A comparative study of PSO, GSA and SCA in parameters optimization of surface grinding process

Bulletin of Electrical Engineering and Informatics
Vol. 8, No. 3, September 2019, pp. 1117~1127
ISSN: 2302-9285, DOI: 10.11591/eei.v8i3.1586  1117
Journal homepage: http://beei.org/index.php/EEI
A comparative study of PSO, GSA and SCA in parameters
optimization of surface grinding process
Teh Muy Shin1
, Asrul Adam2
, Amar Faiz Zainal Abidin3
1,2
Faculty of Manufacturing Engineering, Universiti Malaysia Pahang, Malaysia
3
Faculty of Engineering Technology, Universiti Teknikal Malaysia Melaka, Malaysia
Article Info ABSTRACT
Article history:
Received Jan 2, 2019
Revised Mar 11, 2019
Accepted Mar 29, 2019
The selection of parameters in grinding process remains as a crucial role to
guarantee that the machined product quality is at the minimum production
cost and maximum production rate. Therefore, it is required to utilize more
advance and effective optimization methods to obtain the optimum
parameters and resulting an improvement on the grinding performance. In
this paper, three optimization algorithms which are particle swarm
optimization (PSO), gravitational search, and Sine Cosine algorithms are
employed to optimize the grinding process parameters that may either reduce
the cost, increase the productivity or obtain the finest surface finish and
resulting a higher grinding process performance. The efficiency of the three
algorithms are evaluated and compared with previous results obtained by
other optimization methods on similar studies. The experimental results
showed that PSO algorithm achieves better optimization performance in the
aspect of convergence rate and accuracy of best solution. Whereas in the
comparison of results of previous researchers, the obtained result of PSO
proves that it is efficient in solving the complicated mathematical model of
surface grinding process with different conditions.
Keywords:
Particle swarm optimization
Gravitational search algorithm
Sine Cosine algorithm
Surface grinding process
Production costs
Copyright © 2019 Institute of Advanced Engineering and Science.
All rights reserved.
Corresponding Author:
Asrul Adam,
Faculty of Manufacturing Engineering,
Universiti Malaysia Pahang,
26600 Pekan, Pahang, Malaysia.
Email: asrul@ump.edu.my
1. INTRODUCTION
Grinding process remains as an important conventional machining process that is performed on
various kind of surfaces for the shaping and finishing purpose in order to manufacture precision finished
products [1]. By comparing with other conventional machining processes like turning and milling, grinding
can produce better surface finish with smaller surface roughness. Grinding process involves material removal
and surface transmission operation which having disadvantages in terms of resources, materials, energy
consumptions and power emissions. The accuracy of the surface tolerance of the grinding process is high and
it is capable to machine tough materials. Moreover, the machining process parameter are often decided based
on the operator experience and referred the standard technical process parameters table. In any case, it is
difficult even a skilled operator to achieve the optimum parameter value for every different grinding process.
Consequently, a lot of factors that affect the performance of grinding process had occurred such as poor
surface finish, short tool life, spindle speed exceed machine capacity and inconsistent process temperature.
Therefore, it is crucial to resolve the grinding application issues by increase the process performance.
In general, the optimization problems can be classified into two categories, which are single-
objective [2] and multi-objective [3, 5] problems. By reviewing the previous works with different methods,
it can be seen that many researchers have conducted the optimization of grinding process using combined-
objective functions as one single-objective problem to optimize the process parameters [2, 4-10]. This is

 ISSN: 2302-9285
Bulletin of Electr Eng and Inf, Vol. 8, No. 3, September 2019 : 1117 – 1127
1118
because these optimization methods are having difficulties to expand to real multi objective problems since
they are not introduced to solve the multiple ideal solutions. These optimization methods are lacking from the
drawback that the decision maker needs to have detailed information to make decision on the ranking of the
objective functions. In addition, these basic optimization methods fail to apply in the discontinuous objection
function and consequently being modified or combined to form a new approach in order to solve the multi-
objective optimization problems. In this study, therefore, the single-objective approach is chosen for the
optimization of grinding process.
The motivation of this present works is aimed to evaluate the efficiency of three optimization
algorithms; (1) particle swarm optimization (PSO), (2) gravitational search (GSA), and Sine Cosines (SCA)
algorithms to the surface grinding process problem. The algorithms are utilized for both rough and finish
grinding conditions of the surface grinding process. The findings of this study are analyzed and discussed to
suggest the most efficient algorithm for grinding process problem.
2. MATHEMATICAL FORMULATION OF SURFACE GRINDING PROCESS
There are three sub-objectives that can be considered in this study which are the production cost,
the production rate and the surface finish. The four variables that considered in this paper including the speed
of wheel, , speed of workpiece, , depth of dressing, , and lead of dressing, . In both rough and finish
surface grinding process, the process variables are necessarily to satisfy the standard boundary condition in
order to obtain the boundary value. The parameters bounds for the , , , and, , are set to
[1000,2023], [10,22.7], [0.01,0.137], [0.01,0.137], respectively. The three sub-objective functions would be
separated into two different sections; (1) the rough grinding condition is aimed to minimize the production
cost and maximize the production rate, (2) the finish grinding condition is aimed to minimize the production
cost and reduce the surface roughness.
2.1. Objective functions
The three objective functions that considered in this paper are minimize the production cost
( ⁄ ), maximize the production rate in terms of workpiece removal parameter , ( ⁄ )
and minimize the surface roughness ( ).
2.2. Production cost
The total production cost, in the surface grinding process include the cost specifically related to
the grinding of the part, the cost of non-beneficial time and lastly the cost of material consumption. For the
rough grinding process, it is necessarily to minimize the , considering the various optimization parameter
mentioned above is shown in (1) which is introduced by Wen et. al. [2].
. /. / ( ) . /
.
( )
/ (1)
Where cost per hour labor and administration, =length of workpiece, =empty length of
grinding, =width of workpiece, =empty width of grinding, cross-feed rate, =total thickness of cut,
=down-feed of grinding, =number of spark out grinding, =diameter of wheel, =width of wheel,
=grinding ratio, distance of wheel idling, =number of workpieces loaded on the table, =speed of
wheel idling, =time of loading and unloading workpieces, =time of adjusting machine tool, =batch
size of the workpieces, =total number of workpieces to be ground between two dressing, =total number
of workpieces to be ground during the life of dresser, =cost of wheel per , and =cost of dressing.
2.3. Production rate
The production rate in the rough grinding process is represented by the workpieces removal
rate, . For the rough grinding process, the maximization of production rate in terms of is given
by (2) [2].
(
. / . /
) (2)

Bulletin of Electr Eng and Inf ISSN: 2302-9285 
A comparative study of PSO, GSA and SCA in parameters optimization of surface grinding… (Teh Muy Shin)
1119
Where is the percentage of wheel bond, , is the grind size, and is the workpiece hardness.
2.4. Surface finish
The surface finish of the workpiece is commonly determined to be within a particular value surface
roughness, that is highly affected by the process parameters and the wheel dressing parameter where the
mathematical model is expressed as follow [2],
0 else,
(3)
0
and
. / . / (4)
where, is an average chip thickness during grinding.
2.5. Constraints
The major constraints that involve in the optimization problem in surface grinding process are the
process and variable constraints. The three major process constraints in both rough and finish grinding
process are thermal damage wheel wear parameter, and machine tool stiffness. In addition, surface
finish will exists as a constraint in the rough grinding process in behalf of a specific surface finish needs to be
maintained to obtain the maximum production rate. While in the finish grinding, the production rate exists as
a constraint. The grinding process need a high energy per unit volume of material removed. Then, the energy
will be converted into heat energy which is thought inside the grinding region. The workpiece will be
damaged by the high thermal energy and it prompts to the decreased production rate. The specific energy
is expressed as (5) [2].
. / ( ⁄ ⁄ ) (
⁄
⁄ )
where is an initial waer flat area percentage, and is the wera constant per . The critical specific
energy , which resulting the start of thermal damages, is expressed in term of operating parameter as
shown in Equation (6) [2].
⁄
⁄ ⁄ (6)
Practically, the specific energy must be equal or less than the critical specific energy , or else the
workpiece burn will be occurred. Based on the relationship between the grinding process parameters and the
specific energy, the thermal damage constraint will be specified as Equation (7) [2].
(7)
Wheel wear parameter, is directly related to the grinding process condition and the wheel
dressing preceding is shown as Equation (8) [2].
(
⁄ ⁄
⁄ )
( ( ⁄ )) ( ⁄ )
( ( ⁄ ))
(8)
Where =constant dependant on coolant and wheel grain type. The wheel wear constraint is
obtained from the Equation (2) and (9) is expressed as Equation (9) [2].
(9)
The low machining production rate and poor surface quality may cause the chatter results. Chatter
avoidance is an obvious constraint in the selection of machining parameters. The relationship between
grinding stiffnes, , wheel wear stiffness, and the process parameters is shown in (10) and (11) [2].

 ISSN: 2302-9285
1120
(10)
(11)
In order to avoid chatter during machining, the machine tool stiffness constraint, is given as
(12).
| |
(12)
Where . / , =static machine stiffness and =dynamics machine
characteristics.
In the rough grinding condition, it is necessarily to maintain a finest surface finish with the maximum
production rate. Therefore, the sub-objective of the rough grinding will be the production rate while the
surface finish is a constraint in the rough grinding process. The constraints can be expressed as (13) [2].
(13)
While for the finish grinding condition, it is aimed to obtain an optimum surface finish at a particular
production rate. Hence, the surface finish is chosen as the sub-objective while the production rate is a
constraint. The constraints can be expressed as Equation (14) [2].
(14)
2.6. Combined-objective function model
The sub-objectives of the both rough and finish grinding operation will be normalized to a single-
weighted objective as (15) [2]. By combining the sub-objectives, it may effectively overcome the large
differences of value between the numerical value and the sub-objectives.
( ) (15)
Where =expected total production cost limitation and , , =weighting factors for the
, and respectively. For the rough grinding process, the combined objective function model
which shown in the Equation (15) is modified as Equation (16) since .
( ) (16)
Where = and based on the analysis and researches on previous papers [2].
While for finish grinding process, the combined objective function model which shown in (15) is modified as
(17) since .
( ) (17)
where = and based on the analysis and researches on previous papers [2].
3. OPTIMIZATION ALGORITHM METHODS
3.1. Particle swarm optimization
PSO algorithm is a population-based optimization method that introduced by Kennedy, J and
Eberhart, R. in 1995 [11]. To date, the PSO algorithm has widely been employed in various kind of
applications such feature selection [12], feature scaling [13]-[14], and routing in VLSI [15]. The concept of
PSO involves the changing the velocity of each particle in each level at each step and weighting the
acceleration by the generated random number separately against the ‘ ’ and ‘ ’ locations.
The updates of the particles are expert according to the following (18).

1121
The equation for the new velocity for each particle according to the initial velocity, when the
‘ ’ and ‘ ’ have achieved is shown in (18).
( ) ( ) (18)
Where =inertia coefficient , =initial velocity, =acceleration constant, =initial ,
=individual particle’s position, =initial and , =random number from 0 where
and . The equation for the personal particle’s position, in solution hyperspace is shown in (19).
(19)
The optimum selection of process parameters of PSO like the and as well as is very
important for the convergence of methods. By take into account the particle velocity and position at discrete
time step, the (18) and (19) are substituted to form the non-homogeneous recurrence relation which is shown
in (20).
( ) (20)
where and .
Provide a statement that what is expected, as stated in the "Introduction" chapter can ultimately
result in "Results and Discussion" chapter, so there is compatibility. Moreover, it can also be added the
prospect of the development of research results and application prospects of further studies into the next
(based on result and discussion).
3.2. Gravitational search algorithm
The GSA, proposed by Rashedi et. al. in year 2009 [16], is a law of gravity-based optimization
method which is also has been used in different applications [17-19]. In GSA, the agents are treated as
objects and the performance is calculated by their masses. These objects attract each other by the gravity
force which is caused by a global movement of all objects towards the objects with heavier masses.
The overall force acquired by all other agents calculates the direction of the agent. There are four major
specifications in each mass or agent which including the position, inertia mass, active gravitational mass and
lastly passive gravitational mass. The position of the mass is corresponding to the problem solution while the
inertia and gravitational masses are determined utilizing a fitness function. Particularly, each mass represent a
solution and the inertia and gravitational mass will navigate the algorithm. Theoretically, the masses will be
attracted by the heaviest mass that present an optimum solution in the search space. When considering an
optimization method with masses, the position of the th agents is expressed as (26).
( ) Where (26)
Where is the position of th agent in the th dimension. At certain times ‘ ’, the force acting on
the mass ‘ ’ from mass ‘ ’ is expressed as (27).
( )
( ) ( )
( )
. ( ) ( )/ (27)
Where is the active gravitational mass related to agent , is the passive gravitational mass
related to agent , ( ) is gravitational constant at time , ε is a minor constant and ( ) is the Euclidean
distance between two agents and and expressed as (28).
( ) ‖ ( ) ( )‖ (28)
The total force that acts on the agent in a dimension be a random weighted sum of th
components of the force that exerted from other agents is expressed as Equation (29) where is a
random number in the interval , -.
( ) ∑ ( ) (29)

 ISSN: 2302-9285
1122
The acceleration of the agent at time and the in direction th ( ) by the law of motion is
expressed as Equation (30), where is the inertia weight of the th agent.
( )
( )
( )
(30)
Moreover, the next velocity is considered as a fraction of its current velocity added to its
acceleration. The position and velocity is calculated and expressed as (31) and (32), where is a random
number in the interval [0,1].
( ) ( ) ( ) (31)
( ) ( ) ( ) (32)
The gravitational constant, is initialized and decreased with time to control the accuracy of search
space. Generally, is a function of the primarily value of ( ) and time ( )n which is expressed as follow,
( ) ( ) (33)
The fitness evaluation measures the gravitational and inertia masses. A heavier mass indicates
higher efficiency of the agent. The better agent will walk slowly and higher attractions. The gravitational and
inertia masses are updated as (34), (35) and (36).
(34)
( )
( ) ( )
( ) ( )
(35)
( )
( )
∑ ( ) (36)
( ) indicates the fitness value of the agent at time . For the optimization problem,
the ( ) and ( ) are formulated as (37) and (38).
( ) * + ( ) (37)
( ) * + ( ) (38)
In order to perform better compromise between the exploration and exploitation, the numbers of
agents are reduced with lapse of time in Equation (29). Therefore, a set of agents with larger mass apply their
force to the other is proposed. This policy may decrease the exploration energy and increase the exploitation
ability. Initially, the exploration must be used in the algorithm to prevent trapping in a local optimum.
By a lapse of interactions, the exploration is necessary to fade out while the exploitation is necessary to fade
in. The performance of the GSA will be improved through controlling the exploration and exploitation, only
the agents will attract to each other. is the function of time with the primarily value of at the
starting and reducing with time. Initially, the force is applied on all agents and through a period, is
decreased gradually and lastly there will only an agent is applying force to others. The equation (29) is
modified as (39).
( ) ∑ ( ) (39)
is the set of the first agents with the best fitness value and a larger mass.
3.3. Sine cosines algorithm
Sine and Cosine Algorithm (SCA) is a population-based optimization method which begin the
optimization process with a set of random solutions [20]. The SCA algorithm has been successfully
implemented in various engineering problem including parameter optimization of PID controller [21],
support vector regression [22], short-term hydrothermal scheduling [23], and other fields at present [24].

1123
Basically, SCA is typically divided to two stages which involving the former and exploitation. The position
updating equations for both former and exploitation stages are expressed as follow,
( ) | | (40)
( ) | | (41)
where, is the position of the current solution in -thdimension at -th iteration, , and are random
numbers, is the position of the destination point in -th dimension, and represents the absolute value.
These two equations are combined and expressed as follow,
{
( ) | |
( ) | |
(42)
where, is a random number in the interval , -.
In SCA, there are four major parameters which are as shown in (42). The parameter
defines the next position’s that could be either in the space between the solution and destination or outside it.
The parameter defines the distance of the movement which supposedly toward or outward the destination.
The parameter indicates a random weight for the destination in order to stochastically emphasize ( )
or deemphasize ( ) the effect of destination in characterizing thr distance. Lastly, the parameter
switches equally between the sine and cosine elements in Equation (42).
In order to balance the exploration and exploitation to search for the promising areas of the search
space and converge to the global optimum eventually, the range of sine and cosine in Equations (40) to (42)
is modified as,
(43)
(44)
(45)
where, is the current iteration, is the maximum number of iterations and is a constant. Over the course
of iterations will reduce the range of sine and cosine functions. The SCA explores the search space when the
range of the sine and cosine functions are in the interval , - and , -.
4. EXPERIMENTAL RESULTS AND DISCUSSIONS
4.1. Comparison performance of PSO, GSA, and SCA
A comparison of performance of the best solution at of PSO (blue line), GSA (red
line), and SCA (green line) in both rough and finish grinding conditions are presented in the convergence
plots as shown in Figure 1 and Figure 2. Referring to these figures, PSO algorithm shows better optimization
performance and the obtained results are more consistent and accurate when compared with GSA and SCA.
Generally, GSA and SCA tend to find the global optimum faster than PSO algorithm but the final results are
more crucial because they are reflecting the capability of the algorithms to escape from the bad local optima
as well as locating a close-global optimum. For this situation, PSO algorithm is able to converge the best
solution accurately at maximum iterations of 100, meanwhile for the GSA and SCA, they tend to converge
immediately when is large and maximum iterations of 100 as they are commonly used to optimize an
immerse range of problems with the fixed small population size.

 ISSN: 2302-9285
1124
Figure 1. Convergence curve of rough grinding condition
Figure 2. Convergence plot in finish grinding condition
Next, the optimal solution of these PSO, GSA and SCA algorithms are compared to the results of
previous researches on similar optimization problem in both rough and finish grinding condition. The smaller
value of COF indicates the better optimization. In Table 1, RSFS algorithm shows significant better COF
when compared with the optimization algorithms that done by previous researchers as well as PSO algorithm,
GSA and SCA. The reasons of RSFS shows better performance may due to the parameters of RSFS are tuned
by robust design methodology and providing optimal values for the algorithm. Next, the best solution of PSO
algorithm is having similar COF with HPSO, TDEOA and Mod-εDE. The best solution obtained by PSO in
this paper shows an improvement of results when compared with the previous research paper that presented
by Pawar and Rao (2013) which may influenced by using different value of the parameters and optimzation
problem is achieved by the average best solution of 50 independent runs which resulting a more accurate and
better results. This proves that PSO algorithm is efficient in solving these optimization problems.
Besides, the comparison results of the finish grinding condition for the surface grinding process are
tabulated in Table 2. In Table 2, it can be clearly seen that PSO algorithm shows better optimization
performance when compared to most of the optimization algorithms except EP-PSO. Although EP-PSO
proposes a superior optimization performance with the minimum COF value in finish grinding condition,
nonetheless its optimization performance in rough grinding condition is just average of these entire
optimization algorithms. For this reason, the optimization performance of EP-PSO is slightly inconsistent in
solving the optimization problem with different conditions. On the other hand, SCA shows average best
solution while the best solution of the GSA is quite unpleasant in the comparison of optimization
performance. Briefly, the optimization performance of PSO algorithm is better than GSA and SCA as the
solution implemented by PSO algorithm has significant smaller COF value comparing to GSA and SCA.
This is because the update operators in PSO algorithm are systematic which allow PSO algorithm to refrain

1125
trapping in local optima. Moreover, when the population size of the optimization problem is large, PSO
algorithm was able to evaluate the best solution within a short time. Meanwhile, GSA and SCA required
longer time to evaluate the best solution yet unable to obtain a better solution with smaller value of COF.
Subsequently, the obtained result of PSO algorithm proves that it is efficient in solving the complicated
mathematical model of the surface grinding process in both rough and finish grinding condition.
Table 1. Comparison results of various optimization algorithms in rough grinding condition
Method Vs Vw Doc L CT WRP Ra COF
RSFS [4] 2023 13.150 0.0740 0.137 7.273 23.716 1.7999 -0.2293
HPSO [9] 2023 13.290 0.0730 0.137 7.236 23.641 1.7999 -0.2292
TDEOA [8] 2023 13.170 0.0740 0.167 7.267 23.701 1.8000 -0.2292
PSO [This study] 2023 13.115 0.0745 0.137 7.281 23.730 1.7999 -0.2292
Mod-εDE[6] 2023 12.950 0.0760 0.137 7.261 23.690 1.7999 -0.2292
HCL-PSO [4] 2023 13.860 0.0680 0.137 7.096 23.353 1.7996 -0.2290
EP-PSO [5] 2023 10.490 0.1160 0.082 7.400 24.300 1.8000 -0.2233
SCA [This study] 2023 10.000 0.1040 0.137 8.311 25.090 1.7587 -0.2117
MPEDE [4] 1991 15.110 0.0580 0.128 6.828 22.024 1.7923 -0.2092
GSA [This study] 1983 10.708 0.0932 0.137 6.978 24.108 1.7696 -0.2019
GA [7] 1998 11.300 0.1010 0.065 7.863 22.256 1.7747 -0.1632
QP [2] 2000 19.960 0.0550 0.044 6.204 17.506 1.7477 -0.1275
Rough Grinding Condition: ( , , )
Table 2. Comparison results of the performance of various optimization algorithms in finish
grinding condition
Method Vs Vw Doc L CT WRP Ra COF
EP-PSO [5] 2023 20.22 0.015 0.137 6.700 20.000 0.7600 0.5040
PSO [This study] 2023 22.70 0.0109 0.137 7.13 20.000 0.7934 0.5224
RSFS [4] 2023 22.70 0.0109 0.137 7.130 20.000 0.7934 0.5224
HPSO [16] 2023 22.70 0.0109 0.137 7.130 20.000 0.7934 0.5224
TDEOA [8] 2023 22.66 0.011 0.137 7.136 20.001 0.7931 0.5225
Mod-εDE [6] 2023 22.70 0.011 0.137 7.132 20.005 0.7935 0.5225
MPEDE [4] 2023 22.14 0.012 0.137 7.212 20.020 0.7891 0.5232
ACO [10] 2023 19.36 0.019 0.134 7.694 20.012 0.7644 0.5281
SCA [This study] 2017 17.34 0.041 0.115 8.207 20.334 0.7756 0.5332
GA [7] 1986 21.40 0.024 0.136 7.371 20.616 0.8268 0.5427
HCL-PSO [4] 1965 18.74 0.0338 0.127 7.875 20.260 0.8037 0.5488
GSA [This study] 1944 19.10 0.036 0.127 7.361 20.259 0.8196 0.5531
QP [2] 2000 19.99 0.052 0.091 7.719 20.088 0.8309 0.5547
Finish Grinding Condition: ( n: , , )
5. CONCLUSIONS
In this paper, PSO algorithm is applied to optimize the surface grinding process parameter in both
rough and finish grinding condition. The optimal value of the decision variables which including , ,
and were obtained by using the mathematical model with the purpose of reducing the production cost and
increasing the production rate as well as improving the surface finish. An experiment had conducted to
evaluate the efficiency of these proposed algorithms as well as obtain results that are more reliable According
to the obtained results from the experiment, it can be observed that PSO algorithm shows superior results
when compared to the proposed GSA and SCA. Thus, the experimental results were compared with previous
studies and successful evaluated the performance of PSO algorithm. Throughout the comparison, it can be
found that the obtained results of proposed PSO algorithms shows better optimization performance when
compared to most of the optimization algorithms except EP-PSO. Moreover, the obtained result of this paper
is showing significant improvement when compared with the similar study of PSO algorithm by Pawar and
Rao [25].
Furthermore, the optimization performance of PSO algorithm in aspects of convergence rate and
accuracy of the best solution was successfully achieved. By comparing with GSA and SCA, PSO algorithm is
able to achieve good convergence time of the best solution with higher accuracy. Additionally, the maximum
iteration of 100 is sufficient for PSO algorithm to convergence accurate and consistent best solution and it
meant it could be easily redesign to optimize various process parameters of different conventional machining
processes like turning, milling and so on. Subsequently, the obtained result of PSO algorithm proves that it is
efficient in solving the complicated multi objective optimization model of the surface grinding process in
both rough and finish grinding condition.

 ISSN: 2302-9285
1126
ACKNOWLEDGEMENTS
This research is funded by Internal UMP Research Grant (RDU180394) that awarded by Ministry of
Education Malaysia to Universiti Malaysia Pahang (UMP). Also thanks to Faculty of Manufacturing
Engineering, UMP for supporting this research work through the utilization of faculty’s equipments.
REFERENCES
[1] Tönshoff, H. K., Karpuschewski, B., Mandrysch, T., & Inasaki, I., "Grinding process achievements and their
consequences on machine tools challenges and opportunities," CIRP Annals-Manufacturing Technology, vol. 47,
no. 2, pp. 651-668, 1998.
[2] Wen, X.M., A.A.O. Tay, A.Y.C. Nee, "Micro-Computer-Based Optimization of the Surface Grinding
Process," Journal of Materials Processing Technology, vol. 29, no. 1–3, pp. 75–90, 1992.
[3] Janardhan, M., "An Integrated Evaluation Approach for Modelling and Optimization of Surface Grinding Process
Parameters," Materials Today: Proceedings, vol 2, nom. 4–5, pp. 1622–1633, 2015.
[4] Khalilpourazari, S., & Khalilpourazary, S., "A lexicographic weighted Tchebycheff approach for multi-constrained
multi-objective optimization of the surface grinding process," Engineering Optimization, vol. 49, no. 5, pp. 878–
895, 2017.
[5] X. Lin and H. Li, "Enhanced Pareto Particle Swarm Approach for Multi-Objective Optimization of Surface
Grinding Process," 2008 Second International Symposium on Intelligent Information Technology Application,
Shanghai, 2008, pp. 618-623.
[6] Rana, P., & Lalwani, D. I., "Parameters optimization of surface grinding process using Modified ε constrained
Differential Evolution," Materials Today: Proceedings, vol. 4, no. 9, pp. 10104–10108, 2017.
[7] Saravanan, R., Asokan, P., & Sachidanandam, M., "A multi-objective genetic algorithm (GA) approach for
optimization of surface grinding operations," International Journal of Machine Tools and Manufacture, vol. 42, no.
12, pp. 1327–1334, 2002.
[8] J. Tsai, K. Lee and J. Chou, "Robust Evolutionary Optimal Tolerance Design for Machining Variables of Surface
Grinding Process," in IEEE Transactions on Industrial Informatics, vol. 10, no. 1, pp. 301-312, Feb. 2014.
[9] Zhang, G., Liu, M., Li, J., Ming, W., Shao, X., & Huang, Y., "Multi-objective optimization for surface grinding
process using a hybrid particle swarm optimization algorithm," The International Journal of Advanced
Manufacturing Technology, vol. 71, no. 9, pp. 1861–1872, April 2014.
[10] Baskar, N., Saravanan, R., Asokan, P., & Prabhaharan, G., "Ants colony algorithm approach for multi-objective
optimisation of surface grinding operations," The International Journal of Advanced Manufacturing
Technology, vol. 23, no. 5-6, pp. 311-317, 2004.
[11] Kennedy, J., & Eberhart, R., "Particle Swarm Optimization," Paper presented at the Proceedings of the IEEE
International Conference on Neural Networks (ICW), Perth, Western Australia, 1995.
[12] Asrul Adam, Mohd Ibrahim Shapiai, Mohd Zaidi Mohd Tumari, Mohd Saberi Mohamad, and Marizan Mubin,
"Feature Selection and Classifier Parameters Estimation for EEG Signals Peak Detection Using Particle Swarm
Optimization," The Scientific World Journal, vol. 2014, pp. 1-13, 2014.
[13] K. G. Li, M. I. Shapiai, A. Adam and Z. Ibrahim, "Feature scaling for EEG human concentration using particle
swarm optimization," 2016 8th International Conference on Information Technology and Electrical Engineering
(ICITEE), Yogyakarta, 2016, pp. 1-6.
[14] Adam, A., Ibrahim, Z., Mokhtar, N., Shapiai, M. I., Cumming, P., & Mubin, M., "Improving EEG signal peak
detection using feature weight learning of a neural network with random weights for eye event-related
applications," Sādhanā, vol. 42, no. 5, pp. 641-653, May 2017.
[15] Z. M. Yusof et al., "A Two-Step Binary Particle Swarm Optimization Approach for Routing in VLSI," ICIC
Express Letters," vol. 6., no. 3, pp. 771-776, 2012.
[16] Rashedi, E., Nezamabadi-pour, H., & Saryazdi, S, "GSA: A Gravitational Search Algorithm. Information
Sciences," Information Sciences, vol. 179, no. 13, pp. 2232–2248, 2009.
[17] A. Adam, N. Mokhtar, M. Mubin, Z. Ibrahim, M. Z. M. Tumari and M. I. Shapiai, "Feature Selection and Classifier
Parameter Estimation for EEG Signal Peak Detection Using Gravitational Search Algorithm," 2014 4th
International Conference on Artificial Intelligence with Applications in Engineering and Technology, Kota
Kinabalu, 2014, pp. 103-108.
[18] K. Z. M. Azmi, D. Pebrianti, Z. Ibrahim, S. Sudin and S. W. Nawawi, "Simultaneous Computation of Model Order
and Parameter Estimation for System Identification Based on Gravitational Search Algorithm," 2015 6th
International Conference on Intelligent Systems, Modelling and Simulation, Kuala Lumpur, 2015, pp. 135-140.
[19] Ibrahim, I., Ibrahim, Z., Ahmad, H., Yusof, Z. M., Shapiai, M. I., Nawawi, S. W., & Mubin, M., "An Assembly
Sequence Planning Approach with Binary Gravitational Search Algorithm," In SoMeT, August 2014, pp. 179-193.
[20] Mirjalili, S., "SCA: A Sine Cosine Algorithm for solving optimization problems," Knowledge-Based Systems, vol.
96, pp. 120–133, 2016.
[21] Shashikant, Shaw B., "Comparison of SCA-Optimized PID and P&O-Based MPPT for an Off-grid Fuel Cell
System," In: Nayak J., Abraham A., Krishna B., Chandra Sekhar G., Das A. (eds) Soft Computing in Data
Analytics. Advances in Intelligent Systems and Computing, vol 758, 2019, Springer, Singapore
[22] S. Li, H. Fang, and X. Liu, Parameter optimization of support vector regression based on sine cosine algorithm,"
Expert Systems with Applications, vol. 91, pp. 63–77, 2018.

1127
[23] S. Das, A. Bhattacharya, and A. K. Chakraborty, Solution of short-term hydrothermal scheduling using sine cosine
algorithm," Soft Computing, vol. 6, pp. 1–19, .
[24] Bidyadhar Rout, B.B. Pati, and S. Panda, Modified SCA Algorithm for SSSC Damping Controller Design in Power
System," ECTI Transactions on Electrical Engineering, Electronics, and Communications, vol.16, no.1, pp. 46-63,
2018.
[25] Pawar, P. J., & Rao, R. V., "Parameter optimization of machining processes using teaching–learning-based
optimization algorithm," The International Journal of Advanced Manufacturing Technology, vol. 67, no. 5–8, pp.
995–1006, 2013.
BIOGRAPHIES OF AUTHORS
Teh Muy Shin received his first degree, B.Eng. degree in Manufacturing Engineering from
Universiti Malaysia Pahang, Malaysia in 2018. His research interest involves parameters
optimization on machining processes.
Asrul Adam received the B.Eng. degree and M.Eng. degree in Electrical-Mechatronics and
Electrical Engineering (Computational Intelligence) from Universiti Teknologi Malaysia
(UTM), Malaysia, in 2009 and 2012, respectively. He received his PhD degree in Signal and
System from University of Malaya, Malaysia in 2017. Currently, he is a senior lecturer at the
Faculty of Manufacturing Engineering, Universiti Malaysia Pahang, Pahang, Malaysia where
he is conducting teaching activities in the areas of electronics and mechatronics engineering.
Asrul Adam's research activities concern on the areas of biomedical signals processing,
artificial intelligence, brain computer interface, machine learning, and optimization. He has
published 15 Scopus/ISI-WoS indexed journals, 20 international conferences, and one book
chapter. Asrul Adam recently has 110 total citation with 6 h-index and 234 total citation with 9
h-index based on SCOPUS and Google Scholar, respectively.
Amar Faiz Zainal Abidin received his first degree, B.Eng. (Hons) in Electrical and Electronics
Engineering from The University of Nottingham, Malaysia and M.Eng. from Universiti
Teknologi Malaysia (UTM). He is currently attached as an academic staff with Universiti
Teknikal Malaysia Melaka (UTeM). His research interests include multiobjective optimisation
and nature-inspired optimization algorithms.

A comparative study of PSO, GSA and SCA in parameters optimization of surface grinding process

More Related Content

What's hot (17)

Similar to A comparative study of PSO, GSA and SCA in parameters optimization of surface grinding process (20)

More from journalBEEI (20)

Recently uploaded (20)

A comparative study of PSO, GSA and SCA in parameters optimization of surface grinding process