Download as pdf or txt
Download as pdf or txt
You are on page 1of 26

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

net/publication/358227529

Black hole algorithm: A comprehensive survey

Article in Applied Intelligence · January 2022


DOI: 10.1007/s10489-021-02980-5

CITATIONS READS

8 653

9 authors, including:

Laith Abualigah Mohamed Elsayed Abd Elaziz


Al al-Bayt University 378 PUBLICATIONS 15,341 CITATIONS
368 PUBLICATIONS 12,542 CITATIONS
SEE PROFILE
SEE PROFILE

Ahmad M. Khasawneh Mohammad Al Shinwan


Amman Arab University Applied Science Private University
42 PUBLICATIONS 986 CITATIONS 45 PUBLICATIONS 933 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Hippopotamus (Hippos) Optimization Algorithm View project

Multiobjective Optimization View project

All content following this page was uploaded by Hayfa Abu Addous on 06 April 2022.

The user has requested enhancement of the downloaded file.


Applied Intelligence
https://doi.org/10.1007/s10489-021-02980-5

Black hole algorithm: A comprehensive survey


Laith Abualigah1,2 · Mohamed Abd Elaziz3,4,5,6 · Putra Sumari2 · Ahmad M. Khasawneh1 ·
Mohammad Alshinwan1 · Seyedali Mirjalili7,8 · Mohammad Shehab9 · Hayfa Y. Abuaddous1 · Amir H. Gandomi10

Accepted: 3 November 2021


© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2021

Abstract
This paper provides an in-depth literature review of the Black Hole Algorithm (BHA) which is considered as a recent
metaheuristic. BHA has been proven to be very efficient in different applications. There has been several modifications
and variants of this algorithm in the literature, so this work reviews various variants of the BHA. The applications of BHA
in engineering problems, clustering, task scheduling, image processing, etc. have been thoroughly reviewed as well. This
review article sheds lights on the pros and cons of this algorithm and enables finding a right variant of this algorithm for a
certain application area. The paper concludes with an in-depth future direction.

Keywords Black hole algorithm · Optimization algorithms · Meta-heuristics · Real-word problems · Optimization
problems · Algorithm · Optimization

1 Introduction has prompted and motivated researchers to introduce


or improve nature-inspired optimization algorithms [18].
Optimization is a rule that determines the best decision Moreover, growing computational complexity has encour-
variables to achieve the minimum or maximum value of aged researchers to pioneer new novel nature-inspired opti-
the given objective function [1–3]. Optimization techniques mization algorithms and made it necessary. The devel-
work by examining the search space for finding the opti- opment of AI techniques has brought soft computing
mal solution to the present problem [4]. The optimization approaches not only for the scientific population yet also for
process begins the endeavors of improvements subject to the general society [19].
the objective function, constraints, and parameters [5, 6]. Over the past decades, several researchers have been
By executing the optimization procedure with the underly- motivated, by the demand for solving complex optimization
ing objective function, a minimization will guide towards problems, to introduce new optimization algorithms by
getting the optimum value for desired decision values. different inspirations [20, 21]. Evolutionary and swarming
Besides, there are various single and multi-objective opti- are the most popular of these optimization algorithms.
mization problems with different solutions to address opti- These algorithms including a wide range of methods such
mization problems [7, 8]. Some solutions have constraints as Evolution strategies (ES) [22], evolution programming
according to the appropriateness of the given problems and [23], differential evolution (DE) [24], genetic algorithms
wanted solutions [9–14]. (GA) [25], black hole algorithm [26], multi-verse optimizer
The complexity of optimization problems is growing algorithm [27], salp swarm algorithm [28], ant Lion
exponentially [15, 16]. Many complicated real-world opti- Optimizer [29], moth-flame optimization algorithm [30],
mization problems are presented with challenging struc- krill herd algorithm [31], particle swarm optimization [32],
ture to be solved by gradient-free optimization (classi- and Arithmetic Optimization Algorithm (AOA) [33].
cal algorithms) [17]. The weakness of classical optimiza- In the same context, Hatamlou in [26] introduced a new
tion algorithms in addressing hard optimization problems metaheuristic optimization method inspired by the black
hole phenomenon, namely, the black hole algorithm (BHA).
In general, BHA is similar to other population algorithms
 Laith Abualigah which means, the first step in the optimization process is to
Aligah.2020@gmail.com
start with a set of random solutions (i.e., initial population)
and the underlying objective function that is used to evaluate
Extended author information available on the last page of the article. the decision values. The best solution so far is elected as

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

Fig. 1 The number of published


versions of the Black Hole
Algorithm

the black hole in each iteration, which next begins attracting of published applications, that have used the Black Hole
other candidate solutions toward it, called stars. If the other Algorithm, is given. It is clear from Fig. 2 that the most
candidate solutions reach near to the best-obtained solution, published papers that employed the Black Hole Algorithm
it will be consumed by the best solution. In the first paper, were on engineering applications. It followed by clustering
the effectiveness of the black hole algorithm is tested on applications, scheduling applications, benchmark functions,
several NP-hard problems (i.e., data clustering datasets). machine learning, network applications, image processing,
The results proved that the suggested black hole algorithm parameter estimation, aircraft applications, wireless sensor
beat other competitor optimization algorithms for several network, and software engineering.
test cases. Finally, the number of published papers that have used
The primary purpose of this survey paper is to give the Black Hole algorithm among the years since it was
a global summary of the main procedure of the Black proposed until the second quarter of 2020 are given in
Hole Algorithm and its sources. The variants of the Fig. 3. The most number of published papers were in 2016,
Black Hole Algorithm are presented as groups of basic, followed by 2017, 2015, 2019, 2020, 2018, 2014, and 2013.
modified, binary, hybrid, and multi-objective versions. It is clear that the amount of interest in this algorithm still
The applications that have employed the Black Hole growing and it can be used further for other applications and
Algorithm are presented as groups of benchmark functions, in different modifications.
machine learning applications, clustering applications, The rest structure of this review paper is organized
scheduling applications, image applications, engineering as follows. Section 2 presents the main procedure of the
applications, parameter estimation applications, aircraft Black Hole Algorithm. Section 3 summarizes the variants
applications, software engineering applications, network and modifications of the Black Hole Algorithm. Section 4
applications, wireless sensor network applications, and gives an overview of the applications of the Black Hole
other applications. As well, the results of the Black Algorithm. In Section 5, the results of the conventional
Hole Algorithm for solving the classical benchmark test Black Hole Algorithm are presented and compared with
functions are provided and compared its results with other other similar algorithms. Section 6 shows the discussion
optimization algorithms published in the literature to show and theoretical analysis for the versions and application
for the researchers its general ability in solving this kind of of the Black Hole Algorithm with its main advantage and
problems. Finally, discussions and theoretical analysis are disadvantages. Finally, the conclusion and some potential
provided to show the most applications which have used future work directions are given in Section 7.
this algorithm, and its main advantages and disadvantages.
Some plots are presented to show a clear picture of the
publication’s growth of this algorithm in regards to years, 2 Black hole algorithm: An overview
versions, and other applications [34–36].
Figure 1 shows the number of published versions of In this section, the mathematical formulation of the Black
the Black Hole Algorithm. It is clear from this figure that Hole Algorithm (BHA) is introduced [26]. It is a population-
the most published papers over versions of the Black Hole based metaheuristic that mimics the black hole phenomena
Algorithm are modified versions. It followed by binary, behavior of pulling in nearby stars. BHA, similar to
hybrid, and multi-objective versions. In Fig. 2, the number other MH, starts by generating initial position for a set

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

Fig. 2 The number of published


applications that have used the
Black Hole Algorithm

of starts/solutions inside the search domain. Followed by Thereafter the fitness value (f it) of each star is computed
determining the best fitness value which corresponding to and the best star is allocated as black hole (XBH ).
the best fitness value. The next process is to update the Followed [26], the stars (Xi ) around black hole XBH will
position of the stars by moving them towards the black hole be absorbed by it and this formulated as:
(i.e., best star) and replacing those stars that have position
in specified neighbour of the black hole by new generated Xi (t + 1) = Xi (t) + rand × (XBH − Xi (t)), i = 1, 2, ..., N
stars. The details of BHA are discuss as follows:
The first generation of stars is constructed inside the (2)
search space bounded by lower value LB and upper value
U B. This formulated using the following equation: where t represents the current iteration and rand ∈ [0, 1] is
random number.
Moreover, during moving the start (Xi , i = 1, ..., N)
Xi = LB + rand × (U B − LB), i = 1, 2, ..., N (1) in the direction of the black hole, there is a possibility

Fig. 3 The number of published


papers that have used the Black
Hole algorithm among the years
since it was proposed until the
second quarter of 2020

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

of crossing the event horizon 1 . So, those stars Xi will been underlined as a solution for various field problems,
be sucked by XBH and new stars will be generated. This including civil engineering, artificial intelligence, software
process is defined according to the radius of the event engineering, wireless sensor network, image processing,
horizon and image processing and data clustering.
Ibrahim et al. in [55] anticipated the growth of BHA
f itBH
R = N (3) application in more areas, namely routing in Very Large
i=1 f iti Scale Integration (VSLI), and DNA sequence design. BHA
surpasses GA, ACO, and PSO algorithms since it is faster
In the case of the distance between stars and XBH is less
than they are and better in finding a solution for solving the
than R then these stars died and new stars are created
traveling salesman problem (TSP) [56]. Kumar et al. in [39]
randomly in the search domain. The steps of BHA are given
unveiled that BHA can provide a model for solving wireless
in Algorithm 1.
sensor network optimization problems since it is free from
parameter tuning issues like GA. Moreover, it has a simple
structure with minimal computational complexity; besides,
it offers a preeminent solution and fast convergence.
Feature selection can be perceived as an optimization
problem. The experimental results in [26] showed that
BHA outperforms the other traditional heuristic algorithms
like BB–BC, GSA, PSO, and K-means in solving NP-hard
problems such as data clustering. On the other hand, some
limitations of the original BHA in terms of exploration
(global search) and exploitation (local search) abilities have
been reported along with the slow rate of convergence so
many iterations are needed to achieve the effect [36]; thus,
some modifications need to be done to improve the BHA’s
performance mainly in global search in big problem spaces.

Generally, exploration and exploitation are the main search 3.2 Modified versions of the black hole algorithm
methods used in any optimization algorithm. Each tool
specializes in a different search process to reach the best Chaotic Inertia Weight Black Hole Algorithm (CIWBH)
and optimal solution through several methods. Expanded was proposed by [38] to improve the global search of BHA.
research performs a large-scale exploration process in the Chaos methods are added to the basic version of BHA
process of updating the current solutions. In contrast, a to improve its performance further. Better performance
narrow search conducts a limited investigation to improve outcomes than original BHA and other metaheuristic
the existing solutions to get better results. algorithms, GSA and PSO, were obtained.
The black hole algorithm has been applied to the
aircraft design. For example, Gao et al. [40] have been
3 Variants of black hole algorithm proposed the black hole algorithm and Spencer method to
examine the stability of elevated embankment slope of an
In this section, Black Hole algorithm variables are presented airport. The study shows that the proposed approach very
in detail to show the modifications and applications they efficient to measure the slope stability [40]. In addition,
have been used in. An overview of these studies is given to determine the critical slip surface of soil slope, a
in Table 1. The following subsections are divided into five combination of Spencer approach and the Morgenstern-
main sections as follows. Price algorithm with Black Hole algorithm used. To
verify the performance of the presented algorithm, a set
3.1 Basic versions of the black hole algorithm of numerical experiments has been conducted. The result
shows significant performance in determining the critical
Black Hole optimization algorithm (BHA) has received slip surface [54].
extensive attention from many scholars. Since it has In term of improving the process of grinding production,
the BH algorithm used with the extreme learning machine
(ELM) sensors and several activations approach. The ELM
1 Event Horizon: It is a spherical shape constructed around the black neural network method used to study the different seven
hole activation functions. To optimize the result of ELM the

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

Table 1 Overview of the versions of the Black Hole Algorithm

Proposed Variant Problem Year Reference

BHA Basic Feature selection 2013 [26]


HBHA Hybrid scheduling 2014 [37]
MTLBO–BH Hybrid Optimal operating cost 2014 [37]
CIWBH Modified Test functions 2015 [38]
BHA Basic Wireless sensor network 2015 [39]
MBHA Modified aircraft 2016 [40]
MBHA Modified Test functions 2016 [41]
MBHA Modified Test functions 2016 [42]
BBHA Binary Characterizing irregular customers 2016 [43]
BBHA Binary Set Covering Problem 2016 [44]
BBH Binary Set Covering Problem 2016 [45]
GSBHA Hybrid Test functions 2016 [46]
BOBHA Multi-objective Assignment issue and facility location 2017 [47]
BOBHA Multi-objective Automotive manufacturers 2017 [48]
AMOBH Multi-objective Density assessment problems 2017 [49]
BBHA Binary Set coverage problems 2017 [50]
BBHA Binary Set coverage problems 2017 [50]
HBHA Hybrid Test functions 2017 [51]
BBHA Binary Feature selection 2017 [52]
BHLS Modified Test functions 2017 [53]
MBHA Modified Soil slope 2017 [54]
BHA Basic DNA sequence design 2018 [55]
BHA Basic Traveling salesman problem 2018 [56]
HBHA Hybrid Test functions 2018 [57]
BBHA Hybrid Gene selection 2019 [58]
MBHA Modified Machine learning 2020 [59]
MOBHA Multi-objective Motor performance optimization 2020 [60]

BH algorithm with a combination of golden sine operator a new token range if they surpass the smallest way to the
and Levy flight operator a proposed called GSLBH. The black hole. Further, the modified approach generated in the
simulation result shows high performance for the grinding exploration space chosen in an arbitrary way to evade the
process [59]. potential changes of trapping impact.
To solve and simplify the numerical optimization
problems, Mohammed, S. K., et al. propose a modified 3.3 Binary versions of the black hole algorithm
BH algorithm. The local search implemented with BH
algorithm to extract the nearby solution around the optimal Biological data is composed of redundant and meaningless
solution named (BHLS) [53]. Another approach based on characteristics. These characteristics can lead to algorithms
BH algorithm proposed by Olivares, Rodrigo, et al. [41] modelling misleading and the problem of overfitting.
to evaluate selection functions to self-adaptive Constraint Without a method of selection features, it is challenging
optimization. In this work [41], the effect and significance for current models to identify the trends on data accurately.
of 16 various constructed functions are considered. The The aim of the selected feature is to select a small
proposed method seeks to enhance search tree exploration number of appropriate or important features to optimize the
speed by changing the enumeration approach in real time. classification efficiency. Current methods features selection
The proposed method got better results compared to other suffer from issues such as being trapped in local optima and
methods. becoming computationally expensive. Thus, an effective
In [42] employ the BH algorithm to solve the potential global search technique is required to solve these issues.
problems caused by the trapping impact. The trapping Black Hole Algorithm (BHA) is an effective and recent
passed, and particles near to the black hole are taken with global search technique, influenced by black hole behaviour

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

that is being used to solve many optimization issues. influenced by physical behaviour in black holes to improve
However, BHA’s ability for feature selection has not yet the discovery to optimum values during the search.
been investigated. In many sectors, there are several problems that can be
In [52], Pashaei et al. have designed and implemented addressed with combinatorial optimization. In this context,
a Binary version of the Black Hole Algorithm named the Set Covering Problem is one of the common of them,
BBHA for solving the problem of feature selection in the being utilized in multiple factions of engineering and
area of biological data. The BBHA is an enhancement science, enabling to identify a set of solutions that address
of existing BHA by a suitable binarization. In addition, the needs found in the constraints with the lowest possible
this analysis compared the performance of six state-of- cost [44]. Rubio et al. [45] suggested an OR-Library-
the-art decision tree classifiers (Random Forest (RF), inspired binary black holes (BBH) algorithm for resolving
Bagging, C5.0, C4.5, Boosted C5.0, and CART) to utilize known instances of SPC. It also reproduces black holes
the best one as an evaluator of the proposed scheme. behaviour, using different operators to offer better solutions.
The performance of the proposed scheme is evaluated on The proposed method got better results compared to other
eight accessible biological datasets and is compared with methods.
Particle Swarm Optimization (PSO), Genetic Algorithm
(GA), Simulated Annealing (SA), and Correlation-based 3.4 Hybrid versions of the black hole algorithm
Feature Selection (CFS) in terms of accuracy, sensitivity,
specificity, Matthews Correlation Coefficient (MCC) and Recently, a challenging issue facing the electrical power
Region Under the Receiver Operating Function (ROCC). system operator would be how to handle the power
To validate the BBHA’s applicability and generality, it generating units optimally across a one-day scheduling
integrated with the classifier Naive Bayes (NB) and applied horizon with all the realistic equity, inequalities and
to the text and image domains on additional datasets. complex constraints in mind. Such restrictions include
The set coverage issue requires minimal cost component a balance of load plus transmission losses, valve-point
family from given subgroups, which both cover the whole impact, restricted operating areas, multi-fuel options,
set. Garcı́a et al. [50] introduce a multi-dynamic binary line flow limitations, operating reserve and minimum
black hole algorithm to solve the problem of set coverage. on/off times. The proposed structure first provides a
The proposed technique has the distinctive feature of realistic model for the scheduling of short-term thermal
proposing a generic dynamic binarization approach to generation (STGS) in this respect. It does have a high-
handle the properties of discovery and exploitation. In dimensional, elevated-constraints, non-convex, non-smooth
addition, using on Apache Spark the distributed framework, and non-linear existence and needs to be solved with
implementation of the algorithm has been explored. an effective algorithm. A new strategy to optimization,
In [43], a parallel between the issue of Brazil’s known as gradient-based modified teaching – learning-
commercial losses and the task of characterizing irregular based optimization along with black hole (MTLBO–BH)
customers have been introduced through a recent meta- algorithm, was then suggested to find the optimal operating
heuristic optimization technique called the Black Hole cost [37]. Despite the superior features of the MTLBO and
Algorithm. The experimental configuration is carried out BH algorithms, still suffer from the issue of local optima
over two different data sets (commercial and industrial) trapping. Additionally, this work presented the efficient
supported by a Brazilian electrical utility and demonstrated combination of MTLBO and BH and a novel self-adaptive
the significance of selecting the most important features in wavelet mutation operator for organizing the new robust
the characterization of theft. technique. Although the MTLBO–BH method has many
The Set Covering Problem (SCP) is one of the most features, some issues remain pleasant to address. One of
common problems of combinatorial optimization, and it has these issues is the more robust solution generated which
numerous applications in various engineering, science and can be overcome by the classical gradient-based technique.
other disciplines. It attempted to explore a set of solutions Lastly, the output of the proposed method is being tested on
with the lowest possible cost to address the needs specified various thermal power systems.
in the constraints. In this study, a current Binary Black Holes Gene selection is a useful technique for processing
(BBH)-inspired binary algorithm have been used to solve data in cancer classification, but it is a challenging task
multiple instances of the issue with the known benchmarks because of the extensive search space. The goal of this
obtained from the OR-library [44]. The approach described study is, therefore, to design a hybrid meta-heuristic Binary
emulates the behaviour of these celestial bodies that use Black Hole Algorithm (BBHA) and Binary Particle Swarm
various operators, such as rotation and collapse, to provide Optimization (BPSO) model (4-2) that focuses on gene
effective solutions. It also introduced some enhancements selection. In [58], the BBHA is integrated into the BPSO
in several operators, as well as added others that were also (4-2) scheme to make the BPSO (4-2) efficient and to

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

accelerate the BPSO (4-2) algorithm’s exploration and facility location. Then, the enhanced BH dealing with the
exploitation to optimize the overall performance further. multi-objective supply chain. The result shows that the
This model was compatible with the Recursive Random proposed method improves the performance, reliability, and
Forest (RF-RFE) pre-filtering technique. Sparse Partial flexibility of the supply chain.
Least Squares Discriminant Analysis (SPLSDA); the k- [48] proposed another approach for enhancing the
nearest neighbour and Naive Bayes are the classifications supply chain process to automotive manufacturers based
evaluated in the proposed framework. The performance of on milk runs and supermarkets. The optimization takes
the system proposed was tested on two metrics and three the assignment and facility location to improve the
clinical microarrays. The findings and statistics demonstrate manufacturing processes. The presented approach explores
the BPSO (4-2)-BBHA’s higher performance compared to the best location of barriers, and find out the optimal
the BBHA, the BPSO (4-2) and various state-of-the-art assignment of barriers and machines to reduce the material
techniques in terms of avoiding local minima, convergence handling expenses with consider to order capacity of devices
rate, accuracy and selected gene number. The results and resources.
also reveal that established biologically and statistically A multiobjective individual density assessment approach
relevant genes from the clinical datasets can be successfully with the BH algorithm named (AMOBH) introduced by Wu,
identified by the BPSO (4-2)-BBHA model. Chong, et al. Cell density approach has the features of low
Mohammed et al. [46] used a novel hybrid population- complexity of computational and variety of Pareto frontier,
based method to investigate the solution of optimization and keep the balance of concentration. In this approach, the
problems. A combination of the Black Hole Algorithm Pareto frontier moved to a different objective space. Then,
(BHA) and the Gravitational Search Algorithm (GSA) The Shannon entropy used to determine the evolutionary
is suggested for the Gravitational Search – Black Hole strategies and status. Finally, to estimate the robustness of
Algorithm (GSBHA). The key concept is to use GSA solutions by combining the cell density with dominance
to improve standard BHA. Standard real-parameter test strength evaluation [49].
functions of CEC 2014 are used to compare the hybrid Rao, Jing, et al. propose a multi-objective framework
algorithm with both standard BHA and GSA algorithms in based on the BH algorithm for motor performance
the development of the best solution. The findings show optimization. In this work, the local with the momentum
better efficiency of the hybrid method and better capability gradient technique is used; on the other hand, the
to escape with faster convergence from local optimums than global search is used with the random search approach.
the regular BHA and GSA. Furthermore, to make the optimization effect global, a new
Black Hole Algorithm (BHA) is a relativity theory-based file management method is used [60]. The proposed method
optimization algorithm based on the black hole discovery. got better results compared to other methods.
Previously, BHA introduced white hole operator, which
is based on the opposite of black hole. In [57], a local
with a white hole operator is applied to the BHA. A 4 Applications of black hole algorithm
new enumeration strategies online control system based on
recent metaheuristics inspired by nature: bat algorithm and In this section, Black Hole algorithm applications are
black hole optimization has been presented by [51]. The presented in detail to show the applications that have used
bat algorithm imitates the location capabilities of bats using the Black Hole algorithm. An overview of these studies is
echoes to distinguish the objects in their surrounding areas given in Table 2. The following subsections are divided into
while optimizing the black hole inspires their behaviour on twelve main sections as follows.
the gravitational pull of black holes in space. Through using
various enumeration methods and well-known benchmarks, 4.1 Benchmark functions
different experiments have been performed, where the
proposed solutions significantly outperform existing online Gupta et al. [76] compared between both of the Black Hole
control work. Algorithm (BHA) and Particle Swarm Optimization (PSO)
to prove that the BHA is not better than PSO as the author
3.5 Multi-objective versions of the black hole of BHA claim. Thus, 25 benchmark functions have been
algorithm employed, where they divided into 20 are multimodal, and 5
are unimodal problems. The experiment results showed that
In this paper [47], a new optimization model based on the BHA outperformed the PSO. However, the preference
the BH introduces to find solutions for the supply chain of the other functions was to PSO.
of consignment-store-based. Two factors are taken into Xie et al. [94] focused on BHA by highlighting
account in this approach to solve the assignment issue and its mechanisms of searching (i.e., balancing between

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

Table 2 Overview of the applications of the Black Hole Algorithm

Proposed Variant Application Year Reference

BKmeans Hybrid Clustering applications 2014 [61]


BHA Basic Engineering applications 2014 [62]
BHA Basic Aircraft applications 2014 [63]
BHA Basic Aircraft applications 2014 [64]
BHA Basic Software engineering 2015 [65]
BHA Basic Test functions 2015 [66]
BHA Basic Machine learning 2015 [67]
BHA Modified Clustering applications 2015 [68]
BHA Basic Image applications 2015 [69]
BH-DEO Hybrid Engineering applications 2015 [70]
BHA Basic Engineering applications 2015 [71]
BHA Basic Test functions 2015 [72]
BHA Basic Engineering applications 2015 [73]
BHA Basic Wireless sensor network 2015 [74]
BHA Basic Scheduling applications 2016 [75]
BHA Basic Test functions 2016 [76]
WH-BH Hybrid Test functions 2016 [77]
BHA Basic Test functions 2016 [78]
BHA Basic Machine learning 2016 [79]
MOBHA Multi-objective Scheduling applications 2016 [80]
BHA Basic Image applications 2016 [81]
BHA Basic Software engineering 2016 [82]
BHA Basic Network applications 2016 [83]
BHA Basic Wireless sensor network 2016 [84]
BHA Basic Scheduling applications 2017 [85]
BHA Basic Scheduling applications 2017 [86]
BBPSO Modified Engineering applications 2017 [87]
BHA Basic Parameter estimation 2017 [88]
BHA Basic Network applications 2017 [89]
HBHA Hybrid Clustering applications 2018 [90]
BHA Basic Scheduling applications 2018 [91]
BHA Basic Parameter estimation 2018 [92]
LBH Modified Clustering applications 2019 [93]
GSLBH Modified Test functions 2019 [94]
BHA Basic Machine learning 2019 [95]
BHA Basic Machine learning 2019 [96]
BHACO Modified Scheduling applications 2019 [97]
BH-CDOA Hybrid Engineering applications 2019 [98]
BHA-means Hybrid Clustering applications 2020 [99]
BHA Basic Scheduling applications 2020 [100]
BHA Basic Image applications 2020 [101]
BHA Baisc Engineering applications 2020 [102]

exploration and exploitation). For example, the exploration the final enhanced version of BH, called GSLBH. Finally,
has been enhanced using the Levy flight operator. In 17 benchmark functions were employed to evaluate the
contrast, the golden sine (G-S) operator has been utilized performance of GSLBH. The results illustrated that the
to improve the exploitation. It is worth to mention that GSLBH achieved the highest accuracy and best convergence

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

rate comparing with Golden Sine Algorithm (Gold-SA), methods such as CART, LDA and C5.0 decision tree.
Particle Swarm Optimization (PSO), Firefly Algorithm The experimental results using Matthews’ Correlation
(FA), and Whale Optimization Algorithm (WAO). Coefficient (MCC) and Accuracy for model evaluation
Mohammed et al. [77] proposed a new algorithm that indicated better results compared to other classification
depends on the law of gravity (i.e., white holes and hole- methods, especially in predicting for unbalanced and large
black hole attract stars), namely, WH-BH. The initial datasets. Another research has recommended the use of
population in the WH-BH contain a set of candidate BHA to train Convolutional Neural Networks (CNNs) [96].
solutions; the fitness function is calculated for each solution. The new-trained algorithm can be used to classify cancer
The white whole is responsible for determining the worst disease data. This can be achieved by avoiding local minima
solution, while the black whole determines the best solution. entrapment to improve performance.
where 30 benchmark functions from the CEC2014 test
functions are used. 4.2.2 Clustering
Aliman et al. [66] conducted comparisons between
the most common optimization algorithms; Gravitational Data clustering is a typical unsupervised learning technique
Search Algorithm (GSA), PSO, and BHA algorithms. of data mining, where data are grouped into highly cohesive
Twenty-two benchmark functions form the CEC2014 clusters, which are loosely coupled in terms of similarity.
benchmark dataset has been utilized to evaluate the Partition clustering algorithm like k-means is primarily used
performance of all the above algorithms. It’s worth for document clustering since it can help in identifying
mentioning that the set of benchmark functions divided textual patterns in documents amongst thousands of such
into four unimodal functions, 7 multimodal functions, and textual features. Nevertheless, many shortcomings have
six hybrid functions. The experiments repeated twice times been found regarding initial seed dependency and trapping
to understand the behavior of the selected algorithms and into the optimal local solution. To avoid such kind of
achieve a more concrete conclusion. problems and give optimal results, [90] have suggested
In [72], various optimization algorithms have been solved the implementation of BHA using k-mean as embedding.
using BHA. The nineteen benchmark test functions are Their experimental results, which were evaluated on Purity
used to evaluate the performance of the BHA. The used and Silhouette Index, revealed that the hybrid utilization of
benchmark functions divided into seven unimodal test BHA and k-means overtakes the hybridization of GA and
functions, five multimodal test functions, and seven fixed K-means algorithm and gives the most optimal or near to
dimension test functions. The experiment results showed optimal solution.
that the performance of BHA outperformed the PSO and Li et al. in [68] have embedded BHA in the membrane
GA. structure (P system) in which promising results and higher
Wang et al. [78] improved the ability of BHA in optimum value were obtained comparing to other classical
the exploration search and increased the diversity using algorithms such as PSO, GSA and k-means. Abdulwahab
Euclidean distance. To evaluate the performance of the et al. in [93] have proposed an innovative data clustering
proposed algorithm, a set of benchmark functions are technique, which introduces the levy flight into BH
utilized and conducted a comparison with PSO, ABC, DE, algorithm (LBH). The experimental results showed more
and basic BHA. The results illustrated that the proposed effectiveness and robustness in data clustering comparing to
algorithm achieved the best results compare with other other algorithms such as Big Bang Big Crunch, ABC, PSO,
algorithms. GWO, GSA, BA and Bat algorithm.
Another recent study [99] used the same BHA-means
4.2 Machine learning applications combination and found better performance results compared
to GSA, BHA and k-means. Bisecting k-Means (BKmeans)
4.2.1 Classification which is a variation of K-means, is one of the most well-
known partitional clusters algorithm used in many different
BHA can help in detecting and predicting the right application areas, where the dataset is being partitioned into
knowledge for feature selection rather than examining all k clusters. However, only when the initial partitions are
the features in a database. So, it can enhance knowledge close to the final solution, the Bisecting k-means algorithm
discovery in terms of speed and accuracy [95, 103–105]. produces better performance results since the performance
A novel approach was undertaken to classify medical vastly depends on the initial state as well as reach to
data by [67] for feature selection purposes—this approach the local optimal solution. Thus, many papers have sug-
based on combining BHA with a decision tree (C4.5). gested the use of the binary combination of BHA and
The results were compared to seven eminent classification BKmeans to overcome such problems. In their study [61],

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

such combination was proposed, and they obtained outper- determine the optimal parameters of images to increase the
forming results than using traditional bisecting k-means, quality of results. The experiments result illustrated that the
BHA and PSO. performance of the BHA outperformed the PSO, GA, CS,
and HE. It should be noticed that the output of this work
4.3 Scheduling applications converted the entered image to an improved image with
clear embossed.
The task scheduling issue in the Cloud computing Dhanachandra et al. [101] introduced a hybridization
environment encourages many researchers to introduce algorithm that combines between BHA and clustering
optimization algorithms to solve this issue and make the methods for image segmentation. The clustering method
Cloud system more scalable and execute tasks faster. The responsible for enhancing the performance by recursive
goal of task scheduling is to reduce task compilation time density expectation method of surrounding pixels. While
and enhance resource utilization. To solve this issue [97], BHA used to find two couple of segments, completely-
uses the BH and Ant Colony Optimization (ACO) algorithm matched, the results showed that the hybridizing algorithm
to reduce the execution time and makespan, and improve achieved better results compared with other algorithms.
the task scheduling performance. In addition, Ebadifard A generalized constitutional law is applied in [81]
et al. [85] introduce a framework for the cloud resource for an elastic-plastic constitutive model. Employing this
scheduling approach based on multi-objective scheduling general constitutive law, the model identification image is
BH algorithm and Pareto optimizer method. formulated into a variable back analysis problem that is a
Besides, a combination of BH and the Analytic hierarchy traditional and sophisticated optimization. In this paper, a
process (AHP) decision-making model, used by [91] and [100] black hole method is adopted to optimize the effectiveness
to solve the multi-objective workflow tasks scheduling on of the traditional optimization technique. A new back
cloud virtual machines. The AHP algorithm used to transform analysis approach for model detection of rocks surrounding
the single-objective BH to the multi-objective relationship. underground roadways in the coal mine is suggested,
In terms of enhancing the job scheduling on the parallel integrating the generalized constitutive rule for an elastic-
machines by improving the weighted flow time and plastic constitutive model with black hole method. The
maximum tardiness, several formulations proposed by [86]. elastic-plastic constitutive models for two underground
The BH used to optimize the scheduling tasks based roadways at the Huainan coal mine were back-calculated
on idle time and makespan of two machine flow shops. using this new approach. The findings are compared
Besides, in the setup times and processing time of tasks to those of the standard genetic algorithm, the fast-
formalized in fuzzy values form. A multi-objective BH genetic algorithm, and the immune-continuous ant-colony
presented with a combination of Genetic algorithm and algorithm suggested in the previous study. The results
weighted objectives. The result shows an out-performance demonstrate that the proposed design back analysis method
comparing with other approaches [80]. can dramatically increase the calculation performance and
The BH algorithm used to solve the Set Covering prob- the computation impact and is a handy tool for final analysis
lem, which is an optimization issue in the manufacturing of the rock model in coal mine surrounding underground
applications, for instance, the equipment line balancing, roadways.
production planning. The BH improved with the inclusion
of a transformer function and discretization schemes to deal 4.5 Engineering applications
with the binary nature of the issue [75].
A number of issues have been reported regarding cloud Yepes et al. [102] used one of the recent optimization
computing, such as dynamic resource scaling and power algorithms, namely, the Black Hole Algorithm (BHA), for
consumption. It is thus affecting the operational cost and the sustainable design of counterfort retaining walls. The
efficiency of the cloud system. The efficiency of a cloud authors took advantage of BHA to reduce the CO2 and
can be upgraded, and the cost can be significantly reduced cost emissions in earth-retaining walls. As well as enhance
by using a workload prediction model. Kumar et al. in [79] the emissions are related to the environmental impact of
have achieved improvements in the workload prediction construction. The experiment results illustrate that BHA
model when using a neural network and BHA. achieved better results compare with other algorithms by
reducing CO2 emissions.
4.4 Image applications Sharma et al. [87] enhanced the static voltage stability
to find the optimal load shedding using BAH. The main
Yaghoobi et al. [69] utilized BHA to improve image gray- objective of the work is to get rid of the load mandatory,
level, which consider as an essential technique in image which leads to reviving the stabilization of point of
processing. The authors aim to take advantage of BHA to operation. Thus, BAH has been used to aid in achieving the

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

work’s goal. The results illustrated the efficiency of BAH novel mission planner based on BH algorithm for UCAVs
comparing with Bare Bones PSO (BBPSO). 3D mission planning [64].
Ren et al. [70] have been proposed hybrid BHA with
Differential Evolution (DE), called BH-DEO to solve the 4.5.2 Parameter estimation applications
high dimensional electromagnetic optimization problems.
The authors combined these algorithms to benefit the In [92], used BHA to find the optimal model parameter
advantages of DE in convergence rate. As well as BH in self-potential (SP) data. It is worth mentioning that the
to enhance flexibility and diversity. BH-DEO has been SP data is frequently highly nonlinear. Therefore, local-
compared with other algorithms in the literature to evaluate based search algorithms are close to gradient approaches,
its performance. which lead to employing BHA, where it considers as one
Rodrigues et al. [71] used BHA to solve non-technical of the standard global-based search algorithms. The result
losses characterization. The work aims to conduct a parallel illustrated that BHA accurately locates model uncertainty
through the non-technical losses problem and the mission of and model parameters. As well as, it highly potential as an
patchy consumers’ characterization. Two specific datasets innovative approach for SP data inversion.
have been utilized by the Brazilian electric power company Singh and Shukla [88] utilized BHA to determine
in the experiment to evaluate the performance of BHA as the optimal parameter of the electrochemical machining
well as show the best features that related to non-technical process. In this work, two performance parameters have
losses identification. been used; (i) overcut (OC). (ii) material removal rate
Kazamkar et al. [98] utilized Black Hole-Collective (MRR) to achieve the optimal machining parameter
Decision Optimization (BH-CDOA) and Euclidean Dis- settings. The results showed that the BHA got best results
tance to optimize the integrated Heat and Power Economic comparing with bio-geography based optimization (BBO),
Dispatch (CHPED). The authors aim to solve the compli- artificial bee colony (ABC), and genetic algorithm (GA).
cated inequality chains which lead to a satisfactory process
of CHP units and their capacity limits. Four units system 4.6 Software engineering applications
(2-units of cogeneration, 1-unit of electricity, and 1-unit of
heat generation) form the industrial area have been used to One essential utilization of the BHA is in software
evaluate the performance of BH-CDOA. clustering. Software clustering (modularization) refers to
Hasan et al. [62] used BHA to find the optimal power the process of partitioning system into smaller manageable
flow problem. The authors looked to decrease fuel costs, clusters (subsystems) encompassing highly related modules.
voltage deviation and also enhance the stability of the In this context, the multi-objective BH algorithm using
voltage, which considered as an objective function. There GA and auxiliary archive (MOBHGA) algorithm was
are two operations; starts sucking and star absorption have introduced to cluster android mobile applications [82]. The
been used in experiments. The results showed that the BHA proposed algorithm showed better results comparing to
outperformed the other algorithms in the literature using Two-Archive approach. Another study, which demonstrated
IEEE 57-Bus systems and IEEE 30-Bus. good results, was conducted by [65]. They suggested
Ghaffarzadeh and Heydari [73] utilized BHA to deter- an enhancement to the search-based technique using the
mine the optimal design of digital overcurrent relays. The hybrids of BHA and GA for software clustering.
major objective of this work is minimizing the coordination
of relays. In other words, the authors try to combination con- 4.7 Network applications
suming time of overall main relays for error current of near
to all relays to achieve the lowest value. The standard 33- A new effective and reliable method to cable network
bus radial distribution system has been used to evaluate the diagnosis using Time Domain Reflectometry (TDR) and
performance of BHA. The results illustrated that BHA got an Improved Black Hole (IBH) optimization algorithm
high accuracy with fast processing comparing with Particle have been proposed by [89]. For a given wiring network,
Swarm Optimization (PSO) algorithm. a forward system based on RLCG criteria and Finite
Difference Time Domain (FDTD) model is built to emulate
4.5.1 Aircraft applications the TDR response. To detect, classify and locate defects in
wiring networks, the IBH method is employed to overcome
An improved Blackhole algorithm is introduced to enhance the inverse problem. Research investigations are conducted
the intuitive way planning in the Unmanned combat aerial to demonstrate that the proposed method can be utilized for
vehicle (UCAV) [63]. The introduced approach can estimate real-world systems diagnostics.
the quasi-optimal trajectories and the feasibility for UCAV Heuristic search techniques were widely used in the
flight. Furthermore, in another study, the author presented a major issue of reactive power optimization based on

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

their benefits in multi-objective and multi-dimensional have been introduced to extend the lifetime of WSNs. In
optimization. Pei et al. [83] proposed a novel heuristic their study [84] presented a secure routing mechanism from
search scheme called black hole gravitational search source to the sink node, which is able to detect and defence
algorithm (BHGSA), which is a combination of black single and cooperative Gray Hole and Black Hole attack.
hole algorithm (BHA) and gravitational search algorithm The proposed protocol indicated better performance than
(GSA), to address the issue of reactive power optimization LEACH protocol.
in the distribution network day-ahead. The simulation Wireless Sensor Networks (WSN) nowadays have seve-
result is given to overcome IEEE 33-bus distribution ral utilizes for supervisory operations in a variety of areas
network system’s day-ahead reactive power optimization because of the low cost and easy communication. The res-
issue with wind generators and photovoltaic generator using triction of the resources available in the network nodes is
BHGSA. Compared with the conventional gravitational the key issue affecting the network’s sustainability. Exper-
search algorithm, it provides advantages in the ability to iments demonstrated that the production of energy would
search globally and prevent early convergence. increase with the development of control over the number
An intelligent computing scheme called Enhanced Black and position of the cluster heads on, which leads to an in-
Hole (EBH) is suggested by [106] to which the operation crease of the network lifetime. However, this issue is not ap-
and weight factor of the mutation are added. In EBH, propriate to traditional modelling techniques due to the dy-
the original Black Hole (BH) method applies many elites namic nature of the problem and due to continuously chang-
as role models rather than just one. CEC 2013 test suit ing in cluster heads in any of the network traffic. Farahman-
verifies the efficiency of the EBH method and demonstrates dian et al. [74] employed the black hole algorithm (BHA)
better outcomes than the current BH. Additionally, the EBH to measure the number and position of the cluster head
and other praised techniques can be utilized to overcome optimal optimization. The application of the criteria will be
Wireless Sensor Network (WSN) node coverage issues with based on the average energy consumed by each sensor node
adequate results on 3-D terrain. over each data transmission activity cycle, which lead to the
Biju et al. [107] proposed a new approach for solving the creation of a balance in the energy consumption of cluster
problem of reconfiguration of the radial distribution system heads and thus optimize the overall network lifetime.
(RDS) to reduce the losses and improve system reliability.
Efficient reconfiguration chooses the appropriate set of fee- 4.9 Other applications
ders by adjusting the switching status of sectionalizing and
connection switches in order to enhance the resulting RDS Evolutionary algorithms, demonstrated a better outcome,
voltage profile and minimize power loss. Additionally, the have been employed to solve several optimization issues
effect of DG and the capacitor is also taken into conside- [108]. However, if these methods are implemented, they in-
ration in the problem formulation. In order to measure the troduce the challenge of deciding on the correct criteria.
reliability such as SAIFI, SAIDI, CAIDI, AENS and ASAI, Factors are usually defined before the method is executed
the reconfiguration method is also utilized as a method for and include the probabilities of population size, selection
reducing failure rates. This paper introduced the black hole rate and operator. This step is called as offline control and is
algorithm (BHA) application to overcome the optimal net- known as an optimization issue itself. In comparison, input
work reconfiguration issue. A standard radial bus system criteria or online control is a modification of the original
IEEE 69 is selected for the study. To demonstrate the effi- version of the method. The critical point is to modify the
cacy of the proposed method in choosing the optimal solu- criteria so that the interested algorithm can produce a good
tion, simulations are performed on the test system, and find- rate of convergence, and thus achieve the best efficiency.
ings are compared briefly before and after reconfiguration. Soto et al. [109] proposed an adaptive black hole algorithm
capable of adjusting its population based on solving output
4.8 Wireless sensor network dynamically.
To that end, authors employed free search that emerged
Wireless sensor network (WSNs) is a distributed system as a new technology that uses the problem solver to con-
that consists of a tremendous number of mobile sensor trol and evolve its own criteria and heuristics during solving
nodes (SN) and base stations (BS). WSNs can be subjected to become more productive without an expert user’s knowl-
to numerous types of security threats and denials of edge. In order to verify this method, it solved the set
service attacks due to its dynamic topology, resource covering the issue, which is a classic optimization standard
constraint sensors and highly decentralized infrastructure. for many industrial applications, including line balanc-
Thus, energy utilization becomes a big challenge since these ing, crew scheduling, service deployment and databases.
attacks can result in energy-hole creation in the network. Authors demonstrated promising experimental findings in
Different energy-efficient cluster head selection protocols which the suggested technique is capable of achieving

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

different global optimums for such a well-known instance clear from Table 3 that the Black Hole Algorithm obtained
collection from Beasley’s OR-Library while enhancing var- better results over all other optimization algorithms in
ious modern metaheuristic approaches. terms of the average fitness function value. These results
A new technique for adjusting the classifier using reflect the Black Hole Algorithm’s ability to solve different
metaheuristics based on the black hole algorithm is optimization problems according to the obtained results.
introduced in [110]. The purpose of this study is to obtain Table 4 shows the performance of the Black Hole
results comparable to those obtained with the techniques Algorithm against the compared optimization algorithms
of manual noise elimination. The MAHNOB HCI Tagging in terms of the minimization of the fitness function. As
Database was used for evaluating the introduced scheme. presented in Table 4, the Black Hole Algorithm got better
Findings demonstrate that it obtained an accuracy of 92.56% results in most cases, and it is considered as a robust
over 30 runs using the black hole algorithm to optimize the competitor algorithm in the domain of the optimization
Support Vector Machine feature vector. algorithms, which motivate the future researchers to apply
Mehrani et al. [111] conducted a study aimed at opti- and modify the Black Hole Algorithm for solving different
mizing the portfolio using the black hole method in Tehran optimization problems according to its performance in
security exchange and comparing with Markowitz. Due to solving mathematical optimization problems.
its ability to measure downside risk, the semi-variance app- Table 5 shows the Black Hole Algorithm’s performance
roach is employed in this study, and it is attempted to dec- against the compared optimization algorithms in terms of
rease the appropriate portfolio risk. This work carried out the maximization of the fitness function. As well, in Table 5,
in the period 2012 to 2017. The approach of this research is the Black Hole Algorithm got better results in most cases
implemented in terms of research intent, and it is a des- and comparative results on the other cases. This proved the
criptive-correlative data collection basis that they seek to op- ability of the Black Hole Algorithm to solve the various
timize the portfolio using black hole technique retrospective optimization problems.
approach, post-event and by analyzing the observed data. In regards to the standard deviation (STD) as shown in
Research findings indicate that it concluded with the use of Table 6, the Black Hole Algorithm obtained better results
black hole method that the black hole algorithm has achie- in most cases, and it got comparative results in other cases.
ved results the same as Markowitz ’s results overall years We conclude that the Black Hole Algorithm is compara-
and can be the appropriate pattern for portfolio optimization. tive optimization method, and it can get promising results
in some cases and comparative results in other cases com-
pared to different optimization algorithms published in the
5 Results and discussion literature. The obtained results showed that the Black Hole
Algorithm can efficiently determine the optimal global
5.1 Benchmark functions solutions of 23 benchmark test functions with quick con-
vergence speed. The Friedman ranking test is conducted to
In this section, experiments are conducted on 23 classical show the statistical results for the comparative methods as
benchmark functions to show the effectiveness of the Black shown in Table 7. It is clear the BHA is an excellent com-
Hole Algorithm compared to other optimization methods petitor compared to the other well-known methods in the
published in the literature. The competitor algorithms are literature. BHA got the sixth ranking over 11 test methods.
Genetic Algorithm (GA) [112], Particle Swarm Optimiza-
tion (PSO) [113], Differential Evolution (DE) [114], Bat- 5.2 Engineering design problems
inspired Algorithm (BAT) [115], Grasshopper Optimization
Algorithm (GOA) [116], Dragonfly Algorithm (DA) [117], 5.2.1 Problem 1: Tension/compression spring design
Ant Lion Optimizer (ALO) [118], Multi-verse Optimizer problem
(MVO) [119], Moth–Flame Optimizer (MFO) [120], and
Since Cosine Algorithm (SCA) [121]. The tension/compression spring dilemma is presented to
The experiments are conducted in Matlab software using degrade the entire weight of a assigned spring [122, 123].
computers with Windows 10, 16 GB RAM. The parameter Three principal variables are considered in this dilemma, as
setting of used comparative algorithms is set as given in presented in Fig. 4: wire diameter (d), mean coil diameter
the original papers mentioned above. Moreover, the general (D), and the number of active coils (P ).
parameters such as the number of solutions are fixed to 30, The result of the BHA is compared with other well-
and the maximum number of iterations is 500. Each method known optimization methods, including PSO [124], HS
is executed 30 runs to reasonable comparisons. [125], MPM [126] GSA [127], ES [128], MFO [129], BA
Table 3 shows the average performance of the Black Hole [130], WOA [131], MVO [27], RO [132], HHO [133],
Algorithm against the compared optimization algorithms. It GWO [134], and BHA [26].

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

Table 3 The average results of the competitor optimization algorithms on 23 benchmark test functions

Func. BHA GA PSO DE BAT GOA DA ALO MVO MFO SCA

F1 2.66E+01 1.90E-01 2.19E+00 4.00E-01 4.64E+03 1.28E-01 2.66E+01 3.83E+01 3.95E-01 8.05E+00 7.16E+00
F2 4.16E-03 1.15E-02 4.26E-03 8.42E-02 2.43E+01 2.32E-01 2.65E+00 1.30E+01 1.99E-01 4.21E-01 1.03E-01
F3 1.01E+02 5.57E+02 1.03E+02 2.08E+03 7.94E+03 5.87E+00 3.21E+02 9.03E+02 2.86E+00 1.42E+03 2.46E+02
F4 6.72E+00 4.95E+00 3.52E+00 8.39E+00 4.23E+01 3.12E-01 6.79E+00 1.05E+01 5.08E-01 1.12E+01 4.34E+00
F5 1.64E+02 1.22E+02 3.40E+01 1.65E+02 4.37E+06 1.57E+02 4.79E+03 2.95E+03 3.40E+02 6.68E+02 4.01E+02
F6 6.22E+00 4.50E-01 2.83E+00 3.16E-01 4.95E+03 1.20E-01 2.08E+01 1.20E+01 3.88E-01 6.37E+00 3.20E+00
F7 2.10E-02 2.15E-02 2.68E-02 3.33E-02 1.13E+00 2.38E-02 6.33E-02 1.12E-01 1.25E-02 4.41E-02 2.30E-02
F8 -1.82E+03 -3.63E+03 -1.62E+03 -3.55E+03 -2.12E+61 -2.83E+03 -2.79E+03 -2.40E+03 -2.82E+03 -3.22E+03 -1.96E+03
F9 1.80E+01 6.26E+00 1.25E+01 1.60E+01 4.41E+01 2.24E+01 3.54E+01 2.23E+01 2.70E+01 2.67E+01 1.81E+01
F10 2.86E+00 2.59E-01 9.49E-01 6.93E-01 1.53E+01 1.41E+00 5.08E+00 7.15E+00 1.01E+00 2.49E+00 3.86E+00
F11 3.47E-01 2.61E-01 3.60E+00 6.66E-01 4.57E+01 3.17E-01 9.06E-01 4.09E-01 6.86E-01 9.44E-01 7.33E-01
F12 3.10E-02 3.28E-02 2.25E-01 3.63E-02 6.60E+06 1.27E+00 4.50E+00 8.00E+00 4.95E-01 2.43E+00 1.50E+00
F13 1.57E+00 2.63E-02 1.36E-01 7.85E-02 1.70E+07 3.54E-02 7.75E+00 2.82E+01 4.34E-02 2.68E+00 1.58E+00
F14 2.41E+00 1.08E+00 2.82E+00 1.03E+00 2.28E+01 1.36E+00 3.64E+00 6.42E+00 2.62E+00 3.25E+00 2.41E+00
F15 4.35E-03 3.03E-03 8.91E-04 1.25E-03 1.19E-02 3.71E-03 4.34E-03 4.50E-03 4.35E-03 1.14E-03 1.25E-03
F16 -1.11E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.01E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00
F17 4.24E-01 4.09E-01 3.98E-01 3.98E-01 4.85E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01 4.11E-01
F18 3.45E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00
F19 -3.45E+00 -3.86E+00 -3.86E+00 -3.86E+00 -3.79E+00 -3.86E+00 -3.86E+00 -3.85E+00 -3.86E+00 -3.86E+00 -3.85E+00
F20 -3.13E+00 -3.29E+00 -3.27E+00 -3.32E+00 -3.17E+00 -3.27E+00 -3.24E+00 -3.25E+00 -3.26E+00 -3.23E+00 -2.82E+00
F21 -7.13E+00 -3.34E+00 -5.65E+00 -8.34E+00 -3.82E+00 -7.15E+00 -5.47E+00 -4.84E+00 -5.27E+00 -7.44E+00 -1.98E+00
F22 -5.23E+00 -4.36E+00 -5.25E+00 -8.61E+00 -3.80E+00 -8.15E+00 -6.15E+00 -5.25E+00 -6.62E+00 -6.52E+00 -2.84E+00
F23 -7.34E+00 -7.98E+00 -6.07E+00 -8.74E+00 -3.30E+00 -7.25E+00 -5.04E+00 -5.62E+00 -7.24E+00 -7.63E+00 -2.48E+00

Table 4 The minimum results of the competitor optimization algorithms on 23 benchmark test functions

Func. BHA GA PSO DE BAT GOA DA ALO MVO MFO SCA

F1 2.82E-02 9.12E-03 8.39E-08 9.91E-02 9.39E+02 2.68E-02 2.92E-02 1.79E-03 1.54E-01 1.36E+00 4.02E-02
F2 8.22E-06 3.31E-04 8.64E-06 5.07E-02 1.07E+01 8.75E-03 2.08E-01 1.34E+00 7.77E-02 1.39E-01 1.35E-03
F3 8.45E-02 8.33E+01 8.85E-02 9.14E+02 1.46E+03 4.37E-01 3.00E+01 4.32E+01 5.55E-01 1.12E+02 5.60E+00
F4 1.00E+00 1.13E+00 6.14E-02 5.09E+00 2.47E+01 7.89E-02 1.01E+00 2.32E-01 2.40E-01 4.51E+00 3.56E-01
F5 5.64E+01 8.49E+00 8.14E-04 5.74E+01 4.10E+05 8.26E+00 1.45E+01 2.94E-01 8.82E+00 4.82E+01 1.42E+01
F6 7.01E-01 6.51E-05 2.01E-09 6.73E-02 8.05E+02 2.44E-02 1.04E-01 1.76E-04 6.07E-02 7.33E-01 8.76E-01
F7 2.00E-03 3.41E-03 4.34E-03 9.66E-03 1.07E-01 5.66E-03 5.93E-03 2.81E-03 2.82E-03 8.78E-03 1.90E-03
F8 -2.62E+03 -4.07E+03 -2.42E+03 -3.92E+03 -7.29E+62 -3.52E+03 -3.61E+03 -4.16E+03 -3.72E+03 -3.71E+03 -2.47E+03
F9 1.23E+00 2.51E+00 4.17E+00 9.70E+00 1.78E+01 3.98E+00 7.86E+00 5.02E+00 6.13E+00 6.09E+00 1.24E+00
F10 1.45E-03 1.03E-02 2.76E-05 2.82E-01 1.15E+01 1.44E-01 1.53E+00 3.93E-01 2.29E-01 6.15E-01 1.55E-01
F11 1.25E-01 5.45E-02 1.30E+00 3.80E-01 1.84E+01 1.45E-01 1.68E-01 8.51E-02 4.05E-01 5.79E-01 1.94E-02
F12 1.97E-04 1.89E-06 7.85E-10 5.89E-03 3.52E+01 1.38E-02 5.90E-01 2.59E+00 8.87E-03 2.16E-02 1.56E-01
F13 4.54E-01 7.72E-04 9.29E-09 2.20E-02 8.83E+04 4.17E-03 1.94E-02 7.66E-04 1.14E-02 1.63E-01 4.46E-01
F14 9.98E-01 9.98E-01 9.98E-01 9.98E-01 9.98E-01 9.98E-01 9.98E-01 9.98E-01 9.98E-01 9.98E-01 9.98E-01
F15 3.16E-04 7.14E-04 3.86E-04 7.09E-04 8.97E-04 7.40E-04 5.96E-04 3.16E-04 5.11E-04 6.32E-04 4.04E-04
F16 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00
F17 3.54E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01
F18 2.12E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00
F19 -3.56E+00 -3.86E+00 -3.86E+00 -3.86E+00 -3.86E+00 -3.86E+00 -3.86E+00 -3.86E+00 -3.86E+00 -3.86E+00 -3.86E+00
F20 -3.08E+00 -3.32E+00 -3.32E+00 -3.32E+00 -3.32E+00 -3.32E+00 -3.32E+00 -3.32E+00 -3.32E+00 -3.32E+00 -3.12E+00
F21 -1.02E+01 -5.10E+00 -1.02E+01 -1.02E+01 -1.02E+01 -1.02E+01 -1.02E+01 -1.02E+01 -1.02E+01 -1.02E+01 -4.77E+00
F22 -1.34E+01 -9.23E+00 -1.04E+01 -1.04E+01 -1.04E+01 -2.75E+00 -1.04E+01 -1.04E+01 -1.04E+01 -1.04E+01 -4.95E+00
F23 -1.15E+01 -1.05E+01 -1.05E+01 -1.05E+01 -1.05E+01 -1.05E+01 -1.05E+01 -1.05E+01 -1.05E+01 -1.05E+01 -5.76E+00

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

Table 5 The maximum results of the competitor optimization algorithms on 23 benchmark test functions

Func. BHA GA PSO DE BAT GOA DA ALO MVO MFO SCA

F1 1.87E+02 9.31E-01 3.03E+01 1.04E+00 1.11E+04 2.54E-01 1.97E+02 3.34E+02 8.16E-01 2.58E+01 4.58E+01
F2 6.32E-02 4.73E-02 6.55E-02 1.62E-01 4.03E+01 7.51E-01 6.48E+00 4.24E+01 4.85E-01 9.10E-01 2.89E-01
F3 7.60E+02 1.76E+03 7.29E+02 3.47E+03 2.06E+04 2.07E+01 1.16E+03 2.26E+03 1.71E+01 1.01E+04 3.05E+03
F4 2.04E+01 2.04E+01 9.71E+00 1.34E+01 6.18E+01 9.14E-01 1.94E+01 2.65E+01 1.29E+00 3.96E+01 1.74E+01
F5 2.90E+02 1.04E+03 2.16E+02 2.88E+02 1.13E+07 1.80E+03 6.66E+04 2.85E+04 2.84E+03 3.43E+03 3.24E+03
F6 1.91E+01 3.05E+00 4.77E+01 6.68E-01 1.13E+04 2.98E-01 8.59E+01 1.60E+02 1.01E+00 2.01E+01 1.81E+01
F7 7.27E-02 5.23E-02 6.81E-02 5.77E-02 2.69E+00 1.22E-01 2.12E-01 5.47E-01 2.94E-02 1.37E-01 7.16E-02
F8 -1.58E+03 -3.24E+03 -1.28E+03 -3.19E+03 -1.29E+52 -2.20E+03 -2.17E+03 -1.81E+03 -2.15E+03 -2.40E+03 -1.60E+03
F9 6.59E+01 1.17E+01 2.99E+01 2.48E+01 6.87E+01 6.17E+01 7.50E+01 6.37E+01 5.49E+01 7.08E+01 6.43E+01
F10 4.77E+00 1.67E+00 6.21E+00 1.81E+00 1.85E+01 2.49E+00 1.17E+01 1.37E+01 2.81E+00 3.94E+00 2.00E+01
F11 6.24E-01 9.68E-01 6.00E+00 9.34E-01 8.24E+01 6.44E-01 1.64E+00 2.06E+00 9.91E-01 1.15E+00 1.24E+00
F12 3.75E-01 3.57E-01 2.80E+00 1.06E-01 5.97E+07 5.79E+00 2.16E+01 2.72E+01 2.12E+00 1.18E+01 4.58E+00
F13 1.75E+01 1.13E-01 2.89E+00 1.35E-01 9.02E+07 1.14E-01 1.40E+02 5.37E+02 1.62E-01 1.24E+01 1.35E+01
F14 1.74E+01 1.99E+00 1.55E+01 1.99E+00 4.15E+02 2.98E+00 1.55E+01 1.55E+01 1.74E+01 1.17E+01 5.94E+00
F15 2.40E-02 2.08E-02 2.89E-03 2.87E-03 1.12E-01 2.30E-02 2.04E-02 2.49E-02 2.05E-02 2.24E-03 2.80E-03
F16 -2.17E-01 -1.03E+00 -1.03E+00 -1.03E+00 -2.16E-01 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00 -1.03E+00
F17 4.43E-01 6.09E-01 3.98E-01 3.98E-01 1.47E+00 3.98E-01 3.98E-01 3.98E-01 3.98E-01 3.98E-01 4.84E-01
F18 3.95E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.00E+00 3.02E+00
F19 -3.05E+00 -3.86E+00 -3.86E+00 -3.86E+00 -3.09E+00 -3.68E+00 -3.79E+00 -3.72E+00 -3.86E+00 -3.86E+00 -3.83E+00
F20 -3.22E+00 -3.20E+00 -3.20E+00 -3.30E+00 -1.70E+00 -3.12E+00 -3.07E+00 -3.02E+00 -3.12E+00 -3.16E+00 -1.91E+00
F21 -2.63E+00 -2.63E+00 -2.63E+00 -2.60E+00 -8.67E-01 -2.63E+00 -2.63E+00 -2.63E+00 -2.63E+00 -2.63E+00 -4.96E-01
F22 -1.84E+00 -2.75E+00 -1.84E+00 -3.49E+00 -6.70E-01 -1.04E+01 -1.84E+00 -1.84E+00 -1.84E+00 -1.84E+00 -5.22E-01
F23 -2.34E+00 -2.87E+00 -1.68E+00 -3.45E+00 -1.03E+00 -2.42E+00 -1.86E+00 -1.68E+00 -2.42E+00 -1.86E+00 -9.32E-01

Table 6 The STD results of the competitor optimization algorithms on 23 benchmark test functions

Func. BHA GA PSO DE BAT GOA DA ALO MVO MFO SCA

F1 3.65E+01 2.17E-01 5.50E+00 2.15E-01 2.27E+03 6.79E-02 3.75E+01 8.01E+01 1.68E-01 6.02E+00 1.23E+01
F2 1.11E-02 1.26E-02 1.16E-02 2.27E-02 8.08E+00 2.04E-01 1.47E+00 1.01E+01 8.79E-02 2.04E-01 8.06E-02
F3 1.55E+02 4.58E+02 1.52E+02 6.10E+02 4.60E+03 5.09E+00 2.77E+02 6.06E+02 2.84E+00 2.13E+03 5.42E+02
F4 4.98E+00 3.78E+00 2.35E+00 2.05E+00 1.19E+01 2.01E-01 4.84E+00 5.98E+00 2.28E-01 7.00E+00 3.73E+00
F5 6.20E+01 2.37E+02 5.75E+01 6.19E+01 2.98E+06 3.48E+02 1.19E+04 5.57E+03 7.89E+02 8.27E+02 7.68E+02
F6 4.63E+00 7.12E-01 8.10E+00 1.49E-01 2.79E+03 5.83E-02 2.58E+01 3.04E+01 2.09E-01 4.52E+00 3.22E+00
F7 1.80E-02 1.29E-02 1.41E-02 1.13E-02 7.35E-01 2.20E-02 5.08E-02 1.01E-01 6.32E-03 2.68E-02 1.77E-02
F8 2.30E+02 2.01E+02 2.24E+02 1.88E+02 1.21E+62 3.22E+02 3.81E+02 5.07E+02 3.52E+02 3.11E+02 1.94E+02
F9 1.68E+01 2.46E+00 5.42E+00 3.29E+00 1.52E+01 1.27E+01 1.42E+01 1.14E+01 1.43E+01 1.35E+01 1.46E+01
F10 8.47E-01 3.95E-01 1.23E+00 3.66E-01 1.49E+00 8.87E-01 2.13E+00 3.28E+00 6.44E-01 7.20E-01 6.08E+00
F11 1.22E+00 1.70E-01 1.42E+00 1.23E-01 1.83E+01 1.44E-01 4.59E-01 3.85E-01 1.65E-01 1.53E-01 2.49E-01
F12 9.16E-02 9.11E-02 5.08E-01 2.31E-02 1.14E+07 1.32E+00 4.06E+00 5.71E+00 6.10E-01 2.62E+00 1.28E+00
F13 2.12E+01 2.60E-02 5.42E-01 2.76E-02 1.87E+07 2.28E-02 2.31E+01 9.64E+01 2.45E-02 2.83E+00 2.24E+00
F14 1.66E-01 2.43E-01 2.87E+00 1.66E-01 6.75E+01 6.53E-01 3.03E+00 4.28E+00 3.58E+00 2.59E+00 1.09E+00
F15 4.68E-04 4.44E-03 4.13E-04 4.44E-04 2.12E-02 6.18E-03 6.01E-03 5.12E-03 6.99E-03 4.02E-04 5.10E-04
F16 1.26E-01 4.56E-07 3.16E-16 3.36E-10 1.36E-01 2.88E-10 9.08E-08 9.66E-13 1.08E-05 2.71E-16 4.66E-04
F17 3.47E-02 4.05E-02 0.00E+00 5.14E-07 2.12E-01 4.98E-10 2.13E-07 2.24E-12 2.00E-05 3.15E-12 1.97E-02
F18 3.20E+01 8.67E-05 5.84E-13 4.66E-12 7.32E-04 2.55E-09 8.32E-07 1.04E-11 6.47E-05 4.85E-15 3.12E-03
F19 2.21E-01 2.28E-08 2.42E-15 1.86E-14 2.15E-01 3.20E-02 1.19E-02 2.94E-02 7.52E-05 3.31E-14 6.98E-03
F20 5.47E-02 5.37E-02 5.88E-02 6.11E-03 3.51E-01 7.00E-02 9.14E-02 9.06E-02 7.39E-02 5.35E-02 3.12E-01
F21 3.15E+01 9.00E-01 3.52E+00 2.27E+00 2.81E+00 3.50E+00 2.55E+00 2.65E+00 2.86E+00 3.40E+00 1.48E+00
F22 3.30E+00 1.93E+00 3.30E+00 2.03E+00 2.60E+00 3.26E+00 2.84E+00 3.18E+00 3.73E+00 3.46E+00 1.30E+00
F23 3.45E+00 3.01E+00 3.62E+00 1.76E+00 2.50E+00 3.58E+00 2.52E+00 3.62E+00 3.75E+00 3.54E+00 1.15E+00

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

Table 7 The Freidman raning test results of the competitor optimization algorithms on 23 benchmark test functions

Func. BHA GA PSO DE BAT GOA DA ALO MVO MFO SCA

F1 8 2 5 4 11 1 8 10 3 7 6
F2 1 3 2 4 11 7 9 10 6 8 5
F3 3 7 4 10 11 2 6 8 1 9 5
F4 6 5 3 8 11 1 7 9 2 10 4
F5 4 2 1 5 11 3 10 9 6 8 7
F6 7 4 5 2 11 1 10 9 3 8 6
F7 2 3 6 7 11 5 9 10 1 8 4
F8 1 10 2 9 11 7 5 4 6 8 3
F9 4 1 2 3 11 7 10 6 9 8 5
F10 7 1 3 2 11 5 9 10 4 6 8
F11 3 1 10 5 11 2 8 4 6 9 7
F12 1 2 4 3 11 6 9 10 5 8 7
F13 6 1 5 4 11 2 9 10 3 8 7
F14 4 2 7 1 11 3 9 10 6 8 4
F15 8 5 1 3 11 6 7 10 8 2 3
F16 10 1 1 1 10 1 1 1 1 1 1
F17 10 8 1 1 11 1 1 1 1 1 9
F18 1 2 2 2 2 2 2 2 2 2 2
F19 11 1 1 1 10 1 1 8 1 1 8
F20 10 2 3 1 9 3 7 6 5 8 11
F21 4 10 5 1 9 3 6 8 7 2 11
F22 8 9 6 1 10 2 5 6 3 4 11
F23 4 2 7 1 10 5 9 8 6 3 11

Average 5.3478261 3.6521739 3.7391304 3.4347826 10.26087 3.3043478 6.826087 7.3478261 4.1304348 5.9565217 6.304348
Ranking 6 3 4 2 11 1 9 10 5 7 8

Table 8 presents the achieved results of the comparative Figure 5 illustrates the curves of the objective values, the
methods in determining the tension/compression spring pro- trajectory of the first solution, and the convergence curve
blem. The acquired constraints costs and optimal decision of the BHA method when solving the tension/compression
cost of the optimal solution for all the similar methods are spring design problem. The BHA method reported the
presented in Table 8. It is clear from Table 8 that the BHA is optimal solution very quickly, before iteration number 40,
a superior optimization method compared to other methods according to the given convergence curve in Fig. 5. The
by achieving better results where the optimal variables at x ∗ figures of the decision variables and the trajectory of the
= (d = 0.05172800, D = 0.357644000, P = 11.24454300) first solution are produced to determine the ability of the
with an objective function: F(x ∗ ) = 0.01267400. BHA to investigate the search space.

Fig. 4 Tension/compression
spring design problem

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

Table 8 The algorithm’s results for solving the tension/compression spring design problem

Algorithm Optimal values for variables Optimal

d D N cost

PSO [124] 0.05172800 0.357644000 11.24454300 0.01267470


RO [132] 0.05137000 0.349096000 11.76279000 0.01267880
MPM [126] 0.05000000 0.315900000 14.25000000 0.01283340
WOA [131] 0.05120700 0.345215000 12.00403200 0.01267630
MFO [129] 0.05199446 0.364109320 10.86842186 0.01266690
BA [130] 0.05169000 0.356720000 11.28850000 0.01267000
ES [128] 0.05164300 0.355360000 11.39792600 0.01269800
MVO [27] 0.05251 0.37602 10.33513 0.012790
GSA [127] 0.05027600 0.323680000 13.52541000 0.01270220
HHO [133] 0.05197639 0.363669510 10.89275181 0.01266674
HS [125] 0.05115400 0.349871000 12.07643200 0.01267060
GWO [134] 0.05169 0.356737 11.28885 0.012666
BHA [26] 0.05172800 0.357644000 11.24454300 0.01267400

5.2.2 Problem 2: Pressure vessel design problem the BHA got better results than almost all comparative
methods methods by getting optimal variables at x ∗ = (Ts =
The pressure vessel design dilemma tries to degrade the 0.812500, Th = 0.437500, R = 42.098353, L = 176.637751)
entire cost of an assigned cylindrical pressure vessel [135, with an objective function: F(x ∗ ) = 6059.7258.
136]. In this dilemma, as presented in Fig. 6, four variables Figure 7 represents the curves of the decision values, the
are examined and analyzed, including the width of the shell trajectory of the 1st solution, and the convergence curve
(Ts ), the width of the head (Th ), internal radius (R), and of the proposed method (BHA) when solving the pressure
height of the cylindrical part without studying the head (L). vessel design problem. The BHA archived the optimal
The result of the BHA is analyzed and compared with solution at the beginning of the enhancement process, at
other well-known optimization methods, including Branch- iteration number 160. It is clear that the solution’s diversity
bound [137], OBSCA [121], HS [138], GA [139], CSCA is stable as presented in Fig. 7. Moreover, we concluded that
[140], CPSO [124], PSO-SCA [141], MVO [27], HPSO the BHA method obtained the most fitting parameters and
[142], WOA [131], ES [128], GWO [134], and GSA [127]. the minimum objective function.
In Table 9, the results of the BHA and the other
optimization methods are provided. These methods involve
Branch-bound, OBSCA, HS, GA, CSCA, PSO-SCA, MVO, 6 Discussion and theoretical analysis
HPSO, WOA, ES, GSA, and ACO on this utilization.
The acquired constraints conditions and optimal decision The benefit of meta-heuristic algorithms is confirmed for se-
variables of the optimal solution for all the used algorithms veral reasons. The first reason is that the benefit of the sto-
are presented in Table 9. It is clear from Table 9 that chastic style supports these algorithms to avoid being stuck

Fig. 5 Qualitative results for the tension/compression spring design problem

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

Fig. 6 Pressure vessel design


problem

in local exploitation and focalizing to the near-optimal sol- optimization techniques is due to its origin, abilities, sim-
ution. Directly, the aim is not to obtain the best solution to the plicity, and balance experimentally.
given problem but to obtain the near-optimal solution (with The Black Hole Algorithm is a simple optimization
high fitness value within a reasonable running time). The method over many other ways, as shown in the literature.
main factor in creating this aim is to make the proper stab- The overall results of the Black Hole Algorithm can obtain
ility across search procedures. Exploration propose is to disc- promising results as matching with other similar optimiza-
over the more promising areas in a complex-wide search area. tion methods in the literature. This algorithm has been
Consequently, the exploitation method is to become applied to solve various optimization problems, including
intensified by the local search scheme in an encouraging global optimization problems, engineering problems, clus-
area to suggest better solutions. The optimal accomplish- tering, task scheduling, image processing, and other applica-
ment and effectiveness of a novel optimization technique tions, as found in the literature. Several types of researches
in trade-off these methods, the better technique will be modified the basic procedure of the Black Hole Algorithm
given. The current meta-heuristic algorithms adjust the bal- to solve the optimization problems such as binary, modify,
ance amongst these two approaches in various approaches. hybrid, and multi-objective versions.
They could be further qualified and adjusted for local or The Black Hole Algorithm is free from parameter
for global search procedures. Two or more optimization tuning arguments, which facilitate problem-solving on
techniques are mixed by hybrid its elements to get its essen- the researchers. Researchers typically need to tune the
tial characteristics together and powerful than each one algorithm’s parameters to find the most suitable solutions
alone while averting as much as possible their disadvantages according to the chosen parameters. So, no parameters that
and shortcomings. The second cause, the completion of need tuning have been added to this algorithm.

Table 9 The algorithm’s results for solving the pressure vessel design problem

Algorithm Optimal values Optimal

Ts Th R L cost

Branch-bound [137] 1.125 0.625 48.97 106.72 7982.5


GSA [127] 1.125 0.625 55.9886598 84.4542025 8538.8359
HS [138] 1.125000 0.625000 58.29015 43.69268 7197.730
WOA [131] 0.812500 0.437500 42.0982699 176.638998 6059.7410
MVO [27] 0.8125 0.4375 42.090738 176.73869 6060.8066
PSO-SCA [141] 0.8125 0.4375 42.098446 176.6366 6059.71433
CPSO [124] 0.8125 0.4375 42.091266 176.7465 6061.0777
HPSO [142] 0.8125 0.4375 42.0984 176.6366 6059.7143
CSCA [140] 0.8125 0.4375 42.098411 176.63769 6059.7340
ES [128] 0.8125 0.4375 42.098087 176.640518 6059.74560
GA [139] 0.81250 0.43750 42.097398 176.65405 6059.94634
OBSCA [121] 1.2500 0.0625 59.1593 70.8437 5833.9892
GWO [134] 0.812500 0.434500 42.089181 176.758731 6051.5639
BHA [26] 0.812500 0.437500 42.098353 176.637751 6059.7258

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

Fig. 7 Qualitative results for the


pressure vessel design problem

The most promising results obtained by this algorithm given technique fully. The problem of early convergence
was using the modified versions. This motivates future (premature convergence) in the Black Hole Algorithm
researchers to change the conventional Blak Hole Algorithm typically begins the recovery procedure to be stuck through-
to solve the different problems effectively. The Black out the exploitation search. This effect usually occurs while
Hole Algorithm has not addressed several optimization the solutions divergence decreases, and the solutions can-
problems. The Black Hole Algorithm can solve various not avert falling in the local optima. Furthermore, there are
other real-world applications, such as task scheduling considerable potentials for researchers to employ and apply
in cloud computing environments, several engineering the advantages of the Black Hole Algorithm to approach the
design problems, data mining problems and text clustering complex industry and any real-world problems.
problems, smart home appliances management, brain signal In addition, there is another point has influence on
process, FMRI image tracking, recent CEC benchmark increasing the number of publications of Black Hole
problems, and many others. Algorithm is comparison between it and PSO in [76].
As a result of the Black Hole Algorithm and due to its This comparison concluded that PSO is better than BHA
simplicity and similarity to the real behaviour in nature, that leads the researchers to doubt in the performance of
the Black Hole Algorithm has earned significant attention BHA. This, our, observation has been reached based on
from researchers since it was proposed. This algorithm can the decreasing on the number of publications after this
be classified into four elements: parameters and solutions comparison occurred in 2016 (see Fig. 3). In addition,
initialization, parameter tuning, exploitation search strategy there is another reason that effect on increasing the number
(local), and exploration search strategy (global). Several of publication results from presented other metaheuristic
types of researches have attempted to enhance the perfor- techniques in 2016 and years later such as whale
mance of the Black Hole Algorithm by improving its parts. optimization algorithm, grey wolf optimization, arithmetic
Hence, since the development of Black Hole Algorithm, optimization algorithm and others.
several versions have been proposed. However, among the
searching for the conventional Black Hole Algorithm (basic
version) in the literature, different details can be found. It is 7 Conclusion and potential future works
clear that the main procedure of this algorithm is straight-
forward, and it is accessible to implement compared to This paper introduced a comprehensive review of the Black
other similar optimization algorithms. Additionally, other Hole Algorithm (BHA) that established its high ability
improved implementations of the Black Hole Algorithm to solve different problems at variant applications. The
have been determined with several modifications of the contents of this review paper were collected from the lit-
main procedures of the searching strategies. Thus, the erature that has been published from the beginning of the
general structure of the first implementations of the Black first version of BHA from 2013 until 2020. This literature
Hole Algorithm is introduced in this paper, to present a presented a variant of modifications of BHA, also, its advan-
clear understanding of these modifications and the relation- tages, robustness, and limitations. In addition, we exhaus-
ships between them. Furthermore, this paper analyses the tively summarized different problems that have been solved
improvements in the various components of the Black Hole using BHA such as global optimization, engineering design,
Algorithm to provide a clear procedure for the researchers image processing, machine learning, software engineering,
who are seeking to enhance this algorithm with helpful aircraft, wireless sensor network, traditional networks, and
references and direction. other important applications. Besides that, the BHA is con-
The main difficulty of the Black Hole Algorithm is sidered one of the most metaheuristic techniques that have a
defining the probabilistic convergence properties of the high ability to find suitable solutions for the tested problems
Black Hole Algorithm, which is required to understand the as concluded from the published papers. This paper supports

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

researchers who planning to present work based on BHA 5. Wang S-K, Chiou J-P, Liu C-W (2009) Parameters tuning of
by valuable information about how can be applied BHA, its power system stabilizers using improved ant direction hybrid
differential evolution. Int J Elect Power Energ Syst 31(1):34–42
properties, and its weaknesses. Therefore, the Black Hole
6. Biswas K, Vasant PM, Vintaned JAG, Watada J (2020) A review
Algorithm can be applied to new different applications. of metaheuristic algorithms for optimizing 3d well-path designs.
According to the discussion, we concluded that the Archives of Computational Methods in Engineering
behavior of the Black Hole Algorithm still requires 7. Ide J, Schöbel A (2016) Robustness for uncertain multi-objective
optimization: a survey and analysis of different concepts. OR
improvements, and this open the door to extent BHA.
Spect 38(1):235–271
This can be achieved by combining it with other MH 8. Bandaru S, Ng AH, Deb K (2017) Data mining methods for
techniques, traditional optimization methods, and other knowledge discovery in multi-objective optimization: Part a-
mathematical concepts. Therefore, in future works, we survey. Expert Syst Appl 70:139–159
9. Abualigah L, Diabat A (2020) A novel hybrid antlion optimiza-
recommend the following viewpoints to improve BHA and
tion algorithm for multi-objective task scheduling problems in
apply to different applications. cloud computing environments. Clust Comput 1–19
10. Abualigah L, Khader AT (2017) Unsupervised text feature
• Using the operators of some metaheuristic techniques selection technique based on hybrid particle swarm optimization
to improve the exploration and exploitation abilities of algorithm with genetic operators for the text clustering. J
BHA. Supercomput 73(11):4773–4795
• Modify the features of BHA and apply it to solve 11. Abualigah L, Khader AT, Hanandeh ES (2019) Modified krill
herd algorithm for global numerical optimization problems.
different problems in the Internet of things (IoT) for In: Advances in nature-inspired computing and applications,
example, feature selection, task scheduling at edge Springer, pp 205–221
layer, or cloud layer. 12. Boveiri H, Elhoseny M (2018) A-coa: an adaptive cuckoo opti-
• Enhancing the ability of BHA as a Multiobjective tech- mization algorithm for continuous and combinatorial optimiza-
tion. Neural Comput Applic 1–25
nique and applied to solve real-world problems such 13. Yazdani M, Jolai F (2016) Lion optimization algorithm (loa):
as image segmentation, feature selection, engineering a nature-inspired metaheuristic algorithm. J Computa Des Eng
design, renewable energy. Since most of these applica- 3(1):24–36
tions have conflict objectives. 14. Dhal KG, Das A, Ray S, Gálvez J, Das S (2019) Nature-
• Improving BHA as a many-objective approach to inspired optimization algorithms and their application in multi-
thresholding image segmentation. Archiv Comput Methods Eng
handling problems that consist of more than three 1–34
objectives. 15. Traversa FL, Cicotti P, Sheldon F, Di Ventra M (2018) Evidence
of exponential speed-up in the solution of hard optimization
problems. Complexity
Acknowledgements This research was supported by TPU develop-
ment program Priority 2030. 16. Vempala SS, Wang R, Woodruff DP (2020) The communication
complexity of optimization. In: Proceedings of the Fourteenth
Annual ACM-SIAM Symposium on Discrete Algorithms,
Declarations SIAM, pp 1733–1752
17. Kaur A, Jain S, Goel S (2020) Sandpiper optimization algorithm:
Ethics approval This article does not contain any studies with human a novel approach for solving real-life engineering problems. Appl
participants or animals performed by any of the authors. Intell 50(2):582–619
18. Lones MA (2020) Mitigating metaphors: A comprehensible
guide to recent nature-inspired algorithms. SN Comput Sci
Informed consent Informed consent was obtained from all individual 1(1):49
participants included in the study. 19. Deb S, Gao X-Z, Tammi K, Kalita K, Mahanta P (2019) Recent
studies on chicken swarm optimization algorithm: a review
Conflict of Interests The authors declare that there is no conflict of (2014–2018). Artif Intell Rev 1–29
interest regarding the publication of this paper. 20. Hussain K, Salleh MNM, Cheng S, Shi Y (2019) Metaheuristic
research: A comprehensive survey. Artif Intell Rev 52(4):2191–
2233
21. Abdel-Basset M, Shawky LA (2019) Flower pollination algorithm:
References A comprehensive review. Artif Intell Rev 52(4):2533–2557
22. Beyer H-G, Schwefel H-P (2002) Evolution strategies–a
1. Hatta N, Zain AM, Sallehuddin R, Shayfull Z, Yusoff Y (2019) comprehensive introduction. Nat Comput 1(1):3–52
Recent studies on optimisation method of grey wolf optimiser 23. Hsiao Y-T (2004) Multiobjective evolution programming
(gwo): a review (2014–2017). Artif Intell Rev 52(4):2651–2683 method for feeder reconfiguration. IEEE Trans Power Syst
2. Hussein WA, Sahran S, Abdullah SNHS (2017) The variants of 19(1):594–599
the bees algorithm (ba): A survey. Artif Intell Rev 47(1):67–121 24. Das S, Suganthan PN (2010) Differential evolution: a survey of
3. Gharehchopogh FS, Shayanfar H, Gholizadeh H (2019) A com- the state-of-the-art. IEEE Trans Evolution Comput 15(1):4–31
prehensive survey on symbiotic organisms search algorithms. 25. Mirjalili S (2019) Genetic algorithm. In: Evolutionary algorithms
Artif Intell Rev 1–48 and neural networks. Springer, pp 43–55
4. Karaboga D, Akay B (2009) A comparative study of artificial bee 26. Hatamlou A, hole Black (2013) A new heuristic optimization
colony algorithm. Appl Math Comput 214(1):108–132 approach for data clustering. Inform Sci 222:175–184

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

27. Mirjalili S, Mirjalili S, Hatamlou A (2016) Multi-verse 46. Mohammed SK, Ibrahim Z, Daniyal H, Aziz NAA (2016) A
optimizer: a nature-inspired algorithm for global optimization. new hybrid gravitational search–black hole algorithm. In: The
Neural Comput Applic 27(2):495–513 National Conference for Postgraduate Research, pp 834–842
28. Abualigah L, Shehab M, Alshinwan M, Alabool H (2019) 47. Bányai Á, Bányai T, Illés B (2017) Optimization of consignment-
Salp swarm algorithm: a comprehensive survey. Neural Comput store-based supply chain with black hole algorithm. Complexity
Applic 1–21 48. Veres P, Bányai T, Illés B (2017) Optimization of in-plant
29. Abualigah L, Shehab M, Alshinwan M, Mirjalili S, Abd Elaziz production supply with black hole algorithm. In: Solid State
M (2020) Ant lion optimizer: A comprehensive survey of its Phenomena. vol. 261, Trans Tech Publ, pp 503–508
variants and applications. Archives of Computational Methods in 49. Wu C, Wu T, Fu K, Zhu Y, Li Y, He W, Tang S (2017) Amobh:
Engineering Adaptive multiobjective black hole algorithm, Computational
30. Shehab M, Alshawabkah H, Abualigah L, Nagham A-M (2020) intelligence and neuroscience
Enhanced a hybrid moth-flame optimization algorithm using new 50. Garcı́a J, Crawford B, Soto R, Garcı́a P (2017) A multi dynamic
selection schemes. Eng Comput 1–26 binary black hole algorithm applied to set covering problem.
31. Abualigah LMQ (2019) Feature selection and enhanced krill herd In: International Conference on Harmony Search Algorithm.
algorithm for text document clustering. Springer, Berlin Springer, pp 42–51
32. Abualigah L, Khader AT, Hanandeh ES (2018) A new feature selec- 51. Soto R, Crawford B, Olivares R, Niklander S, Johnson F, Paredes
tion method to improve the document clustering using particle F, Olguı́n E (2017) Online control of enumeration strategies
swarm optimization algorithm. J Comput Sci 25:456–466 via bat algorithm and black hole optimization. Nat Comput
33. Abualigah L, Diabat A, Mirjalili S, Abd Elaziz M, Gandomi AH 16(2):241–257
(2021) The arithmetic optimization algorithm. Comput Method 52. Pashaei E, Aydin N (2017) Binary black hole algorithm for
Appl Mechan Eng 376:113609 feature selection and classification on biological data. Appl Soft
34. Deeb H, Sarangi A, Mishra D, Sarangi SK (2020) Improved blck Comput 56:94–106
hole optimization algorithm for data clustering, Journal of King 53. Mohammed SK, Daniyal H, Subari N, Muhammad B, Musa Z,
Saud University-Computer and Information Sciences Aziz AA, Ibrahim Z, Mohd Azmi KZ, Rahman TA Improving
35. Abdul Aziz NH, Ab Aziz NA, Shapiai MI, Ab Rahman T, Adam the effectiveness of the black hole algorithm using a local search
A, Mokhtar N, Md Yusof Z, Subari N (2020) A survey on technique. International Journal of Simulation–Systems, Science
applications of black hole algorithm & Technology 18 (4)
36. Ibrahim Z, Mohammed SK, Subari N, Ab Aziz NA, Aziz NHA, 54. Gao W (2017) Investigating the critical slip surface of soil
Ab Rahman T, Adam A, Yusof ZM, Shapiai MI, Mokhtar N slope based on an improved black hole algorithm. Soils Found
(2020) A review on fundamental advancements of black hole 57(6):988–1001
algorithm 55. Ibrahim Z, Mohammed SK, Subari N, Adam A, Yusof ZM, Ab
37. Azizipanah-Abarghooee R, Niknam T, Bavafa F, Zare M (2014) Aziz NA, Aziz NHA, Ab Rahman T, Shapiai MI, Mokhtar N
Short-term scheduling of thermal power systems using hybrid (2020) A survey on applications of black hole algorithm
gradient based modified teaching–learning optimizer with black 56. Hatamlou A (2018) Solving travelling salesman problem using
hole algorithm. Elect Power Syst Res 108:16–34 black hole algorithm. Soft Comput 22(24):8167–8175
38. Aslani H, Yaghoobi M, Akbarzadeh-t M-R (2015) chaotic inertia 57. Ibrahim Z, K Mohammed S, Subari N, Adam A, Yusof ZM, Ab
weight in black hole algorithm for function optimization. In: Aziz NA, Aziz NHA, Ab Rahman T, Mokhtar N (2020) Black
2015 international congress on technology, communication and hole white hole algorithm with local search
knowledge (ICTCK). IEEE, pp 123-129 58. Pashaei E, Pashaei E, Aydin N (2019) Gene selection using
39. Kumar S, Datta D, Singh SK (2015) Black hole algorithm and hybrid binary black hole algorithm and modified binary particle
its applications. In: Computational intelligence applications in swarm optimization. Genomics 111(4):669–686
modeling and control. Springer, pp 147–170 59. Xie W, Wang J, Xing C, Guo S, Guo M, Zhu L (2020) Extreme
40. Gao W, Wang X, Dai S, Chen D (2016) Study on stability of high learning machine soft-sensor model with different activation
embankment slope based on black hole algorithm. Environ Earth functions on grinding process optimized by improved black hole
Sci 75(20):1381 algorithm. IEEE Access 8:25084–25110
41. Olivares R, Soto R, Crawford B, Barrı́a M, Niklander S (2016) 60. Rao J, Wu T, Chong W, Li Y, He W (2020) Momentum
Evaluation of choice functions to self-adaptive on constraint multi-objective optimization algorithm based on black hole
programming via the black hole algorithm. In: 2016 XLII Latin algorithm. IOP Conf Series Mater Sci Eng 768:072046.
American Computing Conference (CLEI). IEEE, pp 1–8 https://doi.org/10.1088/1757-899x/768/7/072046
42. Yaghoobi S, Mojallali H (2016) Modified black hole algorithm 61. Eskandarzadehalamdary M, Masoumi B, Sojodishijani O (2014)
with genetic operators. Int J Comput Intell Syst 9(4):652–665 A new hybrid algorithm based on black hole optimization and
43. Ramos CC, Rodrigues D, de Souza AN, Papa JP (2016) On bisecting k-means for cluster analysis. In: 2014 22nd iranian
the study of commercial losses in brazil: a binary black hole conference on electrical engineering (ICEE). IEEE, pp 1075–
algorithm for theft characterization. IEEE Trans Smart Grid 1079
9(2):676–683 62. Hasan Z, El-Hawary ME (2014) Optimal power flow by black
44. Gómez A., Crawford B, Soto R, Jaramillo A, Mansilla S, Salas hole optimization algorithm. In: 2014 IEEE Electrical Power and
J, Olguı́n E (2016) An binary black hole algorithm to solve Energy Conference. IEEE, pp 134–141
the set covering problem. In: 2016 11th iberian conference on 63. Heidari A, Abbaspour R (2014) Improved black hole algorithm
information systems and technologies (CISTI). IEEE, pp 1–5 for efficient low observable ucav path planning in constrained
45. Rubio ÁG, Crawford B, Soto R, Jaramillo A, Villablanca SM, aerospace. Adv Comput Sci Int J 3(3):87–92
Salas J, Olguı́n E (2016) An binary black hole algorithm 64. Heidari A, Abbaspour R (2020) A gravitational black hole
to solve set covering problem. In: International conference algorithm for autonomous ucav mission planning in 3d realistic
on industrial, Engineering and Other Applications of Applied environments, International Journal of Computer Applications 95
Intelligent Systems. Springer, pp 873–883 (9)

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

65. Jeet K, Dhir R (2015) Software architecture recovery using 84. Dongare SP, Mangrulkar R (2016) Optimal cluster head selection
genetic black hole algorithm. ACM SIGSOFT Softw Eng Notes based energy efficient technique for defending against gray hole
40(1):1–5 and black hole attacks in wireless sensor networks. Procedia
66. Aliman MN, Ibrahim Z, Naim F, Nawawi SW, Sudin S (2020) Computer Science 78(C):423–430
Performance evaluation of black hole algorithm, gravitational 85. Ebadifard F, Babamir SM (2017) Optimizing multi objective
search algorithm and particle swarm optimization. Malaysian J based workflow scheduling in cloud computing using black hole
Fund Appl Sci 11 (1) algorithm. In: 2017 3th international conference on web research
67. Pashaei E, Ozen M, Aydin N (2015) An application of black (ICWR). IEEE, pp 102–108
hole algorithm and decision tree for medical problem. In: 86. Jeet K, Sharma S, Nailwal KK (2017) Two-machine fuzzy flow
2015 IEEE 15th international conference on bioinformatics and shop scheduling using black hole algorithm. Global J Pure Appl
bioengineering (BIBE). IEEE, pp 1–6 Math 13(6):1935–1946
68. Li Q, Pei Z (2015) The black hole clustering algorithm based 87. Sharma NK, Varma A, Choube S, Yadav SK (2017) Optimal load
on membrane computing. In: 2015 International symposium on shedding to improve static voltage stability employing black hole
computers & informatics, atlantis press optimization algorithm. In: 2017 6th international conference on
69. Yaghoobi S, Hemayat S, Mojallali H (2015) Image gray- computer applications in electrical engineering-recent advances
level enhancement using black hole algorithm. In: 2015 2nd (CERA). IEEE, pp 341–346
international conference on pattern recognition and image 88. Singh D, Shukla R (2017) Parameter optimization of electro-
analysis (IPRIA). IEEE, pp 1–5 chemical machining process using black hole algorithm. In:
70. Ren Z, He S, Zhang D, Koh C-S (2020) A novel hybrid algorithm Materials Science and Engineering Conference Series. vol 282, p
of black hole and differential evolution for high dimensional 012006
electromagnetic optimal problems 89. Smail M, Bouchekara H, Pichon L, Boudjefdjouf H, Amloune A,
71. Rodrigues D, Ramos CCO, De Souza AN, Papa JP (2015) Black Lacheheb Z (2017) Non-destructive diagnosis of wiring networks
hole algorithm for non-technical losses characterization. In: 2015 using time domain reflectometry and an improved black hole
IEEE 6th latin american symposium on circuits & systems algorithm. Nondestructive Testing Eval 32(3):286–300
(LASCAS). Ieee, pp 1–4 90. Rafi M, Aamer B, Naseem M, Osama M (2018) Solving
72. Farahmandian M, Hatamlou A (2015) Solving optimization document clustering problem through meta heuristic algorithm:
problems using black hole algorithm. J Adv Comput Sci Technol black hole. In: Proceedings of the 2nd International Conference
4(1):68 on Machine Learning and Soft Computing, pp 77–81
73. Ghaffarzadeh N, Heydari S (2015) Optimal coordination of 91. Ebadifard F, Babamir SM (2020) Optimal workflow scheduling
digital overcurrent relays using black hole algorithm. World Appl in cloud computing using ahp based multi objective black hole
Program 5(2):50–55 algorithm
74. Farahmandian M, Hatamlou A (2020) Optimization of energy 92. Warnana DD et al (2018) Black hole algorithm for determining
consumption in clustered nodes of wsn using the black hole model parameter in self-potential data. J Appl Geophys 148:189–
algorithm 200
93. Abdulwahab HA, Noraziah A, Alsewari AA, Salih SQ (2019)
75. Soto R, Crawford B, Figueroa I, Niklander S, Olguı́n E (2016)
An enhanced version of black hole algorithm via levy flight
A black hole algorithm for solving the set covering problem.
for optimization and data clustering problems. IEEE Access
In: International conference on industrial, Engineering and Other
7:142085–142096
Applications of Applied Intelligent Systems. Springer, pp 855–
94. Xie W, Wang J, Tao Y (2019) Improved black hole algorithm
861
based on golden sine operator and levy flight operator. IEEE
76. Gupta H, Gupta A, Gupta SK, Nayak P, Shrivastava T
Access 7:161459–161486
(2016) How effective is black hole algorithm? In: 2016 2nd
95. Khatatneh K (2020) Using black hole algorithm for solving
International Conference on Contemporary Computing and
feature selection problem. International Journal of Advances in
Informatics (IC3I). IEEE, pp 474–478
Electronics and Computer Science 6 (4)
77. Mohammed SK, Ibrahim Z, Daniyal H, Aziz NAA (2016) White 96. Salih SQ (2020) A new training method based on black hole
hole-black hole algorithm. In: The National Conference for algorithm for convolutional neural network. Journal of Southwest
Postgraduate Research, pp 824–833 Jiaotong University 54 (3)
78. Wang T, Liu W, Liu C (2016) Optimization algorithm of black- 97. Jethava AN, Desai MR (2019) Optimizing multi objective based
hole based on euclidean distance. J Shenyang Univ Technol dynamic workflow using aco and black hole algorithm in cloud
38(2):201–205 computing. In: 2019 3rd International Conference on Computing
79. Kumar J, Singh AK (2016) Dynamic resource scaling in cloud Methodologies and Communication (ICCMC). IEEE, pp 1144–
using neural network and black hole algorithm. In: 2016 1147
fifth international conference on eco-friendly computing and 98. Kazamkar D, Ehsandoost SH, Lotfipour A (2019) Combined
communication systems (ICECCS). IEEE, pp 63–67 heat and power economic dispatch (chped) using euclidean
80. Jeet K, Dhir R, Singh P (2016) Hybrid black hole algorithm for distance and black hole-collective decision optimization (bh-
bi-criteria job scheduling on parallel machines. Int J Intell Syst cdoa) algorithm. In: 2019 24th electrical power distribution
Appl 8(4):1–17 conference (EPDC). IEEE, pp 95–99
81. Gao W, Ge M, Chen D, Wang X (2016) Back analysis for rock 99. Pal S, Pal S (2020) Black hole and k-means hybrid clustering
model surrounding underground roadways in coal mine based on algorithm. In: Computational Intelligence in Data Mining.
black hole algorithm. Eng Comput 32(4):675–689 Springer, pp 403–413
82. Jeet K, Dhir R (2016) Software clustering using hybrid multi- 100. Ebadifard F, Babamir SM (2020) Scheduling scientific work-
objective black hole algorithm. In: SEKE, pp 650–653 flows on virtual machines using a pareto and hypervolume based
83. Pei H, Li Y, Liu K (2016) A multi-object black hole gravitational black hole optimization algorithm. J Supercomput 1–54
search algorithm for day-ahead reactive optimization in distribu- 101. Dhanachandra N, Chanu YJ, Singh KM (2020) A new hybrid
tion network. In: 2016 IEEE chinese guidance, navigation and image segmentation approach using clustering and black hole
control conference (CGNCC). IEEE, pp 901–906 algorithm. Computational Intelligence

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Black hole algorithm...

102. Yepes V, Martı́ JV, Garcı́a J (2020) Black hole algorithm for meta-heuristic optimization algorithm. Computers & Industrial
sustainable design of counterfort retaining walls. Sustainability Engineering 157:107250
12(7):2767 123. Hassan MH, Kamel S, Abualigah L, Eid A (2021) Develop-
103. Cano A, Zafra A, Ventura S (2013) Weighted data gravitation ment and application of slime mould algorithm for optimal
classification for standard and imbalanced data. IEEE Trans economic emission dispatch. Expert Systems with Applications
Cybern 43(6):1672–1687 182:115205
104. Peng L, Yang B, Chen Y, Abraham A (2009) Data gravitation 124. He Q, Wang L (2007) An effective co-evolutionary particle
based classification. Inf Sci 179(6):809–819 swarm optimization for constrained engineering design prob-
105. Peng L, Zhang H, Yang B, Chen Y (2014) A new approach for lems. Eng Appl Artif Intell 20(1):89–99
imbalanced data classification based on data gravitation. Inf Sci 125. Lee KS, Geem ZW (2005) A new meta-heuristic algorithm for
288:347–373 continuous engineering optimization: Harmony search theory
106. Pan J-S, Chai Q-W, Chu S-C, Wu N (2020) 3-D terrain node and practice. Comput Methods Appl Mechan Eng 194(36-
coverage of wireless sensor network using enhanced black hole 38):3902–3933
algorithm. Sensors 20(8):2411 126. Belegundu AD, Arora JS (1985) A study of mathematical
107. Biju E (2020) Reliability improvement and loss reduction in programming methods for structural optimization. part i: Theory.
radial distribution system by reconfiguration using black hole Int J Numer Methods Eng 21(9):1583–1599
algorithm, International Journal Of Information And Computing 127. Rashedi E, Nezamabadi-Pour H, Saryazdi S (2009) Gsa: A
Science gravitational search algorithm. Inform Sci 179(13):2232–2248
108. Soto R, Crawford B, Figueroa I, Olivares R, Olguı́n E (2016) The 128. Mezura-Montes E, Coello CAC (2008) An empirical study
set covering problem solved by the black hole algorithm. In: 2016 about the usefulness of evolution strategies to solve constrained
11th iberian conference on information systems and technologies optimization problems. Int J Gen Syst 37(4):443–473
(CISTI). IEEE, pp 1–4 129. Mirjalili S, algorithm Moth-flameoptimization (2015) A novel
109. Soto R, Crawford B, Olivares R, Taramasco C, Figueroa I, nature-inspired heuristic paradigm. Knowl-Based Syst 89:228–
Gómez Á, Castro C, Paredes F (2018) Adaptive black hole 249
algorithm for solving the set covering problem. Mathematical 130. Gandomi AH, Yang X-S, Alavi AH, Talatahari S (2013) Bat
Problems in Engineering algorithm for constrained optimization tasks. Neural Comput
110. Munoz R, Olivares R, Taramasco C, Villarroel R, Soto R, Applic 22(6):1239–1255
Barcelos TS, Merino E, Alonso-Sánchez MF (2018) Using 131. Mirjalili S, Lewis A (2016) The whale optimization algorithm.
black hole algorithm to improve eeg-based emotion recognition. Adv Eng Softw 95:51–67
Computational Intelligence and Neuroscience 132. Kaveh A, Khayatazad M (2012) A new meta-heuristic method:
111. Mehrani K, Mirshahvalad A, Abbasi E (2020) Portfolio ray optimization. Comput Struct 112:283–294
optimization using black hole meta heuristic algorithm, Specialty 133. Wang S, Jia H, Abualigah L, Liu Q, Zheng R (2021) An
Journal of Accounting and Economics 5 improved hybrid aquila optimizer and harris hawks algorithm for
112. Whitley D (1994) A genetic algorithm tutorial. Stat Comput solving industrial engineering optimization problems. Processes
4(2):65–85 9(9):1551
113. Kennedy J, Eberhart R (1995) Particle swarm optimization. In: 134. Mirjalili S, Mirjalili S, Lewis A (2014) Grey wolf optimizer. Adv
Proceedings of ICNN’95-International Conference on Neural Eng Softw 69:46–61
Networks, vol 4, IEEE, pp 1942–1948 135. Wang S, Liu Q, Liu Y, Jia H, Abualigah L, Zheng R, Wu D
114. Price KV (2013) Differential evolution. In: Handbook of (2021) A hybrid ssa and sma with mutation opposition-based
Optimization, Springer, pp 187–214 learning for constrained engineering problems. Computational
115. Alomari OA, Khader AT, Al-Betar MA, Abualigah L (2017) Intelligence and Neuroscience
Gene selection for cancer classification by combining minimum 136. Abd Elaziz M, Elsheikh AH, Oliva D, Abualigah L, Lu S, Ewees
redundancy maximum relevancy and bat-inspired algorithm. Int AA (2021) Advanced metaheuristic techniques for mechanical
J Data Mining Bioinform 19(1):32–51 design problems. Archiv Comput Methods Eng 1–22
116. Abualigah L, Diabat A (2020) A comprehensive survey of 137. Sandgren E (1990) Nonlinear integer and discrete programming
the grasshopper optimization algorithm: results, variants, and in mechanical design optimization. J Mech Des 112(2):223–229
applications. Neural Comput Applic 1–21 138. Mahdavi M, Fesanghary M, Damangir E (2007) An improved
117. Mirjalili S (2016) Dragonfly algorithm: a new meta-heuristic harmony search algorithm for solving optimization problems.
optimization technique for solving single-objective, discrete, and Appl Math Comput 188(2):1567–1579
multi-objective problems. Neural Comput Applic 27(4):1053– 139. Coello CAC (2000) Use of a self-adaptive penalty approach for
1073 engineering optimization problems. Comput Ind 41(2):113–127
118. Mirjalili S (2015) The ant lion optimizer. Advances in 140. Huang F-Z, Wang L, He Q (2007) An effective co-evolutionary
engineering software 83:80–98 differential evolution for constrained optimization. Appl Math
119. Abualigah L (2020) Multi-verse optimizer algorithm: A compre- Comput 186(1):340–356
hensive survey of its results, variants, and applications. Neural 141. Liu H, Cai Z, Wang Y (2010) Hybridizing particle swarm opti-
Comput Applic 1–21 mization with differential evolution for constrained numerical
120. Shehab M, Abualigah L, Al Hamad H, Alabool H, Alshinwan and engineering optimization. Appl Soft Comput 10(2):629–640
M, Khasawneh AM (2019) Moth–flame optimization algorithm: 142. He Q, Wang L (2007) A hybrid particle swarm optimization with
variants and applications. Neural Comput Applic 1–26 a feasibility-based rule for constrained optimization. Appl Math
121. Elaziz MA, Oliva D, Xiong S (2017) An improved opposition- Comput 186(2):1407–1422
based sine cosine algorithm for global optimization. Expert Syst
Appl 90:484–500
122. Abualigah L, Yousri D, Abd Elaziz M, Ewees AA, Al- Publisher’s note Springer Nature remains neutral with regard to
qaness MA, Gandomi AH (2021) Aquila optimizer: A novel jurisdictional claims in published maps and institutional affiliations.

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


L. Abualigah et al.

Affiliations
Laith Abualigah1,2 · Mohamed Abd Elaziz3,4,5,6 · Putra Sumari2 · Ahmad M. Khasawneh1 ·
Mohammad Alshinwan1 · Seyedali Mirjalili7,8 · Mohammad Shehab9 · Hayfa Y. Abuaddous1 · Amir H. Gandomi10

Mohamed Abd Elaziz


abd el aziz m@yahoo.com
Putra Sumari
putras@usm.my
Ahmad M. Khasawneh
a.khasawneh@aau.edu.jo
Mohammad Alshinwan
mohmdsh@aau.edu.jo
Seyedali Mirjalili
ali.mirjalili@gmail.com
Mohammad Shehab
moh.shehab12@gmail.com
Hayfa Y. Abuaddous
haddose@aau.edu.jo
Amir H. Gandomi
gandomi@uts.edu.au
1 Faculty of Computer Sciences and Informatics, Amman Arab
University, Amman 11953, Jordan
2 School of Computer Sciences, Universiti Sains Malaysia,
11800, Pulau Pinang, Malaysia
3 Department of Mathematics, Faculty of Science, Zagazig
University, Zagazig, 44519, Egypt
4 Artificial Intelligence Research Center (AIRC), Ajman University,
Ajman 346, United Arab Emirates
5 Department of Artificial Intelligence Science & Engineering,
Galala University, Suez, Egypt
6 School of Computer Science and Robotics, Tomsk Polytechnic
University, Tomsk, Russia
7 Centre for Artificial Intelligence Research and Optimization,
Torrens University Australia, 90 Bowen Terrace, Fortitude Valley
QLD 4006, Australia
8 Yonsei Frontier Lab, Yonsei University, Seoul, South Korea
9 Information Technology, The World Islamic Sciences
and Education University, Amman, Jordan
10 Faculty of Engineering and Information Technology, University
of Technology Sydney, Ultimo, NSW 2007, Australia

Content courtesy of Springer Nature, terms of use apply. Rights reserved.


Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

onlineservice@springernature.com

View publication stats

You might also like