2020 International Conference on High Performance Big Data and Intelligent Systems (HPBD&IS)

TIGER Algorithm
Saman Darvish Kermany
IEEE Member PhD
Melbourne, Austarlia
[email protected]

Abstract— TIGER Algorithm (TA) is a metaheuristic by Xing and Gao in 2014 [5] names 134 of Clever
algorithm that has been designed for finding global and some Algorithms. There are several basic metaheuristic algorithms
noticeable local optimization points. TIGER algorithm is the for example, in Particle Swarm Optimisation (PSO) each
abbreviation of “Target Identify with Gross Error Reduction search process has an initially random velocity within the
algorithm” but it follows the Tigers’ life and behaviors which search space, and carries out moves by adding this to its
include; Birth, Migration, Migration of cubs (children), Death, current position at each iteration [6]. At each iteration of the
Suicide, Prowl, Rivalry mating and signing. Optimization Genetic Algorithm (GA), search points with relatively high
problems would be solved in a noticeable time if they are so objective value are selected from the existing population [7].
complicated, but this algorithm tries to solve them as fast as
Each search process in Artificial Bee Colony Algorithm
possible and the main difference between TA and other
(ABC) generates local moves in the direction of the current
metaheuristic algorithms is that TIGER algorithm doesn’t
need to initialize the population and it adjusts the population
position of another, randomly selected; search process, and
by producing new tigers or deleting them for solving only improving moves are accepted [8]. Bacterial Foraging
optimization problems efficiently. In this article some Optimization (BFO) is designed in a way that in each search
definitions of this algorithm are described but, just some of process carries out sequential moves in the same direction
them are used in optimization for simplicity. TIGER algorithm until this no longer leads to improvement [9]. In Harmony
could be used for engineering design or business planning and Search (HS) at each iteration, a single new solution is created
other cases. from a randomly selected existing solution [10]. There are
some new metaheuristic algorithms such as Ant Lion
Keywords—TIGER Algorithm (TA), Metaheuristic Optimizer (ALO) [11], Bat Algorithm (BA) [12], Brain
algorithms, optimization, Data-Clustering Storm Optimization (BSO) [13], Cuckoo Optimization
Algorithm (COA) [14], and Teacher-Learning Based
I. INTRODUCTION Optimization (TLBO) [15] which are compared based on
Commonalities and citations in [16]. The TA optimization
The word "Tiger" is taken from the Greek word "Tigris", process is like Firefly Algorithm (FA) [17], Gravitational
which is possibly derived from a Persian source meaning Search Algorithm (GSA) [18], and Charged System Search
"arrow", a reference to the animal's speed and also the origin (CSS) [19]. In each iteration in these three algorithms, all
for the name of the Tigris river [1]. The TA could be used for search processes carry out moves towards the current
solving optimization problems and this algorithm is a positions of other search processes. The degree of movement
metaheuristic algorithm. TA is the abbreviation of “Target towards each point is calculated using an inverse-square law
Identify with Gross Error Reduction algorithm” too; this is weighted by its objective value, causing an intensification of
for variable values in time cases and it means, there is a search towards points with higher objective values. Some of
method that by using error reduction this algorithm could recent metaheuristic algorithms are Grey Wolf Optimizer
follow optimum variable points. TA or tiger search is a new (GWO) [20] that each search process carried out moves
algorithm and it is based on tiger life that includes; Birth, around the edges of a hypercube centred around a target
Migration, Migration of cubs (children), Death, Suicide, search point, Whale Optimization Algorithm (WOA) [21]
Prowl, Rivalry mating and signing. All these are definitions carries out moves in a hypercube around a target search point
of the TA and tiger hunt that is a model for searching. It is and iteratively moves towards this target either by shrinking
important to say that some parts of this algorithm look like the hypercube or through a spiral motion, and Red deer
other Metaheuristic algorithms else especially the firefly algorithm (RDA) [22]. Finding an optimal set of features for
algorithm, but some parts of this algorithm are based on new a better rate of classification is surveyed in [23]. In some
ideas such as the migration of children, birth, death, and articles for achieving accuracy and having a minimum
suicide [2-4]. Trying to solve NP-hard “Non-deterministic processing time for a query and the reduction of the
Polynomial time-hard” optimization problems is the last aim, processing load a Hybrid PSO- Grammatical Evolution (GE)
but is it possible? In 1997, the publication of the “no free methods has been proposed [24]. The role and performance
lunch theorems for optimization” by D.H Wolpert and W.G. of different traditional classification and nature-inspired
Macready has shown that perhaps it couldn’t be done but it computing techniques are explained in [25].
worth trying [3]. The TA has two methods for searching, first
for variable time’s problem and some definitions are The rest of this study is organized as follows. Section II
explained in this article and second for optimization investigates the inspiration of the algorithm originated from
problems which are explained in detail. In this article, just Tigers along with their behaviours in nature. Section III
solving optimization problems are considered. presents the TA formulation along with its parameters.
Section IV not only tests the proposed algorithm with some
In recent years, some of the new metaheuristic algorithms examples but also the performance of the proposed algorithm
have investigated different sources of inspiration within the is validated by comparing the results of solving a standard
biological and natural systems. It is difficult to know how benchmark function with some known optimization
these algorithms work and understand their relationships to algorithms. Finally, Section V summarizes the conclusions
other metaheuristics. Over a hundred nature-inspired and future directions of this work.
algorithms have been published since 2020. The review book

978-1-7281-6512-7/20/$31.00 ©2020 IEEE

Authorized licensed use limited to: Netaji Subhas University of Technology New Delhi. Downloaded on February 19,2025 at 11:42:52 UTC from IEEE Xplore. Restrictions apply.
 II. BASIC DEFINITIONS OF TIGER ALGORITHM There are two kinds of tigers in this metaheuristic
When a tiger wants to hunt, approaches to prey and algorithm, males and females. Suppose that there
whenever prey starts running, the tiger directly chases prey. are M males and F females (however at first M could be
When prey scapes, the tiger corrects its direction in a way equal to F). They try to find goals and they will breed new
that the shortest path is chosen. This new direction is not a tigers if a male finds a female in his area and this new tiger
direct path to prey, it is toward a point that prey would has all memories (often the best) of its parents but could
probably be there furthermore than any point else. This can migrate to other places (memory is something that is
be done by deleting primary errors (direct path toward prey) proposed, but it is not used for formulation). The tigers will
and correcting path with new experience. Whenever prey suicide or die if their population becomes very crowded (by
changes its path, the tiger corrects its path too and lessens produce or migration new tigers) and because of that speed
their distance. Finally, the tiger could probably get prey even of the algorithm doesn’t get slow. The tigers will suicide or
if the prey runs faster because the tiger runs less path than die if they are very old and weak in finding good goals.
prey. If prey continuously changes its direction (dynamic Another method for reducing tigers is deleting weak tigers
functions with variable values in time) tiger will lock on prey in crowded areas that are used in this article. Some new cubs
and chase it. will randomly migrate to new locations to find better new
A Tiger is trying to hunt a zebra (goal) as shown in goals and populate in good points. This algorithm tries to
(Figure 1) find global and noticeable local optimum points of a
function. Increasing of new tigers will help in finding
I. Tiger approaches to zebra by searching for a certain
further optimum points. This method increases the
area. If the target is constant in time, so it is not a
probability of finding global points and decreases impaction
dynamic optimization problem, and TA is used for
solving. in local points. Whenever a tiger finds an optimum area, it
starts prowling to find the best optimum point. If TA could
II. Vector (A), is a vector that is chosen for hunting the get optimum points in a pre-defined time, it will finish its
zebra, but it will change to another vector (C) if search and show optimum values.
zebra changes its location to (D) because it is This algorithm is based on tiger life behaviors which are
obvious that zebra wouldn’t be in a permanent explained as follow:
position. So the next position of zebra won’t be in
elongation of the vector (B) in the close future. A. Birth and Breed
After acquiring vector (D) (it could be done with When a couple (male and female tigers) approach each
reducing gross errors and approximately finding other in an optimum area, they will produce new children if
vector (D)), this vector will be corrected (vector they stay there in a given enough time and any progress
(B)) and vector (C) will be chosen. doesn’t happen for finding new better points. These new
III. Tiger tries to hunt zebra by locking on target and it cubs migrate to new random points. In the real-world, some
reduces errors in every step. tigers migrate even 120 Km to new places. These new
children could be from 2 to 6 cubs and this depends on
IV. Hunting the goal is the last aim, but the tiger multiplicity and population of all tigers. Finding new points
continues its search if zebra changes its location and is something that could occasion this process.
locks on the target.
This method is used only for solving dynamic problems B. Migration of Children
which would be future directions of this article. TA which is New children randomly spread in the area of the
described in this article is just for finding global points. optimization problem and this migration could be useful for
Similar to many metaheuristic algorithms, in this algorithm, mobility of algorithm. If the TA could not get the optimum
there are some methods for finding the goals. It is based on point in a pre-defined time, migration is something that
Tiger life that includes; Birth, Migration, Migration of would help to find better points.
children, Death, Suicide, Prowl, Rivalry mating, and
C. Migration
signing.
If there are not enough tigers that could not produce new
tigers in a defined time, some tigers will randomly add to the
algorithm. This method could help the algorithm to be
dynamic and increase its ability.

D. Death
Birth and migration could cause that the number of tigers
to increase and this could collapse searching. So, the
algorithm needs to decrease its population. Some tigers will
die and this could equilibrium algorithm. The couple tigers
will die that are so old and have not gotten any good goals
compared to other tigers.

E. Suicide
Sometimes, death is not enough to lessen the tigers in the
algorithm so suicide could help to decrease the number of
Figure. 1 Explanation of the tiger behaviour for chasing preys

Authorized licensed use limited to: Netaji Subhas University of Technology New Delhi. Downloaded on February 19,2025 at 11:42:52 UTC from IEEE Xplore. Restrictions apply.
 tigers. If an old tiger is in a crowded area and could not get
2
good goals the male or female tiger will suicide. This could D ( x ) = D o e −η x (3)
help the algorithm produces better new tigers and stabilizes
searching time trend.
Amount of absorbing as for preys is denoted by φ
F. Prowl coefficient. So:
Some tigers congest around local or global optimum 2
points and for finding the optimum points they prowl around φ = φ 0 e −η x (4)
these global optimum points to get the best values.

G. Rivalry Mating φ0 is absorption in x = 0 , by using another

If males and females tigers are so crowded in an area approximation:
they will choose the best mate respectively. It means that the
best male chooses the best female and this could cause better φ0
search. φ = (5)
1 +ηx2
H. Signing
Some tigers migrate to other areas, so they may migrate Distance between two tigers i and j that are in xi and x j
to an area that they were before. They never come back to is explained as:
an area that they were before by signing. But this method
increases the usage of memory and this could retard d
searching. So, for NP-hard problems it is not recommended.
 (x ) 2
d ij = x i − x j = i ,k − x j ,k (6)
k =1
III. FORMULATION
The main difference between TA and other metaheuristic Where xi , k is the kth component of the spatial coordinate
algorithms is that TA doesn’t need to initialize the
population and it adjusts the population by producing new xi of ith tiger in 2-D case, so:
tigers or deleting them for solving optimization problems by
using Birth, migration, death, and suicide.
For example, if any progress doesn't happen in a defined
time which shows the numbers of tigers aren't enough, the d ij = (x i − x j )
2
+ (y i − y j )
2
(7)
algorithm will automatically produce new tigers
by Birth and migration, but if it is so crowded and algorithm The movement of a tiger i is attracted to another more
couldn't progress in a defined time then, the algorithm will attractive (better) tiger j is determined by:
automatically reduce the tigers by Death and Suicide.
Basic formulas are described in the following. These
formulas are simplified for reducing the time of computation.
xi = xi + φ 0e
− η x ij2
(x j − x i ) + ωξ i
(8)
TA is based on the distance between tigers and their prey.
If a tiger is far from prey, it would hardly access to prey. In formula (8) coefficient ω and ξ i for simplicity are
This could be explained as a formula that distance is reverse
between 0 to 1 and TA works with these values. At first,
power two:
numbers of male and female tigers are equal.
Di In general there are some methods for stopping criteria.
D( x ) = (1) The number of generation, time limit, and other common
x2
methods could be used for this algorithm. For controlling
the population in TA, in every define iteration, it is possible
In this formula Di is a coefficient that distances between to increase the number of tigers based on a defined ratio.
prey and tiger and amount of them would determine this. x The parameters of the algorithm are adjusted in a way
is distance between tiger and prey and D(x) is a function that the tigers’ dispersal in the searching area could be
steady at first and it would help to avoid sticking in local
that shows a comparison between tigers. If η determine as points. If a female tiger approaches a male tiger, a cub will
absorption coefficient between tigers and preys, there could produce and place it in a random area. If some define male
be a formula: or female tiger gather in a close area, the weaker ones will
die.
D = D o e −η x (2) In this algorithm, all search processes carry out moves
towards the current positions of other search processes in
Do is a coefficient that is the primary distance between each iteration. The degree of movement towards each point
is calculated using an inverse-square law weighted by its
tiger and prey. In formula (1) for avoiding zero division, two objective value, causing an intensification of search towards
later formulas with Gaussian approximation are used and points with higher objective values.
then:

Authorized licensed use limited to: Netaji Subhas University of Technology New Delhi. Downloaded on February 19,2025 at 11:42:52 UTC from IEEE Xplore. Restrictions apply.
 IV. IMPLEMENTATION OF ALGORITHM final optimum points are sorted and print by the program
For explaining the implementation of this algorithm, a and it is possible to find the best optimum point.
function is considered which graph of this function is shown
in (Figure 2):
2
−( y− 4)2 ) 2
−( y− 4)2 )
z = e (−(x−4) + e(−(x+ 4)
(9)
(− x2 − ( y + 4)2 ) (− x2 − y2 )
+ 2×e + 2×e

Number of Tigers
Iteration
Figure. 4 The number of tigers and the number of male and female tigers.
Male tigers increase from 30 to 37 and female tigers increase from 30 to 34

Figure. 2 Graph of function in formula (9) with 4 local optimum points

The final location of 71 tigers is shown in (Figure 3), the

initialized number of tigers was 60 (30 males & 30 females)
and after 200 iterations optimum point is found. However y
before 200 iterations this algorithm could find optimum
point but for this function stopping criteria is defined as 200.

y Figure. 5 Trace of tigers movement and random location of new cubs. The
density of tigers in global optimum points is further than in other areas.

One of the objectives of this algorithm is to find a way to

solve the problems with equal optimum points. For example
a function is considered in formula (10) which has 4 equal
optimum points and most of the metaheuristic algorithms
could not find all optimum points because they try to find
the global optimum point not optimum points. In (Figure 6)
the graph of function in formula (10) with 4 equal optimum
x points is shown:
Figure. 3 The final location of 71 tigers after 200 iterations that was
(x + y )× e (− 0 . 0625 × (x + y ))
2 2
initialised with 60 tigers (30 males in blue & 30 females in red) z = (10)
In (Figure 4) the number of tigers and the number of male
and female tigers are shown. Male tigers increase from 30 to In (Figure 7) the final location of 58 tigers is shown, the
37 and female tigers increase from 30 to 34 after 200 initialized number of tigers was 50 (25 males & 25 females)
iterations. The algorithm starts with 60 tigers that it could be and after 100 iterations optimum points are found.
an even random number and here it is finished with 71 tigers. In (Figure 8) the number of tigers and the number of male
The algorithm is almost steady after this number of tigers. and female tigers are shown. Male and female tigers
In (Figure 5) Trace of tigers movement and random increase from 25 to 29 after 100 iterations. The algorithm
location of new cubs are shown. The density of tigers in starts with 50 tigers that it could be an even random number
global optimum points is further than in other areas. The and here it is ended with 58 tigers.

Authorized licensed use limited to: Netaji Subhas University of Technology New Delhi. Downloaded on February 19,2025 at 11:42:52 UTC from IEEE Xplore. Restrictions apply.
 could be used for solving these kinds of problems in this
algorithm.

Figure. 6 Graph of function in formula (10) with 4 equal optimum points

Figure. 9 All points that are solitary are filtered

x
Figure. 7 The final location of 58 tigers after 100 iterations that was
initialised with 50 tigers (25 males & 25 females)

Figure. 10 All points are clustered in 4 main set and it shows that there are
4 optimum points. In this example, amount of all this 4 points are equal
Number of Tigers

Scale

Iteration
Figure. 8 The number of tigers and the number of male and female tigers.
Male and female tigers increase from 25 to 29 after 100 iterations. Data

In summary, silhouette is a method that shows the size of Figure. 11 clustering of points (29 data)
the sets and which set is greater than other sets. The
silhouette of data that is shown in (Figure 9-12) describes
that all 4 points could be a global optimum point. It is a way
that could find global optimum points and it is an ability that

Authorized licensed use limited to: Netaji Subhas University of Technology New Delhi. Downloaded on February 19,2025 at 11:42:52 UTC from IEEE Xplore. Restrictions apply.
 Figure. 13 Graph of Ackley’s function

Figure. 12 Silhouette of data shows that all 4 points could be global

optimum point Graph of Ackley’s function in 2-D (n=2) for finding
When a new searching algorithm or metaheuristic maximum using TA is shown in (Figure 14) (TA is designed
algorithm is designed, some standard tests should be done to for finding maximum but with multiply function at a
compare to other algorithms. One of these tests is Ackley’s negative, it easily changes to find minimum)
function test [3]. However at least there are 13 other
functions test but in this article only Ackley’s function test
is considered, other suggestion function tests are:
1) De Jong’s functions
2) Easom’s function
3) Equality-Constrained function
4) Griewank’s function
5) Michealwicz’s function
6) Perm functions
7) Rastrigin’s function
8) Rosenbrock’s function
9) Schwefel’s function
10) Six-hump camel back function
11) Shubert’s function
12) Xin-She Yong’s functions
13) Zakharov’s functions

Formula (11) is Ackley’s function formula and (Figure 13)

shows graph of Ackley’s function [4]. Figure. 14 Graph of Ackley’s function in 2-D (n=2) for finding maximum
using TA (TA is designed for finding maximum but with multiply at a
1 1
n negative it easily changes to find minimum)
(−
5 n  x i2 )
f ( x ) = − 20 × e i =1

(11) The final location of 50 tigers after 5 iterations that was

n
1 initialised with 50 tigers (25 males & 25 females) is shown
(
n  cos( 2 π x i ))
−e i =1
+ 20 + e in (Figure 15). The final points are around the optimum
points. Whereas tigers don’t have enough time to breed or
die (every 10 iterations algorithm let them do) the number of
Where n=1,2,… and − 32 . 768 ≤ x i ≤ 32 . 768 tigers doesn’t change.
for i = 1 , 2 ,..., n . This function has the global
minimum f * = 0 at x * = ( 0 , 0 ,..., 0 ) .

Authorized licensed use limited to: Netaji Subhas University of Technology New Delhi. Downloaded on February 19,2025 at 11:42:52 UTC from IEEE Xplore. Restrictions apply.
 y

x Figure. 16 Finding the best optimum points by genetic algorithm with 50

Figure. 15 The final location of 50 tigers after 5 iterations that was initial gen and 51 iterations, using Gatool in Matlab program [26]
initialised with 50 tigers (25 males & 25 females)

This test shows that the TA could find the optimum point
of a function very fast. Genetic algorithm, annealing search,
and pattern search are considered for solving Ackley’s
function in 2-D for comparing to TA. In table (1) some
algorithms are compared and as it shows TA could solve
Ackley’s function in 5 iterations. TA starts searching with
50 tigers (25 male & 25 female).
TABLE I
COMPARING TIGER ALGORITHM TO OTHER ALGORITHMS

Searching results
Algorithm
X Y Best Z No.
iteration
Genetic 0.008 -0.004 2.9828 50 gen 51
Annealing -0.02 - 2.9918 Start 5 1019
Pattern Search 0 - 2.9805 Start 5 24
Tiger (Male) -0.0043 0.0021 -0.0301 25 5
Tiger (Female) -0.0079 0.0057 -0.0301 25 5
Description: the function that is introduced in the Matlab program
Figure. 17 Finding the best optimum points by Annealing search starts
[26] as Ackley’s function has a minimum as 3 instead of 0, but X from (X=5) and 1019 iterations, using Optimization tool in Matlab program
and Y are zero, and Z=3 instead of Z=0 is not important for [26]
comparing. In the TA there are 50 initialize tigers and for
comparing male and female tigers are shown separately.

(Figure 16) shows iterations and best values of genetic

algorithm for finding Ackley’s function optimum point,
likely (Figures 17) and (Figures 18) show the best values of
Annealing search and Pattern Search.

Figure. 18 Finding the best optimum points by pattern search starts from
(X=5) and 24 iterations, using Optimization tool in Matlab program [26]

TA could be used for engineering design or business

planning and other cases.

Authorized licensed use limited to: Netaji Subhas University of Technology New Delhi. Downloaded on February 19,2025 at 11:42:52 UTC from IEEE Xplore. Restrictions apply.
 V. CONCLUSIONS [6] R. Eberhart, and J. Kennedy, "Particle swarm optimization." pp.
1942-1948. 1995.
A metaheuristic algorithm named tiger algorithm or tiger [7] H. John, “Holland, Adaptation in natural and artificial systems.
search is introduced. This algorithm is based on tiger life 1992,” Ann Arbor, MI: University of Michigan Press, 1975.
which includes; Birth, Migration, Migration of children, [8] D. Karaboga, An idea based on honey bee swarm for numerical
optimization, Technical report-tr06, Erciyes university,
Death, Suicide, Prowl, Rivalry mating, and signing. These engineering faculty, computer …, 2005.
new definitions are introduced in this article and by using [9] K. M. Passino, “Biomimicry of bacterial foraging for distributed
these methods, this algorithm could efficiently find the optimization and control,” IEEE control systems magazine, vol.
optimum or optimum points of a function. The tiger 22, no. 3, pp. 52-67, 2002.
[10] Z. W. Geem, J. H. Kim, and G. V. Loganathan, “A new
algorithm doesn’t need to initialize the population and it heuristic optimization algorithm: harmony search,” simulation,
adjusts the population by producing new tigers with Birth vol. 76, no. 2, pp. 60-68, 2001.
and Migration or deleting them by Death and Suicide. In [11] S. Mirjalili, “The ant lion optimizer,” Advances in engineering
this algorithm, all search processes carry out moves towards software, vol. 83, pp. 80-98, 2015.
[12] X. S. Yang, and A. H. Gandomi, “Bat algorithm: a novel
the current positions of other search processes in each approach for global engineering optimization,” Engineering
iteration. The degree of movement towards each point is computations, 2012.
calculated using an inverse-square law weighted by its [13] Y. Shi, "Brain storm optimization algorithm." pp. 303-309.
objective value, causing an intensification of search towards [14] R. Rajabioun, “Cuckoo optimization algorithm,” Applied soft
computing, vol. 11, no. 8, pp. 5508-5518, 2011.
points with higher objective values. [15] R. V. Rao, V. J. Savsani, and D. Vakharia, “Teaching–learning-
Optimization problems would be solved in a noticeable based optimization: a novel method for constrained mechanical
time if they are so complicated, but this algorithm tries to design optimization problems,” Computer-Aided Design, vol.
solve them as fast as possible by adjusting the population 43, no. 3, pp. 303-315, 2011.
[16] M. A. Lones, “Mitigating metaphors: A comprehensible guide
and the main difference between this algorithm and other to recent nature-inspired algorithms,” SN Computer Science,
metaheuristic algorithms is that tiger algorithm doesn’t need vol. 1, no. 1, pp. 49, 2020.
to initialize the population and it adjusts the population by [17] X.-S. Yang, "Firefly algorithms for multimodal optimization."
producing new tigers or deleting them for solving pp. 169-178. 2009.
[18] E. Rashedi, H. Nezamabadi-Pour, and S. Saryazdi, “GSA: a
optimization problems efficiently. This algorithm could find gravitational search algorithm,” Information sciences, vol. 179,
most of the optimum global and noticeable local points. no. 13, pp. 2232-2248, 2009.
Unlike most metaheuristic algorithms, this algorithm finds [19] A. Kaveh, and S. Talatahari, “A novel heuristic optimization
global and local optimum points and clusters them. A method: charged system search,” Acta Mechanica, vol. 213, no.
3-4, pp. 267-289, 2010.
function may have various equal optimum points and this [20] S. Mirjalili, S. M. Mirjalili, and A. Lewis, “Grey wolf
algorithm finds them by data clustering method. Compare to optimizer,” Advances in engineering software, vol. 69, pp. 46-
other metaheuristic algorithms, TA finds the global 61, 2014.
optimum point in less irritation. [21] S. Mirjalili, and A. Lewis, “The whale optimization algorithm,”
Advances in engineering software, vol. 95, pp. 51-67, 2016.
The tiger algorithm which is described in this article is [22] A. M. Fathollahi-Fard, M. Hajiaghaei-Keshteli, and R.
just for finding global points but, by using the "Gross Error Tavakkoli-Moghaddam, “Red deer algorithm (RDA): a new
Reduction" technic which is explained in section II of this nature-inspired meta-heuristic,” Soft Computing, pp. 1-29, 2020.
article, this algorithm can be used for solving dynamic [23] M. Sharma, and P. Kaur, “A Comprehensive Analysis of
Nature-Inspired Meta-Heuristic Techniques for Feature
problems which would be future directions of this article. Selection Problem,” Archives of Computational Methods in
Engineering, pp. 1-25, 2020.
REFERENCES [24] S. Meera, and C. Sundar, “A hybrid metaheuristic approach for
[1] H. G. Liddell, E. A. Barber, H. S. Jones, R. McKenzie, and R. efficient feature selection methods in big data,” Journal of
Scott, A Greek-English lexicon, Rev. and augm. throughout / Ambient Intelligence and Humanized Computing, pp. 1-9, 2020.
ed., Oxford: Clarendon Press, 1990. [25] S. Sharma, G. Singh, and D. Singh, “Role and Performance of
[2] E. Alba, Optimization techniques for solving complex problems, Different Traditional Classification and Nature-Inspired
2009. Computing Techniques in Major Research Areas,” EAI
[3] X.-S. Yang, Engineering optimization : an introduction with Endorsed Transactions on Scalable Information Systems, vol. 6,
metaheuristic applications, Hoboken, N.J.: John Wiley, 2010. no. 21, 2019.
[4] X.-S. Yang, Nature-inspired metaheuristic algorithms, 2010. [26] MATLAB help (2015). Program & Help version 8.5.0.197613
[5] B. Xing, W.-J. Gao, and SpringerLink, Innovative (R2015a)
Computational Intelligence: A Rough Guide to 134 Clever
Algorithms, 2014.

Authorized licensed use limited to: Netaji Subhas University of Technology New Delhi. Downloaded on February 19,2025 at 11:42:52 UTC from IEEE Xplore. Restrictions apply.