This article appeared in a journal published by Elsevier.

The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elseviers archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright

Author's personal copy

Journal of Manufacturing Systems 32 (2013) 197205

Contents lists available at SciVerse ScienceDirect

Journal of Manufacturing Systems

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

Technical Paper

Normal-boundary intersection based parametric multi-objective optimization of

green sand mould system
T. Ganesan a , P. Vasant b, , I. Elamvazuthi c
a

Chemical Engineering Department, University Technology Petronas, Malaysia

Fundamental & Applied Sciences Department, University Technology Petronas, Malaysia
c
Electrical & Electronic Engineering Department, University Technology Petronas, Malaysia
b

a r t i c l e

i n f o

Article history:
Received 27 April 2012
Received in revised form 26 October 2012
Accepted 30 October 2012
Available online 20 November 2012
Keywords:
Multi-objective (MO)
Green sand mould system
Normal Boundary Intersection (NBI)
Genetic algorithm (GA)
Particle Swarm Optimization (PSO)
Uniform spread
Pareto frontier

a b s t r a c t
In manufacturing engineering optimization, it is often that one encounters scenarios that are multiobjective (where each of the objectives portray different aspects of the problem). Thus, it is crucial for
the engineer to have access to multiple solution choices before selecting of the best solution. In this work,
a novel approach that merges meta-heuristic algorithms with the Normal Boundary Intersection (NBI)
method is introduced. This method then is used generate optimal solution options to the green sand
mould system problem. This NBI based method provides a near-uniform spread of the Pareto frontier in
which multiple solutions with gradual trade-offs in the objectives are obtained. Some comparative studies
were then carried out with the algorithms developed and used in this work and that from some previous
work. Analysis on the performance as well as the quality of the solutions produced by the algorithms is
presented here.
2012 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

1. Introduction
Most issues encountered when dealing with emerging technologies in engineering are multi-objective (MO) in nature [1,2].
Strategies in multi-objective optimization (MO) can be crudely
classied into two classes. First being methods that use the concept of Pareto optimality to trace the non-dominated solutions
at the Pareto curve (for instance, Zitzler and Thieles [3] Strength
Pareto Evolutionary Algorithm (SPEA) and Non-dominated Sorting Genetic Algorithm II (NSGA-II) by Deb et al. [4]). The other
type of methods is known as the weighted (or scalarization) techniques. In these methods, the objective functions in the problem
are aggregated into a single objective function which is then solved
for various scalar (weight) values. Some known scalarization techniques include the Weighted Sum method [5,6], Goal Programming
[7] and Normal-Boundary Intersection (NBI) method [8]. Using
these techniques, the scalars (or weights) are used to consign relative trade-offs to the objectives during the aggregation procedure.
Hence, alternative near-optimal solution options are generated for
various values of the scalars.
A unique and ideal solution that explains all the features of a MO
problem in engineering are rarely encountered [9,10]. Nevertheless

Corresponding author.
E-mail address: vasant0001@yahoo.com (P. Vasant).

in more practical scenarios, the decision maker (DM) is only interested in a single optimal solution. To select this unique optimum,
the DM utilizes some supplementary knowledge which is usually
very heuristic and too complex to be represented mathematically
[11]. Therefore, it is very useful for the DM to have access to numerous solution options with a variety of signicance with respect to
the objectives prior to the selection the best optimal solution. See
[1,12,13] for more detail investigations and explanations on MO
techniques in engineering optimization.
In optimization problems of this kind, it is required that the solution method caters for the multiobjective nature of the problem.
Thus, in this work the MO issue is tackled using the NBI method for
geometrical trade-offs of the weights while the GA-PSO is used to
iteratively improve the solutions for each respective weight. This
work aims to generate a series of Pareto-optimal solutions that
obtain a near-complete trade-off among the objective functions for
the green mould sand system. This problem was presented and
solved in Surekha et al. [14] by the application of genetic algorithm (GA) and Particle Swarm Optimization (PSO) techniques in
conjunction with the Weighted Sum approach.
The difference between sand mould and green sand mould is
that green sand mould has green compression strength, permeability, hardness and bulk density requirements where as sand mould
has the same properties without the green constraints. In green
mould systems, the quality of the product obtained from the moulding process is very dependent on the physical properties of the

0278-6125/$ see front matter 2012 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jmsy.2012.10.004

Author's personal copy

198

T. Ganesan et al. / Journal of Manufacturing Systems 32 (2013) 197205

moulding sand (such as hardness, permeability, green compression

strength and bulk density). Incorrect proportions of the mentioned
properties may lead to casting defects such as poor surface nish,
blowholes, scabs and pinhole porosity. Controllable variables such
as percentage of water, percentage of clay, grain neness number
and number of strokes heavily inuence the physical properties of
the moulded sand. Hence, by characterizing these parameters as the
decision variables and the mould sand properties as the objective
function, the MO optimization problem was formulated in Surekha
et al. [14]. The purpose of this formulation is for the identication of best controllable parameters for optimal nal-product of
the moulding process. A more comprehensive study on the optimization and model formulation of mould systems can be seen in
[15,16].
In this work, the green mould sand system is optimized further
using genetic algorithms (GA), Particle Swarm Optimization (PSO)
and a hybrid GA-PSO in conjunction with the Normal Boundary
Intersection (NBI) method to generate a series of Pareto-optimal
solutions. Comparison studies were then performed on the optimal
solutions obtained in this work against those obtained in Surekha
et al. [14].
Genetic algorithms (GA) were introduced by Holland in the
nineties [17]. GAs belong to the group of stochastic search methods (such as simulated annealing [18] and some forms of branch
and bound). While most stochastic search techniques operate on a
distinct solution for a particular problem, GAs operates on a population of solutions. In recent times, GAs have been widely applied
in engineering scenario (see [19,20]). For a more comprehensive
text on GAs refer to [21]. Particle Swarm Optimization (PSO) is an
optimization method developed based on the movement and intelligence of swarms. PSO was developed by Kennedy and Eberhart
[22] in 1995. Lately, PSO has been applied to a variety of areas
including optimization problems in engineering [23] as well as
economic dispatch problems. Many works have done on the application of meta-heuristic techniques for modelling and optimization
of manufacturing systems [2426].
This paper is organized as follows. In Section 2 of this paper,
the standard meta-heuristic techniques are presented, and this is
followed by description on the Scalarisation Technique and Proposed Algorithms in Section 3. The real world application problem
on green sand mould system is illustrated in Section 4. Section 5 discusses computational results and nally, the concluding remarks
are given in Section 6.

some (user specied) tness condition. The candidate optimal solutions obtained by this algorithm are achieved as a result of particles
which are in motion (swarming) through the tness landscape. In
the beginning, some candidate solutions are selected by the PSO
algorithm. These solutions can be randomly selected or be established with the aid of some a priori facts. Next, the evaluation of
the particles position and velocity (which are also the candidate
solutions) relative to the tness function is carried out. Consequently, in conjunction with the tness function a condition is
introduced; where if the tness function is not fullled, then the
algorithm updates the individual and social terms by the aid of a
user-specied update rule. Following this, the velocity and the position of the particles are updated. This recursive course of action is
iterated until the tness function is satised by all candidate solutions and solutions have thus converged into a x position. It is
essential to note that the velocity and position updating rule is critical to the optimization capabilities of this method. The velocity of
each particle in motion (swarming) is updated using the following
equation.

vi (t + 1) = wvi (t) + c1 r1 [xi (t) xi (t)] + c2 r2 [g(t) xi (t)]

where each particle is identied by the index i, vi (t) is the particle

velocity and xi (t) is the particle position with respect to iteration
(t). The parameters w, c1 , c2 , r1 and r2 are usually dened by the
user.
3. Scalarisation Technique and Proposed Algorithms
3.1. Scalarisation technique: Normal Boundary Intersection (NBI)
method
The NBI method was rst introduced by Das and Dennis [8].
This method is a geometrically inspired scalarization approach for
solving MO problems. In contrast to the Weighted Sum method,
the NBI approach has the ability to nd a near-uniform spread
of Pareto-optimal solution options in the frontier. This makes the
NBI approach a more interesting alternative as compared to the
Weighted Sum method when solving non-convex MO problem.
The green mould system problem is presented as the following:
Min F(x) subject to
X = {x : g(x) = 0; h(x) 0, 1 x 4}
F = (f1 , f2 , f3 , f4 )

2. Standard meta-heuristic techniques

2.1. Genetic algorithm (GA)
A genetic algorithm (GA) was applied in conjunction to the NBI
approach for the MO optimization of the green sand mould system.
GAs are categorized as a class of population-based search and optimization algorithms [27,28]. An N-point crossover operator was
used to create new offspring for each successive generation. To
avoid the solution from getting stagnant at the local minima, an
N-bit ip mutation operator was used.
2.2. Particle Swarm Optimization (PSO)
The PSO algorithm introduced in 1995 (by Kennedy and Eberhart
[22]) springs from two distinct frames of ideas. The rst concept
was based on the examination of swarming (or ocking) behaviours
of certain species of organisms (such as birds, ants, bees and reies). The second idea was sprung from the study of evolutionary
computations. The PSO algorithm searches the search space for
candidate solutions and evaluates these solutions with respect to

(1)

(2)

where F* is the utopia point for this MO problem. Let the individual
minimum be denoted as xi and be obtained for i [1, 4]. The convex
hull of the individual minima is generated in this fashion. Thus, the
representation of the simplex from the convex hull is as follows:
 = { Y :  = F(xi );

Y = i : 1 x 4}

(3)

4

= 1. The formulation of
where  forms a 4 by 4 matrix and
i=1 i
the NBI -sub problem is as the following:
Max(X,t) t

subject to

 Y + tn = F(x) and

xX

(4)

where t is some dened distance parameter, and n is the normal

vector at the point towards the utopia point. The NBI scalarization
method nds the maximum distance, t in the direction of the normal vector, n between a point on the simplex and the origin (or
the utopia point). Next, the scalarization is carried out. The scalars,
Y are varied thoroughly to generate a near-uniform spread of the
Pareto frontier. The procedures of which this method is executed
are as follows:

Author's personal copy

T. Ganesan et al. / Journal of Manufacturing Systems 32 (2013) 197205
Table 1
Genetic algorithm (GA) settings.
Parameters

Values

Length of individual string

No. of individuals in the population
Probability of mutation
Probability of recombination
Initial string of individuals
Bit type of individuals string
Cross-over type
Mutation type
Selection type

6 bit
6
0.3333
0.5
Random
Real-coded
N-point
N-bit ip
Tournament

Step 3: By recombination from the current population, create offspring for the next generation.
Step 4: Mutate offspring for this generation.
Step 5: The parent selection to create the next generation is done
by tournament selection.
Step 6: The next population of n individuals is chosen.
Step 7: Set new population to current population.
Step 8: Assess the tness of each offspring in the generation.
Step 9: If the stopping criterion are satised halt program and print
solutions, else go to Step 3.
Algorithm 2.

The MO maximization problem is reformulated as a minimization

problem by inverting the objective functions.
Obtain the local minima of individual objective functions.
The multi-objective (MO) problem is then reformulated as a
single-objective -sub problem by using the value of the local
optimums.
The single-objective -sub problem is then solved.
The inverse transformation is then performed to re-obtain the
best maximal values of the objective functions.
The Pareto front for multiple scalarization values of is obtained.
The optimal value of the solutions is then selected from the
obtained Pareto front.
Thus, the single-objective -sub problem is solved using GA, PSO
or the hybrid GA-PSO where these algorithms are termed as NBIPSO
(NPSO), NBIGA (NGA) and NBI Hybrid GA-PSO (NHPSO).
Algorithm 1.

Genetic algorithm (GA)

The GA scheme applied in this work is as the following. The

parameter settings initialized prior to the execution of the GA used
in this work are shown in Table 1. The owchart of the GA algorithm
is shown in Fig. 1.
Step 1: Initialize a random chromosome for n individuals in the
population.
Step 2: Assign tness conditions to each of the n individuals in the
population.

Particle Swarm Optimization (PSO)

These parameter settings for this algorithm are usually constrained as the following:
0 w 1.2
0 c1 2
0 c2 2

(5)

0 r1 1
0 r2 1
The term wvi (t) in Eq. (1) (also referred to as the inertial
term) maintains the particles motion in the same direction as
its original vector. The inertial coefcient w serves as a dampener or an accelerator during the movement of the particles.
The term c1 r1 [xi (t) xi (t)] better known as the cognitive component functions serves as the memory. This component ensures
that the particle tends to return to the position in the search
space where the particle had a very high value of the tness
function. The term c2 r2 [g(t) xi (t)] (also known as the social
component) function as mover of the particles to the position
where the swarm has visited in the previous iterations. Next,
the particles position is then computed as is shown in the
following:
xi (t + 1) = xi (t) + vi (t + 1)

Randomly initialize
a population of n
individuals.

Generate offspring
by crossover/recombination

Offspring
Mutation

Perform
parent
selection

NO
Set new
population
to current
population

(6)

Until all candidate solutions are at their highest tness positions

and the termination criterion is satised, these iterations are then

START
Fitness criterions are
assigned to each of
the n individuals.

199

Evaluate
offspring
fitness

Termination
criterion
satisfied?
YES

STOP

Fig. 1. Algorithm ow for GA.

Author's personal copy

200

T. Ganesan et al. / Journal of Manufacturing Systems 32 (2013) 197205

sustained. The algorithm of the PSO method used in this work is

shown as follows:

Step 7: If the tness conditions are satised, stop program and

print solutions, else go to Step 3.

Step 1: Set number of particles, i and the initialize parameter settings w, c1 , c2 , r1 , r2 , no

Step 2: Randomly initialize particles position xi (n) and velocity
vi (n)
Step 3: Calculate inertial and social components of the particles
Step 4: Compute position xi (n + 1) and velocity vi (n + 1) of the particles at next iteration
Step 5: If the swarm evolution time, n > no + T, update position xi
and velocity vi and go to Step 3, else proceed to Step 6
Step 6: Proceed with the evaluation of the tness of each particle
in the swarm.

where no is some constant, n is the swarm iteration and T is the overall program iteration. However, in the event during this iterative
process the position of all the particles converges, the solutions are
feasible with respect to the specied ranges, no further optimization of the objective function occurs and all the decision variables
are non-negative (for the problem at hand) then it can be said that
the tness criterion are met. Hence the candidate solutions are
at their ttest and the program is stopped and the solutions are
printed. The initialization parameters for this algorithm is shown
in Table 2 and the workow is provided in Fig. 2.

START

Initialize no of
particles, i

Initialize
algorithm
parameters

Randomly initialize
position xi(n) and
velocity vi(n)
n = n +1

Compute inertial and

social influence

Compute position xi(n+1)

and velocity vi(n+1) at
next iteration

T = T +1

NO
Is fitness criterion
satisfied?

Evaluate
fitness of the
swarms

YES

YES

STOP
Fig. 2. Algorithm ow for PSO.

Is n > no+T ?

NO

Author's personal copy

T. Ganesan et al. / Journal of Manufacturing Systems 32 (2013) 197205
Table 2
Particle Swarm Optimization (PSO) settings.
Parameters

Values

Initial parameter (c1 , c2 , r1 , r2 , w)

Number of particles
Initial social inuence (s1 , s2 , s3 , s4 , s5 , s6 )

(1, 1.2, 0.5, 0.5, 0.8)

6
(1.1, 1.05, 1.033,
1.025, 1.02, 1.017)
(3, 4, 5, 6, 7, 8)

Initial personal inuence (p1 , p2 , p3 , p4 , p5 , p6 )

201

Step 14: If the swarm evolution time, n > no + T, update position xi

and velocity vi and go to Step 12, else proceed to Step 6
Step 15: Evaluate tness swarm based on the objective function
of the -sub problem for each scalarization.
Step 16: If tness criterion satised, halt and print solutions, else
go to Step 12.

4. Application data
Algorithm 3.

Hybrid GA-PSO

The hybrid GA-PSO in this work was developed by using both

techniques at different sections of the NBI approach. As mentioned
in the introduction section, meta-heuristic techniques (such as GA
and PSO) can be incorporated into the NBI approach at two segments. First being the use of these techniques as a tool to search for
the local minima of the individual objective functions. Secondly,
using these techniques for solving the -sub problem by varying
the weights. In contrast to the pure GA and PSO approach (where
both segments in the NBI approach are embedded with the PSO
or GA technique), the hybrid approach uses GA to nd the local
minima and PSO for solving the -sub problem. The parameter settings in the hybrid method are identical to the settings employed in
the pure methods. The hybridization procedure in Fig. 3 shows the
mechanism of placement of the GA and PSO algorithms in the NBI
sections.
The algorithm for the hybrid GA-PSO approach is as the following:
Step 1: Initialize a random chromosome for n individuals in the
population.
Step 2: Assign tness conditions to each of the n individuals in the
population.
Step 3: By recombination from the current population, create offspring for the next generation.
Step 4: Mutate offspring for this generation.
Step 5: The parent selection to create the next generation is done
by tournament selection.
Step 6: The next population of n individuals is chosen.
Step 7: Set new population = current population.
Step 8: Assess the tness of each offspring in the generation.
Step 9: If the stopping criterion are satised halt program and print
solutions, else go to Step 3.
Step 10: Set no of particles, i and the initialize parameter settings
w, c1 , c2 , r1 , r2 , no
Step 11: Randomly initialize particles position xi (n) and velocity
vi (n)
Step 12: Calculate inertial and social components of the particles
Step 13: Compute position xi (n + 1) and velocity vi (n + 1) at next
iteration

The responses of the mould heavily inuence the quality of the

nal product of the green sand mould system. In Surekha et al. [14],
these responses are represented mathematically as the objective
functions. The responses are; green compression strength (f1 ), permeability (f2 ), hardness (f3 ) and bulk density (f4 ). These objectives
on the other hand are inuenced by on the process (or decision)
variables which are; the grain neness number (A), percentage of
clay content (B), percentage of water content (C) and number of
strokes (D). The objective functions and the range of the decision
variables are shown as follows:
f1 = 17.2527 1.7384A 2.7463B + 32.3203C + 6.575D
+ 0.014A2 + 0.0945B2 7.7857C 2 1.2079D2 + 0.0468AB
0.1215AC 0.0451AD + 0.5516BC + 0.6378BD + 2.689CD)
(7)

f2 = 1192.51 15.98A 35.66B + 9.51C 105.66D + 0.07A2

+ 0.45B2 4.13C 2 + 4.22D2 + 0.11AB + 0.2AC + 0.52AD
+ 1.19BC + 1.99BD 3.1CD

(8)

f3 = 38.2843 0.0494A + 2.4746B + 7.8434C + 7.774D

+ 0.001A2 0.00389B2 1.6988C 2 0.6556D2 0.0015AB
0.0151AC 0.0006AD 0.075BC 0.1938BD + 0.65CD
(9)

f4 = 1.02616 + 0.01316A 0.00052B 0.06845C + 0.0083D

0.00008A2 + 0.0009B2 + 0.0239C 2 0.00107D2
0.00004AB 0.00018AC + 0.00029AD 0.00302BC
0.00019BD 0.00186CD

(10)

52 A 94
8 B 12
1.5 C 3

(11)

3D5

Fig. 3. Flow of the pure and hybrid GA and PSO techniques in the NBI method.

To obtain the size distributions of the silica sand and the

grain neness number, sieve analysis tests were carried out in
Parappagoudar et al. [29]. Similarly, the authors also conducted
gelling index tests for the determination the strength of clay. Next,
experiments were conducted by varying the combination of the
parameters using the central composite design. The mathematical
model of the green mould system was developed where; the objective functions as given in Eqs. (7)(10) and the constraints as given

Author's personal copy

202

T. Ganesan et al. / Journal of Manufacturing Systems 32 (2013) 197205

Table 3
The comparison of the best solutions obtained by the algorithms.
Description

Algorithms
PSO [14]

NGA

NPSO

NHPSO

Objective function
54.9377
f1
53.679
f2
89.4473
f3
f4
1.5888

GA [14]

55.4112
107.895
84.7936
1.5079

61.5992
60.2611
89.0263
1.58366

61.3174
120.022
88.8441
1.51525

58.2195
135.478
88.3809
1.50958

Decision variable
93.9998
A
11.9999
B
C
2.6546
D
4.9998

52.0001
11.9998
2.8452
4.9999

73.3421
11.9024
2.05415
4.00906

54.5778
11.574
2.54539
4.18636

52.7462
11.9231
2.1876
3.80256

in Eq. (11). The MO optimization problem statement for the green

mould system problem is shown as follows:
Max(f1 , f2 , f3 , f4 ) subject to
52 A 94
8 B 12

(12)

1.5 C 3
3D5
The algorithms used in this work were programmed using the
C++ programming language on a personal computer (PC) with an
Intel dual core processor running at 2 GHz.
5. Results and discussion
The comparison of the best candidate solutions obtained by the
NBI-Genetic Algorithm (NGA), NBI-Particle Swarm Optimization
(NPSO) and the NBI-hybrid GA and PSO (HNPSO) methods in this
work and by the PSO and GA methods (Weighted-Sum) in Surekha
et al. [14] is shown in Table 3. The Pareto frontiers of the objectives
obtained by the NGA method are presented in Fig. 4.
The best solution candidate in Table 3 was obtained by
the NGA method at the weights (objective function trade-offs)

(1 , 2 , 3 4 ) = (0.4, 0.1, 0.4, 0.1). The best solution candidate

obtained by the NPSO method at the weights (objective function
trade-offs) (1 , 2 , 3 4 ) = (0.1, 0.4, 0.2, 0.3). In Table 3 the solution was obtained by the NPSO method at the weights (objective
function trade-offs) (1 , 2 , 3 4 ) = (0.1, 0.1, 0.5, 0.3). The Pareto
frontiers of the objectives obtained by the NGA, NPSO and NHPSO
methods are presented in Figs. 46, respectively.
It can be observed in Table 3 that the NGA and NPSO methods outperform the GA and PSO method from Surekha et al. [14].
However, a new optima is achieved by the NPSO method (see
Table 3) since it outperforms the NGA method. The NHPSO method
compromises on the objectives f1 , f3 and f4 while maximizing the
objective f2 very effectively. Thus, it can be said that the NPSO
method in this work outweighs the overall optimization capabilities of NHPSO and NGA. The computational time taken for the
algorithms in the previous work (see Surekha et al. [14]) for the GA
and PSO are, respectively, 0.021 and 0.013 s. In this work, the computational time taken for the NGA, NSPO and the NHPSO algorithms
are 282.404, 21.683 and 31.48 s respectively.
In Surekha et al. [14], the GA and the PSO method was used in
conjunction with the Weighted Sum method on an Intel Pentium
IV processor (single core). As mentioned previously, the algorithms
presented in this work: NGA, NPSO, and NHPSO were executed
on an Intel dual-core processor which is more superior than the
machine used in Surekha et al. [14]. However, it can be seen that the
computational time for NGA, NPSO and NHPSO algorithms are far
greater as compared to the GA and PSO in [14]. Although the algorithms NGA, NPSO and NHPSO are executed on a superior machine,
these algorithms seem to be computationally inferior as compared
to the GA and PSO [14] algorithms. This can be mainly attributed
to the complexity of the NBI scheme which is incorporated into the
algorithms presented in this work. The NBI scheme (see Methodology Section A) requires the application of the algorithms twice,
rst to obtain the individual optima then to solve the -sub problems for each of the scalarization. Thus, although the NGA, NPSO
and NHPSO may produce excellent results, it does compromise in
terms of computational due to the complexity of the algorithm. It is
also observed similar to the results in Surekha et al. [14], the NPSO
performs better and computationally more efcient as compared

Fig. 4. Pareto frontiers of the objectives obtained by the NGA method.

Author's personal copy

T. Ganesan et al. / Journal of Manufacturing Systems 32 (2013) 197205

203

Fig. 5. Pareto frontiers of the objectives obtained by the NPSO method.

to the NGA algorithm. Due to the incorporation of the GA segment

in the NHPSO, thus it is computationally more expensive than the
NPSO but more efcient than the NGA.
In Surekha et al. [14], the Weighted Sum method produces
a progression of Pareto efcient solutions although the spread
of solutions are not well distributed. In this work, using the NBI
method, the spread of Pareto efcient solutions are near-uniformly
spaced. The spread of the Pareto efcient solutions are vital in
MO scenarios. This is because a uniform solution spread gives a
more gradual change in the relative signicance of the objectives

in the alternative solutions. However, both methods (NBI and

Weighted Sum) do not guarantee Pareto optimality (only in the
weak sense [30]). The NGA, NPSO and NHPSO algorithms performed stable computations during the search of the individual
minima as and while solving the -sub problems. The stopping
criteria used in the algorithms used in this work was the maximum number of function evaluations (which was pre-dened
to 50). All Pareto-efcient solutions produced by the algorithms
developed in this work were feasible and no constraints were
compromised.

Fig. 6. Pareto frontiers of the objectives obtained by the NHPSO method.

Author's personal copy

204

T. Ganesan et al. / Journal of Manufacturing Systems 32 (2013) 197205

The advantages of using the NPSO algorithm as compared to the

other algorithms used in this work is that it produces highly effective results in terms of optimization of the parameters. Besides,
among the algorithms used in this work it can be said that the NPSO
has the lowest execution time. However, although the NPSO performs well relative to algorithms used in this work, it can be clearly
seen that the execution time is much higher than the one obtained
by Surekha et al. [14] using the Weighted-Sum method.
The NGA method can be said to be the second best optimizer
as compared to the NPSO method. Since the NGA method uses an
evolutionary background the diversication of the search space is
high and thus resulting in high computational time as compared
with the NPSO and NHPSO. Besides, in comparison with the NHPSO
algorithm the NGA method produces much superior results.
As for the NHPSO, this hybrid is high in terms of algorithmic
complexity and due to the GA component is performs inferior to
NPSO method in terms of computational time. However, the effectiveness of the overall optimization of all the objectives is not as
satisfactory as the NGA or the NPSO method. This method optimizes
the second objective to a very high degree while compromising on
the other objectives. Thus, it also performs poorly in terms of the
overall optimization as compared with the NGA and NPSO methods.
6. Conclusions
In this work, a new local maximum was achieved using the NPSO
method. More Pareto-efcient solution options to the green mould
system MO optimization problem were obtained. Besides, using
the NGA, NPSO and NHPSO algorithms, the solution spread of the
frontier was near-uniformly distributed. This work also produces
results of testing the green mould sand problem with a hybrid
algorithm. In the future, other meta-heuristic algorithms such as
Genetic Programming (GP) [31], Analytical Programming (AP) [32],
Hybrid Neuro-GP [33], Hybrid Neuro-PSO [34], Hybrid Tabu [35],
MO evolutionary algorithm [36,37] and Articial Immune Systems
(AIS) [38,39] should be applied in conjunction with the NBI method.
During these numerical experiments, the spacing metric should be
measured and compared for the observation the uniformity of the
spreads with respect to the algorithms. Besides, convergence and
diversity metrics should also be utilized to compare the performance of the algorithms. More large-scale MO problems should
be studied using the NGA, NPSO and NHPSO method for a better understanding of the mentioned algorithms performance and
efciency.
Acknowledgements
This work was supported by STIRF Grant (STIRF CODE NO:
90/10.11) of University Technology Petronas (UTP), Malaysia. The
authors sincerely thank the anonymous reviewers for their valuable
and constructive comments and suggestions for the improvement
of this research paper.
References
[1] Eschenauer H, Koski J, Osyczka A. Multicriteria design optimization. Berlin:
Springer-Verlag; 1990.
[2] Statnikov RB, Matusov JB. Multicriteria optimization and engineering. New
York: Chapman and Hall; 1995.
[3] Zitzler E, Thiele L. Multiobjective optimization using evolutionary
algorithmsa comparative case study. In: Eiben AE, Bck T, Schoenauer
M, Schwefel H-P, editors. Parallel problem solving from nature, V. Berlin,
Germany: Springer; 1998. p. 292301.
[4] Deb K, Pratap A, Agarwal S, Meyarivan T. A fast and elitist multi-objective
genetic algorithm: NSGA-II. IEEE Transactions on Evolutionary Computation
2002;6(2):18297.
[5] Fishburn PC. Additive utilities with incomplete product set: applications to
priorities and assignments. Baltimore, MD, USA: Operations Research Society
of America (ORSA); 1967.

[6] Triantaphyllou E. Multi-criteria decision making: a comparative study. Dordrecht, The Netherlands: Kluwer Academic Publishers (now Springer); 2000.
p. 320.
[7] Luyben ML, Floudas CA. Analyzing the interaction of design and control. I.
A multiobjective framework and application to binary distillation synthesis.
Computational Chemical Engineering 1994;18(10):18933.
[8] Das I, Dennis JE. Normal-boundary intersection: a new method for generating the Pareto surface in nonlinear multicriteria optimization problems. SIAM
Journal of Optimization 1998;8(3):63157.
[9] Van Sickel JH, Venkatesh P, Lee KY. Analysis of the Pareto front of a multiobjective optimization problem for a fossil fuel power plant. In: IEEE PES
general meeting. 2008.
[10] Kim IY, de Weck OL. Adaptive weight-sum method for biobjective optimization:
Pareto front generation. In: AIAA structures, structural, dynamics, and materials
conference. 2002.
[11] Sendn JOH, Alonso AA, Banga JR. Efcient and robust multi-objective optimization of food processing: a novel approach with application to thermal
sterilization. Journal of Food Engineering 2010;98:31724.
[12] Sandgren E. Multicriteria design optimization by goal programming. In: Adeli
H, editor. Advances in design optimization. London: Chapman & Hall; 1994. p.
22565 [Chapter 23].
[13] Stanikov RB, Matusov JB. Multicriteria optimization and engineering. New
York: Chapman and Hall; 1995.
[14] Surekha B, Kaushik LK, Panduy AK, Vundavilli PR, Parappagoudar MB.
Multi-objective optimization of green sand mould system using evolutionary algorithms. International Journal of Advance Manufacturing Technology
2011;58(14):19.
[15] Kumar S, Satsangi PS, Prajapati DR. Optimization of green sand casting process parameters of a foundry by using Taguchi method. International Journal of Advance Manufacturing Technology 2010;55(14):
2334.
[16] Rosenberg RS. Simulation of genetic populations with biochemical properties.
Ph.D. thesis, University of Michigan; 1967.
[17] Holland JH. Adaptation in natural and articial systems: an introductory analysis with applications to biology. In: Control and articial intelligence. USA: MIT
Press; 1992.
[18] Kirkpatrick S, Gelatt CD, Vecchi MP. Optimization by simulated annealing. Science 1983;220:67180.
[19] Li CL. A feature-based approach to injection mould cooling system design.
Computer-Aided Design 2001;33(14):107390.
[20] Li CL. Automatic synthesis of cooling system design for plastic injection mould.
In: ASME 2001 design engineering technical conferences and computers and
information in engineering conference, 27th design automation conference.
2001.
[21] Gen M, Cheng R. Genetic algorithms and engineering optimization. John Wiley
& Sons; 1997, 432 pp.
[22] Kennedy J, Eberhart R. Particle swarm optimization. In: IEEE proceedings of the
international conference on neural networks. 1995. p. 19428.
[23] Phuangpornpitak N, Prommee W, Tia S, Phuangpornpitak W. A study of particle swarm technique for renewable energy power systems. In: PEA-AIT
international conference on energy and sustainable development: issues and
strategies, Thailand. 2010. p. 17.
[24] Chryssolouris G, Lee M, Domroese M. The use of neural networks in determining operational policies for manufacturing systems. Journal of Manufacturing
Systems 1991;10(2):16675.
[25] Williams NR. Neural networks vs. regression in localized scan correction for robotic drilling applications. Journal of Manufacturing Systems
2005;24(1):2134.
[26] El Maraghy H, Patel V, Abdallah IB. Scheduling of manufacturing systems under
dual-resource constraints using genetic algorithms. Journal of Manufacturing
Systems 1991;10(2):16675.
[27] Lee TS, Ting TO, Lin YJ. An investigation of grinding process optimization via
evolutionary algorithms. In: Proceedings of the 2007 IEEE swarm intelligence
symposium. 2007.
[28] Sardinas RQ, Santana MR, Brindis EA. Genetic algorithm based multi-objective
optimization of cutting parameters in turning process. Engineering Applied
Articial Intelligence 2006;19:12733.
[29] Parappagoudar MB, Pratihar DK, Datta GL. Non-linear modeling using central
composite design to predict green sand mould properties. Proceedings IMechE
B: Journal of Engineering Manufacture 2007;221:88194.
[30] Shukla PK. On the normal boundary intersection method for generation of efcient front. In: Shi Y, et al., editors. ICCS 2007, Part I. LNCS, vol. 4487. Berlin
Heidelberg: Springer-Verlag; 2007. p. 3107.
[31] Koza JR. Genetic programming: on the programming of computers by means
of natural selection. USA: MIT Press; 1992.
[32] Zelinka I. Analytic programming by means of soma algorithm. Mendel02.
In: Proc 8th international conference on soft computing Mendel02. 2002. p.
93101. ISBN 80-214-2135-5.
[33] Ganesan T, Vasant P, Elamvazuthi I. Optimization of nonlinear geological
structure mapping using hybrid neuro-genetic techniques. Mathematical and
Computer Modelling 2011;54(11):291322.
[34] Ganesan T, Vasant P, Elamvazuthi I. Hybrid neuro-swarm optimization
approach for design of distributed generation power systems. Neural Computing and Applications 2012, http://dx.doi.org/10.1007/s00521-012-0976-4.
[35] Vasant P, Ganesan T, Elamvazuthi I. Hybrid Tabu search hopeld recurrent ANN
fuzzy technique to the production planning problems: a case study of crude

Author's personal copy

T. Ganesan et al. / Journal of Manufacturing Systems 32 (2013) 197205
oil in renery industry. International Journal of Manufacturing, Materials, and
Mechanical Engineering 2012;2(1):4765.
[36] Qu BY, Suganthan PN. Multi-objective evolutionary algorithms based on the
summation of normalized objectives and diversied selection. Information
Sciences 2010;180:317081.
[37] Li K, Kwong S, Cao J, Li M, Zheng J, Shen R. Achieving balance between proximity
and diversity in multi-objective evolutionary algorithm. Information Sciences
2011;182:22042.

205

[38] De Castro LN, Timmis J. Articial immune systems: a new computational intelligence approach. Springer; 2002, 380 pp.
[39] Yang D, Jiao L, Gong M, Liu F. Articial immune multi-objective SAR image
segmentation with fused complementary features. Information Sciences
2011;181:2797812.