Materials Today: Proceedings xxx (xxxx) xxx

Contents lists available at ScienceDirect

Materials Today: Proceedings

journal homepage: www.elsevier.com/locate/matpr

Parametric optimization of parameters affecting dimension precision of

FDM printed part using hybrid Taguchi-MARCOS-nature inspired
heuristic optimization technique
Abhishek Mohanty a, Keshab Singh Nag a, Dilip Kumar Bagal a,⇑, Abhishek Barua b, Siddharth Jeet b,
Siba Sankar Mahapatra c, Hemanta Cherkia b
a
Department of Mechanical Engineering, Government College of Engineering, Kalahandi, Bhawanipatna, Odisha, India
b
Department of Mechanical Engineering, Centre for Advanced Post Graduate Studies, BPUT, Rourkela, Odisha, India
c
Department of Mechanical Engineering, National Institute of Technology, Rourkela, Odisha, India

a r t i c l e i n f o a b s t r a c t

Article history: Fused Deposition Modelling is a fast emerging technology due to its capacity to generate usable compo-
Received 23 May 2021 nents with multiple geometrical designs in a fair period of time without the usage of any tooling or
Received in revised form 9 June 2021 human interaction. The features and reliability of FDM fabricated parts are highly dependent on a small
Accepted 14 June 2021
number of processing variables and their settings. The current study examines the relationship between
Available online xxxx
five significant processing constraints. i.e. raster angle, part orientation, air gap, layer thickness and raster
width and what effect do they have on the dimensional accuracy of the fabricated part. Twenty-seven
Keywords:
experiments were piloted and configured using Taguchi’s architecture and recently formulated
ABS M30
Dimensional accuracy
MARCOS Method. Here, the Genetic Algorithm Optimization, Simulated Annealing Algorithm
Fused deposition modelling Optimization, Particle Swarm Optimization, Grey-Wolf Optimization Algorithm, Moth Flame
MARCOS method Optimization Algorithm, Whale Optimization Algorithm, Jaya Algorithm Optimization, Sunflower
Nature-based meta-heuristical optimization Optimization Algorithm, Lichtenberg Algorithm Optimization and Forensic Based Investigation
Optimization approaches as ten different optimizations have been utilized to predict the optimal setting
of the experiment. A comparative inspection of these nature-inspired algorithms in FDM printed part was
performed in this study which reported part orientation as the most significant element.
Ó 2021 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the 2nd International Con-
ference on Functional Material, Manufacturing and Performances.

1. Introduction layer to generate a part. Because of their poor mechanical strength,

these engineered parts are often unsuitable in a variety of manu-
Manufacturing product improvement cycle time reduction is a facturing applications. As a result, numerous analysts have com-
vital understanding in industry in order to remain competitive in mitted significant determinations to advance the properties of
the open market, and attention has changed away from traditional these parts [1,2]. Fig. 1 depicts the FDM process in action.
product design approaches and toward quick fabrication activities Sood et al. [3–5] scrutinized the effect of process constraints on
similar to rapid prototyping (RP). Fused deposition modelling the dimensional ac-curacy of FDM printed part by employing Grey-
(FDM) is a RP approach that constructs shapes or components by Taguchi method, artificial neural network (ANN) and reported
stacking and binding melted deposits in a single direction. This layer thickness of 0.254 mm, part orientation of 00, raster angle
process utilizes very slender pre - heated thermoplastic strands of 00, raster width of 0.4564 mm, and air gap of 0.008 mm will pro-
as a feed to a nozzle tip that deposits the thermoplastic layer by vide the best outcomes. After studying the mechanical properties,
they found that raster angle as the most influential parameter
⇑ Corresponding author. [6,7]. For investigating sliding wear and surface finish they used
E-mail addresses: [email protected] (A. Mohanty), keshabsinghnag@g- response surface methodology for getting optimum result [8].
mail.com (K.S. Nag), [email protected] (D.K. Bagal), rahulbarua69@gmail. Mishra et al. [9–11] studied the mechanical properties of FDM
com (A. Barua), [email protected] (S. Jeet), [email protected] (S.S. Mahapa- processed ABS part and reported raster deposition angle of 300
tra), [email protected] (H. Cherkia).

https://doi.org/10.1016/j.matpr.2021.06.216
2214-7853/Ó 2021 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the 2nd International Conference on Functional Material, Manufacturing and Performances.

Please cite this article as: A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al., Parametric optimization of parameters affecting dimension precision of FDM printed part using
hybrid Taguchi-MARCOS-nature inspired heuristic optimization technique, Materials Today: Proceedings, https://doi.org/10.1016/j.matpr.2021.06.216
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

mization, Simulated Annealing Algorithm Optimization, Particle

Swarm Optimization, Grey-Wolf Optimization Algorithm, Moth
Flame Optimization Algorithm, Whale Optimization Algorithm,
Jaya Algorithm Optimization, Sunflower Optimization Algorithm,
Lichtenberg Algorithm Optimization and Forensic Based Investiga-
tion Optimization approaches as ten different optimizations have
been utilized to predict the optimal setting of the experiment. A
comparative inspection of these nature-inspired algorithms in
FDM printed part was performed in this study.

2. Experimental methodology

2.1. Material used

Test specimen is fabricated using ABS M30 (Acrylonitrile

Butadiene Styrene) which comprises of about 90–100% of styrene,
acrylonitrile and butadiene resins along with some % of tallow, wax
and mineral oil. It is about 25–70% stronger than a regular ABS
Fig 1. FDM process.
material. ABS M300 s layer bonding is significantly stronger than
standard ABS, making it suitable for development, systematic pro-
helped in reduction of anisotropic characteristic of the FDM totyping, product construction, and production components, as per
printed parts. properties shown in Table 1 [17].
For investigating sliding wear performance and dimensional
precision, Equbal et al. employed Taguchi based Artificial Neural
2.2. Specimen fabrication
Network (ANN) and reported part orientation as the most influen-
tial contributing factor for FDM printed part [12,13].
CAD models of the test specimens were modelled in CATIA V5
Sahu et al. [14] studied the dimensional accuracy of FDM
R21 and its STL files was exported to Insight FDM software. For
printed ABS P400 part using Fuzzy Inference System where they
the fabrication of the specimens, variables are fed according to
found that part orientation plays a crucial part in maintaining
the experiment plan. From a filament spool, bendable ABS M30 fil-
the dimensional precision of the ABS P400 part. Similar finding
ament is loaded to the FDM machine’s nozzle head. The FORTUS
was also reported by Padhi et al. [15] by using Taguchi method
400mc device was used to create all specimens for their respective
and ANN.
distinctive measurement. Fig. 2 shows the dimension of the speci-
Vijayaraghavan et al. [16] used multi-gene genetic program-
men where l = 20 mm, w = 20 mm, t = 3 mm, d = 12 mm and
ming for studying the wear strength of FDM built part and reported
h = 450. Fig. 3 shows the fabrication process of specimen using
decrease in wear strength with an up-surge in layer thickness and
FDM.
raster width and rise in wear strength when air gap in-creases.
Barua et al. [17,18] reported part orientation as the most signif-
icant factor in improving surface finish and mechanical strength of Table 1
ABS part printed using FDM by using Taguchi based MOORA, Desir- Properties of ABS M30 (commercially available) [17].
ability and Utility Concept. Similarly Bagal et. al. [19] also reported Parameter Value
part orientation as the most significant factor in improving
Elongation at Break 7%
mechanical strength of ABS part printed using FDM by using Tagu- Rockwell Hardness 109.5 HRC
chi based Utility Concept. Yield Strength (Tensile) 32 MPa
Jeet et al. [20,21] suggested some of the new methods for pro- Density 1040 kg/m3
cessing polymers using free form fabrication techniques where Ultimate Strength (Tensile) 36 MPa
Young’s Modulus 2.413 GPa
FDM plays a significant role in processing of ABS material. Thermo- Layer thickness  0.25 mm
plastics such as hydroxyapatite (HA) reinforced polylactic acid
(PLA), polystyrene, expandable polystyrene (EPS) along with
acrylonitrile–butadienestyrene (ABS) were employed by Prakash
et al. [22], Singh et al. [23] and Sandhu et. al. [24,25] for parametric
appraisal of the 3D printed parts.
Based on the findings of previous investigation it is clear that
none of the re-searcher has taken more than four different dimen-
sion of a part and optimized the printing parameters of FDM
though they have used some advanced optimization techniques.
In this current exploration, prominence on assessment of dimen-
sional accuracy of part printed by fused deposition modelling
where Horizontal length (l), vertical length (w), thickness (t) and
hole diameter (d) and Angle of inclination (h) were considered.
Experiments have been funneled using Taguchi’s design and opti-
mized using recently formulated Taguchi based MARCOS Method.
From the experimental data, the process constraints such as raster
angle, part orientation, air gap, layer thickness and raster width
which substantially impact the dimensional accuracy of processed
part are to be optimized. In the later part, Genetic Algorithm Opti- Fig 2. Dimension of the specimen (mm).

2
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

2.3. Experimental design along with specimen testing

Five FDM process parameters are defined as significant factors

in this study. viz. raster angle, air gap, part orientation, layer thick-
ness, and raster width. All of the considerations were assigned a
standard based on the apparatus manufacturer’s recommended
lowest and highest settings, as well as experience and real-world
industrial applications. Fixed parameters and control factors are
depicted in Tables 2 and 3 respectively.
Horizontal length (l), vertical length (w), thickness (t) and hole
diameter (d) were measured using digital Vernier calliper with
0.01 mm least count. Angle of inclination (h) of the specimen
was calibrated using optical profile projector as shown in Fig. 4.

2.4. Design of criteria weights with Pair-Wise comparison

Fig 3. Fabrication process of specimen using FDM [18].
The relative effect of output responses was determined using
the geometric mean of the AHP equation. The pair-wise compar-
ison matrix was computed using Saaty’s nine-point choice scale,
as seen in Table 4. Equations (1) and (2) were being used to assess
Table 2
Fixed Parameters [3,17]. output response weights. [26–33]

Parameter Value !1=n

Y
n
Part fill style Perimeter or raster GMi ¼ bij ð1Þ
Observable surface Normal raster j¼1
Perimeter to raster air gap 0.0000 mm
Counter width 0.4064 mm Xn
Part internal style Solid normal wj ¼ GMi = j¼1
GMi ð2Þ
X Y & Z contract feature 1.0038
Output response weights were obtained as 0.51 for Dl, 0.076 for
Dw, 0.0846 for Dt, 0.25 for Dd, 0.0788 for Dh after the consistency
check.

Table 3
Control Factors [4,18]. 2.5. Measurement alternatives and ranking according to compromise
solution (MARCOS)
Constraints Representation Levels
1 2 3
MARCOS method was proposed by Željko Stević et. al in 2020
Layer thickness A 0.127 mm 0.178 mm 0.254 mm [34]. Initially, MARCOS method was developed for a viable supplier
Part Orientation B 00 150 300
assortment in the healthcare industry of Herzegovina and Bosnia
Raster angle C 00 300 600
Raster width D 0.4064 mm 0.4654 mm 0.5064 mm which includes the following steps:
Air Gap E 0.000 mm 0.004 mm 0.008 mm Step 1: Development of decision-making matrix (eqn (3)) and
an extended decision-making matrix containing the ideal solution

Fig 4. Measurement of different dimensions of the FDM printed part.

3
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

Table 4 2.6. Genetic algorithm

Pair-wise comparison table between norms.

Dl Dw Dt Dd Dh This approach is based on the natural encroachment method,

Dl 1 5 5 3 7 which is used to speed up responses to complicated update prob-
Dw 0.2 1 1 0.2 1 lems. The wellness work in this case re-establishes sole arithmeti-
Dt 0.2 1 1 0.33 1 cal fitness, which is comparable to the potency or limit of the
Dd 0.33 5 3 1 3 organism that gene addresses. Two guardians are chosen, and their
Dh 0.14 1 1 0.33 1
chromosomes are melded with half and half and alter instruments.
The combination is becoming increasingly useful for rapidly scout-
ing the hunting grounds. Just a few random pursuits are available
(IS) and anti-ideal solution (AIS) where ‘IS’ is the best alternative
via transformation [35–37].
and ‘AIS’ is the worst one according to eqn (4) and (5).
2 3
xIS1 xIS2 :: xISj 2.7. Simulated annealing
6 x :: x1j 7
6 11 x12 7
6 7
6 x21 x22 :: x2j 7 It’s a deterministic scheme that represents the road to harden-
X¼6
6 ::
7 ð3Þ
:: 7
ing (restrained cooling of fluid metal), with a specific ultimate goal
6 :: :: 7
6 7 of achieving the smallest linear regards in the loss function. The
4 xi1 xi2 :: xij 5
conserving marvel is surrendered by implementing a temperature
xAIS1 xAIS1 :: xAISj that is close to a boundary and offers the Boltzmann probability
where circulation concept [38–40].

IS ¼ maxxij for benefit and minxij for cost ð4Þ 2.8. Particle swarm optimization

AIS ¼ minxij for benefit and maxxij for cost ð5Þ PSO is an inhabitant-based stochastic augmentation method.
Step 2: Normalization of the extended initial matrix by applying The enhancement technique is ordained with masses of sporadic
equations (6) and (7): game plans and searches for optima by invigorating ages. The pos-
sible plans called particles fly over the concerned universe by suc-
xij
nij ¼ for benefit ð6Þ ceeding in contemporary faultless particles. Only the slightest bit
xis of food in the district was used as the gander. None of the feathered
xis animals had foggiest thoughts about food. In progress issues, each
nij ¼ for cost ð7Þ feathered animal in the chase space is alluded to as an atom. All
xij
elements were surveyed by the well-being ability to be enhanced
Step 3: Calculation of the weighted matrix where weight coef- and the speed of the particles. The issue is instated with a get-
ficients of the criterion wj calculated using AHP method [9–13], together of unpredictable particles and a short time later hunts
equation (8). for targets by energizing ages [40].
v ij ¼ nij  wj ð8Þ
2.9. Whale optimization
Step 4: Utility degree computation of alternatives ki (eqn (9)
and (10)). It’s a naturally leaded metaheuristic optimization method
Si inspired by humpback whale hunting conduct. They use an unu-
þ
ki ¼ ð9Þ sual bubble-net feeding system to search and hunt for food, in
SISi
which bubbles are produced by enclosing or crossing a ‘90 -shaped
Si path. This probing efficiency is seen in arithmetical terms as a dual

ki ¼ ð10Þ degree [41].
SAISi
where Si ði ¼ 1; 2; ::; mÞ epitomizes the addition of elements of the 2.9.1. Probing and enclosing target
weighted matrix V (eqn (11)). Equations (15) and (16) can be used to show the goal
X
n examination:
Si ¼ v ij ð11Þ
D ¼ jC:X rand  X j ð15Þ
i¼0

Step 5: Computation of utility function of alternatives f ðki Þ (eqn X ðt þ 1Þ ¼ X rand  A:D ð16Þ
(12)).
þ  A and C are the coefficient vectors in this case
ki þ ki
f ðki Þ ¼ ð12Þ A ¼ 2:a:r  a ð17Þ
1f ðkþ Þ 1f ðk Þ
1 þ f kþ i þ f k i
ðiÞ ðiÞ
 þ C ¼ 2:r ð18Þ
where f ki is the utility function on the subject of ideal solution,
  Here, ‘a’ = linearly decreasing from 2 to 0, and ‘r’ = any value
while f ki is the utility function on the subject of anti-ideal solu-
between 0 and 1.
tion (13 and 14).
  þ D ¼ jC:X  ðt Þ  XðtÞj ð19Þ
þ kj
f kj ¼ þ  ð13Þ
ki þ ki X ðt þ 1Þ ¼ X  ðt Þ  A:D ð20Þ

  
ki In lieu of A  1, Equations (17) and (18) for searching, Equations
f ki ¼ þ  ð14Þ (19) and (20) are cast-off, where X = position vector, t = current
ki þ ki
iteration, and X* = preeminent value of the position vector.
4
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

value of f(x), whereas the foulest entrant obtains the worst value
f(x) with the worst incomplete contender resolutions. Xj:k:i is the
value of the jth variable for the kth contender through ith reitera-
tion; at that time, this value is improved as per the subsequent
Equation (26).
     
X 1j:k:i ¼ X j:k:i þ r 1:k:l X j:best:i  X j:k:i   r 2:k:l X j:worst:i  X j:k:i  ð26Þ

2.12. Moth flame optimization

Seyedali Mirjalili [46] formulated the Moth-Flame Optimization

(MFO) process whose flow data begins by randomly spawning
moths inside the solution space and for each moth, position (fit-
ness values) are resolute and marking of best position is done using
flame. Following that, a spiral movement feature is used to update
the moths’ positions in order to attain better positions tagged by a
Fig 5. WOA Bubble net searching mechanism.
flame, as well as updating the current best individual positions
until the termination criteria are met.
2.9.2. Spirally pass on locus
Point apprising is signified by Equation (21): 2.13. Sunflower optimization algorithm
 
X ðtÞ  A:D; if p < 0:5
X ð t þ 1Þ ¼ ð21Þ Guilherme Ferreira Gomes et. al. [47] proposed the sunflower
D:ebl : cos ð2plÞ þ X  ðt Þ; if p ¼ 0:5 optimization which is based on the cycle of a sunflower which fol-
where ‘l’ is a quantity between 1 and 1; ‘p’ is an arbitrary numeric lows the sun. The algorithm begins with the generation of a uni-
value and 1; and ‘b’ is a continuous for retelling the wrapped ver- form or random population of flowers. The appraisal of each
sion (see Fig. 5). individual allows for the determination of who will be turned into
the sun (best solution). And, like sunflowers, all the other entities
will orient themselves towards the sun and take spontaneous
2.10. Grey Wolf Optimization (GWO)
moves in a certain direction.
It is a native electrification metaheuristic advance technique
gained by hunting grey wolves. GWO anticipated the five main 2.14. Lichtenberg Algorithm Optimization
appraisal of Grey wolves linked to ’a’ from the beginning, who
are the trailblazer wolves controlling and leading the whole pack Lichtenberg Algorithm is a modern hybrid metaheuristic that
of Grey wolves, including the guiding pursuing process. The reply incorporates trajectory and population algorithm patterns pro-
from prior wolves is offered to pioneer by the ’b’ wolf, who is a posed by João Luiz Junho Pereira et. al. in 2020 [48]. Inspired by
transcending contender to be a. ’d’ wolves lead the wolves of the thunderstorms and, more specifically, radial intra-cloud lighting,
last level, ’x’ wolves, who are in charge of preserving the wolf where the Lichtenberg figures and the strength of fractals can be
pack’s security and unchanging quality [42–44]. used. The algorithm generates Lichtenberg figures in the quest
The distances from a; andd wolves, Da , D, and Dd ,to each of the space using diffusion-limited aggregation theory and random
‘ scales and rotations for every iteration. The lowest value point of
lasting wolves X are computed using Equation (22), using which each iteration is the trigger point of the next figure, and points of
‘ ‘ ‘
the consequence of a; andd wolves on the target viz. X1 ; X2 and X3 this form are chosen for computation of the objective function.
can be computed as used in Eq. (23). As a result, the population is allocated according to the size of
      the statistic, which can shrink or expand the search space by nearly
‘  ‘ ‘ ‘ ‘  ‘ ‘ ‘ ‘  ‘ ‘ ‘
Da ¼ C 1 : X a  X ; D ¼ C 2 : X  X ; Dd ¼ C 3 : X d  X  ð22Þ 0%–100%, giving the algorithm great scope and improving the qual-
ity of its solutions.

‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘ ‘
X 1 ¼ X a  A1 : Da ; X 2 ¼ X  A2 : D ; X 1 ¼ X d  A1 : Dd ð23Þ 2.15. Forensic-based investigation algorithm

‘ ‘ ‘ ‘ ‘ ‘ Forensic-based investigation algorithm (FBI) was formulated by

A ¼ 2 a : r 1  a; C ¼ 2: r 2 ð24Þ Jui-Sheng Chou and Ngoc-Mai Nguyen in 2020 [49]. This algorithm
is inspired by suspect investigation–location–pursuit practice used
‘ ‘ ‘ ‘
X ðt þ 1Þ ¼ ðX 1 þ X 2 þ X 3 Þ=3 ð25Þ by police officers for solving any case and catch the culprit.
At first a case is opened (Problem statement) followed by the
The values of the regulatory constraints of the system, that is, a, Analysis of findings (Initialization of parameters) where a team
A, and C, are computed using Equation (25). evaluates the evidence and attempts to link it to the impressions
they already have about the investigation in order to identify
2.11. Jaya algorithm potential suspects (Best solution). Unit members create some
depending on the analysis of conclusions and test them in a new
It is centered on the notion that the revelation attained for a direction in the third phase of the inquiry. Following the establish-
problematic must transfer only in the direction of the best solution ment of lines of investigation and objectives, the team decides on
only 73,74. Let f(x) be an unbiased function for either minimization the next steps to take. The final move is to pursue the perpetrator,
or maximization. At iteration i, m no. of design variables n no. of which will continue until one prevailing and reasonably complete
aspirant solution [45]. The preeminent entrant achieves the best theory about anything surely happened emerges. It comes to a
5
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

close when a criminal has been detected (best value) and a deci- component specimen after fabrication where Dl = |20 - li|, Dw = |
sion on their case has been taken. 20 - wi|, Dt = |3 - ti|, Dd = |12 – di| and Dh = |45 - hi| (where
i = 1, 2, . . .., 27).
3. Results and discussions Computational details of MARCOS method used for optimizing
the output responses is depicted in Table 7 respectively. Experi-
As seen in Table 5, all of the research tests were made using mental data on change in dimension is computed using the benefit
Taguchi’s experimental design. In the FDM setup, 27 experimental criterion, as seen in the experimental layout (see Table 5).
runs are carried out in line with the specification of the experi-
ment. Table 6 shows the horizontal length (l), vertical length (w), 3.1. Parametric optimization using nature-inspired meta heuristic
thickness (t), hole diameter (d), and angle of inclination (h) of the optimization algorithms

Table 5
After optimizing with the MARCOS algorithm, a regression
L27 experimental runs. equation was created using the utility function values of the alter-
natives, which will be used as a fitness function in various nature-
Expt. No. A B C D E
based optimization techniques in the later stage (Equation (27) :
1. 0.127 0 0 0.4064 0 Minimize-
2. 0.127 0 0 0.4064 0.004
3. 0.127 0 0 0.4064 0.008
4. 0.127 15 30 0.4654 0
Output ¼ 5:99 þ 0:26x1  0:01418x2  0:00573x3
5. 0.127 15 30 0.4654 0.004 þ 28:65x4  46:2x5 þ 2:52x21 þ 0:000195x22
6. 0.127 15 30 0.4654 0.008
7. 0.127 30 60 0.5064 0 þ 0:000104x23  31:56x24 þ 5408x25  60:3x1 x5
8. 0.127 30 60 0.5064 0.004
9. 0.127 30 60 0.5064 0.008 þ 0:413x2 x5  0:061x3 x5  6:2x4 x5 ð27Þ
10. 0.178 0 30 0.5064 0
11. 0.178 0 30 0.5064 0.004 According to previous studies, meta-heuristic optimization
12. 0.178 0 30 0.5064 0.008 techniques provide more accurate outcomes than mathematical
13. 0.178 15 60 0.4064 0 optimization techniques. As a consequence, nature-based meta-
14. 0.178 15 60 0.4064 0.004 heuristic optimization techniques are used in this analysis to find
15. 0.178 15 60 0.4064 0.008
a more precise FDM printing constraint. For performing a concur-
16. 0.178 30 0 0.4654 0
17. 0.178 30 0 0.4654 0.004 rent optimization of total output parameters, ten different
18. 0.178 30 0 0.4654 0.008 nature-based metaheuristic optimization algorithms were
19. 0.254 0 60 0.4654 0 employed for obtaining optimal printing conditions for FDM. The
20. 0.254 0 60 0.4654 0.004
algorithms were coded in MATLAB R2019a software.
21. 0.254 0 60 0.4654 0.008
22. 0.254 15 0 0.5064 0 In order to achieve a compound score of the responses, ten dif-
23. 0.254 15 0 0.5064 0.004 ferent nature-based metaheuristic optimization algorithms were
24. 0.254 15 0 0.5064 0.008 applied to find out the consistency of the output. These methods
25. 0.254 30 30 0.4064 0 are Genetic Algorithm Optimization, Simulated Annealing Algo-
26. 0.254 30 30 0.4064 0.004
rithm Optimization, Particle Swarm Optimization, Grey-Wolf Opti-
27. 0.254 30 30 0.4064 0.008
mization Algorithm, Moth Flame Optimization Algorithm, Whale

Table 6
Output responses for experimental runs.

Expt. No. l, mm Dl, mm W, mm Dw, mm t, mm Dt, mm d, mm Dd, mm h, 0

Dh, 0

1. 20.06 0.06 20.04 0.04 3.14 0.14 11.86 0.14 45.25 0.25
2. 20.05 0.05 20.04 0.04 3.21 0.21 11.80 0.20 45.40 0.40
3. 20.08 0.08 20.03 0.03 3.22 0.22 11.79 0.21 45.30 0.30
4. 20.16 0.16 20.07 0.07 3.20 0.20 11.85 0.15 45.51 0.51
5. 20.15 0.15 20.08 0.08 3.28 0.28 11.80 0.20 45.66 0.66
6. 20.18 0.18 20.06 0.06 3.29 0.29 11.78 0.22 45.56 0.56
7. 20.15 0.15 20.13 0.13 3.24 0.24 11.72 0.28 45.51 0.51
8. 20.13 0.13 20.14 0.14 3.32 0.32 11.67 0.33 45.65 0.65
9. 20.17 0.17 20.12 0.12 3.33 0.33 11.65 0.35 45.55 0.55
10. 20.04 0.04 20.04 0.04 3.15 0.15 11.85 0.15 45.46 0.46
11. 20.03 0.03 20.05 0.05 3.23 0.23 11.80 0.20 45.61 0.61
12. 20.06 0.06 20.03 0.03 3.24 0.24 11.78 0.22 45.50 0.50
13. 20.07 0.07 20.05 0.05 3.23 0.23 11.93 0.07 45.34 0.34
14. 20.06 0.06 20.06 0.06 3.30 0.30 11.87 0.13 45.49 0.49
15. 20.09 0.09 20.04 0.04 3.31 0.31 11.86 0.14 45.39 0.39
16. 20.07 0.07 20.16 0.16 3.18 0.18 11.81 0.19 45.28 0.28
17. 20.06 0.06 20.17 0.17 3.26 0.26 11.76 0.24 45.43 0.43
18. 20.09 0.09 20.15 0.15 3.27 0.27 11.74 0.26 45.33 0.33
19. 20.07 0.07 20.03 0.03 3.12 0.12 11.88 0.12 45.10 0.10
20. 20.06 0.06 20.04 0.04 3.20 0.20 11.83 0.17 45.25 0.25
21. 20.09 0.09 20.02 0.02 3.21 0.21 11.81 0.19 45.15 0.15
22. 20.10 0.10 20.07 0.07 3.19 0.19 11.77 0.23 45.19 0.19
23. 20.09 0.09 20.08 0.08 3.27 0.27 11.72 0.28 45.33 0.33
24. 20.12 0.12 20.06 0.06 3.28 0.28 11.71 0.29 45.23 0.23
25. 20.12 0.12 20.08 0.08 3.28 0.28 11.69 0.31 45.33 0.33
26. 20.11 0.11 20.09 0.09 3.35 0.35 11.64 0.36 45.48 0.48
27. 20.14 0.14 20.07 0.07 3.36 0.36 11.62 0.38 45.37 0.37

6
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

Optimization Algorithm, Jaya Algorithm Optimization, Sunflower Raster angle 300, Raster width 0.5064 mm and Air Gap 0.004 mm
Optimization Algorithm, Lichtenberg Algorithm Optimization and which can be reported as the optimum configuration for printing
Forensic Based Investigation Optimization (see Figs. 6–15) which of ABS part using FDM for achieving a superior dimensional
are tabulated according to the chronological order of the formula- precision.
tion year along with the fitness value and optimal printing param-
eter setting in Table 8.
Astonishingly, the predicted outcomes are similar in the case of
having all ten algorithms employed: Genetic Algorithm Optimiza-
tion, Simulated Annealing Algorithm Optimization, Particle Swarm
Optimization, Grey-Wolf Optimization Algorithm, Moth Flame
Optimization Algorithm, Whale Optimization Algorithm, Jaya Algo-
rithm Optimization, Sun-flower Optimization Algorithm, Lichten-
berg Algorithm Optimization and Forensic Based Investigation
Optimization as Layer thickness 0.127 mm, Part Orientation 300,

Table 7
Computational details of MARCOS method.

Run No. Si Ki Kiþ f ðKÞ f ðKþÞ f ðK i Þ

1. 0.54 0.54 2.69 0.83 0.17 0.5206
2. 0.39 0.39 1.93 0.83 0.17 0.3734
3. 0.42 0.42 2.09 0.83 0.17 0.4041
4. 0.33 0.33 1.63 0.83 0.17 0.3147
5. 0.25 0.25 1.24 0.83 0.17 0.2403
6.. 0.26 0.26 1.30 0.83 0.17 0.2521 Fig 7. Convergence Plot of Simulated Annealing.
7. 0.27 0.27 1.36 0.83 0.17 0.2641
8. 0.22 0.22 1.09 0.83 0.17 0.2115
9. 0.23 0.23 1.13 0.83 0.17 0.2196
10. 0.45 0.45 2.23 0.83 0.17 0.4325
11. 0.35 0.35 1.76 0.83 0.17 0.3401
12. 0.35 0.35 1.73 0.83 0.17 0.3353
13. 0.43 0.43 2.13 0.83 0.17 0.4119
14. 0.31 0.31 1.56 0.83 0.17 0.3026
15. 0.33 0.33 1.67 0.83 0.17 0.3238
16. 0.42 0.42 2.10 0.83 0.17 0.4074
17. 0.30 0.30 1.52 0.83 0.17 0.2951
18. 0.32 0.32 1.62 0.83 0.17 0.3131
19. 0.89 0.89 4.47 0.83 0.17 0.8661
20. 0.47 0.47 2.33 0.83 0.17 0.4517
21. 0.62 0.62 3.11 0.83 0.17 0.6018
22. 0.50 0.50 2.49 0.83 0.17 0.4812
23. 0.33 0.33 1.66 0.83 0.17 0.3211
24. 0.40 0.40 1.98 0.83 0.17 0.3824
25. 0.32 0.32 1.60 0.83 0.17 0.3093
26. 0.25 0.25 1.23 0.83 0.17 0.2389
27. 0.28 0.28 1.38 0.83 0.17 0.2672
AIS 0.20 Fig 8. Convergence Plot of Particle Swarm Optimization.
IS 1.00

Fig 6. Convergence Plot of Genetic Algorithm. Fig 9. Convergence Plot of Grey Wolf Optimization.

7
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

Fig 13. Convergence Plot of Sunflower Optimization.

Fig 10. Convergence Plot of Moth Flame Optimization.

Fig 11. Convergence Plot of Whale Optimization. Fig 14. Convergence Plot of Lichtenberg Algorithm Optimization.

Fig 12. Convergence Plot of Jaya Optimization. Fig 15. Convergence Plot of Forensic Based Investigation Optimization.

8
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

3.2. Most significant factor S = 0.6002, R-Sq = 97.16%, R-Sq(adj) = 95.38%

It can also be noted that the estimated value residual error of
Table 9 shows the ANOVA for all dimensions where factor B, the completed experiment is relatively low, at around 2.84 percent.
part orientation with 45.20% is the most significant controlled This demonstrates not just the efficiency of experimental data fit-
parameters for fabrication of FDM processed part followed by fac- ting, but also the correctness and precision of the experimental
tor E, air gap with 19.9% contribution, factor A, Layer thickness runs. In Fig. 16, it can be noted that all the values of the normality
with 15.13% contribution, factor C, Raster angle with 11.81% contri- plot are very close to the normality line with an R-squared value of
bution and factor D, raster width with 5.12% of contribution when 97.16%. The standard residual value is equally divided in the versus
minimization of dimension error was concurrently measured. fits and the histogram also curves like a bell curve of a normal dis-

Table 8
Different Optimization method with fitness value and optimal printing parameter.

Sl. Optimization Methods Year Total Population Maximum Fitness Optimal Printing Parameter Setting
No. Formulated Iteration Value
A B C D E
1 Genetic Algorithm Optimization 1950 200 1000 0.070715 0.127 30 28.79 0.5064 0.004
2 Simulated Annealing Algorithm 1970 (Maximum Temperature is 5000 0.07762 0.127 28.48 27.91 0.5064 0.004
Optimization 100 °C)
3 Particle Swarm Optimization 1995 100 1000 0.070712 0.127 29.99 28.81 0.5064 0.004
4 Grey-Wolf Optimization Algorithm 2014 100 20 0.070743 0.127 30 29.33 0.5064 0.004
5 Moth Flame Optimization Algorithm 2015 100 50 0.070711 0.127 30 28.8 0.5064 0.004
6 Whale Optimization Algorithm 2016 100 20 0.070711 0.127 30 28.8 0.5064 0.004
7 Jaya Algorithm Optimization 2016 500 300 0.070711 0.127 30 28.8 0.5064 0.004
8 Sunflower Optimization Algorithm 2019 100 100 0.070711 0.127 30 28.8 0.5064 0.004
9 Lichtenberg Algorithm Optimization 2020 100 100 0.070711 0.127 30 28.83 0.5064 0.004
10 Forensic Based Investigation 2020 100 50 0.0707 0.127 29.99 28.84 0.5064 0.004
Optimization

Table 9
ANOVA for SN ratios.

Source DOF Adj MS F-test P-test % Contribution

A 2 30.680 15.3400 42.58 15.13
B 2 91.618 45.8092 127.16 45.20
C 2 23.932 11.9662 33.22 11.81
D 2 10.373 5.1865 14.40 5.12
E 2 40.341 20.1705 55.99 19.90
Residual Error 16 5.764 0.3603 2.84
Total 26 202.709

Fig 16. Residual Plot of Utility value.

9
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

Table 10
Actual Setting and Predicted Optimal setting.

Sl. No. Optimization Methods Fitness Value Optimal Printing Parameter Setting Actual Printing Parameter Setting
A B C D E A B C D E
1 Genetic Algorithm Optimization 0.070715 0.127 30 28.79 0.5064 0.004 0.127 30 30 0.5064 0.004
2 Simulated Annealing Algorithm Optimization 0.07762 0.127 28.48 27.91 0.5064 0.004 0.127 30 30 0.5064 0.004
3 Particle Swarm Optimization 0.070712 0.127 29.99 28.81 0.5064 0.004 0.127 30 30 0.5064 0.004
4 Grey-Wolf Optimization Algorithm 0.070743 0.127 30 29.33 0.5064 0.004 0.127 30 30 0.5064 0.004
5 Moth Flame Optimization Algorithm 0.070711 0.127 30 28.8 0.5064 0.004 0.127 30 30 0.5064 0.004
6 Whale Optimization Algorithm 0.070711 0.127 30 28.8 0.5064 0.004 0.127 30 30 0.5064 0.004
7 Jaya Algorithm Optimization 0.070711 0.127 30 28.8 0.5064 0.004 0.127 30 30 0.5064 0.004
8 Sunflower Optimization Algorithm 0.070711 0.127 30 28.8 0.5064 0.004 0.127 30 30 0.5064 0.004
9 Lichtenberg Algorithm Optimization 0.070711 0.127 30 28.83 0.5064 0.004 0.127 30 30 0.5064 0.004
10 Forensic Based Investigation Optimization 0.0707 0.127 29.99 28.84 0.5064 0.004 0.127 30 30 0.5064 0.004

tribution, which indicates the significance of the ANOVA model. [4] A.K. Sood, R.K. Ohdar, S.S. Mahapatra, Parametric appraisal of mechanical
property of fused deposition modelling processed parts, Mater. Des. 31 (1)
Table 10 shows the Actual Setting and Predicted Optimal setting.
(2010) 287–295.
[5] A.K. Sood, V. Chaturvedi, S. Datta, S.S. Mahapatra, Optimization of process
parameters in fused deposition modeling using weighted principal component
4. Conclusions analysis, J. Adv. Manufacturing Systems 10 (2011) 241–259.
[6] A.K. Sood, R.K. Ohdar, S.S. Mahapatra, Experimental investigation and
empirical modelling of FDM process for compressive strength improvement,
ABS M30 parts were processed by the help of FDM in this anal- J. Adv. Res. 3 (1) (2012) 81–90.
ysis. The MARCOS Method, which was recently developed, was [7] A.K. Sood, S.S. Mahapatra, R.K. Ohdar, Weighted principal component approach
used to improve the process parameters. Purposeful relationship for improving surface finish of ABS plastic parts built through fused deposition
modelling process, Int. J. Rapid Manufacturing 2 (1/2) (2011) 4, https://doi.org/
between different dimensions for FDM fabricated parts has been 10.1504/IJRAPIDM.2011.040687.
established with optimization techniques. A few of the critical [8] A.K. Sood, A. Equbal, V. Toppo, R.K. Ohdar, S.S. Mahapatra, An investigation
results taken from laboratory investigations for choosing the best on sliding wear of FDM built parts, CIRP J. Manuf. Sci. Technol. 5 (1) (2012)
48–54.
mix of method constraints for the FDM component are as follows:- [9] S.B. Mishra, R. Malik, S.S. Mahapatra, Effect of external perimeter on flexural
strength of FDM build parts, Arabian J. Sci. Eng. 42 (11) (2017) 4587–4595.
1) The best levels of process constraints for minimization of [10] S.B. Mishra, K. Abhishek, M.P. Satapathy, S.S. Mahapatra, Parametric appraisal
of compressive strength of fdm build parts, Mater. Today: Proc. 4 (9) (2017)
dimensional error/ maximization of dimensional accuracy
9456–9460.
for FDM processed part are part orientation of 300, layer [11] S.B. Mishra, S.S. Mahapatra, Improvement in tensile strength of FDM built
thickness of 0.127 mm, raster angle of 300, air gap of parts by parametric control, Appl. Mech. Mater. 592 (2014) 1075–1079.
[12] A. Equbal, A.K. Sood, V. Toppo, R.K. Ohdar, S.S. Mahapatra, Prediction and
0.004 mm and raster width of 0.5064 mm according to all
analysis of sliding wear performance of fused deposition modelling-processed
ten nature inspired meta-heuristic optimization method. ABS plastic parts, Proc. Inst. Mech. Eng., Part J: J. Eng. Tribol. 224 (12) (2010)
2) It is interesting to see that the factor setting obtained from 1261–1271.
all ten nature inspired meta-heuristic optimization are very [13] A. Equbal, A.K. Sood, S.S. Mahapatra, Prediction of dimensional accuracy in
fused deposition modelling: a fuzzy logic approach, Int. J. Productivity Quality
similar to each other thus providing an optimal setting. Manag. 7 (1) (2011) 22, https://doi.org/10.1504/IJPQM.2011.037730.
3) In compared to other process factors, component orientation [14] R.K. Sahu, S.S. Mahapatra, A.K. Sood, A study on dimensional accuracy of fused
contribution is the most important controllable parameter deposition modeling (FDM) processed parts using fuzzy logic, J. Manuf. Sci.
Prod. 13 (3) (2013) 183–197.
for managing the dimensional correctness of an FDM-built [15] S.K. Padhi, R.K. Sahu, S.S. Mahapatra, H.C. Das, A.K. Sood, B. Patro, A.K. Mondal,
item.. Optimization of fused deposition modeling process parameters using a fuzzy
inference system coupled with Taguchi philosophy, Adv. Manufacturing 5 (3)
(2017) 231–242.
As a consequence, this study broadens the area of Fused Depo- [16] V. Vijayaraghavan, A. Garg, J.S.L. Lam, B. Panda, S.S. Mahapatra, Process
sition Modeling features optimization to include more process lim- characterisation of 3D-printed FDM components using improved evolutionary
itations and influences on convoluted geometrical components, computational approach, Int. J. Adv. Manufacturing Technol. 78 (5-8) (2015)
781–793.
enabling for faster realisation of enhanced component construction
[17] A. Barua, S. Jeet, D.K. Bagal, P.K. Agrawal, A.K. Pattanaik, Comparative analysis
superiority. based on MCDM optimization of printing parameters affecting compressive
and tensile strength of fused deposition modelling processed parts, Int. J.
Technical Innov. Modern Eng. Sci. 5 (2) (2019) 383–392.
Declaration of Competing Interest [18] A. Barua, S. Jeet, H. Cherkia, D.K. Bagal, B.B. Sahoo, Parametric optimization of
FDM processed part for improving surface finish using MOORA technique and
desirability function analysis, Int. J. Appl. Eng. Res. 14 (13) (2019) 1–7.
The authors declare that they have no known competing finan- [19] D.K. Bagal, A. Barua, A.K. Pattanaik, S. Jeet, D. Patnaik, Parametric Optimization
cial interests or personal relationships that could have appeared Based on Mechanical Characterization of Fused Deposition Modelling
to influence the work reported in this paper. Fabricated Part Using Utility Concept, in: S. Singh, C. Prakash, S.
Ramakrishna, G. Krolczyk (Eds.), Advances in Materials Processing. Lecture
Notes in Mechanical Engineering, Springer, Singapore, 2020, https://doi.org/
10.1007/978-981-15-4748-5_3.
References [20] Jeet, S., Barua, A., & Kar, S. (2018). Free-Form Fabrication-An Emerging Trend in
Engineering. Proceedings of the Advances in Robotics, Mechanical Engineering
[1] S.S. Mahapatra, A.K. Sood, Bayesian regularization-based Levenberg– and Communication (ARMEC – 2018), Grenze Scientific Society, 78-84.
Marquardt neural model combined with BFOA for improving surface finish [21] S. Jeet, A. Barua, S. Kar, An overwiew on machining of engineering polymers,
of FDM processed part, Int. J. Adv. Manuf. Technol. 60 (9-12) (2012) 1223– Int. J. Sci. Eng. Res. 9 (4) (2018) 46–50.
1235. [22] C. Prakash, G. Singh, S. Singh, W.L. Linda, H.Y. Zheng, S. Ramakrishna,
[2] S.K. Panda, S. Padhee, A.K. Sood, S.S. Mahapatra, Optimization of fused R. Narayan, Mechanical reliability and in vitro bioactivity of 3D-printed
deposition modelling (FDM) process parameters using bacterial foraging porous polylactic acid-hydroxyapatite scaffold, J. Mater. Eng. Perform. (2021)
technique, Intelligent Information Manage. 01 (02) (2009) 89–97. 1–11.
[3] A.K. Sood, R.K. Ohdar, S.S. Mahapatra, Parametric appraisal of fused deposition [23] S. Singh, G. Singh, K. Sandhu, C. Prakash, R. Singh, Investigating the optimum
modelling process using the grey Taguchi method, Proc. Inst. Mech. Eng., Part parametric setting for MRR of expandable polystyrene machined with 3D
B: J. Eng. Manufacture 224 (1) (2010) 135–145. printed end mill tool, Mater. Today: Proc. 33 (2020) 1513–1517.

10
A. Mohanty, Keshab Singh Nag, Dilip Kumar Bagal et al. Materials Today: Proceedings xxx (xxxx) xxx

[24] K. Sandhu, S. Singh, C. Prakash, Analysis of angular shrinkage of fused filament [35] D.K. Bagal, A. Barua, S. Jeet, P. Satapathy, D. Patnaik, MCDM optimization of
fabricated poly-lactic-acid prints and its relationship with other process parameters for wire-EDM machined stainless steel using hybrid RSM-TOPSIS,
parameters, IOP Conference Series: Materials Science and Engineering 561 genetic algorithm and simulated annealing, Int. J. Eng. Adv. Technol. 9 (1)
(2019) 012058, https://doi.org/10.1088/1757-899X/561/1/012058. (2019) 366–371.
[25] K. Sandhu, G. Singh, S. Singh, R. Kumar, C. Prakash, S. Ramakrishna, G. Królczyk, [36] S. Jeet, A. Barua, D.K. Bagal, A.K. Pattanaik, P.K. Agrawal, S.N. Panda, Multi-
C.I. Pruncu, Surface characteristics of machined polystyrene with 3d printed parametric optimization during drilling of aerospace alloy (UNS A97068) using
thermoplastic tool, Materials 13 (12) (2020) 2729, https://doi.org/ hybrid RSM-GRA, GA and SA, Int. J. Manage. Technol. Eng. 9 (2019) 2501–2509.
10.3390/ma13122729. [37] A. Barua, S. Jeet, B. Parida, B.B. Sahoo, D.K. Bagal, A. Samantray, Virtual
[26] D.K. Bagal, A.K. Patra, S. Jeet, A. Barua, A.K. Pattanaik, D. Patnaik, MCDM optimization of motorcycle sprocket material by using FEA and Taguchi
Optimization of Karanja Biodiesel Powered CI Engine to Improve Performance coupled TOPSIS-GA-SA, Int. J. Adv. Scientific Res. Manage. 3 (9) (2018) 54–63.
Characteristics Using Super Hybrid Taguchi-Coupled WASPAS-GA, SA, PSO [38] S. Jeet, A. Barua, B. Parida, B.B. Sahoo, D.K. Bagal, Multi-objective optimization
Method, in: S. Bag, C.P. Paul, M. Baruah (Eds.), Next Generation Materials and of welding parameters in GMAW for stainless steel and low carbon steel using
Processing Technologies. Springer Proceedings in Materials, 9, Springer, hybrid RSM-TOPSIS-GA-SA approach, Int. J. Technical Innov. Modern Eng. Sci. 4
Singapore, 2021, pp. 491–503. https://doi.org/10.1007/978-981-16-0182-8_ (2018) 683–692.
36. [39] B. Sahoo, A. Barua, S. Jeet, D.K. Bagal, Multi objective optimization of WEDM
[27] D.K. Bagal, A. Giri, A.K. Pattanaik, S. Jeet, A. Barua, S.N. Panda, MCDM process parameters using hybrid RSM-GRA-FIS, GA and SA approach, Int. J. Res.
Optimization of Characteristics in Resistance Spot Welding for Dissimilar Advent Technol. 6 (7) (2018) 1752–1761.
Materials Utilizing Advanced Hybrid Taguchi Method-Coupled CoCoSo, EDAS [40] S. Jeet, A. Barua, D.K. Bagal, S. Pradhan, D. Patnaik, A.K. Pattanaik, Comparative
and WASPAS Method, in: S. Bag, C.P. Paul, M. Baruah (Eds.), Next Generation investigation of CNC turning of nickel-chromoly steel under different cutting
Materials and Processing Technologies, Springer Proceedings in Materials, 9, environment with a fabricated portable mist lubricator: a super hybrid
Springer, Singapore, 2021, pp. 475–490. https://doi.org/10.1007/978-981-16- Taguchi-WASPAS-GA-SA-PSO approach. Advanced Manufacturing Systems
0182-8_36. and Innovative Product Design. Lecture, in: Notes in Mechanical Engineering,
[28] D.K. Bagal, S.K. Panda, A. Barua, S. Jeet, A.K. Pattanaik, D. Patnaik, Parametric Springer, Singapore, 2021, pp. 515–531.
appraisal of CNC micro-drilling of aerospace material (PMMA) using taguchi- [41] Bagal D.K., Parida B., Barua A., Naik B., Jeet S., Singh S.K., Pattanaik A.K.,
based EDAS method, Adv. Mech. Processing Design (2021) 449–458. Mechanical Characterization of Hybrid Polymer SiC Nano Composite Using
[29] D.K. Bagal, S.K. Mahapatra, A. Barua, S. Jeet, A.K. Pattanaik, D. Patnaik, Multi- Hybrid RSM-MOORA-Whale Optimization Algorithm. In IOP Conference
parametric optimization of Wire-EDM of Inconel 718 super alloy using Series: Materials Science and Engineering, 2020, 970(1).
Taguchi-coupled WASPAS method, Adv. Mech. Processing Design (2021) [42] S.R. Mangaraj, D.K. Bagal, N. Padhi, S.N. Panda, A. Barua, S. Jeet, Experimental
459–467. study of a portable plasma arc cutting system using hybrid RSM-nature
[30] D.K. Bagal, B. Naik, B. Parida, A. Barua, S. Jeet, S.K. Singh, A.K. Pattanaik, inspired optimization technique, Mater. Today: Proc. (2021).
Comparative mechanical characterization of M30 concrete grade by fractional [43] D.K. Pradhan, B. Sahu, D.K. Bagal, A. Barua, S. Jeet, S. Pradhan, Application of
replacement of Portland Pozzolana cement with industrial waste using CoCoSo progressive hybrid RSM-WASPAS-grey wolf method for parametric
and CODAS methods, IOP Conference Series: Materials Science and optimization of dissimilar metal welded joints in FSSW process, Mater.
Engineering 970 (2020) 012015, https://doi.org/10.1088/1757-899X/970/1/ Today: Proc. (2021).
012015. [44] S. Mirjalili, S.M. Mirjalili, A. Lewis, Grey wolf optimizer, Adv. Eng. Softw. 69
[31] B. Naik, S. Paul, S.P. Mishra, S.P. Rout, A. Barua, D.K. Bagal, Performance analysis (2014) 46–61.
of M40 grade concrete by partial replacement of Portland Pozzolana cement [45] K. Abhishek, V.R. Kumar, S. Datta, S.S. Mahapatra, An integrated multi-
with marble powder and fly ash using Taguchi-EDAS method, JASC: J. Appl. Sci. response optimisation route combining principal component analysis, fuzzy
Computations 6 (2019) 733–743. inference system, nonlinear regression and JAYA algorithm: a case
[32] B. Naik, S. Paul, A. Barua, S. Jeet, D.K. Bagal, Fabrication and strength analysis experimental study on machining of GFRP (epoxy) composites, Int. J. Ind.
of hybrid jute-glass-silk fiber polymer composites based on hybrid Syst. Eng. 32 (4) (2019) 497–525.
Taguchi-WASPAS method, Int. J. Manage. Technol. Eng. IX (IV) (2019) 3472– [46] S. Mirjalili, Moth-flame optimization algorithm: a novel nature-inspired
3479. heuristic paradigm, Knowl.-Based Syst. 89 (2015) 228–249.
[33] A. Barua, S. Jeet, D.K. Bagal, P. Satapathy, P.K. Agrawal, Evaluation of [47] G.F. Gomes, S.S. da Cunha, A.C. Ancelotti, A sunflower optimization (SFO)
mechanical behavior of hybrid natural fiber reinforced nano SiC particles algorithm applied to damage identification on laminated composite plates,
composite using hybrid Taguchi COCOSO method, Int. J. Innovative Technol. Eng. Comput. 35 (2) (2019) 619–626.
Exploring Eng. 8 (10) (2019) 3341–3345. [48] J.L.J. Pereira, M. Chuman, S.S. Cunha Jr, G.F. Gomes, Lichtenberg optimization
[34] Ž. Stević, D. Pamučar, A. Puška, P. Chatterjee, Sustainable supplier selection in algorithm applied to crack tip identification in thin plate-like structures, Eng.
healthcare industries using a new MCDM method: measurement of Comput. 38 (1) (2020) 151–166.
alternatives and ranking according to compromise solution (MARCOS), [49] A.M. Shaheen, A.R. Ginidi, R.A. El-Sehiemy, S.S.M. Ghoneim, A forensic-based
Comput. Ind. Eng. 140 (2020) 106231, https://doi.org/10.1016/ investigation algorithm for parameter extraction of solar cell models, IEEE
j.cie.2019.106231. Access 9 (2021) 1–20.

11