Residential high-power load
with the sparrow search algorithm in Luoyang River Basin, prediction based on optimized LSTM network. In: 2020 Inter-
China. Environ Earth Sci 80(16):521 national Conference on Artificial Intelligence and Computer
102. Li G, Hu T, Bai D (2021) BP neural network improved by spar- Engineering (ICAICE)
row search algorithm in predicting debonding strain of FRP- 120. Xiong J, Liang W, Liang X, Yao J (2022) Intelligent quantifi-
strengthened RC beams. Adv Civil Eng 2021:9979028 cation of natural gas pipeline defects using improved sparrow
103. Wu Z, Wang B (2021) An ensemble neural network based on search algorithm and deep extreme learning machine. Chem Eng
variational mode decomposition and an improved sparrow search Res Des 183:567–579
algorithm for wind and solar power forecasting. IEEE Access 121. Zhang Z, Han Y (2022) Discrete sparrow search algorithm
9:166709–166719 for symmetric traveling salesman problem. Appl Soft Comput
104. Zhao Y, Pan D, Xu Z, Wang S, Li J, Xu L (2021) An artifi- 118:108469
cial neural network optimized by sparrow search algorithm for 122. Li L-L, Xiong J-L, Tseng M-L, Yan Z, Lim MK (2022) Using
predicting the boiling point temperature of working fluid. In: multi-objective sparrow search algorithm to establish active dis-
2021 3rd International Conference on Artificial Intelligence and tribution network dynamic reconfiguration integrated optimiza-
Advanced Manufacture (AIAM) tion. Expert Syst Appl 193:116445
105. Wang Y, Tuo J (2020) Blood glucose prediction based on empiri- 123. Fathy A, Alanazi TM, Rezk H, Yousri D (2022) Optimal energy
cal mode decomposition and SSA-KELM. In: 2020 Chinese management of micro-grid using sparrow search algorithm.
Automation Congress (CAC) Energy Rep 8:758–773
106. Wu H, Zhang A, Han Y, Nan J, Li K (2022) Fast stochastic con- 124. Kathiroli P, Kanmani S (2022) Multi-objective sparrow search
figuration network based on an improved sparrow search algo- algorithm-based clustering and routing in wireless sensor net-
rithm for fire flame recognition. Knowl-Based Syst 245:108626 works. In: Proceedings of International Conference on Intelligent
107. Li X, Ma X, Xiao F, Xiao C, Wang F, Zhang S (2022) Time- Cyber-Physical Systems. Springer, Singapore
series production forecasting method based on the integration 125. Nguyen T, Nguyen T, Vu QH, Huynh TTB, Nguyen BM (2021)
of Bidirectional Gated Recurrent Unit (Bi-GRU) network and Multi-objective sparrow search optimization for task scheduling
Sparrow Search Algorithm (SSA). J Petrol Sci Eng 208:109309 in fog-cloud-blockchain systems. In: 2021 IEEE International
108. Chen G, Tang B, Zeng X, Zhou P, Kang P, Long H (2022) Short- Conference on Services Computing (SCC)
term wind speed forecasting based on long short-term memory 126. Li J, Wang W, Chen G, Han Z (2022) Spatiotemporal assessment
and improved BP neural network. Int J Electr Power Energy Syst of landslide susceptibility in Southern Sichuan, China using SA-
134:107365 DBN, PSO-DBN and SSA-DBN models compared with DBN
109. Pan Z, Lu W, Wang H, Bai Y (2022) Recognition of a linear model. Adv Space Res 69(8):3071–3087
source contamination based on a mixed-integer stacked chaos 127. Meng Z, Zhang Y, Zhu B, Pan Z, Cui L, Li J, Fan F (2022)
gate recurrent unit neural network–hybrid sparrow search algo- Research on rolling bearing fault diagnosis method based on
rithm. Environ Sci Pollut Res 29:33528 ARMA and optimized MOMEDA. Measurement 189:110465

13
454 F. S. Gharehchopogh et al.

128. Fathy A, Rezk H, Yousri D, Kandil T, Abo-Khalil AG (2022) 145. Liu T, Liu H, Zheng M, Tan C (2021) SSA-based WSN cluster-
Real-time bald eagle search approach for tracking the maximum ing routing algorithm for power grid. In: 2021 2nd Information
generated power of wind energy conversion system. Energy Communication Technologies Conference (ICTC)
249:123661 146. Karthick S, Gomathi N (2021) Sparrow search algorithm-based
129. Zhao J, Wang X, Tian H, Shu G (2022) Optimization strategy and resource management in internet of things (IoT). EAI Endorsed
capacity planning for coordinated operation of regional energy Trans 22(37):1–11
internet system based on sparrow search algorithm. Int J Green 147. Miao B, Tan M, Hu G, Du B (2021) Shape optimization of
Energy. https://​doi.​org/​10.​1080/​15435​075.​2021.​20056​07 developable bézier-like surfaces with multiple shape parame-
130. Wang Y, Wang W, Tao G, Li H, Zheng Y, Cui J (2022) Opti- ters using sparrow search algorithm. In: 2021 6th International
mization of the semi-sphere vortex generator for film cooling Conference on Image, Vision and Computing (ICIVC)
using generative adversarial network. Int J Heat Mass Transf 148. Cao G, Wang H, Huang Y, Lu T, Wang C, Li H (2021)
183:122026 Research on single-stage photovoltaic grid-connected control
131. Chen M, Liang Z, Cheng Z, Zhao J, Tian Z (2022) Optimal based on SSA optimized LADRC. In: 2021 China Automation
scheduling of FTPSS With PV and HESS considering the online Congress (CAC). 2021.
degradation of battery capacity. IEEE Trans Transp Electrif 149. Shi M, Liang Y, Qin L, Zheng Z, Huang Z (2021) Prediction
8(1):936–947 method of ball valve internal leakage rate based on acoustic
132. Wang Y, Wang H, Wen J, Lun Y, Wu J (2020) Obstacle avoidance emission technology. Flow Meas Instrum 81:102036
of UAV based on neural networks and interfered fluid dynami- 150. Xu L, Wang H, Liu Y, Xue W, Li N (2021) PID control for
cal system. In: 2020 3rd International Conference on Unmanned aeroengine based on sparrow search algorithm. In: 2021 China
Systems (ICUS) Automation Congress (CAC)
133. Zhu D, Huang Z, Xie L, Zhou C (2022) Improved particle swarm 151. Song J, Jin L, Xie Y, Wei C (2021) Optimized XGBoost based
based on elastic collision for DNA coding optimization design. sparrow search algorithm for short-term load forecasting. In:
IEEE Access 10:63592 2021 IEEE International Conference on Computer Science,
134. Hui X, Guangbin C, Shengxiu Z, Xiaogang Y, Mingzhe H (2022) Artificial Intelligence and Electronic Engineering (CSAIEE)
Hypersonic reentry trajectory optimization by using improved 152. Xing Z, Yi C, Lin J, Zhou Q (2021) Multi-component fault
sparrow search algorithm and control parametrization method. diagnosis of wheelset-bearing using shift-invariant impulsive
Adv Space Res 69(6):2512–2524 dictionary matching pursuit and sparrow search algorithm.
135. Wang J, Cui Q, He M (2022) Hybrid intelligent framework for Measurement 178:109375
carbon price prediction using improved variational mode decom- 153. Abdulhammed OY (2022) Load balancing of IoT tasks in the
position and optimal extreme learning machine. Chaos Solitons cloud computing by using sparrow search algorithm. J Super-
Fractals 156:111783 comput 78(3):3266–3287
136. Kumar Ganti P, Naik H, Kanungo Barada M (2022) Environ- 154. Tuerxun W, Chang X, Hongyu G, Zhijie J, Huajian Z (2021)
mental impact analysis and enhancement of factors affecting the Fault diagnosis of wind turbines based on a support vector
photovoltaic (PV) energy utilization in mining industry by spar- machine optimized by the sparrow search algorithm. IEEE
row search optimization based gradient boosting decision tree Access 9:69307–69315
approach. Energy 244:122561 155. Gai J, Zhong K, Du X, Yan K, Shen J (2021) Detection of
137. Feng Z-k, Duan J-f, Niu W-j, Jiang Z-q, Liu Y (2022) Enhanced gear fault severity based on parameter-optimized deep belief
sine cosine algorithm using opposition learning, adaptive evolu- network using sparrow search algorithm. Measurement
tion and neighborhood search strategies for multivariable param- 185:110079
eter optimization problems. Appl Soft Comput 119:108562 156. Sharma M, Sharma M, Sharma S (2021) Desert sparrow optimi-
138. Chang L, Wang R, Zhang Y (2022) Decoding SSVEP patterns zation algorithm for the bicriteria flow shop scheduling problem
from EEG via multivariate variational mode decomposition- with sequence-independent setup time. Oper Res 22:4353
informed canonical correlation analysis. Biomed Signal Process 157. Xu L, Cai D, Shen W, Su H (2021) Denoising method for Fiber
Control 71:103209 Optic Gyro measurement signal of face slab deflection of con-
139. Phani Raghav L, Seshu Kumar R, Koteswara Raju D, Singh AR crete face rockfill dam based on sparrow search algorithm and
(2022) Analytic hierarchy process (AHP)—swarm intelligence variational modal decomposition. Sens Actuators A 331:112913
based flexible demand response management of grid-connected 158. Wu M, Yang D, Yang Z, Guo Y (2021) Sparrow search algorithm
microgrid. Appl Energy 306:118058 for solving flexible jobshop scheduling problem. In: Advances in
140. Zafar MH, Khan UA, Khan NM (2021) A sparrow search opti- Swarm Intelligence. Springer, Cham
mization algorithm based MPPT control of PV system to har- 159. Kathiroli P (2021) An efficient cluster-based routing using Spar-
vest energy under uniform and non-uniform irradiance. In: 2021 row Search Algorithm for heterogeneous nodes in Wireless Sen-
International Conference on Emerging Power Technologies sor Networks. In: 2021 International Conference on Communica-
(ICEPT) tion information and Computing Technology (ICCICT)
141. Jebaraj L, Sakthivel S (2022) A new swarm intelligence opti- 160. Tolba MA, Bulatov RV, Burmeyster MV (2021) A robust meth-
mization approach to solve power flow optimization problem odology approach based sparrow search algorithm for the incor-
incorporating conflicting and fuel cost based objective functions. poration of rdgs to improve the distribution grid performance.
e-Prime Adv Electr Eng Electron Energy 2:100031 In: 2021 International Ural Conference on Electrical Power Engi-
142. Wang Y, Ding S, Wang L, Du S (2022) A manifold p-spec- neering (UralCon)
tral clustering with sparrow search algorithm. Soft Comput 161. Zhang Q, Zhang Y, Zhu X (2021) A novel node localization
26(4):1765–1777 algorithm based on sparrow search for WSNs. In: 2021 IEEE
143. Zhou S, Xie H, Zhang C, Hua Y, Zhang W, Chen Q, Gu G, Sui X 11th International Conference on Electronics Information and
(2021) Wavefront-shaping focusing based on a modified sparrow Emergency Communication (ICEIEC)
search algorithm. Optik 244:167516 162. Singh AR, Ding L, Raju DK, Raghav LP, Kumar RS (2022) A
144. Kumar KK, Reddy GN (2021) The sparrow search algorithm for swarm intelligence approach for energy management of grid-
optimum position of wind turbine on a wind farm. Int J Renew connected microgrids with flexible load demand response. Int J
Energy Res 11(4):1–8 Energy Res 46(4):4301–4319

13
Advances in Sparrow Search Algorithm: A Comprehensive Survey 455

163. Wu Y, Zhou W, Gu X, Wu W, Yang G (2022) A fault diagnosis Communications, Signal Processing, and Systems. Springer,
method based on support vector machine optimized by sparrow Singapore
search algorithm. In: Proceedings of 2021 Chinese Intelligent
Systems Conference. Springer, Singapore Publisher's Note Springer Nature remains neutral with regard to
164. Decoderz, (2020) Sparrow search algorithm (SSA): a swarm jurisdictional claims in published maps and institutional affiliations.
intelligence optimization algorithm for the application to solve
practical engineering examples. Transpire Online 2020(1):1–15 Springer Nature or its licensor holds exclusive rights to this article under
165. Wu D, Yuan C (2022) Correction to: Threshold image segmenta- a publishing agreement with the author(s) or other rightsholder(s);
tion based on improved sparrow search algorithm. Multimedia author self-archiving of the accepted manuscript version of this article
Tools Appl is solely governed by the terms of such publishing agreement and
166. Wang Z, Wang S, Tang H (2022) Wireless sensor network applicable law.
coverage optimization based on sparrow search algorithm. In:

13