You are on page 1of 10

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 Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright

com/locate/jmansys Technical Paper Normal-boundary intersection based parametric multi-objective optimization of green sand mould system T.12.Author's personal copy Journal of Manufacturing Systems 32 (2013) 197–205 Contents lists available at SciVerse ScienceDirect Journal of Manufacturing Systems journal homepage: www. 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.10]. the DM utilizes some supplementary knowledge which is usually very heuristic and too complex to be represented mathematically [11]. Malaysia b a r t i c l e i n f o a b s t r a c t In manufacturing engineering optimization. In this work. This problem was presented and solved in Surekha et al. alternative near-optimal solution options are generated for various values of the scalars. the objective functions in the problem are aggregated into a single objective function which is then solved for various scalar (weight) values. the scalars (or weights) are used to consign relative trade-offs to the objectives during the aggregation procedure. I. Goal Programming [7] and Normal-Boundary Intersection (NBI) method [8]. Some known scalarization techniques include the Weighted Sum method [5. University Technology Petronas. hardness and bulk density requirements where as sand mould has the same properties without the green constraints.2]. Malaysia Fundamental & Applied Sciences Department. Nevertheless ∗ Corresponding author. Using these techniques. [4]). Published by Elsevier Ltd.jmsy. This method then is used generate optimal solution options to the green sand mould system problem. the decision maker (DM) is only interested in a single optimal solution.doi.1016/j. Published by Elsevier Ltd. In these methods. Malaysia c Electrical & Electronic Engineering Department. See [1. it is often that one encounters scenarios that are multiobjective (where each of the objectives portray different aspects of the problem). University Technology Petronas. it is crucial for the engineer to have access to multiple solution choices before selecting of the best solution. All rights reserved.004 . a novel approach that merges meta-heuristic algorithms with the Normal Boundary Intersection (NBI) method is introduced. In green mould systems. P.∗ . The other type of methods is known as the weighted (or scalarization) techniques.2012. A unique and ideal solution that explains all the features of a MO problem in engineering are rarely encountered [9. it is very useful for the DM to have access to numerous solution options with a variety of significance with respect to the objectives prior to the selection the best optimal solution. Elamvazuthi c a Chemical Engineering Department. Vasant b. in more practical scenarios. The difference between sand mould and green sand mould is that green sand mould has green compression strength. University Technology Petronas. it is required that the solution method caters for the multiobjective nature of the problem. To select this unique optimum. Thus. Vasant). Some comparative studies were then carried out with the algorithms developed and used in this work and that from some previous work. © 2012 The Society of Manufacturing Engineers.10. Analysis on the performance as well as the quality of the solutions produced by the algorithms is presented here.6]. Strategies in multi-objective optimization (MO) can be crudely classified into two classes. 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 1.13] for more detail investigations and explanations on MO techniques in engineering optimization. http://dx. Thus. E-mail address: vasant0001@yahoo. Ganesan a .com (P. In optimization problems of this kind. [14] by the application of genetic algorithm (GA) and Particle Swarm Optimization (PSO) techniques in conjunction with the Weighted Sum approach. Hence. Introduction Most issues encountered when dealing with emerging technologies in engineering are multi-objective (MO) in nature [1. 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. 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. All rights reserved. First being methods that use the concept of Pareto optimality to trace the non-dominated solutions at the Pareto curve (for instance. permeability. 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. Zitzler and Thiele’s [3] Strength Pareto Evolutionary Algorithm (SPEA) and Non-dominated Sorting Genetic Algorithm II (NSGA-II) by Deb et al.elsevier.org/10. Therefore.

The procedures of which this method is executed are as follows: . the velocity and the position of the particles’ are updated. In the beginning. Thus. A more comprehensive study on the optimization and model formulation of mould systems can be seen in [15. 1 ≤ x ≤ 4} ∗. The green mould system problem is presented as the following: Min F (x) subject to X = {x : g (x) = 0. and n is the normal vector at the point towards the utopia point. [14]. This paper is organized as follows. f ∗. The candidate optimal solutions obtained by this algorithm are achieved as a result of particles which are in motion (swarming) through the fitness landscape. The convex minimum be denoted as xi hull of the individual minima is generated in this fashion. f ∗) F ∗ = (f1 2 3 4 T (2) where F* is the utopia point for this MO problem. The first concept was based on the examination of swarming (or flocking) behaviours of certain species of organisms (such as birds.Author's personal copy 198 T.t ) t subject to x∈X (4) · Y + tn = F (x) and where t is some defined distance parameter. GAs operates on a population of solutions. Lately. 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. Next. / Journal of Manufacturing Systems 32 (2013) 197–205 moulding sand (such as hardness. the scalarization is carried out. In this work. vi (t) is the particle velocity and xi (t) is the particle position with respect to iteration (t). then the algorithm updates the individual and social terms by the aid of a user-specified update rule. the evaluation of the particles’ position and velocity (which are also the candidate solutions) relative to the fitness function is carried out. Incorrect proportions of the mentioned properties may lead to casting defects such as poor surface finish. the green mould sand system is optimized further using genetic algorithms (GA).28]. 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. Hence. It is essential to note that the velocity and position updating rule is critical to the optimization capabilities of this method. 2. ants. The velocity of each particle in motion (swarming) is updated using the following equation. permeability. 4]. in conjunction with the fitness function a condition is introduced. Controllable variables such as percentage of water. the concluding remarks are given in Section 6. green compression strength and bulk density). the standard meta-heuristic techniques are presented. Particle Swarm Optimization (PSO) is an optimization method developed based on the movement and intelligence of swarms. The second idea was sprung from the study of evolutionary computations. GAs are categorized as a class of population-based search and optimization algorithms [27. [14]. c1 .1. Genetic algorithms (GA) were introduced by Holland in the nineties [17]. Comparison studies were then performed on the optimal solutions obtained in this work against those obtained in Surekha et al. This makes the NBI approach a more interesting alternative as compared to the Weighted Sum method when solving non-convex MO problem. GAs have been widely applied in engineering scenario (see [19. percentage of clay. and this is followed by description on the Scalarisation Technique and Proposed Algorithms in Section 3. An N-point crossover operator was used to create new offspring for each successive generation.1. For a more comprehensive text on GAs refer to [21].2. The parameters w. where if the fitness function is not fulfilled. Scalarisation Technique and Proposed Algorithms 3. While most stochastic search techniques operate on a distinct solution for a particular problem. In contrast to the Weighted Sum method. 2. GAs belong to the group of stochastic search methods (such as simulated annealing [18] and some forms of branch and bound).16]. vi (t + 1) = wvi (t ) + c1 r1 [ˆ xi (t ) − xi (t )] + c2 r2 [g (t ) − xi (t )] (1) where each particle is identified by the index i. These solutions can be randomly selected or be established with the aid of some a priori facts. 3. Y are varied thoroughly to generate a near-uniform spread of the Pareto frontier. Let the individual ∗ and be obtained for i ∈ [1. Standard meta-heuristic techniques 2. scabs and pinhole porosity. by characterizing these parameters as the decision variables and the mould sand properties as the objective function. bees and fireflies). f ∗. Many works have done on the application of meta-heuristic techniques for modelling and optimization of manufacturing systems [24–26]. The PSO algorithm searches the search space for candidate solutions and evaluates these solutions with respect to some (user specified) fitness condition. Section 5 discusses computational results and finally. the MO optimization problem was formulated in Surekha et al. The formulation of where forms a 4 by 4 matrix and i=1 i the NBI ␤-sub problem is as the following: Max(X. The scalars. Y = ˇi : 1 ≤ x ≤ 4} 4 (3) ˇ = 1. some candidate solutions are selected by the PSO algorithm. r1 and r2 are usually defined by the user. Scalarisation technique: Normal Boundary Intersection (NBI) method The NBI method was first introduced by Das and Dennis [8]. The purpose of this formulation is for the identification of best controllable parameters for optimal final-product of the moulding process. c2 .20]). PSO has been applied to a variety of areas including optimization problems in engineering [23] as well as economic dispatch problems. The real world application problem on green sand mould system is illustrated in Section 4. This method is a geometrically inspired scalarization approach for solving MO problems. To avoid the solution from getting stagnant at the local minima. In recent times. PSO was developed by Kennedy and Eberhart [22] in 1995. The NBI scalarization method finds the maximum distance. an N-bit flip mutation operator was used. This recursive course of action is iterated until the fitness function is satisfied by all candidate solutions and solutions have thus converged into a fix position. Consequently. Following this. blowholes. Particle Swarm Optimization (PSO) The PSO algorithm introduced in 1995 (by Kennedy and Eberhart [22]) springs from two distinct frames of ideas. the NBI approach has the ability to find a near-uniform spread of Pareto-optimal solution options in the frontier. In Section 2 of this paper. t in the direction of the normal vector. Next. h(x) ≤ 0. Ganesan et al. the representation of the simplex from the convex hull is as follows: ={ ·Y : ∗ = F (xi ). grain fineness number and number of strokes heavily influence the physical properties of the moulded sand. n between a point on the simplex and the origin (or the utopia point).

• The optimal value of the solutions is then selected from the obtained Pareto front.2 0 ≤ c1 ≤ 2 0 ≤ c2 ≤ 2 0 ≤ r1 ≤ 1 0 ≤ r2 ≤ 1 The term wvi (t) in Eq. the particles’ position is then computed as is shown in the following: xi (t + 1) = xi (t ) + vi (t + 1) (6) (5) The GA scheme applied in this work is as the following. (1) (also referred to as the inertial term) maintains the particle’s motion in the same direction as its original vector. Algorithm flow for GA. Algorithm 1. of individuals in the population Probability of mutation Probability of recombination Initial string of individuals Bit type of individual’s string Cross-over type Mutation type Selection type Values 6 bit 6 0. Step 7: Set new population to current population. Next. 1. Step 8: Assess the fitness of each offspring in the generation. The flowchart of the GA algorithm is shown in Fig. Step 9: If the stopping criterion are satisfied halt program and print solutions. create offspring for the next generation. Thus. • The inverse transformation is then performed to re-obtain the best maximal values of the objective functions. Step 5: The parent selection to create the next generation is done by tournament selection. 1. Offspring Mutation NO Set new population to current population Evaluate offspring fitness Termination criterion satisfied? YES Perform parent selection STOP Fig. Algorithm 2. The term c1 r1 [ˆ xi (t ) − xi (t )] better known as the cognitive component functions serves as the memory. Generate offspring by crossover/recombination Randomly initialize a population of n individuals. 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.5 Random Real-coded N-point N-bit flip Tournament 199 Step 3: By recombination from the current population. 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 fitness function. Genetic algorithm (GA) These parameter settings for this algorithm are usually constrained as the following: 0 ≤ w ≤ 1. Until all candidate solutions are at their highest fitness positions and the termination criterion is satisfied. the single-objective ˇ-sub problem is solved using GA. Particle Swarm Optimization (PSO) • The MO maximization problem is reformulated as a minimization problem by inverting the objective functions. • Obtain the local minima of individual objective functions. Parameters Length of individual string No. • The Pareto front for multiple scalarization values of ˇ is obtained.Author's personal copy T. The inertial coefficient w serves as a dampener or an accelerator during the movement of the particles’. The parameter settings initialized prior to the execution of the GA used in this work are shown in Table 1. PSO or the hybrid GA-PSO where these algorithms are termed as NBIPSO (NPSO). • The multi-objective (MO) problem is then reformulated as a single-objective ˇ-sub problem by using the value of the local optimums. else go to Step 3. Step 1: Initialize a random chromosome for n individuals in the population. . / Journal of Manufacturing Systems 32 (2013) 197–205 Table 1 Genetic algorithm (GA) settings. Ganesan et al. Step 4: Mutate offspring for this generation. Step 2: Assign fitness conditions to each of the n individuals in the population.3333 0. these iterations are then START Fitness criterions are assigned to each of the n individuals. Step 6: The next population of n individuals is chosen. NBIGA (NGA) and NBI Hybrid GA-PSO (NHPSO). • The single-objective ˇ-sub problem is then solved.

n is the swarm iteration and T is the overall program iteration. / Journal of Manufacturing Systems 32 (2013) 197–205 sustained. stop program and print solutions. Algorithm flow for PSO. else proceed to Step 6 Step 6: Proceed with the evaluation of the fitness of each particle in the swarm. 2. n > no + T. in the event during this iterative process the position of all the particles converges. The initialization parameters for this algorithm is shown in Table 2 and the workflow is provided in Fig. However. update position xi and velocity vi and go to Step 3. i Initialize algorithm parameters Randomly initialize position xi(n) and velocity vi(n) n = n +1 Compute inertial and social influence T = T +1 Compute position xi(n+1) and velocity vi(n+1) at next iteration NO Is fitness criterion satisfied? YES Evaluate fitness of the swarms YES Is n > no+T ? NO STOP Fig. START Initialize no of particles. r2 . i and the initialize parameter settings w. 2. the solutions are feasible with respect to the specified 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 fitness criterion are met. Step 7: If the fitness conditions are satisfied. The algorithm of the PSO method used in this work is shown as follows: Step 1: Set number of particles. r1 . Hence the candidate solutions are at their fittest and the program is stopped and the solutions are printed. . 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. Ganesan et al. where no is some constant. c2 . else go to Step 3. c1 .Author's personal copy 200 T.

Step 9: If the stopping criterion are satisfied halt program and print solutions.5.00302BC − 0. r2 . 1. Step 10: Set no of particles.033.66D + 0. sieve analysis tests were carried out in Parappagoudar et al.00389B2 − 1.5 ≤ C ≤ 3 3≤D≤5 To obtain the size distributions of the silica sand and the grain fineness number.45B2 − 4.00008A2 + 0.6556D2 − 0. 3 shows the mechanism of placement of the GA and PSO algorithms in the NBI sections. The objective functions and the range of the decision variables are shown as follows: f1 = 17.1CD (8) f3 = 38. r1 . s2 . 1. permeability (f2 ).51 − 15. p6 ) Values (1.Author's personal copy T. update position xi and velocity vi and go to Step 12.0151AC − 0. Ganesan et al. w) Number of particles Initial social influence (s1 . Next. 0.05.001A2 − 0.0006AD − 0.00107D2 − 0.00019BD − 0.2079D2 + 0.025. As mentioned in the introduction section.0009B2 + 0.0451AD + 0.014A2 + 0.1215AC − 0.6378BD + 2. / Journal of Manufacturing Systems 32 (2013) 197–205 Table 2 Particle Swarm Optimization (PSO) settings. Secondly. The parameter settings in the hybrid method are identical to the settings employed in the pure methods. else go to Step 3. 4. else go to Step 12. [29]. In Surekha et al.19BC + 1. hardness (f3 ) and bulk density (f4 ). Application data Algorithm 3. green compression strength (f1 ). the objective functions as given in Eqs. Step 4: Mutate offspring for this generation.00052B − 0.2527 − 1. The responses are. Hybrid GA-PSO The responses of the mould heavily influence the quality of the final product of the green sand mould system. The hybridization procedure in Fig.51C − 105. meta-heuristic techniques (such as GA and PSO) can be incorporated into the NBI approach at two segments. the grain fineness number (A).07A2 + 0. Step 16: If fitness criterion satisfied. [14]. Flow of the pure and hybrid GA and PSO techniques in the NBI method.4746B + 7. Step 2: Assign fitness conditions to each of the n individuals in the population.5.22D2 + 0. r1 . i and the initialize parameter settings w. The algorithm for the hybrid GA-PSO approach is as the following: Step 1: Initialize a random chromosome for n individuals in the population. experiments were conducted by varying the combination of the parameters using the central composite design. Parameters Initial parameter (c1 . Step 8: Assess the fitness of each offspring in the generation. 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 f2 = 1192. c2 . else proceed to Step 6 Step 15: Evaluate fitness swarm based on the objective function of the ˇ-sub problem for each scalarization.0945B2 − 7.075BC − 0. s4 .99BD − 3.2843 − 0. s5 . 1. In contrast to the pure GA and PSO approach (where both segments in the NBI approach are embedded with the PSO or GA technique). These objectives on the other hand are influenced by on the process (or decision) variables which are.5516BC + 0.7463B + 32. Step 7: Set new population = current population.3203C + 6. using these techniques for solving the ˇ-sub problem by varying the weights. p4 .00186CD (10) 52 ≤ A ≤ 94 8 ≤ B ≤ 12 1. c1 .0015AB − 0. .1. these responses are represented mathematically as the objective functions. 5.66B + 9. the hybrid approach uses GA to find the local minima and PSO for solving the ˇ-sub problem. n > no + T.01316A − 0.0468AB − 0. 1.689CD) (7) The hybrid GA-PSO in this work was developed by using both techniques at different sections of the NBI approach. (7)–(10) and the constraints as given (11) Fig.0239C 2 − 0. Step 3: By recombination from the current population. 3.575D + 0. halt and print solutions.7857C 2 − 1.0494A + 2. First being the use of these techniques as a tool to search for the local minima of the individual objective functions. 1. 1. create offspring for the next generation.6988C 2 − 0. p2 . p3 . The mathematical model of the green mould system was developed where. 0.13C 2 + 4.00004AB − 0. Step 5: The parent selection to create the next generation is done by tournament selection.11AB + 0. c2 .65CD (9) f4 = 1.774D + 0.2. s6 ) Initial personal influence (p1 .06845C + 0. Similarly.00029AD − 0. 4.8434C + 7.7384A − 2.2AC + 0. the authors also conducted gelling index tests for the determination the strength of clay.1938BD + 0.98A − 35. s3 .8) 6 (1. 6. percentage of clay content (B). Step 6: The next population of n individuals is chosen. p5 .017) (3.02.0083D − 0.02616 + 0. r2 . percentage of water content (C) and number of strokes (D). 0.00018AC + 0. 7. 8) 201 Step 14: If the swarm evolution time.52AD + 1.

the algorithms presented in this work: NGA. first to obtain the individual optima then to solve the ˇ-sub problems for each of the scalarization.574 2. although the NGA. 0. 4–6.4. 4. Ganesan et al.5992 60. / Journal of Manufacturing Systems 32 (2013) 197–205 Table 3 The comparison of the best solutions obtained by the algorithms.9998 2. [14].4. respectively.3).9024 2. NPSO. The Pareto frontiers of the objectives obtained by the NGA method are presented in Fig.5778 11. NPSO and NHPSO may produce excellent results.3421 11.5888 Decision variable 93. ˇ2 .9231 2.51525 54.3174 120. Although the algorithms NGA.9999 B C 2.1). 0. a new optima is achieved by the NPSO method (see Table 3) since it outperforms the NGA method.9998 PSO [14] 55. Results and discussion The comparison of the best candidate solutions obtained by the NBI-Genetic Algorithm (NGA).6546 D 4.9998 A 11. [14]) for the GA and PSO are. However. ˇ2 .50958 52.021 and 0. ˇ2 .1.1. 4. 0.4473 f3 f4 1. [14]. The NHPSO method compromises on the objectives f1 .54539 4.5 ≤ C ≤ 3 3≤D≤5 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. f3 .9999 NGA 61.478 88. the NPSO performs better and computationally more efficient as compared Fig. [14]. As mentioned previously. it does compromise in terms of computational due to the complexity of the algorithm.5. these algorithms seem to be computationally inferior as compared to the GA and PSO [14] algorithms.013 s.679 f2 89. the computational time taken for the NGA. [14] is shown in Table 3.022 88. The Pareto frontiers of the objectives obtained by the NGA. and NHPSO were executed on an Intel dual-core processor which is more superior than the machine used in Surekha et al.2195 135. f4 ) subject to 52 ≤ A ≤ 94 8 ≤ B ≤ 12 1.683 and 31. it can be said that the NPSO method in this work outweighs the overall optimization capabilities of NHPSO and NGA. NPSO and NHPSO are executed on a superior machine. ˇ3 ˇ4 ) = (0.80256 in Eq.8441 1. This can be mainly attributed to the complexity of the NBI scheme which is incorporated into the algorithms presented in this work.00906 NPSO 61.2611 89. It can be observed in Table 3 that the NGA and NPSO methods outperform the GA and PSO method from Surekha et al. ˇ3 ˇ4 ) = (0. 21.Author's personal copy 202 T.4112 107. The computational time taken for the algorithms in the previous work (see Surekha et al. 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.4. ˇ3 ˇ4 ) = (0. The best solution candidate in Table 3 was obtained by the NGA method at the weights (objective function trade-offs) (12) (ˇ1 .0001 11. NPSO and NHPSO algorithms are far greater as compared to the GA and PSO in [14]. the GA and the PSO method was used in conjunction with the Weighted Sum method on an Intel Pentium IV processor (single core).1876 3. 0.48 s respectively. it can be seen that the computational time for NGA.2. 0.0263 1. 0. Description Algorithms GA [14] Objective function 54. It is also observed similar to the results in Surekha et al. In Table 3 the solution was obtained by the NPSO method at the weights (objective function trade-offs) (ˇ1 . In this work.9377 f1 53. f2 . 0. The best solution candidate obtained by the NPSO method at the weights (objective function trade-offs) (ˇ1 .5079 52. In Surekha et al.58366 73. 0. The NBI scheme (see Methodology – Section A) requires the application of the algorithms twice.7936 1.1. The MO optimization problem statement for the green mould system problem is shown as follows: Max(f1 . Thus. Pareto frontiers of the objectives obtained by the NGA method. NSPO and the NHPSO algorithms are 282.1. f3 and f4 while maximizing the objective f2 very effectively.895 84. 0.7462 11.404. 0. 5. However.3). Thus. (11).18636 NHPSO 58.8452 4. [14]. NPSO and NHPSO methods are presented in Figs. .3809 1. respectively.05415 4.

Fig. [14]. Pareto frontiers of the objectives obtained by the NPSO method. Due to the incorporation of the GA segment in the NHPSO. . / Journal of Manufacturing Systems 32 (2013) 197–205 203 Fig. This is because a uniform solution spread gives a more gradual change in the relative significance of the objectives in the alternative solutions. both methods (NBI and Weighted Sum) do not guarantee Pareto optimality (only in the weak sense [30]). In this work. 6. The NGA. In Surekha et al. However. Pareto frontiers of the objectives obtained by the NHPSO method. the Weighted Sum method produces a progression of Pareto efficient solutions although the spread of solutions are not well distributed. using the NBI method. the spread of Pareto efficient solutions are near-uniformly spaced. 5. Ganesan et al.Author's personal copy T. All Pareto-efficient solutions produced by the algorithms developed in this work were feasible and no constraints were compromised. The spread of the Pareto efficient solutions are vital in MO scenarios. to the NGA algorithm. thus it is computationally more expensive than the NPSO but more efficient than the NGA. The stopping criteria used in the algorithms used in this work was the maximum number of function evaluations (which was pre-defined to 50). NPSO and NHPSO algorithms performed stable computations during the search of the individual minima as and while solving the ˇ-sub problems.

Lin YJ. 1998. A multiobjective framework and application to binary distillation synthesis. p. p. USA: Operations Research Society of America (ORSA). Vasant P. Normal-boundary intersection: a new method for generating the Pareto surface in nonlinear multicriteria optimization problems. Lee KY. Parappagoudar MB.18(10):18–933.1007/s00521-012-0976-4. 320. 2002. John Wiley & Sons. 2000. 6. in comparison with the NHPSO algorithm the NGA method produces much superior results. other meta-heuristic algorithms such as Genetic Programming (GP) [31]. A feature-based approach to injection mould cooling system design. [9] Van Sickel JH. Journal of Manufacturing Systems 2005. Additive utilities with incomplete product set: applications to priorities and assignments. This method optimizes the second objective to a very high degree while compromising on the other objectives. In: AIAA structures. More Pareto-efficient solution options to the green mould system MO optimization problem were obtained. Ganesan et al. editor. editors. Engineering Applied Artificial Intelligence 2006. . [8] Das I. Genetic algorithms and engineering optimization. Analytic programming by means of soma algorithm. 2007. Germany: Springer. In: Adeli H. [35] Vasant P. 1995. it also performs poorly in terms of the overall optimization as compared with the NGA and NPSO methods. MD. 1995. the effectiveness of the overall optimization of all the objectives is not as satisfactory as the NGA or the NPSO method. Tia S. In: PEA-AIT international conference on energy and sustainable development: issues and strategies. p. Besides. Gelatt CD. Thus. Multicriteria optimization and engineering. / Journal of Manufacturing Systems 32 (2013) 197–205 [6] Triantaphyllou E. In: IEEE PES general meeting. University of Michigan. regression in localized scan correction for robotic drilling applications. Floudas CA. A study of particle swarm technique for renewable energy power systems. Analysis of the Pareto front of a multiobjective optimization problem for a fossil fuel power plant. Berlin Heidelberg: Springer-Verlag. In: ASME 2001 design engineering technical conferences and computers and information in engineering conference. Computational Chemical Engineering 1994. [18] Kirkpatrick S. http://dx. [21] Gen M. Thiele L.org/10. Malaysia. Matusov JB. Automatic synthesis of cooling system design for plastic injection mould. Meyarivan T. Science 1983.39] should be applied in conjunction with the NBI method. Optimization of green sand casting process parameters of a foundry by using Taguchi method. ISBN 80-214-2135-5. Neural networks vs. vol. p. International Journal of Advance Manufacturing Technology 2011. 1967. More large-scale MO problems should be studied using the NGA. Patel V. although the NPSO performs well relative to algorithms used in this work. International Journal of Advance Manufacturing Technology 2010. Banga JR. Advances in design optimization. USA: MIT Press.doi. SIAM Journal of Optimization 1998. 1942–8. Baltimore. among the algorithms used in this work it can be said that the NPSO has the lowest execution time. [33] Ganesan T. Part I. Phuangpornpitak W. dynamics. In: Shi Y. [19] Li CL. Lee M. 1992. [11] Sendín JOH. [15] Kumar S. In: IEEE proceedings of the international conference on neural networks. Optimization by simulated annealing. it can be clearly seen that the execution time is much higher than the one obtained by Surekha et al.D. Parallel problem solving from nature. Particle swarm optimization. [30] Shukla PK. Multicriteria design optimization. New York: Chapman and Hall. p. In: Proc 8th international conference on soft computing Mendel’02. [17] Holland JH. p. Brindis EA.98:317–24. Dennis JE. Hybrid Neuro-GP [33]. LNCS. convergence and diversity metrics should also be utilized to compare the performance of the algorithms. [2] Statnikov RB. Multi-criteria decision making: a comparative study. Hybrid neuro-swarm optimization approach for design of distributed generation power systems. [20] Li CL. Vecchi MP. Schwefel H-P. As for the NHPSO. In: Eiben AE. [31] Koza JR. the spacing metric should be measured and compared for the observation the uniformity of the spreads with respect to the algorithms. Multi-objective optimization of green sand mould system using evolutionary algorithms. Pratap A. MO evolutionary algorithm [36. Mathematical and Computer Modelling 2011. [10] Kim IY. Panduy AK. The NGA method can be said to be the second best optimizer as compared to the NPSO method. [4] Deb K.6(2):182–97. Efficient and robust multi-objective optimization of food processing: a novel approach with application to thermal sterilization. 27th design automation conference. [7] Luyben ML. Non-linear modeling using central composite design to predict green sand mould properties. Dordrecht. [23] Phuangpornpitak N.10(2):166–75. Elamvazuthi I. Multicriteria optimization and engineering. Abdallah IB. The Netherlands: Kluwer Academic Publishers (now Springer). Computer-Aided Design 2001. ICCS 2007. Vundavilli PR. Osyczka A. [14] Surekha B. Thailand. editors. References [1] Eschenauer H. Simulation of genetic populations with biochemical properties. A fast and elitist multi-objective genetic algorithm: NSGA-II. 1992.33(14):1073–90. Genetic programming: on the programming of computers by means of natural selection. Datta GL. Analytical Programming (AP) [32].10(2):166–75.37] and Artificial Immune Systems (AIS) [38. Berlin: Springer-Verlag. Genetic algorithm based multi-objective optimization of cutting parameters in turning process. Optimization of nonlinear geological structure mapping using hybrid neuro-genetic techniques. Kaushik LK. Analyzing the interaction of design and control. However. [27] Lee TS. However. [29] Parappagoudar MB.220:671–80. Cheng R. Besides.8(3):631–57. 1–7. p. Agarwal S. a new local maximum was achieved using the NPSO method. Prommee W. thesis. Since the NGA method uses an evolutionary background the diversification of the search space is high and thus resulting in high computational time as compared with the NPSO and NHPSO. [13] Stanikov RB. An investigation of grinding process optimization via evolutionary algorithms.11) of University Technology Petronas (UTP). Multiobjective optimization using evolutionary algorithms—a comparative case study. [16] Rosenberg RS. 1967. V. The authors sincerely thank the anonymous reviewers for their valuable and constructive comments and suggestions for the improvement of this research paper. Hybrid Tabu search hopfield recurrent ANN fuzzy technique to the production planning problems: a case study of crude 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.24(1):21–34. Neural Computing and Applications 2012. Elamvazuthi I.. 1990. NPSO and NHPSO method for a better understanding of the mentioned algorithm’s performance and efficiency.58(1–4):1–9. 310–7. Scheduling of manufacturing systems under dual-resource constraints using genetic algorithms. the solution spread of the frontier was near-uniformly distributed. Bäck T. 1994. 2001. Vasant P. Ph. New York: Chapman and Hall. de Weck OL. Multicriteria design optimization by goal programming. In: Proceedings of the 2007 IEEE swarm intelligence symposium. Mendel’02. [28] Sardinas RQ.Author's personal copy 204 T.19:127–33. Adaptation in natural and artificial systems: an introductory analysis with applications to biology. [25] Williams NR. [26] El Maraghy H. IEEE Transactions on Evolutionary Computation 2002. Prajapati DR. Proceedings IMechE B: Journal of Engineering Manufacture 2007. USA: MIT Press. [32] Zelinka I. [12] Sandgren E. et al. Journal of Food Engineering 2010.221:881–94. Domroese M. Conclusions In this work. [5] Fishburn PC. 2002. 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. Matusov JB. [22] Kennedy J. and materials conference. Berlin. Elamvazuthi I. Schoenauer M. Venkatesh P. Santana MR. In: Control and artificial intelligence. NPSO and NHPSO algorithms.54(11):2913–22. Adaptive weight-sum method for biobjective optimization: Pareto front generation. Hybrid Neuro-PSO [34].55(1–4): 23–34. Acknowledgements This work was supported by STIRF Grant (STIRF CODE NO: 90/10. Pratihar DK. Journal of Manufacturing Systems 1991. [14] using the Weighted-Sum method. [24] Chryssolouris G. 2007. Journal of Manufacturing Systems 1991. Besides. London: Chapman & Hall. During these numerical experiments. Besides. 1997. 432 pp. On the normal boundary intersection method for generation of efficient front. 2010. Eberhart R. 2008. I. Ganesan T. This work also produces results of testing the green mould sand problem with a hybrid algorithm. using the NGA. 292–301. The use of neural networks in determining operational policies for manufacturing systems. 4487. [3] Zitzler E. 1995. Hybrid Tabu [35]. 93–101. Koski J. In the future. 225–65 [Chapter 23]. Ting TO. Alonso AA. [34] Ganesan T. Satsangi PS. structural.

. Gong M. Liu F. Cao J. [36] Qu BY. Timmis J. Kwong S.181:2797–812. [37] Li K.2(1):47–65. Information Sciences 2010. Ganesan et al. Jiao L. Achieving balance between proximity and diversity in multi-objective evolutionary algorithm. International Journal of Manufacturing. Suganthan PN. / Journal of Manufacturing Systems 32 (2013) 197–205 oil in refinery industry. Li M. Springer.182:220–42. and Mechanical Engineering 2012. Materials. [39] Yang D. 205 [38] De Castro LN.180:3170–81. Zheng J. Information Sciences 2011.Author's personal copy T. Artificial immune systems: a new computational intelligence approach. Shen R. 2002. Information Sciences 2011. Multi-objective evolutionary algorithms based on the summation of normalized objectives and diversified selection. 380 pp. Artificial immune multi-objective SAR image segmentation with fused complementary features.