See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/322666619

A multi-objective genetic algorithm for design optimisation of simple and

double harmonic motion cams

Article in International Journal of Design Engineering · January 2017

DOI: 10.1504/IJDE.2017.10010608

CITATIONS READS

0 325

2 authors:

R. K. Jana Prasun Bhattacharjee

Indian Institute of Management Raipur Jadavpur University
79 PUBLICATIONS 484 CITATIONS 3 PUBLICATIONS 3 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Multi objective optimisation View project

All content following this page was uploaded by R. K. Jana on 04 September 2020.

The user has requested enhancement of the downloaded file.

Int. J. Design Engineering, Vol. 7, No. 2, 2017 77

A multi-objective genetic algorithm for design

optimisation of simple and double harmonic
motion cams

R.K. Jana*
Indian Institute of Management Raipur,
GEC Campus,
Sejbahar, CG 492015, India
Email: [email protected]
*Corresponding author

Prasun Bhattacharjee
Department of Mechanical Engineering,
Jadavpur Univrsity,
Kolkata, WB 700032, India
Email: [email protected]

Abstract: In this paper, a multi-objective genetic algorithm is developed to

solve the design optimisation problem for cam-follower mechanism with
simple and double harmonic profiles. Torque and jerk are two most important
objectives decisive to be optimised for a successful cam follower system plan.
The proposed approach offers a set of non-dominated solutions for design
variables like angle of rotation, maximum lift of the follower and the angle for
maximum lift for rational operation. The problem is also solved by MATLAB
multi-objective genetic algorithm toolbox. The solutions obtained by the
proposed approach are found to be beneficial over the MATLAB solutions.

Keywords: cam-follower; design optimisation, multi-objective optimisation;

genetic algorithm; Pareto optimality.

Reference to this paper should be made as follows: Jana, R.K. and

Bhattacharjee, P. (2017) ‘A multi-objective genetic algorithm for design
optimisation of simple and double harmonic motion cams’, Int. J. Design
Engineering, Vol. 7, No. 2, pp.77–91.

Biographical notes: R.K. Jana is an Assistant Professor of Decision Science

and Systems Area at the Indian Institute of Management Raipur, India. He
obtained his PhD in Mathematics (Operations Research) from the IIT
Kharagpur, India. He received postdoctoral research fellowships from the
George Mason University, USA and National University of Singapore. He has
won three best paper awards from international conferences held in India and
the USA. He is a senior member of Operational Research Society of India, a
member of Decision Sciences Institute, USA and Indian Statistical Institute,
Kolkata. His research interest includes optimisation, artificial intelligence
techniques, business decision making.

Prasun Bhattacharjee completed his undergraduate studies on Mechanical

Engineering from the Kalyani Government Engineering College and Post
Graduate studies on Industrial Engineering and Management. He is currently a

Copyright © 2017 Inderscience Enterprises Ltd.

78 R.K. Jana and P. Bhattacharjee

PhD Scholar of the Department of Mechanical Engineering, Jadavpur

University, Kolkata, India. His research areas of interest are experimental
design, mathematical modelling of mechanical systems, multi-criterion
decision making and artificial intelligence. He has also work experience in
software application development.

1 Introduction

Cam-follower mechanism is one of the important machineries employed in a range of

automotive and mechanical devices. Rotational motion can be efficiently converted to
reciprocating motion with the help of cam-follower mechanism (Lampinen, 2003). The
essential motion of the follower can be obtained by accurate design of the cam contour. It
has been meticulously employed in various mechanical systems (Jensen, 1987). In
precedent times, cam profile optimisation has been one of the focus-areas of research in
the field of mechanical engineering design. Spline contour and polynomial curve
optimisation has remained in the centre of attention for most of these studies (Tesar and
Matthew, 1976; Sandgren and West, 1989; Tsay and Huey, 1993; Yoon and Rao, 1993;
Neamtu et al., 1998). Competent cam shape optimisation aspires to optimise the action of
follower while gratifying geometric and physical limitations. Optimisation of cam-
follower mechanism is subjected to constraints (Lampinen, 2003) such as diverse
material strength, kinetic, dynamic, manufacturing and geometric restrictions.
A two-step optimisation cam profile design scheme was proposed by Jeon et al.
(1989). In the initial step, valve lift area was maximised without resulting unusual valve
motions by fulfilling all the specified constraints like cam-event angle, maximum valve
acceleration and cam displacements at start and end of the cam-event angle. In the next
step, slight variations of the cam designed in the first step were done to minimise the
acceleration while retaining the maximised valve lift area and gratifying constrains
attained in the previous step. The optimal relationships among valve cam profile, cam
speed and stiffness using a valve train finite element representation for single-mass
dynamic and vibration analysis has been ascertained by Li et al. (1989). Lu et al. (1997)
explored difficulties in valve train design while taking into consideration objectives and
constraints of cam curve optimisation.
The above literature suggests that the classical optimisation techniques have been
applied to solve different types of cam-follower optimisation problems. However, there
are some known drawbacks of these techniques. Genetic algorithms (GAs), proposed by
Holland (1975), are stochastic search techniques having the ability to overcome these
drawbacks. They have been used successfully in various domains (Goldberg 1989; Jana
and Biswal, 2004, 2006; Jana and Sharma 2010; Sharma and Jana 2009a, 2009b).
However, there are only some limited studies available on cam design optimisation in the
literature. Xiao and Zu (2009) proposed a GA for solving polynomial and B-spline
curves.
The above studies focused on single objective cam-follower design optimisation. GAs
is known for their capability of dealing multiple and conflicting objectives. This category
of GAs is known as multi-objective genetic algorithms (MOGAs) (Horn et al., 1993;
 A multi-objective genetic algorithm for design optimisation 79

Srinivas and Deb, 1994). MOGAs have been used for solving a variety of scientific,
engineering and industrial applications involving multiple objectives (Coello and
Lamont, 2004; Coello Coello, 2006, Dimopoulos and Zalzala, 2000). Researchers have
used MOGAs to solve multi-objective optimisation problems to obtain a set of non-
dominated solutions that represents the tradeoffs between conflicting objectives. The
most prominent MOGA techniques include SPEA2 by Zitzler et al. (2000), PESA-2 by
Corne et al. (2001) and NSGA-2 by Deb et al. (2000).
Qin and Chen (2014) obtained the Pareto front for the cam-follower design
optimisation involving positive and negative acceleration and lift area for invariable
acceleration profile. They also mentioned that this area is yet to be explored fully.
Unfortunately, for the cam-follower design optimisation, MOGAs have not been used
much. This motivates us to design a MOGA for solving design optimisation problem that
optimises transmitted torque and jerk of a cam-roller type follower mechanism for simple
harmonic and double harmonic cam-follower mechanism.

2 Description of cam profiles

In this section, simple harmonic motion (SHM) and double harmonic motion (DHM)
cam-follower profiles are presented. To derive the profiles, the following notations are
used:
h maximum rise of the follower, mm

β cam angle for rise h, radian

θ cam angle rotation, radian

t time for cam for rotating angle θ, sec
ω rotational speed of cam, rad/sec

φ a dimensionless parameter that specifies the reciprocal of another non-dimensional

and non-zero number that implies the fractional relation of angle of rotation for
maximum lift of the follower with half of the angular revolution of the cam.

2.1 SHM cam profile

The follower motion is directly reliant on cam contour formation and the displacement of
the follower can be expressed mathematically as: y = f(θ). The cam rotates at a fixed
angular velocity, θ = ωt. With increasing speed of the cam-follower machinery, it turns
out to be acute essential to examine displacement, velocity, acceleration as well as jerk
values.
SHM cam has a cosine acceleration curve. It has a smooth acceleration profile but
discontinuity at the dwell ends due to infinite jerk. Motion equations for SHM cam are as
follows:
80 R.K. Jana and P. Bhattacharjee

h hφ
Displacement , y = (1 − cos φθ ); Velocity, y ′ = (ω sin φθ );
2 2
hφ2 2 hφ3 3
Acceleration, y ′′ = (ω cos φθ ); and Jerk y ′′′ = (ω sin φθ )
2 2
π
A plot of acceleration for SHM cam is shown in Figure 1 considering h = 1 mm, β =
2
and ω = 10 rad/sec.

Figure 1 Plot of acceleration for SHM curves

2.2 DHM cam profile

DHM cam curve is uneven as derived by taking together the discrepancy between two
harmonic motions, one being four times the amplitude and half of the frequency of the
other. Due to absence of the high shock and vibration at the commencement of the stroke,
it turns out to be beneficial compared to SHM curve. The rate of change in acceleration
i.e., jerk is moderately small at the start of the stroke which consequences into smooth
action at the primary point. Thus, a larger cam becomes necessary to offer enough least
cam curvature. For dwell-rise-return-dwell (D-R-R-D) cams, the limit of the sudden
change in acceleration at the maximum rise point allows only controlled cam speeds.
Motion equations for DHM cam are as follows:
 A multi-objective genetic algorithm for design optimisation 81

hφ ⎛ 1 ⎞
Velocity ( y ′) = ω ⎜ sin φθ − sin 2φθ ⎟ ;
2 ⎝ 2 ⎠
hφ2 2
Acceleration( y ′′) = ω (cos φθ − cos 2φθ );
2
hφ3 3
and Jerk ( y ′′′) = ω (sin 2φθ − sin φθ )
2
π
A plot of acceleration for DHM cam is shown in Figure 2 considering h = 1 mm, β =
2
and ω = 10 rad/sec.
Figure 2 Plot of acceleration for DHM curve

3 Multi-objective design optimisation problem formulation

In this section, the multi-objective problem formulation considering transmitted torque
and jerk as objectives is discussed.

3.1 Notation
The following notations are used in this paper:
82 R.K. Jana and P. Bhattacharjee

y Displacement of the follower, mm

dy′ d y 2
y′′ = = Acceleration of the follower, mm/sec2
dθ dθ 2
N Normal component of velocity, mm/sec
αp Pressure angle, deg
τ Torque, Newton-mm
m Mass of follower, kg
dy
y′ = Velocity of the follower, mm/sec
dθ
dy′′ d 2 y
y′′′ = = Jerk of the follower, mm/sec3
dθ dθ 2
Fn Force normal to cam surface, Newton
r Distance between cam centre to roller follower centre, mm
L External cam on cam, Newton
L0 = my′ + L Total load on cam, Newton

3.2 Formulation of transmitted torque objective

Computation of torque is a significant component of any rotating device design. The
adept running of the machine, the amount of the motor drive and the measures of the
driving shaft and allied elements are consequences of successful torque formulation.
Torque determination is exclusively vital for machinery dealing with elevated loads and
high pressure angles. From Figure 3, we can write N = rωsinαp, y′ = rωtanαp,
L0 = Fncosαp and τ = Fnsinαp (Rothbart, 2004).
Transmitted torque to the follower is maximised for more resourceful use of the
power delivered to the cam. Mathematically, this objective for SHM cam can be
expressed as follows:
⎧ hφ2 ⎫ ⎧ hφ ⎫
max : ω2 ⎨m ∗ cos φθ + L ⎬ ∗ ⎨ sin φθ ⎬ N − m
⎩ 2 ⎭ ⎩ 2 ⎭
Also, the transmitted torque objective for DHM cam can be expressed as follows:

⎧ hφ 1 ⎫ ⎧ hφ2 ⎫
max : ω2 ⎨m ∗ sin φθ − sin 2φθ + L ⎬ ∗ ⎨ (cos φθ − cos 2φθ ) ⎬ N − m
⎩ 2 2 ⎭ ⎩ 2 ⎭

3.3 Formulation of jerk objective

Jerk of the follower is an effective objective of a cam-follower mechanism as it is
essential to be minimised to reduce the consequences of impact stresses and vibrations.
The objective function for the jerk for SHM curve can be stated as:
hφ3
min : ω3 sin φθ m / sec3
2
 A multi-objective genetic algorithm for design optimisation 83

Also, the objective function for the jerk for DHM curve can be stated as:
hφ3
min : ω3 (sin 2φθ − sin φθ ) m / sec3
2

Figure 3 Translating roller follower-cam system velocity and forces

3.4 Constraints
The variables involved for this multi-objective problem are maximum rise of the follower
(h), rotational angle of the cam (θ) and dimensionless parameter that specify the
reciprocal of another non-dimensional and non-zero number that implies the fractional
relation of angle of rotation for maximum lift of the follower with half of the angular
revolution of the cam (φ). The proposed algorithm is aimed to search the rotational angle
of the cam for full revolution of the cam, i.e., within a range of 0 to 2π, maximum rise of
the follower within a range of 10 mm to 20 mm. The rotational angle of the cam for the
maximum rise of the follower is considered within a range of higher than zero and equals
to or less than 2π for bi-symmetrical D-R-R-D cam profiles.

4 The proposed MOGA

The techniques used for solving multi-objective programming problems aim to find a set
of non-dominated or Pareto-optimal solutions. In this section, the proposed real coded
84 R.K. Jana and P. Bhattacharjee

MOGA is presented to find non-dominated solutions for the multi-objective design

optimisation of simple and DHM cams involving transmitted torque and jerk of the cam-
roller type follower mechanism as objectives. The algorithm of the proposed MOGA is
presented next.

4.1 The MOGA algorithm

1 Initialise MOGA parameters: population dimension, iteration counter, probability of
crossover, probability of mutation and number of non-dominated solutions.
2 Initialise the population randomly.
3 Calculate fitness of each chromosome.
4 Perform the arithmetic crossover process as follows:
• Generate a number at random within 0 to 1. If the randomly generated number is
less than the probability of crossover, then choose the parent for crossover.
• Begin the crossover process.
• Verify the feasibility of children.
• If the children are feasible, then include them in the new population.
5 Perform mutation operation as follows:
• Generate a uniform random number within 0 to 1. If randomly generated
number is less than the probability of mutation, then select the chromosome for
mutation.
• Begin the mutation method.
• Check the feasibility of the newly generated chromosome.
• If generated chromosome is feasible then include it in the new population.
6 Check the fitness of the new individuals generated from crossover and mutation.
7 Perform dominance check. If a chromosome is not dominated, then add it to the pool.
Select all such non-dominated chromosomes.
8 If enough number of individuals required for Pareto front is obtained then stop, else
continue.
9 Choose the best compromised solution according to the decision-maker’s preference.

5 Results and discussions

Results and discussions are presented in this section. We first apply the proposed MOGA
to two well-known multi-objective test functions – Binh and Korn function (Binh and
Korn, 1997) and Chankong and Haimes function (Chankong and Haimes, 1983). Then it
is applied to the cam design optimisation problems corresponding to SHM and DHM
motion cams. We also solve all the mentioned problems using MATLAB MOGA
toolbox. The results obtained from both the methods are compared.
 A multi-objective genetic algorithm for design optimisation 85

The proposed MOGA is coded in MATLAB. From the computational experiences, it

has been found that the proposed approach performs best for the following parameter
values: population size = 100, probability of crossover = 0.6, probability of
mutation = 0.02 and generation number = 100. For the MATLAB MOGA toolbox, we
have used the default parameter settings.

5.1 Case 1: Binh and Korn function (Binh and Korn, 1997)
The Binh and Korn test function is defined as follows:
min : f1 ( x, y ) = 4 x 2 + 4 y 2
min : f 2 ( x, y ) = ( x − 5) 2 + ( y − 5) 2
subject to g1 ( x, y ) = ( x − 5) 2 + y 2 ≤ 25
g 2 ( x, y ) = ( x − 8) 2 + ( y + 3) 2 ≥ 7.7
0 ≤ x ≤ 5, 0 ≤ y ≤ 3

The Pareto fronts obtained using the proposed MOGA is shown in Figure 4(a) and the
Pareto front for MATLAB MOGA toolbox is shown in Figure 4(b).

5.2 Case 2: Chankong and Haimes function (Chankong and Haimes, 1983)
The Chankong and Haimes test function is defined as follows:
min : f1 ( x, y ) = 2 + ( x − 2) 2 + ( y − 1) 2
min : f 2 ( x, y ) = 9 x − ( y − 1) 2
subject to g1 ( x, y ) = x 2 + y 2 ≤ 225
g 2 ( x, y ) = x − 3 y + 10 ≤ 0
− 20 ≤ x ≤ 20, − 20 ≤ y ≤ 20

The Pareto fronts obtained using the proposed MOGA is shown in Figure 5(a) and the
Pareto front for MATLAB MOGA toolbox is shown in Figure 5(b).

5.3 Case 3: for SHM cam

Using the proposed method, the Pareto front for transmitted torque and jerk for SHM cam
is obtained for m = 1 kg, L = 0 N (inertial load consideration only) and ω = 100 rad/sec.
The Pareto front of the same problem is also obtained using MATLAB MOGA toolbox.
Both the Pareto fronts are shown in Figure 6.

5.4 Case 4: for DHM cam

Using the proposed method, the Pareto front for transmitted torque and jerk for DHM
cam is obtained for m = 1 kg, L = 0 N (inertial load consideration only) and ω = 100
rad/sec. The Pareto front of the same problem is also obtained with same parameter
setting using MATLAB MOGA toolbox. Both the Pareto fronts are shown in Figure 7.
86 R.K. Jana and P. Bhattacharjee

Figure 4 (a) Pareto front using MOGA (b) Pareto front using MATLAB toolbox

(a)

(b)
 A multi-objective genetic algorithm for design optimisation 87

Figure 5 (a) Pareto front using MOGA (b) Pareto front using MATLAB toolbox

(a)

(b)
88 R.K. Jana and P. Bhattacharjee

Figure 6 Comparison of Pareto fronts for transmitted torque and jerk for SHM profile (see online
version for colours)

Figure 7 Comparison of Pareto fronts for transmitted torque and jerk for DHM profile cam
(see online version for colours)

This study centred on designing a cam-follower method based on maximising transmitted

torque and minimising jerk. For enhanced performance of the cam-follower system, jerk
is must be minimised to decrease the unfavourable effects of shocks. At the same time,
the transmitted torque is to be maximised to amplify the effectiveness of torque
transmission. As the higher jerk values results immense impact stresses on the
mechanism, there should be adequate stress-bearing actions to diminish the adverse
effects. A comparative study for these two considered objectives has been carried out for
better understanding of the effect on the design optimisation process. The results obtained
by the proposed MOGA for SHM cam show that the transmitted torque ranges between
 A multi-objective genetic algorithm for design optimisation 89

1,500 and 2,200 N-m and the jerk ranges between 20,000 to 31,000 m/sec3. On the other
hand, the results obtained by the proposed MOGA for DHM cam show that the
transmitted torque ranges between 1,500 to 1,900 N-m and the jerk ranges between 4,000
to 10,000 m/sec3. Thus, for any camshaft the torque value within the obtained range, the
transmitted torque value of the DHM profile is almost double of that of the SHM profile.
The speed of the camshaft is assumed for this study as 100 rad/sec. The crankshaft speed
for four-stroke automotive engines is twice the speed of the crankshaft and thus the
crankshaft speed used for this study is 200 rad/sec or 1900 r.p.m. (approximately).
For the SHM cam, the MATLAB solutions show that the transmitted torque ranges
between 1,500 to 2,200 N-m and the jerk value ranges between 15,000 to 31,000 m/sec3.
For the DHM cam, MATLAB MOGA toolbox provides only one solution with
transmitted torque value 2,700 N-m and jerk 50 m/sec3. It has also been obtained that the
number of solution achieved by MATLAB MOGA toolbox is fewer compared to
solutions achieved by the proposed method. This is due to more concentration of
solutions obtained by the MATLAB toolbox at both ends. These are the situations where
infinite jerks are experienced and MATLAB MOGA toolbox is unable to explore other
solutions in the remaining search domain. Moreover, there is no exclusive provision in
MATLAB toolbox to find a specific number of Pareto solutions before the termination of
the search. This feature is present in the proposed approach.

6 Conclusions

This study has been aimed on designing a cam-follower mechanism based on maximising
transmitted torque and minimising jerk. A MOGA has been designed for this purpose.
The Pareto optimal solutions found by the proposed approach are found to be realistic
and implementable in practical applications. The problem is also solved by MATLAB
MOGA toolbox. The MOGA toolbox could provide only 04 Pareto optimal solutions for
the SHM and 01 Pareto optimal solutions for the DHM cam-follower mechanism within a
given torque range of 3,000 N-m for both the cases. On the other hand, the proposed
MOGA could obtain 100 Pareto optimal solutions. It is found that the DHM profile offers
a quite higher transmitted torque at a low jerk value compared to SHM profile which
facilitates the previous cam-follower mechanism to be fit for high-speed function.
This study will open novel prospects of research. Cam design problem is essentially
multi-objective in nature. However, this problem is not often worked out using
exclusively multi-objective solution procedures. The proposed approach may be directly
applied to other cam design efforts having different objectives.

Acknowledgements

The authors would like to thank the anonymous reviewers and the Editor-in-Chief for
their helpful comments and suggestions that have improved the presentation and quality
of the paper.
90 R.K. Jana and P. Bhattacharjee

References
Binh, T. and Korn, U. (1997) ‘MOBES: a multiobjective evolution strategy for constrained
optimization problems’, in Proceedings of the Third International Conference on Genetic
Algorithms (Mendel97), Brno, Czech Republic, pp.176–182.
Chankong, V. and Haimes, Y.Y. (1983) Multiobjective Decision Making Theory and Methodology,
Elsevier Science, New York.
Coello Coello, C. (2006) ‘Evolutionary multi-objective optimization: a historical view of the field’,
Computational Intelligence Magazine, IEEE, Vol. 1, No. 1, pp.28–36.
Coello, C.A.C. and Lamont, G.B. (2004) Applications of Multi-objective Evolutionary Algorithms,
World Scientific, NJ, USA.
Corne, D., Jerram, N., Knowles, J. and Oates, M. (2001) ‘PESA-II: region-based selection in
evolutionary multiobjective optimization’, in Proc. of the Genetic and Evolutionary
Computation Conference (GECCO2001), Cite-seer.
Deb, K., Agrawal, S., Pratap, A. and Meyarivan, T. (2000) ‘A fast elitist non-dominated sorting
genetic algorithm for multi-objective optimization: NSGA-II’, in Schoenauer, M. et al. (Eds.):
Parallel Problem Solving from Nature PPSN VI. PPSN 2000. Lecture Notes in Computer
Science, Vol. 1917, pp.849–858, Springer, Berlin, Heidelberg.
Dimopoulos, C. and Zalzala, A. (2000) ‘Recent developments in evolutionary computation for
manufacturing optimization: problems, solutions and comparisons’, IEEE Transactions on
Evolutionary Computation, Vol. 4, No. 2, pp.93–113.
Goldberg, D.E. (1989) Genetic Algorithms in Search, Optimization and Machine Learning,
Addison-Wesley, Reading, MA.
Holland, J.H. (1975) Adaptation in Natural and Artificial Systems, University of Michigan Press,
Ann Arbor, MI.
Horn, J., Nafpliots, N. and Goldberg, D.E. (1993) ‘A niched Pareto genetic algorithm for multi-
objective optimization’, IEEE World Congress on: Computational Intelligence, Vol. 1,
pp.82–87.
Jana, R.K. and Biswal, M.P. (2004) ‘Stochastic simulation-based genetic algorithm for chance
constraint programming problems with some discrete random variables’, International Journal
of Computer Mathematics, Vol. 81, No. 12, pp.1455–1463.
Jana, R.K. and Biswal, M.P. (2006) ‘Genetic based fuzzy goal programming for multiobjective
chance constrained programming problems with continuous random variables’, International
Journal of Computer Mathematics, Vol. 83, No. 2, pp.171–179.
Jana, R.K. and Sharma, D.K. (2010) ‘Genetic algorithm based fuzzy goal programming for class of
chance constrained programming problems’, International Journal of Computer Mathematics,
Vol. 87, No. 4, pp.733–742.
Jensen, P.W. (1987) Cam Design and Manufacture, 2nd ed., Marcel Dekker Inc., New York, USA.
Jeon, H.S., Park, K.J. and Park, Y.S. (1989) ‘An optimal cam profile design considering dynamic
characteristics of a cam-valve system’, Experimental Mechanics, Vol. 29, No. 4, pp.357–363.
Lampinen, J. (2003) ‘Cam shape optimisation by genetic algorithm’, Computer-Aided Design,
Vol. 35, No. 8, pp.727–737.
Li, H.Z., Yuan, Z.C. Le, J.B. and Liu, J.C. (1989) ‘Development of valve cam design’, Chin.
Internal Combustion Engine Eng., Vol. 10, No. 1, pp.32–37.
Lu, J.Q., Xu, X. and Li, Y.D. (1997) ‘The optimum design method for valve train’, Trans. CSICE,
Vol. 15, No. 1, pp.120–127.
Neamtu, M., Pottmann, H. and Schumaker, L.L. (1998) ‘Design NURBS cam profiles using
trigonometric splines’, ASME Journal of Mechanical Design, Vol. 120, No. 2, pp.175–180.
Qin, W. and Chen, Y. (2014) ‘Study on optimal kinematic synthesis of cam profiles for engine
valve trains’, Applied Mathematical Modelling, Vol. 38, Nos. 17–18, pp.4345–4353.
Rothbart, H.A. (2004) Cam Design Handbook, McGraw-Hill, USA.
 A multi-objective genetic algorithm for design optimisation 91

Sandgren, E. and West, R.L. (1989) ‘Shape optimization of cam profiles using a B-spline
representation’, ASME Journal of Mechanisms, Transmissions and Automation in Design,
Vol. 111, No. 2, pp.195–201.
Sharma, D.K. and Jana, R.K. (2009a) ‘Fuzzy goal programming based genetic algorithm approach
to nutrient management for rice crop planning’, International Journal of Production
Economics, Vol. 121, No. 1, pp.224–232.
Sharma, D.K. and Jana, R.K. (2009b) ‘A hybrid genetic algorithm model for transhipment
management decisions’, International Journal of Production Economics, Vol. 122, No. 2,
pp.703–713.
Srinivas, N. and Deb, K. (1994) ‘Multi-objective optimization using non-dominated sorting in
genetic algorithms’, Evolutionary Computation, Vol. 2, No. 3, pp.221–248.
Tesar, D. and Matthew, G.K. (1976) The Dynamic Synthesis, Analysis and Design of Modelled
Cam Systems, Lexington Books, D.C. Heath and Co., Lexington, Mass., USA.
Tsay, D.M. and Huey, C.O. (1993) ‘Application of rational B-splines to the synthesis of
cam-follower motion programs’, ASME Journal of Mechanical Design, Vol. 115, No. 3,
pp.621–626.
Xiao, H. and Zu, J.W. (2009) ‘Cam profile optimization for a new cam drive’, Journal of
Mechanical Science and Technology, Vol. 23, No. 10, pp.2592–2602.
Yoon, K. and Rao, S.S. (1993) ‘Cam motion synthesis using cubic splines’, ASME Journal of
Mechanical Design, Vol. 115, No. 3, pp.441–446.
Zitzler, E., Deb, K. and Thiele, L. (2000) ‘Comparison of multi-objective evolutionary algorithms:
empirical results’, Evolutionary Computation, Vol. 8, No. 2, pp.173–195.

View publication stats