TU
"OOVBM
*OUFSOBUJPOBM
$POGFSFODF
PO
*OGPSNBUJPO
BOE
4DJFODFT

"*$*4

A review on the recent application of Jaya

optimization algorithm
1st Mohammad Alsajri
Faculty of Computer Systems and
Software Engineering, University
Malaysia Pahang,Pahang, Malasysia
mohammad.cs88@gmail.com
2nd Mohd Arfian Ismail
Faculty of Computer Systems and
Software Engineering, University
Malaysia Pahang, Pahang, Malaysia
arfian@ump.edu.my
3rd Saba Abdul-baqi Salamn
Department of Computer Science,
College of Education,
Al-Iraqia University, Baghdad, Iraq
saba.a.salamn1@gmail.com

Abstract— Jaya is a new metaheuristic that in recent years, convergence. This situation is encountered when a
has been applied to numerous intractable optimization population diversity is decreased, leading to an inevitable
problems. The important difference between Jaya and other
local optimum convergence. Some of the proposed
optimization algorithms is that Jaya does not require the
tuning of its parameters (a process needed in the other approaches to prevent this premature convergence have
algorithms to escape unwanted convergence). Another been discussed by Pandey et al. [1]. Just like the crossover
different is the efficiency of Jaya in always choosing the best and mutation operators in the GA, other factors such as
solution. In this paper, a review of the recent application of population size, elite size, inertia weight (in PSO),
Jaya algorithm in the field of optimization problem was maximum number of generations, acceleration rate (in
reviewed. PSO), harmony memory (in HS), onlooker bees, employed
bees, scout bees (in ABC), number of improvisation (in
Keywords—Jaya, optimization algorithm, metaheuristic HS), and pitch adjusting rate
algorithm, application of Jaya
(in HS) also contributes to premature convergence. The
INTRODUCTION performance of these algorithms depends on a proper
control and tuning of these specific parameters as they
This paper focused on the recent applications of the novel
control the ability of the algorithms towards finding the
algorithm called Jaya. Several forms of population-based
global optimum. Several studies have reported on the
heuristics have been developed; such heuristics are
control of these specific algorithmic parameters and how
categorized into evolutionary (EAs) and swarm
they help in balancing exploration with exploitation
intelligence (SI)-based algorithms. Both categories of
capabilities of algorithms. Some of the existing approaches
heuristics have been applied in several real-life problems
towards having a balanced exploration and exploitation
where the conventional algorithms may not provide
capabilities include the trial
efficient solutions. Some of the evolutionary-based
and error method, use of parameter-free algorithms (less
heuristics are Artificial Immune Algorithm (AIA),
efficient but strong), following the already existing
Evolutionary Programming (EP), Bacteria Foraging
guidelines such as guidelines from the existing literature),
Optimization (BFO), Evolutionary Strategy (ES), Genetic
use of a pre-existing experience, as well as the use of
Programming (GP), Differential Evolution (DE), and
mathematical models. Furthermore, the importance of
Genetic Algorithm (GA), while the swarm intelligence-
Taguchi, full factorial, and orthogonal array approaches in
based heuristics include Shuffled Frog Leaping (SFL),
finding the tuned parametric values have been reported in
Artificial Bee Colony (ABC), Particle Swarm
the literature. The ongoing discussion has portrayed the
Optimization (PSO), Ant Colony Optimization (ACO),
importance of appropriate parametric tuning and control
and Firefly (FF) algorithms. Other nature-inspired
which undoubtedly, is a difficult and time-consuming
algorithms Gravitational Search (GS), Lion Search (LS),
process. The process is even more complicated by a lack
Biogeography-Based Optimization (BBO), are Harmony
of pre-knowledge in preparing and conducting a proper
Search (HS), and Grenade Explosion Method (GEM).
computational experimental setup. In view of the
These heuristics possess good exploration capabilities and
foregoing, there is a need for a new approach and as such,
have been successfully applied in solving different forms
the Teaching Learning Based Optimization (TLBO) has
of optimization and search problems. The efficiency of
been proposed. The TLBO requires no specific control
these algorithms mainly depends on their parameters
parameter and depends mainly on the population size and
which basically govern their search ability and contribute
a maximum number of iterations. There are 2 phases
majorly to their exploration and exploitation capabilities.
(teacher phase and learner phase) in the TLBO algorithm.
The GA for instance depends on crossover operator which
However, this review is focused on the Jaya algorithm, a
serve as its main exploration operator. It also uses a
new optimization algorithm that is similar to TLBO. It is
mutation operator as an exploitation operator. The function
one of the specific parameter-free algorithms and was
of these operators depends on the specific probabilities that
proposed by Venkata Rao [2]. Jaya is discussed in detail in
are pre-defined before executing a search process. A
this paper with a focus on exploring the full description of
proper tuning of the crossover and mutation operators in
the algorithm and why it should be used in optimization
the GA has been reported to cause the GA search to be
processes.
trapped in a local optimum, a process known as premature


¥
*&&& 
%0*
"J$*4
 JAYA ALGORITHM
Jaya is a robust and simple optimization
framework which has been deployed as a benchmarking
function for both unconstrained and constrained problems
[3]. Despite being a parameter-free framework like TLBO,
it differed from TLBO by requiring no learning phase (i.e.
works only with the teacher phase). Jaya works by
establishing the solution to problems through moving to
the best solution and avoiding the bad solution. This
performance depends only on few control parameters such
as the maximum number of generations, number of design
variables, and the population size and does not require any
specific algorithmic control parameter which may require
to be tuned prior to the computational phase. The working
principle of Jaya is depicted as follows:
Algorithm-1 Jaya algorithm
1- Initialize the population size, number of designed
variables, and termination criteria
2- Repeat Steps 1-5 until the termination criteria is
met
3- Check for the best and worst solutions
4- Modify the solution
Y'j,k,I = Yj,k,I + r1,k,I (Yj,best,I - | Yj,k,I |) – r2,k,I(Yj,worst,I
- | Yj,k,I |)
5- Update the previous solution if Y'j,k,I > Yj,k,I,
otherwise, do not update the previous solution.
6- Display the established optimum solution.
The Jaya algorithm starts by initializing the population
size, setting the number of designed variables and the
termination criteria. Often, the maximum number of
iterations serves as the termination criteria in this
algorithm. The minimization and maximization of an
objective function f (Y) is the primary aim in Jaya.
Consider a situation where m number of design variables
(i.e. j 1, 2,....,m ) and
n number of candidate solution (i.e. k 1, 2,..,n ) exists at
ith iteration. Assume that ‘best’ represents the best-found
candidate solution for f (Y), represented as f (Y)best.
Likewise, let ‘worst’ be the worst found solution for f (Y),
denoted as f (Y) worst. Let Y j k i represents the jth design
variable for the kth candidate at ith iteration and let Y' j k i be
the modification carried out in Y j k i (calculated from Step
4 in algorithm-1). Let X j, best, I and Y j, worst, I be the value of
j for the best and worst solutions respectively. However, r1,
j, i and r2, ,j, i are two random numbers in the range of [0,1]
used to represent the jth variable at ith iteration while r1,k,I
(Yj,best,I - | Yj,k,I |) and – r2,k,I(Yj,worst,I - | Yj,k,I |) are
respectively used to indicate the movement of the search
either to the best solution or towards avoiding the worst
solution. If Y'j,k,I > Yj,k,I,, then, update the previous
solution, else, do not update. A loop was introduced at step
2 of the algorithm which repeats steps 3 to 5 to maintain a
list of all the accepted f (Y) values. These values are used
as the input for the (i+1)th iteration. Step 4 clearly indicates
the capability of the Jaya algorithm to move towards the
best solution and avoid the bad solution. This capability

accounts for the success of the Jaya algorithm, hence, its
name, Jaya, which was derived from a Sanskrit word that
means “victory”.
WHY JAYA ?
There are many reasons why Jaya is considered as the best
solution for many optimization problems, such as:
1. Its simplicity makes it simple to convert codes in
any programming language and simple to
execute.
2. Jaya algorithm does not need any specific
algorithm parameter that needs to be tuned before
the computational experiment.
3. Algorithm-specific parameters make the
computational experiment of any algorithm time
consuming and as such, Jaya is time-efficient.
4. The victorious nature of Jaya makes it to always
choose the best solution.
RECENT APPLICATION OF JAYA
Jaya has found numerous research applications in recent
years. Gong [4] proposed a novel E-Jaya algorithm for the
enhancement of the performance of the original Jaya
framework. In the proposed E-Jaya, the average of the
better and worse groups was used to establish the optimal
solution. The proposed E-Jaya was reported to achieve a
better solution accuracy compared to the original Jaya. In
the E-Jaya, the swarms’ behaviors (the better and the
worse swarms) were considered instead of considering
only the behaviors of the best and the worst individuals.
The E-Jaya was used to assess 12 benchmark functions
with different dimensionality. Similarly, an effective
demand-side management scheme for residential HEMS
has been proposed by [5] for the prevention of peak
creation while reducing electricity bill. In this report, 3
algorithms (JAYA, SBA and EDE) were implemented to
achieve some objectives while the TOU pricing scheme
was deployed for the computation of electricity bill. The
study showed Jaya to work sufficiently in achieving low
electricity bill and reducing the PAR while ensuring
customer satisfaction. The SBA was further reported to
perform better than Jaya and EDE in ensuring user comfort
as it inversely related with the electricity bill. An improved
Jaya (IJAYA) has been developed by Yu et al. [6] for an
accurate and steady estimation of different PV model
parameters.
A self-adaptive weight was incorporated in IJAYA for the
adjustment of the propensity of reaching the best solution
and avoiding the bad solution while searching. The
function of the weight is to help the framework reach the
possible search region early and to execute the local search
later. Additionally, the algorithm incorporates a learning
strategy which is derived from the experience of other
individuals. This learning strategy was randomly
employed for the improvement of the population diversity.
Furthermore, a chaotic learning method was suggested for
the enhancement of the quality of the best solution in every
generation. IJAYA algorithm is easy to implement as it
does not contain any tunable parameter. The performance
emphasis was placed on only two design variables
(bandwidth and delay) for the determination of the optimal
route from the available network routes. This is because a
single metric is not enough for the prediction of paths from
the source to several destinations. A centralized MILP

Table 1
Algorithm Jaya TLBO GA PSO

Conception Means victories Teacher and student Natural selection Inspired by swarm
always chose the teaching-learning process behaviors
best solution process
No. of parameters 2 2 Phases each phase More than 3 More than 3
2 parameters

Parameter tuning Not needed Not Needed Needed Needed

Computational Not complex Complex Complex Complex

complexity

Table 2
Test for best solution Test for mean solution

Algorithms Friedman Normalized Rank Algorithms Friedman Normalized Rank

value value value value
TLBO 44.5 1.36 2 TLBO 40.5 1.52 2

GA 92 2.83 4 GA 88 3.32 4

PSO 57 1.75 3 PSO 59 2.22 3

JAYA 32.5 1 1 JAYA 26.5 1 1

of the algorithm was evaluated using single and double

diodes, as well as PV module model parameter
identification problems and the results showed IJAYA to
achieve better accuracy and reliability measures compared
to the benchmarking frameworks. Thus, IJAYA was
recommended as a good framework for solving PV
models-related parameter identification problems.
Sudhakar [7] introduced a multi-objective Jaya framework
for the improvement of the performance of ad hoc
networks. The work focused only on the minimization of
bandwidth and delay of the path achieved from the
topology of the network. The optimal path in this
technique was selected from the path currently available in
the network. Previous studies have concentrated on single
parameter metrics for the determination of the quality of a
link and for link prediction as well. In this study, the
formulation has been proposed for large-scale urban traffic
light scheduling problem (LUTLSP) [8] with the aim of
and Jaya) to successfully address the problem in real time.
minimizing the total vehicle delay time within a time
frame. The LUTLSP employed 3 algorithms HS, WCA,
The experiment considered 11 situations with varying
numbers of intersections. Real traffic data was sourced
from Singapore and solved using the optimizers. For the
WCA and Jaya algorithms, new discretization strategies
were suggested in order to improve their capability in
solving LUTLSPs. Some improvements were also made
on the WCA and Jaya algorithms to improve their search
mechanism. Furthermore, a feature-based search operator
was developed based on the feature of LUTLSP and
applied to the evaluated optimizers to enhance their
solution and performance quality. Table 1 is a comparison


table that shows the difference between Jaya and the other
algorithms. Table 2 shows the superiority of Jaya over
other algorithms in optimization problems. The table

shows the Friedman rank test solutions obtained for the

‘Best’ and ‘Mean’ solutions [2]
V. CONCLUSION
In this paper, the Jaya algorithm and its working principle
were reviewed. Some of the reasons that make Jaya a good
choice when handling optimization problems were also
provided. From Table 1, it was observed that Jaya is a
simple optimization algorithm compared to TLBO, GA,
and PSO while Table 2 showed the superiority of Jaya over
other algorithms (TLBO, GA, PSO). The table showed the
best and mean solutions of Friedman rank test. The
application of Jaya in engineering and optimization
problems also recorded good performances. From the
foregoing, Jaya is a good and simple algorithm which can
achieve good solutions to several optimization problems.
ACKNOWLEDGMENT
The authors would like to express their appreciation for the
support of the sponsors from RDU Grant Vot No.
RDU1603115 and RDU180307 form Universiti Malaysia
Pahang. Our gratitude also goes to the editor and
anonymous reviewers who reviewed this paper.

REFERENCES
[1] H. M. Pandey, A. Chaudhary, and D. Mehrotra, "A
comparative review of approaches to prevent premature
convergence in GA," Applied Soft Computing, vol. 24, pp.
1047-1077, 2014.
[2] R. Rao, "Jaya: A simple and new optimization algorithm for
solving constrained and unconstrained optimization
problems," International Journal of Industrial Engineering
Computations, vol. 7, no. 1, pp. 19-34, 2016.
[3] H. M. Pandey, "Jaya a novel optimization algorithm: What,
how and why?," in Cloud System and Big Data Engineering
(Confluence), 2016 6th International Conference, 2016, pp.
728-730: IEEE.
[4] C. Gong, "An Enhanced Jaya Algorithm with a Two Group
Adaption," International Journal of Computational
Intelligence Systems, vol. 10, no. 1, pp. 1102-1115, 2017.
[5] O. Samuel, N. Javaid, S. Aslam, and M. H. Rahim, "JAYA
optimization based energy management controller for smart
grid: JAYA optimization based energy management
controller," in Computing, Mathematics and Engineering
Technologies (iCoMET), 2018 International Conference on,
2018, pp. 1-8: IEEE.
[6] K. Yu, J. Liang, B. Qu, X. Chen, and H. Wang, "Parameters
identification of photovoltaic models using an improved
JAYA optimization algorithm," Energy Conversion and
Management, vol. 150, pp. 742-753, 2017.
[7] T. Sudhakar and H. H. Inbarani, "Intelligent Path Selection in
Wireless Networks using Jaya Optimization," Procedia
Computer Science, vol. 133, pp. 976-983, 2018.
[8] K. Gao, Y. Zhang, A. Sadollah, A. Lentzakis, and R. Su,
"Jaya, harmony search and water cycle algorithms for solving
large-scale real-life urban traffic light scheduling problem,"
Swarm and Evolutionary Computation, vol. 37, pp. 58-72,
2017.