1

Introduction

Optimization algorithms are becoming increasingly popular in multi-

engineering design activities, primarily because of the availability and
affordability of high speed computers. They are extensively used in those
engineering design problems where the emphasis is on maximizing or
minimizing a certain goal. For example, optimization algorithms are routinely
used in aerospace design activities to minimize the overall weight, simply
because every element or component adds to the overall weight of the
aircraft. Thus, the minimization of the weight of aircraft components is of
major concern to aerospace designers. Chemical engineers, on the other hand,
are interested in designing and operating a process plant for an optimum rate
of production. Mechanical engineers design mechanical components for the
purpose of achieving either a minimum manufacturing cost or a maximum
component life. Production engineers are interested in designing optimum
schedules of various machining operations to minimize the idle time of
machines and the overall job completion time. Civil engineers are involved
in designing buildings, bridges, dams, and other structures in order to achieve
a minimum overall cost or maximum safety or both. Electrical engineers are
interested in designing communication networks so as to achieve minimum
time for communication from one node to another.
All the above-mentioned tasks involve either minimization or maximiza-
tion (collectively known as optimization) of an objective. It is clear from the
spectrum of the above problems that it is difficult to discuss the formulation
of various engineering optimization problems in a single book. Fortunately,
a designer specialized in a particular design is usually more informed about
different factors governing that design than anyone else. Thus, as far as the
formulation of the optimal problem is concerned, the designer can acquire it
with some practice. However, every designer should know a few aspects of
the formulation procedure which would help him or her to choose a proper
optimization algorithm for the chosen optimal design problem. This requires

1
2 Optimization for Engineering Design: Algorithms and Examples

a knowledge about the working principles of different optimization methods.

In subsequent chapters, we discuss various optimization methods which would
hopefully provide some of that knowledge. In this chapter, we demonstrate the
optimal problem formulation procedures of four different engineering optimal
design problems.

1.1 Optimal Problem Formulation

In many industrial design activities, a naive optimal design is achieved by
comparing a few (limited up to ten or so) alternative design solutions created
by using a priori problem knowledge. In such an activity, the feasibility
of each design solution is first investigated. Thereafter, an estimate of the
underlying objective (cost, profit, etc.) of each design solution is computed
and the best design solution is adopted. This naive method is often followed
because of the time and resource limitations. But, in many cases this method
is followed simply because of the lack of knowledge of the existing optimization
procedures. But whatever may be the reason, the purpose of either achieving
a quality product or of achieving a competing product is never guaranteed
to be fulfilled with the above naive method. Optimization algorithms
described in this book provide systematic and efficient ways of creating
and comparing new design solutions in order to achieve an optimal design.
Since an optimization algorithm requires comparison of a number of design
solutions, it is usually time-consuming and computationally expensive. Thus,
the optimization procedure must only be used in those problems where there
is a definite need of achieving a quality product or a competitive product. It is
expected that the design solution obtained through an optimization procedure
is better than other solutions in terms of the chosen objective—cost, efficiency,
safety, or others.
We begin our discussion with the formulation procedure by mentioning
that it is almost impossible to apply a single formulation procedure for all
engineering design problems. Since the objective in a design problem and
the associated design parameters vary from product to product, different
techniques need to be used in different problems. The purpose of the
formulation procedure is to create a mathematical model of the optimal design
problem, which then can be solved using an optimization algorithm. Since
an optimization algorithm accepts an optimization problem in a particular
format, every optimal design problem must be formulated in that format. In
this section, we discuss different components of that format.
Figure 1.1 shows an outline of the steps usually involved in an optimal
design formulation process. As mentioned earlier, the first step is to realize
the need for using optimization in a specific design problem. Thereafter, the
designer needs to choose the important design variables associated with the
design problem. The formulation of optimal design problems involves other
considerations, such as constraints, objective function, and variable bounds.
As shown in the figure, there is usually a hierarchy in the optimal design
Introduction 3

process; although one consideration may get influenced by the other. We

discuss all these aspects in the following subsections.

Figure 1.1 A flowchart of the optimal design procedure.

1.1.1 Design Variables

The formulation of an optimization problem begins with identifying

the underlying design variables, which are primarily varied during the
optimization process. A design problem usually involves many design
parameters, of which some are highly sensitive to the proper working of
the design. These parameters are called design variables in the parlance of
optimization procedures. Other (not so important) design parameters usually
remain fixed or vary in relation to the design variables. There is no rigid
guideline to choose a priori the parameters which may be important in a
problem, because one parameter may be more important with respect to
minimizing the overall cost of the design, while it may be insignificant with
respect to maximizing the life of the product. Thus, the choice of the important
parameters in an optimization problem largely depends on the user. However,
it is important to understand that the efficiency and speed of optimization
algorithms depend, to a large extent, on the number of chosen design variables.
In subsequent chapters, we shall discuss certain algorithms which work very
efficiently when the number of design variables is small, but do not work that
well for a large number of design variables. Thus, by selectively choosing the
design variables, the efficacy of the optimization process can be increased. The
first thumb rule of the formulation of an optimization problem is to choose as
4 Optimization for Engineering Design: Algorithms and Examples

few design variables as possible. The outcome of that optimization procedure

may indicate whether to include more design variables in a revised formulation
or to replace some previously considered design variables with new design
variables.

1.1.2 Constraints
Having chosen the design variables, the next task is to identify the constraints
associated with the optimization problem. The constraints represent some
functional relationships among the design variables and other design
parameters satisfying certain physical phenomenon and certain resource
limitations. Some of these considerations require that the design remain in
static or dynamic equilibrium. In many mechanical and civil engineering
problems, the constraints are formulated to satisfy stress and deflection
limitations. Often, a component needs to be designed in such a way that
it can be placed inside a fixed housing, thereby restricting the size of the
component. There is, however, no unique way to formulate a constraint in
all problems. The nature and number of constraints to be included in the
formulation depend on the user. In many algorithms discussed in this book,
it is not necessary to have an explicit mathematical expression of a constraint;
but an algorithm or a mechanism to calculate the constraint is mandatory. For
example, a mechanical engineering component design problem may involve
a constraint to restrain the maximum stress developed anywhere in the
component to the strength of the material. In an irregular-shaped component,
there may not exist an exact mathematical expression for the maximum stress
developed in the component. A finite element simulation software may be
necessary to compute the maximum stress. But the simulation procedure and
the necessary input to the simulator and the output from the simulator must
be understood at this step.
There are usually two types of constraints that emerge from most
considerations. Either the constraints are of an inequality type or of an
equality type. Inequality constraints state that the functional relationships
among design variables are either greater than, smaller than, or equal to,
a resource value. For example, the stress (σ(x)) developed anywhere in a
component must be smaller than or equal to the allowable strength (Sallowable )
of the material. Mathematically,
σ(x) ≤ Sallowable .
Most of the constraints encountered in engineering design problems are of this
type. Some constraints may be of greater-than-equal-to type: for example,
the natural frequency (ν(x)) of a system may required to be greater than
2 Hz, or mathematically, ν(x) ≥ 2. Fortunately, one type of inequality
constraints can be transformed into the other type by multiplying both sides
by −1 or by interchanging the left and right sides. For example, the former
constraint can be transformed into a greater-than-equal-to type by either
−σ(x) ≥ −Sallowable or Sallowable ≥ σ(x).
Introduction 5

Equality constraints state that the functional relationships should exactly

match a resource value. For example, a constraint may require that the
deflection (δ(x)) of a point in the component must be exactly equal to 5 mm,
or mathematically,
δ(x) = 5.
Equality constraints are usually more difficult to handle and, therefore, need
to be avoided wherever possible. If the functional relationships of equality
constraints are simpler, it may be possible to reduce the number of design
variables by using the equality constraints. In such a case, the equality
constraints reduce the complexity of the problem, thereby making it easier
for the optimization algorithms to solve the problem. In Chapter 4, we
discuss a number of algorithms which are specially designed to handle equality
constraints. Fortunately, in many engineering design optimization problems,
it may be possible to relax an equality constraint by including two inequality
constraints. The above deflection equality constraint can be replaced by two
constraints:
δ(x) ≥ 4,

δ(x) ≤ 6.

The exact deflection requirement of 5 mm is relaxed by allowing it to deflect

anywhere between 4 mm to 6 mm. Although this formulation is inexact as
far as the original requirement is concerned, this flexibility allows a smoother
operation of the optimization algorithms. Thus, the second thumb rule in the
formulation of optimization problems is that the number of complex equality
constraints should be kept as low as possible.

1.1.3 Objective Function

The third task in the formulation procedure is to find the objective function
in terms of the design variables and other problem parameters. The common
engineering objectives involve minimization of overall cost of manufacturing,
or minimization of overall weight of a component, or maximization of net
profit earned, or maximization total life of a product, or others. Although
most of the above objectives can be quantified (expressed in a mathematical
form), there are some objectives that may not be quantified easily. For
example, the esthetic aspect of a design, ride characteristics of a car suspension
design, and reliability of a design are important objectives that one may be
interested in maximizing in a design, but the exact mathematical formulation
may not be possible. In such a case, usually an approximating mathematical
expression is used. Moreover, in any real-world optimization problem, there
could be more than one objective that the designer may want to optimize
simultaneously. Even though a few multi-objective optimization algorithms
exist in the literature (Chankong and Haimes, 1983), they are complex and
computationally expensive. Thus, in most optimal design problem, multiple
objectives are avoided. Instead, the designer chooses the most important
6 Optimization for Engineering Design: Algorithms and Examples

objective as the objective function of the optimization problem, and the

other objectives are included as constraints by restricting their values within
a certain range. For example, consider an optimal truss structure design
problem. The designer may be interested in minimizing the overall weight of
the structure and simultaneously be concerned in minimizing the deflection of
a specific point in the truss. In the optimal problem formulation, the designer
may like to use the weight of the truss (as a function of the cross-sections
of the members) as the objective function and have a constraint with the
deflection of the concerned point to be less than a specified limit. In general,
the objective function is not required to be expressed in a mathematical form.
A simulation package may be required to evaluate the objective function. But
whatever may be the way to evaluate the objective function, it must be clearly
understood.
The objective function can be of two types. Either the objective function
is to be maximized or it has to be minimized. Unfortunately, the optimization
algorithms are usually written either for minimization problems or for
maximization problems. Although in some algorithms, some minor structural
changes would enable to perform either minimization or maximization, this
requires extensive knowledge of the algorithm. Moreover, if an optimization
software is used for the simulation, the modified software needs to be
compiled before it can be used for the simulation. Fortunately, the duality
principle helps by allowing the same algorithm to be used for minimization
or maximization with a minor change in the objective function instead of
a change in the entire algorithm. If the algorithm is developed for solving a
minimization problem, it can also be used to solve a maximization problem by
simply multiplying the objective function by −1 and vice versa. For example,
consider the maximization of the single-variable function f (x) = x2 (1 − x)
shown by a solid line in Figure 1.2. The maximum point happens to be
at x∗ = 0.667. The duality principle suggests that the above problem is
equivalent to minimizing the function F (x) = −x2 (1 − x), which is shown by
a dashed line in Figure 1.2. The figure shows that the minimum point of the
function F (x) is also at x∗ = 0.667. Thus, the optimum solution remains the
same. But once we obtain the optimum solution by minimizing the function
F (x), we need to calculate the optimal function value of the original function
f (x) by multiplying F (x) by −1.

1.1.4 Variable Bounds

The final task of the formulation procedure is to set the minimum and the
maximum bounds on each design variable. Certain optimization algorithms
do not require this information. In these problems, the constraints completely
surround the feasible region. Other problems require this information in order
to confine the search algorithm within these bounds. In general, all N design
variables are restricted to lie within the minimum and the maximum bounds
as follows:
(L) (U )
xi ≤ xi ≤ xi for i = 1, 2, . . . , N .
Introduction 7

Figure 1.2 Illustration of the duality principle. The maximum point of f (x) is the
same as the minimum point of F (x).

(L)
In any given problem, the determination of the variables bounds xi and
(U )
xi may be difficult. One way to remedy this situation is to make a guess
about the optimal solution and set the minimum and maximum bounds so
that the optimal solution lies within these two bounds. After simulating the
optimization algorithm once, if the optimal solution is found to lie within
the chosen variable bounds, there is no problem. On the other hand, if any
design variable corresponding to the optimal solution is found to lie on or
near the minimum or the maximum bound, the chosen bound may not be
correct. The chosen bound may be readjusted and the optimization algorithm
may be simulated again. Although this strategy may seem to work only with
linear problems, it has been found useful in many real-world engineering
optimization problems.
After the above four tasks are completed, the optimization problem
can be mathematically written in a special format, known as nonlinear
programming (NLP) format. Denoting the design variables as a column
vector1 x = (x1 , x2 , . . . , xN )T , the objective function as a scalar quantity
f (x), J inequality constraints as gj (x) ≥ 0, and K equality constraints as

hk (x) = 0

1
The representation of the design variables in the above column vector helps
to achieve some matrix operations in certain multivariable optimization methods
described in Chapters 3 and 4.
8 Optimization for Engineering Design: Algorithms and Examples

we write the NLP problem:




Minimize f (x) 







subject to 


gj (x) ≥ 0, j = 1, 2, . . . , J; 
(1.1)




hk (x) = 0, k = 1, 2, . . . , K; 





(L) (U ) 

xi ≤ xi ≤ xi , i = 1, 2, . . . , N.

Note that the above formulation can represent a formulation for

maximization problems by using the duality principle and can represent
a formulation for problems with the lesser-than-equal-to type inequality
constraints by using the techniques described earlier. However, the
optimization algorithm used to solve the above NLP problem depends on
the type of the objective function and constraints. It is important to note
that the constraints must be written in a way so that the right-side of the
inequality or equality sign is zero.
It is worth mentioning here that all the above four tasks are not
independent of each other. While formulating the objective function, the
designer may decide to include or delete some constraints. In many problems,
while the constraints are being formulated, it is necessary to add some
artificial design variables, which make the overall formulation easier. The
update of the design variables, the constraints, the objective function, and
the variable bounds may continue for a few iterations until the designer is
finally satisfied with a reasonable formulation. Certain possible iterations are
shown in Figure 1.1. We may mention here that this update also depends on
the knowledge of the optimization algorithms to be used to solve the problem.
But this requires some practice of using optimization algorithms before such
input may be incorporated into the formulation procedure. Nevertheless, after
the optimization problem is formulated, an optimization algorithm is chosen
and an optimal solution of the NLP problem is obtained. We now illustrate
the above four steps of the formulation procedure in four engineering optimal
design problems.

1.2 Engineering Optimization Problems

Optimization problems can be found in most engineering disciplines. Because
of the variety of engineering design problems, it is not possible to discuss the
formulation of every optimization problem that is encountered in engineering
design in a single book. However, in this section, we categorize different
engineering problem-solving tasks in which an optimization method can be
applied to find a solution.
Introduction 9

1.2.1 Design and Manufacturing

A majority of the optimization problems arise from design and manufacturing

areas. Since designs have to work efficiently, have to be cost-effective, have
to have least weight in certain applications, and have to satisfy many other
criteria, it makes perfect sense to apply an optimization algorithm to find such
a solution. Similarly, a manufacturing process involves various parameters
and operations that must be optimized to produce a competitive product or
to achieve a quality product. Hence, the use of an optimization procedure is
essential in most manufacturing problem-solving tasks. To illustrate, here we
only discuss two design problem-solving tasks and one manufacturing process
optimization task.

Optimal design of a truss structure

A truss structure is used in many civil engineering applications including

bridges, buildings, and roofs. There are two different types of optimization
problems in a truss structure design. Firstly, the topology of the truss
structure (the connectivity of the elements in a truss) could be an optimization
problem. In this problem, the objective is to find the optimal connectivity of
truss elements so as to achieve the minimal cost of materials and construction.
Secondly, once the optimal layout of the truss is known, the determination of
every element cross-section is another optimization problem. In this problem,
the objective is to find the optimal cross-sections of all elements in order to
achieve a minimum cost of materials and construction. Although both these
problems attempt to achieve the same objective, the search space and the
optimization algorithm required to solve each problem are different. Here, we
discuss the latter problem only. However, there exist certain algorithms which
can be used to solve both the above problems simultaneously. We discuss more
about these algorithms in Chapter 6.
Consider the seven-bar truss structure shown in Figure 1.3. The loading
is also shown in the figure. The length of the members AC = CE = ℓ = 1 m.
Once the connectivity of the truss is given, the cross-sectional area and the
material properties of the members are the design parameters. Let us choose
the cross-sectional area of members as the design variables for this problem.
There are seven design variables, each specifying the cross-section of a member

Figure 1.3 A typical seven-bar truss structure.

10 Optimization for Engineering Design: Algorithms and Examples

(A1 to A7 ). Using the symmetry of the truss structure and loading, we observe
that for the optimal solution, A7 = A1 , A6 = A2 , and A5 = A3 . Thus, there
are practically four design variables (A1 to A4 ). This completes the first task
of the optimization procedure.
The next task is to formulate the constraints. In order for the truss to
carry the given load P = 2 kN, the tensile and compressive stress generated
in each member must not be more than the corresponding allowable strength
Syt and Syc of the material. Let us assume that the material strength for all
elements is Syt = Syc = 500 MPa and the modulus of elasticity E = 200 GPa.
For the given load, we can compute the axial force generated in each element
(Table 1.1). The positive force signifies tensile load and the negative force
signifies compressive load acting on the member.

Table 1.1 Axial Force in Each Member of the Truss

Member Force Member Force

AB − P2 csc θ BC + P2 csc α

AC + P2 cot θ BD − P2 (cot θ + cot α)

Thereafter, the axial stress can be calculated by dividing the axial load
by the cross-sectional area of that member. Thus, the first set of constraints
can be written as
P csc θ
≤ Syc ,
2A1
P cot θ
≤ Syt ,
2A2
P csc α
≤ Syt ,
2A3
P
(cot θ + cot α) ≤ Syc .
2A4
In the above truss structure, tan θ = 1.0 and tan α = 2/3. The other set of
constraints arises from the stability consideration of the compression members
AB, BD, and DE. Realizing that each of these members is connected by pin
joints, we can write the Euler buckling conditions for the axial load in members
AB and BD (Shigley, 1986):
P πEA21
≤ ,
2 sin θ 1.281ℓ2
P πEA24
(cot θ + cot α) ≤ .
2 5.76ℓ2
Introduction 11

In most structures, deflection is a major consideration. In the above truss

structure, let us assume that the maximum vertical deflection at C is
δmax = 2 mm. By using Castigliano’s theorem (Timoshenko, 1986), we obtain
the deflection constraint:
( )
P ℓ 0.566 0.500 2.236 2.700
+ + + ≤ δmax .
E A1 A2 A3 A4
All the above constraints are of less-than-equal-to type. Once the constraints
are formulated, the next task is to formulate the objective function. In this
problem, we are interested in minimizing the weight of the truss structure.
Since we have assumed the same material for all members, the minimization
of the total volume of material will yield the same optimal solution as the
minimization of the total weight. Thus, we write the objective function as
Minimize 1.132A1 ℓ + 2A2 ℓ + 1.789A3 ℓ + 1.2A4 ℓ.
The fourth task is to set some lower and upper bounds for the four cross-
sectional areas. We may choose to make all four areas lie between 10 and
500 mm2 . Thus, the variable bounds are as
10 × 10−6 ≤ A1 , A2 , A3 , A4 ≤ 500 × 10−6 .
In the following, we present the above truss structure problem in NLP form,
which is suitable for solving by using an optimization algorithm described in
Chapter 4:
Minimize 1.132A1 ℓ + 2A2 ℓ + 1.789A3 ℓ + 1.2A4 ℓ
subject to
P
Syc − ≥ 0,
2A1 sin θ
P
Syt − ≥ 0,
2A2 cot θ
P
Syt − ≥ 0,
2A3 sin α
P
Syc − (cot θ + cot α) ≥ 0,
2A4
πEA21 P
2
− ≥ 0,
1.281ℓ 2 sin θ
πEA24 P
2
− (cot θ + cot α) ≥ 0,
5.76ℓ 2
( )
P ℓ 0.566 0.500 2.236 2.700
δmax − + + + ≥ 0,
E A1 A2 A3 A4

10 × 10−6 ≤ A1 , A2 , A3 , A4 ≤ 500 × 10−6 .

12 Optimization for Engineering Design: Algorithms and Examples

This shows the formulation of the truss structure problem (Deb et al.,
2000). The seven-bar truss shown in Figure 1.3 is statically determinate
and the axial force, stress, and deflection were possible to compute exactly.
In cases where the truss is statically indeterminate and large (for hand
calculations), the exact computations of stress and deflection may not be
possible. A finite element software may be necessary to compute the stress
and deflection in any member and at any point in the truss. Although similar
constraints can then be formulated with the simulated stresses and deflections,
the optimization algorithm which may be used to solve the above seven-bar
truss problem may not be efficient to solve the resulting NLP problem for
statically indeterminate or large truss problems. The difficulty arises due to
the inability to compute the gradients of the constraints. We shall discuss
more about this aspect in Chapter 4.

Optimal design of a car suspension

The comfort in riding a car largely depends on the suspension characteristics.

The car body is usually supported by a suspension coil spring and a damper
at each wheel (Figure 1.4).

Figure 1.4 A two-dimensional model of a car suspension system.

In some cars, the axle assembly is directly supported on the wheel. The
tyre of the wheel can also be assumed to have some stiffness in the vertical
direction. A two-dimensional dynamic model of a car suspension system is
shown in Figure 1.5. In this model, only two wheels (one each at rear
and front) are considered. The sprung mass of the car is considered to be
supported on two axles (front and rear) by means of a suspension coil spring
and a shock absorber (damper). Each axle contains some unsprung mass
which is supported by the tyre.
In order to formulate the optimal design problem, the first task is to
identify the important design variables. Let us first identify all the design
parameters that could govern the dynamic behaviour of the car vibration. In
the following, we list all these parameters:
Introduction 13

Figure 1.5 The dynamic model of the car suspension system. The above model
has four degrees-of-freedom (q1 to q4 ).

Sprung mass ms , Front coil stiffness kf s ,

Front unsprung mass mf u , Rear coil stiffness krs ,
Rear unsprung mass mru , Front tyre stiffness kf t ,
Rear damper coefficient αr , Rear tyre stiffness krt ,
Front damper coefficient αf , Axle-to-axle distance ℓ,
Polar moment of inertia of the car J.

We may consider all the above parameters as design variables, but the
time taken for convergence of the optimization algorithm may be too much.
In order to simplify the formulation, we consider only four of the above
parameters—front coil stiffness kf s , rear coil stiffness krs , front damper
coefficient αf , and rear damper coefficient αr —as design variables. We keep
the other design parameters as constant:

ms = 1000 kg, mf u = 70 kg, mru = 150 kg,

kf t = 20 kg/mm, krt = 20 kg/mm, J = 550 kg-m2 ,

ℓ1 = 1.6 m, ℓ2 = 1.6 m, ℓ = 3.2 m.

The parameters ℓ1 and ℓ2 are the horizontal distance of the front and rear
axles from the centre of gravity of the sprung mass. Using these parameters,
differential equations governing the vertical motion of the unsprung mass at
the front axle (q1 ), the sprung mass (q2 ), and the unsprung mass at the
14 Optimization for Engineering Design: Algorithms and Examples

rear axle (q4 ), and the angular motion of the sprung mass (q3 ) are written
(Figure 1.5):

q̈1 = (F2 + F3 − F1 )/mf u , (1.2)

q̈2 = −(F2 + F3 + F4 + F5 )/ms , (1.3)

/
q̈3 = [(F4 + F5 )ℓ2 − (F2 + F3 )ℓ1 ] J, (1.4)

q̈4 = (F4 + F5 − F6 )/mru , (1.5)

where the forces F1 to F6 are calculated as follows:

F1 = kf t d1 , F2 = kf s d2 , F3 = αf d˙2 ,

F4 = krs d4 , F5 = αr d˙4 , F6 = krt d3 .

The parameters d1 , d2 , d3 , and d4 are the relative deformations in the

front tyre, the front spring, the rear tyre, and the rear spring, respectively.
Figure 1.5 shows all the four degrees-of-freedom of the above system (q1 to
q4 ). The relative deformations in springs and tyres can be written as follows:

d1 = q1 − f1 (t),

d2 = q2 + ℓ1 q3 − q1 ,

d3 = q4 − f2 (t),

d4 = q2 − ℓ2 q3 − q4 .

The time-varying functions f1 (t) and f2 (t) are road irregularities as functions
of time. Any function can be used for f1 (t). For example, a bump can be
modelled as f1 (t) = A sin(πt/T ), where A is the amplitude of the bump and T
the time required to cross the bump. When a car is moving forward, the front
wheel experiences the bump first, while the rear wheel experiences the same
bump a little later, depending upon the speed of the car. Thus, the function
f2 (t) can be written as f2 (t) = f1 (t − ℓ/v), where ℓ is the axle-to-axle distance
and v is the speed of the car. For the above bump, f2 (t) = A sin(π(t−ℓ/v)/T ).
The coupled differential equations specified in Equations (1.2) to (1.5) can
be solved using a numerical integration technique (for example, a fourth-
order Runge-Kutta method can be used) to obtain the pitching and bouncing
dynamics of the sprung mass ms . Equations can be integrated for a time range
from zero to tmax .
After the design variables are chosen, the next task is to formulate the
constraints associated with the above car suspension problem. In order to
simplify the problem, we consider only one constraint. The jerk (the rate
of change of the vertical acceleration of the sprung mass) is a major factor
concerning the comfort of the riding passengers. The guideline used in car
Introduction 15

industries suggests that the maximum jerk experienced by the passengers

should not be more than about 18 m/s3 . Mathematically,

max q2′′′ (t) ≤ 18.

When the four coupled differential equations (1.2) to (1.5) are solved, the
above constraint can be computed by numerically differentiating the vertical
movement of the sprung mass (q2 ) thrice with respect to time.
The next task is to formulate the objective function. In this problem, the
primary objective is to minimize the transmissibility factor which is calculated
as the ratio of the bouncing amplitude q2 (t) of the sprung mass to the road
excitation amplitude A. Thus, we write the objective function as

max abs q2 (t)

Minimize .
A
The above objective function can be calculated from the solution of the
four differential equations mentioned earlier. A minimum value of the
transmissibility factor suggests the minimum transmission of road vibration
to the passengers. This factor is also directly related to the ride characteristics
as specified by the ISO standard. Thus, the optimized design of the above car
suspension system would provide the minimum transmissibility of the road
vibration to the passengers with a limited level of jerk.
Finally, a minimum and maximum limit for each design variable can be
set. This may require some previous experience with a car suspension design,
but the following limits for the above car may include the optimal solution:

0 ≤ kf s , krs ≤ 2 kg/mm,

0 ≤ αf , αr ≤ 300 kg/(m/s).

Thus, the above optimal car suspension design problem can be written in
NLP form as follows (Deb and Saxena, 1997):

max abs q2 (t)

Minimize A

subject to

18 − max q2′′′ (t) ≥ 0,

0 ≤ kf s , krs ≤ 2,

0 ≤ αf , αr ≤ 300.

Metal cutting problem

A steel bar is to be machined on a CNC lathe, using P 20 carbide tool (Sardiñas
et al., 2006). The lathe has a 10 kW motor and a transmission efficiency
of 75%. A maximum allowable cutting force of 5, 000 N is applied. It is
16 Optimization for Engineering Design: Algorithms and Examples

considered that the operation will remove 219,912 mm3 of material. The set-
up time, tool-change time and the time during which the tool does not cut
have been assumed as 0.15, 0.20 and 0.05 min, respectively. The objectives are
minimization of operation time (Tp ) and used tool life (ξ). The multi-objective
optimization problem formulation is given below (Deb and Datta, 2012):
Minimize F1 (x) = Tp (x),

Minimize F2 (x) = ξ(x),

subject to
P (x)
g1 (x) ≡ 1 − ≥ 0,
ηP max
Fc (x)
g2 (x) ≡ 1 − ≥ 0,
Fcmax
R(x)
g3 (x) ≡ 1 − ≥ 0,
Rmax
xmin
i ≤ xi ≤ xmax
i ,
where
( 0.20
)
1+ T (x)
Tp (x) = 0.15 + 219,912 + 0.05,
M RR(x)

219,912
ξ(x) = × 100,
M RR(x) T (x)

5.48(109 )
T (x) = ,
v 3.46 f 0.696 a0.460

6.56(103 )f 0.917 a0.1.10

Fc (x) = ,
v 0.286
vFc
P (x) = , M RR(x) = 1,000vf a,
60,000

125f 2
R(x) = .
rn
The objective ξ(x) is considered as the part of the whole tool life which is
consumed in the machining process, hence an operating condition that will
minimize this objective will use the machining task optimally. Here, R is the
surface roughness and rn is the nose radius of the tool. The constant parameter
values are given below:

P max = 10 (kW), Fcmax = 5,000 N, Rmax = 50 µm,

rn = 0.8 mm, η = 0.75.

Introduction 17

Following variable bounds can be used for cutting speed, feed, and depth of
cut:
vmin = 250 m/min, vmax = 400 m/min,
fmin = 0.15 mm/rev, fmax = 0.55 mm/rev,
amin = 0.5 mm, amax = 6 mm.

1.2.2 Modelling
The next major application area of optimization methods is in modelling of
systems and processes that are often faced in engineering studies. Modelling
refers to a task that is used to describe a system or process in a manner
that can be later analyzed to gather further information about the system
or process. Often a crisp and clear idea about the working principle in any
reasonable scientific detail is usually not known for most engineering systems.
Take, for example, modelling a blast furnace in which steel is produced from
raw materials, such as iron ore, sinter, coke, etc. Given all the input materials
and their sizes and compositions that are fed into a blast furnace, the quality
of the produced steel after some hours of its operation cannot be written in
mathematical form in any accurate way. This is because the quality of the raw
materials is not uniform across their charges. Moreover, all heat and mass
transfer phenomena and chemical reactions may not be known appropriately.
In such a scenario, a mathematical modelling of such a system becomes a
difficult task. However, steel has been produced for more than 100 years and
there exists a plethora of data (input raw material information versus output
quality of steel production) with most steel plants. In such a scenario, there
are two types of modelling tasks that can be performed:
(i) Manipulation of governing equations: Whatever scientific
knowledge is available about the system or process, they can be used
by using some modification parameters. For example, if the following
system of (hypothetical) partial differential equations (PDEs) describes
a particular process:
∂E
= a11 x2 + a12 y,
∂x
∂E
= a21 xy 2 + a12 y 3 ,
∂y
then the difference between the predicted E values (obtained by solving
above equations) and the actual E values (obtained from practice) can
be taken care of by modifying the above PDEs as follows:
∂E
= β11 a11 x2 + β12 a12 y,
∂x
∂E
= β21 a21 xy 2 + β22 a12 y 3 .
∂y
18 Optimization for Engineering Design: Algorithms and Examples

In the above PDEs, the β-terms are unknown. We can then formulate
an optimization procedure to find a set of values of β-terms so that the
difference between simulated and observed E values is minimum:
( )2
Minimize f (β) = E observed − E simulated ,

subject to
E simulated = Solution (PDEs).
For a given point β in the search space, the above PDEs can be
numerically (or exactly, if possible) solved and simulated E values can
be found as a function of x and y. These simulated values can then be
compared with observed values from practice and the objective function
f (β) can be calculated. The optimization algorithm should drive the
search towards a solution β ∗ that would correspond to the minimum
possible difference. The advantage with such a procedure is that at the
end a mathematical expression of E(x, y) can be known and a further
analysis can be performed to optimize its performance and get useful
insights about an optimal operation of a system or a plant.
(ii) Use of non-mathematical techniques: Most recent modelling
tasks are performed using a non-mathematical procedure in which
instead of arriving at a mathematical function of the model, the
modelling information is stored in a network of entities or by some other
means. One such common approach is to use artificial neural networks
(ANN) (Haykin, 1999). In an ANN, a network of connectivities
between input and output layers through a series of hidden layers is
established. This is done by first choosing a network of connectivities
and then employing an ANN learning algorithm to arrive at optimal
connection weight for each connection link by minimizing the error
between ANN and observed output data. This process is called the
training phase of ANN. One popular approach is the use of back-
propagation learning algorithm (Rumelhart and McClelland, 1986).
The learning optimization task is achieved through an optimization
procedure. Both the tasks of finding the optimal connectivity and
the optimal connection weights can be achieved by other more
sophisticated ANN methodologies. However, at the end of the ANN
learning process, the optimal network is capable of providing output
value of a new input set accurately, provided the new test set
interpolates the training data well. Thus, such an ANN model can
be used to perform an analysis of the system or process for a better
understanding of the system.
Here, we illustrate a mathematical modelling task as an optimization problem.
Optimal design of an ammonia reactor
In an ammonia reactor design problem, feed gas containing nitrogen,
hydrogen, methane, argon, and a small percentage of ammonia enters the
bottom of the reactor (Figure 1.6).
Introduction 19

Figure 1.6 A typical ammonia synthesis reactor.

Thereafter, the feed gas rises till it reaches the top of the reactor. Then,
while moving downward, the nitrogen and hydrogen present in the feed gas
undergo reaction to form ammonia in the presence of a catalyst placed in the
reactor. The production of ammonia depends on the temperature of the feed
gas, the temperature at the top of the reactor, the partial pressures of the
reactants (nitrogen and hydrogen), and the reactor length. The modelling
task attempts to arrive at a mathematical governing reaction equations so as
to correctly predict certain problem parameters.
In this problem, for a given reactor length x, we identify three design
parameters—the molar flow rate of nitrogen per unit catalyst area NN2 , the
feed gas temperature Tf , and the reacting gas temperature Tg . In order
to maintain the energy balance of reactions in the reactor, three coupled
differential equations must be satisfied (Murase, Roberts, and Converse, 1970;
Upreti and Deb, 1997). First, the decrease in the feed gas temperature must
be according to the heat loss to the reaction gas:

dTf US1
=− (Tg − Tf ). (1.6)
dx WCpf
In Equation (1.6), U is the overall heat transfer coefficient, S1 is the
surface area of the catalyst tubes per unit reactor length, W is the total mass
flow rate, and Cpf is the specific heat capacity of the feed gas. Secondly, the
change in the reaction gas temperature must be according to the heat gain
from the feed gas and heat generated in the reaction:
dTg US1 (−∆H)S2
=− (Tg − Tf ) + fa
dx WCpg WCpg
( )
1.5pN2 pH2 pNH3
× K1 − K2 , (1.7)
pNH3 1.5pH2
20 Optimization for Engineering Design: Algorithms and Examples

where the parameters K1 and K2 are as follows:

K1 = 1.78954 × 104 exp [−20800/(RTg )],

K2 = 2.5714 × 1016 exp [−47400/(RTg )].
The parameter S2 denotes the cross-sectional area of the catalyst zone, ∆H
is the heat of reaction, Cpg is the specific heat capacity of the reacting gas, fa
is the catalyst activity, and R is the ideal gas constant. The parameters pN2 ,
pH2 , and pNH3 are the partial pressures of nitrogen, hydrogen, and ammonia,
respectively. Thirdly, the mass balance of nitrogen yields
( )
dNN2 1.5pN2 pH2 pNH3
= −fa K1 − K2 . (1.8)
dx pNH3 1.5pH2
The partial pressures in the above differential equations are computed as
286NN2
pN2 = ,
2.598NN2 + 2NN2

pH2 = 3pN2 ,

286(2.23NN20 − 2NN2 )
pNH3 = ,
2.598NN20 + 2NN2

where NN20 is the molar flow rate of nitrogen per unit catalyst area at the top
of the reactor. We use the following parameter values:

Cpf = 0.707 kcal/(kg K), S1 = 10 m,

Cpg = 0.719 kcal/(kg K), S2 = 0.78 m,

∆H = −26, 600 kcal/kmol nitrogen, R = 1.987 kcal/(kmol K),

U = 500 kcal/(m2 h K), W = 26, 400 kg/h,

fa = 1.

Note that all the above three ordinary differential equations (ODEs)
(1.6)–(1.8) are coupled to each other. In order to solve these equations, we
use the following boundary conditions:

Tf (x = 0) = T0 ,

Tg (x = 0) = 694 K,

NN2 (x = 0) = 701.2 kmol/(m2 h).

The three constraints (Equations (1.6) to (1.8)) can be used to compute the
three parameters of the problem. By comparing these three parameters with
Introduction 21

their recorded values from an actual reactor, the chemical process can be
modelled. There are various practical reasons why the solutions of the above
ODEs usually do not match with the plant values. By using the objective
of the model optimization problem as minimization of the difference between
simulated parameter values and their actual plant values, we hope to represent
the model of the reactor with the ODEs. Thus, the NLP problem is as follows:
( )2
plant plant 2
Minimize NN2 − NN2 + (Tf − Tf ) + (Tg − Tgplant )2 ,

subject to
dTf US1
= −β1 (Tg − Tf ),
dx WCpf

dTg US1 (−∆H)S2

= −β1 (Tg − Tf ) + fa
dx WCpg WCpg
( )
1.5pN2 pH2 pNH3
× β2 K 1 − β3 K 2 ,
pNH3 1.5pH2
( )
dNN2 1.5pN2 pH2 pNH3
= −fa β4 K1 − β5 K 2 .
dx pNH3 1.5pH2
Thus, the variables for the optimization problem are βi for i = 1, 2, 3, 4, 5. If
the positivity or negativity of some terms in the expressions must be ensured,
adequate variable bounds can be added for the respective βi variable.

1.2.3 Data Fitting and Regression

Data fitting and regression analysis are activities commonly used by scientists,
engineers and managers. We often employ a software to help us find a fitted
curve on a set of data. If thought carefully, the software formulates and solves
an optimization problem to arrive at the fitted curve. For a set of given data
points (say (xi,p , y i,p ) for i = 1, 2, . . . , K) involving input variable set x and
output y, any curve y = f (x) is assigned an objective function as the sum of
the squared difference between the y i,p and f (xi,p ):

∑K ( )
i,p 2
i=1 f (x ) − y
Minimize i,p
. (1.9)

The decision variables are the coefficients of the fitted curve f (x). For a
single-input linear regression (y = ax + b), there are two decision variables a
and b. The minimization procedure will find a and b values that will make the
overall error between the fitted line and the supplied data points minimum.
The same idea can be extended to fit a curve of any sort. The regression task
can be made more generic, by keeping the nature of the curve y = f (x) also
as a part of the variable set. This domain then falls under the category of
modelling of systems which is discussed in the previous subsection.
22 Optimization for Engineering Design: Algorithms and Examples

1.2.4 Control Systems

An optimal control problem involves decision parameters that change with

time or distance or some other varying quantities. The optimization task is
to find variations of decision parameters so as to minimize a performance
metric or maximize quality of output quantities of the system. We consider
an example here.
Overhead cranes are used in the industries, workshops, factories, docks,
and other places to transport heavy components from one place to another. In
order to increase the productivity, such control operations must be optimized
to find what speed the overhead trolley must be moved so that the supplied
energy to the crane and the overall operation time are minimum.
Figure 1.7 shows a schematic model of the crane (Deb and Gupta, 2004).
In this simplified model, the cable connecting the trolley and the hanging
load is considered of fixed length; however, in practice such cables are varied
in length while moving in order to lower or raise the load (Dissanayake and
Fang, 2001). The system has two degrees-of-freedom: (i) x denoting the linear
motion of the trolley along the x-direction and (ii) α denoting the angular
motion of the hanging mass.

Figure 1.7 A schematic of the overhead crane consisting of a trolley and a swaying
load.

In the control system optimization problem, there are two time-varying

control parameters: (i) a time-varying force F (t) applied to the trolley in
the direction of the destination (along positive x-direction) and (ii) a time-
dependent scheme of lowering the load from a high to a low altitude (with a
difference of lT ). Contrary to design optimization problems, here, the variable
vector for the optimization problem is time-varying. Thus, instead of a single
Introduction 23

value of force F , we need to identify a profile of F as a function of time t that

would optimize the system.
By considering the force balances in x and y directions of all forces acted
on the trolley, we obtain the following two equations:
M ẍ = F + 2T sin(α) − µN,

N = M g + 2T cos(α),

where T is twice the tension in each cable, M is the mass (in kg) of the trolley,
and ẍ is the acceleration of the trolley in the x-direction. Performing a similar
task for the hanging load (of mass m kg), we have the following two equations:
−2T sin(α) − cẋ1 = mx¨1 ,

2T cos(α) − cẏ1 − mg = my¨1 ,

where c is the coefficient of damping arising due to several factors on to the

cable and the hanging mass. The variables x1 and y1 are the displacement
of the hanging load in the x and y directions, respectively. In addition, the
following relationships between trolley and the hanging load motions can be
written with variables along x and y directions:

x1 = x + lo sin(α), y1 = −lo cos(α),

ẋ1 = ẋ + lo α̇ cos(α), ẏ1 = lo α̇ sin(α),

ẍ1 = ẍ + lo α̈ cos(α) − lo α̇2 sin(α). ÿ1 = lo α̈ sin(α) + lo α̇2 cos(α).

Here, lo is the length of the cable, α̇ and α̈ are the angular velocity and
acceleration of the cable. By eliminating T , x1 and y1 from the above
expressions, we get the following two equations of motion of the trolley and
the hanging mass:
[
ẍ = F − cẋ sin2 (α) + mlo sin(α)α̇2 + mg sin(α) cos(α)
]
−f (mlo cos(α)α̇2 − cẋ sin(α) cos(α) − mg sin2 (α))

/(M + m sin2 (α) − f m sin(α) cos(α)), (1.10)

cos(α)
α̈ = − (ẍ + rẋ + g tan(α)) − rα̇, (1.11)
lo
where r is the ratio of c to m. These two equations can be solved using a
numerical integration technique and the variation of x and α with time t can
be found.
A little thought of the problem makes it clear that the two objectives
(i) total energy supplied to the system and (ii) the total time for the block-
load system to reach the desired position and stabilize are the two conflicting
24 Optimization for Engineering Design: Algorithms and Examples

objectives. The supplied energy will be minimum for the case of moving ever
slowly towards the destination. But such a solution will require quite a long
time to complete the task. On the other hand, reaching the destination with
a large velocity and suddenly stopping at the destination would be a quick
way to reach the destination; however, some time needs to be elapsed for
the sway of the load to diminish. Although such a solution may not be the
quickest overall time solution, there would exist a solution with a reasonable
velocity which would minimize the overall time. Ideally, this is a two-objective
problem, but for the sake of our discussion here, we may formulate a single-
objective time minimization of the crane maneuvering problem:

Minimize Time = T ,
subject to

(x(T ), y(T ))T = (L, lT )T ,

α(T ) ≤ ϵ.

Values x(T ) and α(T ) can be found by solving the differential equations given
in Equations (1.10) and (1.11) from suitable initial conditions. The parameter
ϵ is a user-specified small sway angle for a stable system at time T . The
decision variables of this problem are F (t) and ℓ(t).

1.2.5 Inverse Problems

Many problems in practice are solved as an inverse problem, as the forward

problem in such cases is difficult to formulate and solve. For example, in
a number of practical applications it is not only desirable but necessary
to estimate the internal structure of a given object without destroying the
object’s internal structure. For example, in medical science applications, one
is interested in diagnosis of the health of internal organs in a non-invasive
manner. Similarly, an engineer is interested in estimating the presence of
internal defects/inclusions inside a structure by what is called non-destructive
evaluation (NDE) methods. Reconstruction of the cross-sectional view of an
object, without destroying the object’s structure, forms the basis for computed
tomography. Tomographic reconstruction is an inverse problem wherein the
task is to obtain the parameters of a model from observed or measured data.
Tomography makes use of various physical phenomena for collection of
data which serves as input to the reconstruction algorithm chosen. The
selection of a specific physical phenomena depends on the materials in the
specimen being evaluated. Ultrasonic tomography (UT) has been in use for a
long period now and offers the most convenient method for flaw detection in
a material (Rose, 2002). To simulate the time of flight (TOF) of ultrasound
rays, the specimen under consideration is represented as a grid of certain
integer values corresponding to different materials. A number of ultrasound
sources and detectors are positioned around the specimen in what is termed
Introduction 25

as modified cross-hole geometry (Khare et al., 2007). For any configuration,

the sources are actuated in sequence one at a time, and from each of these
ultrasound rays travel to each of the detectors, giving us actual TOF data.
For example, considering 6 sources, 6 detectors and 6 configurations, we
get R = 216 readings, which is essentially the input to our reconstruction
algorithm. The ray coverage for this case is shown in Figure 1.8(a). The
TOF is taken to be the arrival time corresponding to the first peak of the
signal sensed by the detector as shown in Figure 1.8(b). If experimental data
is available for these set of sources and detectors, the same can be used.
Otherwise, the TOF can be calculated by using the following procedure.

Figure 1.8 (a) Ray coverage, (b) Ultrasound signal indicating TOF.

The simulation of TOF assumes that the ultrasound rays follow straight
paths from source to the detector. This assumption is reasonable when the
impedance mismatch within the specimen is small.
A solution to this problem is a configuration in which some cells are
occupied by foreign materials or voids (Kodali et al., 2008). The task of the
optimization is then to identify the cells with foreign materials or voids that
minimize the error between actual TOF and simulated TOF. The simulated
TOF for a ray r originating from j-th source and terminating at the k-th
detector is estimated using Equation (1.12):
∑M (r)
lm (j, k)
TOFsimulated (r) = , (1.12)
m=1
vm

where
M = number of cells intercepted by the ray,
(r)
lm (j, k) = length of ray intercepted by m-th cell along the ray path,
vm = velocity of propagation of ultrasound through the m-th cell.

The process of reconstruction is an optimization task which consists of two

major steps, the first is acquisition of TOF data and the second is using the
26 Optimization for Engineering Design: Algorithms and Examples

acquired data to reconstruct the specimen under consideration. The following

optimization problem can be formulated:
R (
∑ )2
Minimize Error = TOFsimulated (r) − TOFactual (r) , (1.13)
r=1

where R is the number of rays that are considered. The optimization procedure
will then enable finding the size and location of voids or foreign material
embedded in a component.

1.2.6 Scheduling and Routing

Scheduling and routing problems are different kinds of optimization tasks

than what we have discussed so far. They are popularly faced in practice.
For example, cases such as finding an optimal schedule for a classroom time-
tabling, or an examination scheduling, or airline time-tabling, or nurse time-
tabling in hospitals are some such problems that are commonly faced in
practice. Job-shop scheduling in which the different machining tasks (or jobs)
need to be performed on a set of machines and the optimization task is to
find a schedule of tasks so that overall time of completion of the machining
process is minimum.
A travelling salesperson problem is described in Figure 1.9. A salesperson
has to travel to various cities to sell a product. The salesperson will travel
to each city only once and must visit all given cities. In such a problem, the
task is to travel to all given cities in such a way that will minimize the overall
distance covered by the salesperson. Instead of representing variables as xi ,
in scheduling problems, the edge between two quantities is important. For
example, if the salesperson has to travel between three cities A, B, and C,
alternative solutions are as follows: ABC, ACB, BAC, BCA, CAB and CBA.
The solution ABC means that the salesperson will start from city A and then
move to city B and then to city C. In these solutions, the absolute location of
city A, B, or C is unimportant, rather the edge between A and B or B and A is
an important matter. This book does not address the scheduling optimization
problems; however, more information can be found in related texts (Brucker,
2010; Pinedo, 2008).

Optimal design of a transit schedule

Figure 1.10 shows a typical transit system network. The solid lines represent
different routes, the points on the lines represent the stops and the circled
intersections of the routes represent the transfer stations. The problem is to
determine schedules for the routes such that the transit system provides the
best level of service (LOS) to its passengers, within the resources available.
One of the good measures of the LOS is the amount of time passengers wait
during their journey—the lesser the waiting time, the better is the LOS
(Chakroborty et al., 1995). On any transit network, either the passengers
wait to board the vehicle at the station of origin or they wait at a transfer
Introduction 27

Figure 1.9 A typical schedule for a travelling salesperson problem with some
Indian cities.

station at which they transfer from one vehicle to another. For example, a
passenger wishing to travel from station A to station B (in the network shown
in Figure 1.10) will have to wait at station A to board a vehicle on Route 1.
Further, the passenger will have to wait at transfer station C to board a vehicle
on Route 3 (which will take him/her to the destination). We will refer to the
wait at station A as the initial wait time (IWT) and the wait at station C
as the transfer time (TT). A good schedule is one which minimizes the sum
of IWT and TT for all passengers. Thus, the optimization problem involves
finding a schedule of vehicles on all routes (arrival and departure times) such
that the total waiting time for the passengers is minimum.
28 Optimization for Engineering Design: Algorithms and Examples

Figure 1.10 A typical transit system network.

The design variables in this problem are the arrival time aki and departure
time dki for the k-th vehicle at the i-th route. Thus, if in a problem, there are
a total of M routes and each route has K vehicles, the total number of design
variables is 2MK. In addition, there are a few more artificial variables which
we shall discuss later.
The constraints in this problem appear from different service-related
limitations. Some of these constraints are formulated in the following:
Minimum stopping time: A vehicle cannot start as soon as it stops; it has to
wait at the stop for a certain period of time, or

(dki − aki ) ≥ smin for all i and k.

Maximum stopping time: A vehicle cannot stop for more than a certain period
of time even if it means increasing the total transfer time on the network, or

(dki − aki ) ≤ smax for all i and k.

Maximum allowable transfer time: No passenger on the transit network

should have to wait more than a certain period of time T at any transfer
station. This can be enforced by checking all possible differences between
departure and arrival times and limiting those values to T . Mathematically,
this is difficult to achieve. We simplify the formulation of this constraint
k,l
by introducing a new set of variables δi,j between the k-th vehicle of
the i-th route and the l-th vehicle of the j-th route. These variables can
take either a zero or a one. A value of zero means that the transfer of
passengers between those two vehicles is not feasible. A value of one means
otherwise. Consider the arrival and departure times of vehicles in two different
routes at a particular station, as shown in Figure 1.11. The parameters
aki and dki are the arrival and departure times of the k-th vehicle in the
i-th route. A passenger from the k-th vehicle in the i-th route can only
transfer to a vehicle in the j-th route, which is arriving at the station after
aki . According to the figure, the transfer of a passenger from the k-th vehicle
Introduction 29

Figure 1.11 Transfers from the k-th vehicle in the i-th route to three consecutive
vehicles in the j-th route.

in the i-th route is not possible to the (l − 1)-th vehicle in the j-th route,
because the departure time of the latter vehicle (dl−1
j ) is earlier than that of
k,l−1 k,l
aki . Thus, the parameter δi,j takes a value zero, whereas the parameter δi,j
takes a value one. In order to simplify the model, we assume that transfers to
vehicles departing after the l-th vehicle in the j-th route are also not possible.
k,q
All parameters δi,j for q = (l + 1), (l + 2), . . . are also zero. Thus, between
any two vehicles, the following condition must be satisfied:
k,l
(dlj − aki )δi,j ≤T for all i, j, k and l.

It is clear that the left side expression of the above condition is zero for those
transfers that are not feasible. Since transfers only to the next available vehicle
k,l
are assumed, only one δi,j (for l = 1, 2, . . .) is one and the rest all are zeros
for fixed values of i, j, and k. Mathematically,
∑ k,l
δi,j = 1 for all i, j and k.
l

k,l
The introduction of the artificial variables δi,j makes the formulation easier,
but causes a difficulty. Many optimization algorithms cannot handle discrete
k,l
design variables efficiently. Since the artificial design variables δi,j can only
take a value zero or one, another set of constraints is added to enforce the
binary values:
k,l
(dlj − aki ) + M (1 − δi,j )≥0 for all i, j, k and l,

where M is a large positive number. The above constraint ensures that the
k,l
variable δi,j always takes a value one whenever a transfer is possible and the
value zero whenever a transfer is not possible. This constraint is derived purely
from the knowledge of the available optimization algorithms. There may be
other ways to formulate the concept of feasible transfers, but inclusion of such
artificial design variables often makes the understanding of the problem easier.
Maximum headway: The headway between two consecutive vehicles should be
less than or equal to the policy headway, hi , or

(ak+1
i − aki ) ≤ hi for all i and k.
30 Optimization for Engineering Design: Algorithms and Examples

The objective function consists of two terms: the first term represents
the total transfer time (TT) over all the passengers and the second term
represents the initial waiting time (IWT) for all the passengers. The objective
is to minimize the following function:
∑∑∑∑ k,l l
δi,j (dj − aki )wi,j
k
i j k l

∫ k−1
i −ai
ak
∑∑ [ ]
+ vi,k (t) (aki − ak−1
i ) − t dt.
i l 0

k
The parameter wi,j is the number of passengers transferring from the k-th
vehicle of the i-th route to the j-th route. The first term is obtained by
summing the individual transfer time (dlj − aki ) over all passengers for all
the vehicles for every pair of routes. The parameter vi,k (t) is the number of
passengers arriving at the stop for the k-th vehicle in the i-th route at a given
time t. Since the arrival time for passengers can be anywhere between t = 0
and t = (aki −ak−1
i ) (the headway), the initial waiting time also differs from one
passenger to another. For example, a passenger arriving at the stop just after
the previous vehicle has left has to wait for the full headway time (aki − ak−1
i )
before the next vehicle arrives. On the other hand, a passenger arriving at the
stop later has to wait for a shorter time. The calculation of the second term
assumes that passengers arrive at the stop during the time interval ak−1 i to
aki according to the known time-varying function vi,k (t), where t is measured
from ak−1
i . Then the quantity
∫ k−1
i −ai
ak [ ]
vi,k (t) (aki − ak−1
i ) − t dt
0

gives the sum of the initial waiting times for all passengers who board the
k-th vehicle of the i-th route. We then sum it over all the routes and vehicles
to estimate the network total of the IWT. Thus, the complete NLP problem
can be written as follows:
∑ ∑ ∑ ∑ k,l l
Minimize δi,j (dj − aki )wi,j
k
i j k l

∫ k−1
i −ai
ak
∑∑ [ ]
+ vi,k (t) (aki − ak−1
i ) − t dt
i l 0

subject to
smax − (dki − aki ) ≥ 0 for all i and k,

(dki − aki ) − smin ≥ 0 for all i and k,

k,l
T − (dlj − aki )δi,j ≥0 for all i, j, k and l,
Introduction 31

k,l
(dlj − aki ) + M (1 − δi,j )≥0 for all i, j, k and l,

hi − (ak+1
i − aki ) ≥ 0 for all i and k,
∑ k,l
δi,j = 1 for all i, j and k.
l

k,l
In the above NLP problem, the variables δi,j are binary variables taking
only a value zero or a one and other variables aki and dki are real-valued.
Thus, a mixed integer programming technique described in Chapter 5 or
genetic algorithms described in Chapter 6 can be used to solve the above
NLP problem (Chakroborty et al., 1995).

1.2.7 Data Mining

Data mining activities have become an inseparable part of any scientific work,
mainly due to the availability and ease of generation of data. When the data
generation task is expensive or tedious, data is usually generated in parallel
among various research groups and they are shared by scientists. In such an
activity, the first task is often to cluster the data into several categories to
establish which data from what source are related to each other. Optimization
is usually a tool employed for such a clustering activity. Here, we discuss such
a procedure.
Let us consider a two-dimensional data shown in Figure 1.12. When the
data is achieved, they are all jumbled together. One way to formulate an
optimization problem is to first decide on a number of clusters that the data
may be divided to. Then, for each cluster, a few parameters are used variables
for the optimization task. The parameters describe an affinity function for a
cluster. Say, for example, we define a simple affinity function for the i-th
cluster, as follows:

Figure 1.12 Clustering procedure is illustrated.

32 Optimization for Engineering Design: Algorithms and Examples

(x − xi )2 (y − yi )2
ci (x, y) = + .
a2i b2i

This is an equation of an ellipse with centre at (xi , yi ) and major and minor
axis values as ai and bi , respectively. Next, each data point (x, y) is tested for
each cluster affinity function Ci . The point is associated with that cluster for
which the affinity function value is minimum. This procedure will associate
every data point with a particular cluster. Point A (in the figure) may get
associated with Cluster 2.
Next, an intra-cluster distance (Diintra ) between all associated points of
each cluster is computed as follows:
 
1  ∑ √
Diintra = (x − xi )2 + (y − yi )2 . (1.14)
|Ci |
(x,y)∈Ci

Further, an inter-cluster distance (Diinter ) is computed as follows:

 
∑K ∑
K √
1
Diinter = (K )  (xi − xj )2 + (yi − yj )2 . (1.15)
2 i=1 j=1,j̸=i

Then, the following optimization problem is solved:

Diintra
Minimize ,
Dinter
i

subject to

(xi , yi ) ∈ R,
where R is the specified range in which the centres of ellipses are supposed to
lie. The minimization will ensure that intra-cluster points are as close to the
centre as possible and all cluster centres are as far away from each other as
possible, thereby achieving the goal of clustering.
Prediction is another data mining activity which is used on a daily basis by
weather forecasters, share market predictors, etc. The task here is to model
the past few days (or weeks or months) of data and then extrapolate the
model to predict how the future is going to be for the next hours, days or
weeks. Since the modelling activity can be achieved through an optimization
problem-solving (as discussed before), the prediction activity can be achieved
using an optimization procedure.

1.2.8 Intelligent System Design

Many systems in practice are required to be designed in a manner that

may show certain intelligence while performing a task. Since optimization
Introduction 33

algorithms work at finding the optimal solution in the search space, which
is not any arbitrary solution, they can be ideal candidates for arriving at an
intelligent solution. For example, consider the robot path planning problem in
which the task is to design the ‘brain’ of the robot which will help it to avoid
obstacles that may be moving and still reach a destination with minimum
time or using minimum energy or optimizing some other criterion.
The use of an optimization method to solve such problems causes an inher-
ent problem. The optimal path depends on how the obstacles will be moving
in a real scenario and unfortunately this information may not be known a
priori. One way to address such a task is to first develop an optimal ‘rule
base’ by performing an off-line optimization on several training cases. Once
the optimal rule-base is developed, the robot can then utilize it to navigate
in a real scenario. If sufficient training cases are considered, a fairly robust
rule-base can be developed. Figure 1.13 depicts the above-discussed off-line
optimization-based approach (Pratihar et al., 1999). On a set of instantia-
tions, an optimization algorithm is applied to find a knowledge base using
rules or by other means. The optimization task would find a set of rules or

Figure 1.13 Off-line optimization procedure is illustrated.

classifiers which will determine the nature of the outcome based on the vari-
able values at any time instant. In the following, we describe the procedure
in the context of an on-line robot navigation problem. There are essentially
two parts of the problem:
(i) Learn to find any obstacle-free path from point A to B, and
(ii) Learn to choose that obstacle-free path which takes the robot in a
minimum possible time.
This process of avoiding an object can be implemented using a rule of the
following sort:
If an object is very near and is approaching, then turn right to the original path.
34 Optimization for Engineering Design: Algorithms and Examples

Figure 1.14 A schematic showing condition and action variables for the robot
navigation problem.

The angle and distance of a robot from a critical obstacle and its decision
on deviation, if any, are illustrated in Figure 1.14. Because of the imprecise
definition of the deviation in this problem, it seems natural to use a fuzzy logic
technique (Mendel, 2000) here. Table 1.2 shows a set of possible fuzzy rules.
The (1,1) position of the table suggests that if an obstacle is very near (VN)
to the robot and it is left (L) of the robot, the decision is to move ahead (A).

Table 1.2 A possible rule-base

Angle

L AL A AR R

VN A AR AL AL A
Distance

N A A AL A A

F A A AR A A

VF A A A A A

This way, the rule-base shows every possible decision that the robot should
take in navigating in a real scenario. Finding what decision to take in every
combination of angle and distance is the task of an off-line optimization
algorithm. To extend the scope, the presence or absence of a rule for every
combination of angle and distance may be considered as another kind of
decision variable. The fuzzy rules involve membership functions that define
the attributes such as the extent of distance values for VN, N, etc. and also A,
L, etc. Triangular membership functions are commonly used for this purpose.
Introduction 35

This involves another set of decision variables as the width of the triangles
(such as parameters b1 and b2 shown in Figure 1.15). Thus, when a rule-

Figure 1.15 A possible set of membership functions.

base (Table 1.2) is specified completely, it can be used to simulate a robot

for different training scenarios. The overall time required from start to finish
can then be recorded from the simulation process and whether the robot
collides with any obstacle or not can be determined. The resulting off-line
optimization problem then becomes as follows:
Minimize Time (rule-base),

subject to
Path (rule-base) is obstacle-free.
Note that both Time and Path are functions of the rule-base and the elements
of the rule-base are decision variables of the optimization task.

1.3 Classification of Optimization Algorithms

The above optimization problems reveal the fact that the formulation of
engineering design problems could differ from problem to problem. Certain
problems involve linear terms for constraints and objective function but
certain other problems involve nonlinear terms for them. In some problems,
the terms are not explicit functions of the design variables. Unfortunately,
there does not exist a single optimization algorithm which will work in all
optimization problems equally efficiently. Some algorithms perform better on
36 Optimization for Engineering Design: Algorithms and Examples

one problem, but may perform poorly on other problems. That is why the
optimization literature contains a large number of algorithms, each suitable to
solve a particular type of problem. The choice of a suitable algorithm for an
optimization problem is, to a large extent, dependent on the user’s experience
in solving similar problems. This book provides a number of optimization
algorithms used in engineering design activities.
Since the optimization algorithms involve repetitive application of certain
procedures, they need to be used with the help of a computer. That is why the
algorithms are presented in a step-by-step format so that they can be easily
coded. To demonstrate the ease of conversion of the given algorithms into
computer codes, most chapters contain a representative working computer
code. Further, in order to give a clear understanding of the working of
the algorithms, they are hand-simulated on numerical exercise problems.
Simulations are performed for two to three iterations following the steps
outlined in the algorithm sequentially. Thus, for example, when the algorithm
suggests to move from Step 5 to Step 2 in order to carry out a conditional
statement, the exercise problem demonstrates this by performing Step 2 after
Step 5. For the sake of clarity, the optimization algorithms are classified into
a number of groups, which are now briefly discussed.
Single-variable optimization algorithms. Because of their simplicity, single-
variable optimization techniques are discussed first. These algorithms provide
a good understanding of the properties of the minimum and maximum
points in a function and how optimization algorithms work iteratively to
find the optimum point in a problem. The algorithms are classified into two
categories—direct methods and gradient-based methods. Direct methods do
not use any derivative information of the objective function; only objective
function values are used to guide the search process. However, gradient-based
methods use derivative information (first and/or second-order) to guide the
search process. Although engineering optimization problems usually contain
more than one design variable, single-variable optimization algorithms are
mainly used as unidirectional search methods in multivariable optimization
algorithms.
Multivariable optimization algorithms. A number of algorithms for uncons-
trained, multivariable optimization problems are discussed next. These
algorithms demonstrate how the search for the optimum point progresses in
multiple dimensions. Depending on whether the gradient information is used
or not used, these algorithms are also classified into direct and gradient-based
techniques.
Constrained optimization algorithms. Constrained optimization algorithms
are described next. These algorithms use the single-variable and multivariable
optimization algorithms repeatedly and simultaneously maintain the search
effort inside the feasible search region. Since these algorithms are mostly used
in engineering optimization problems, the discussion of these algorithms covers
most of the material of this book.
Introduction 37

Specialized optimization algorithms. There exist a number of structured

algorithms, which are ideal for only a certain class of optimization
problems. Two of these algorithms—integer programming and geometric
programming—are often used in engineering design problems and are
discussed. Integer programming methods can solve optimization problems
with integer design variables. Geometric programming methods solve
optimization problems with objective functions and constraints written in
a special form.
There exist quite a few variations of each of the above algorithms. These
algorithms are being used in engineering design problems since sixties. Because
of their existence and use for quite some years, we call these algorithms as
traditional optimization algorithms.
Nontraditional optimization algorithms. There exist a number of other search
and optimization algorithms which are comparatively new and are becoming
popular in engineering design optimization problems in the recent past. Two
such algorithms—genetic algorithms and simulated annealing—are discussed
in this book.
We have put together about 36 different optimization algorithms. Over
the years, researchers and practitioners have modified these algorithms to suit
their problems and to increase the efficiency of the algorithms. However, there
exist a few other optimization algorithms—stochastic programming methods
and dynamic programming method—which are very different than the above
algorithms. Because of the space limitation and occasional use of these
algorithms in engineering design problems, we have not included them in
this book. A detailed discussion of these algorithms can be found elsewhere
(Rao, 1984).
Many engineering optimization problems contain multiple optimum
solutions, among which one or more may be the absolute minimum or
maximum solutions. These absolute optimum solutions are known as global
optimal solutions and other optimum solutions are known as local optimum
solutions. Ideally, we are interested in the global optimal solutions because
they correspond to the absolute optimum objective function value. An
algorithm that has a perspective of finding an optimal solution close to the
initial starting solution is known as a local search algorithm. Most traditional
optimization algorithms described in this book have this local search property.
Starting with a single initial point, the algorithms try to find an optimal
solution for which either the derivative vector is close to a zero vector or the
KKT necessary conditions are satisfied. These algorithms do not guarantee
finding the global optimal solution. However, nontraditional algorithms, such
as genetic algorithms and simulated annealing algorithm, are found to have a
better global perspective than the traditional methods. The global optimality
issues are discussed in Chapter 6.
Moreover, when an optimal design problem contains multiple global
solutions, designers are interested in finding not only just one global optimum
solution, but as many as possible for various reasons. The optimization
38 Optimization for Engineering Design: Algorithms and Examples

algorithms that attempt to find multiple optimal solutions are called

multimodal optimization algorithms (Deb and Goldberg, 1989). There are
several practical reasons for finding multiple optimal solutions. Firstly,
a design suitable in one situation may not be valid in another situation.
Secondly, it is also not possible to include all aspects of the design in
the optimization problem formulation. Thus, there always remains some
uncertainty about the obtained optimal solution. Thirdly, designers may not
be interested in finding the absolute global solution, instead may be interested
in a solution which corresponds to a marginally inferior objective function
value but is more amenable to fabrication. Thus, it is always prudent to know
about other equally good solutions for later use. However, if the traditional
methods are used to find multiple optimal solutions, they need to be applied
a number of times, each time starting from a different initial solution and
hoping to achieve a different optimal solution each time. Genetic algorithms
described in Chapter 6 allow an easier way to find multiple optimal solutions
simultaneously in a single simulation.
Another class of optimization problems deals with simultaneous
optimization of multiple objective functions. In formulating an optimal design
problem, designers are often faced with a number of objective functions.
For example, the truss structure problem described earlier should really be
reformulated as the minimization of both the weight of the truss and the
deflection at the point C. Multiobjective optimization problems give rise to
a set of optimal solutions known as Pareto-optimal solutions (Chankong and
Haimes, 1983), all of which are equally important as far as all objectives
are concerned. Thus, the aim in these problems is to find as many Pareto-
optimal solutions as possible. Because of the complexity involved in the
multiobjective optimization algorithms, designers usually choose to consider
only one objective and formulate other objectives as constraints. Genetic
algorithms described in Chapter 6 demonstrate one way to handle multiple
objectives and help find multiple Pareto-optimal solutions simultaneously.
Many engineering design problems have a multi-level structure. For
example, in a bi-level optimization problem, there are two levels of
optimization problems. An optimal solution to the lower-level optimization
problem becomes a feasible solution to the upper-level optimization problem.
In such a nested optimization task, in which the evaluation of a solution on
the upper-level optimization problem requires solving another optimization
problem (lower-level problem), unless the optimal solution of the lower-level
problem is found, the evaluation of the upper-level solution is meaningless.
A broom-balancing problem is a good example of a bi-level optimization
problem. The strategy to control the motion of the platform for minimum
energy requirement is the upper-level optimization problem and once a
strategy is given, finding whether the broom will get balanced in a minimum
time is the lower-level optimization problem. Thus, unless the broom is
balanced and it does it in the minimum possible time, the strategy cannot be
evaluated to find the energy it demands in the upper-level. Such problems
are common in engineering and efficient optimization algorithms are needed
Introduction 39

to solve such problems (Deb and Sinha, 2010). Tri-level and higher-level
optimization problems are also possible.
Most optimization problems discussed in this chapter are deterministic
in nature, meaning that the objective function, constraint function or
parameters are fixed from the beginning till the end of the optimization run.
These problems are called deterministic optimization problems. However, as
discussed somewhat in Section 1.2.8, some problems may have dynamically
changing objective function or constraint function or parameters. While the
optimization run is in process, one or a few of the functions and parameters
may change. This happens usually if a system depends on the environment
(ambient temperature or humidity) and the process continues over most of the
day. The ambient condition becomes different from morning to noon. Thus,
the solution that was optimal during morning may not remain optimal in
the noon time. Such problems are called stochastic optimization problems,
or sometimes dynamic optimization problems as well. Although an off-line
optimization, as discussed in Section 1.2.8, is one approach, recent dynamic
optimization methodologies (Deb et al., 2007) are other viable and pragmatic
approaches.
At the end of the optimization process, one obvious question may arise:
Is the obtained solution a true optimum solution? Unfortunately, there is
no easy answer to this question for an arbitrary optimization problem. In
problems where the objective functions and constraints can be written in
simple, explicit mathematical forms, the Kuhn-Tucker conditions described
in Chapter 4 may be used to check the optimality of the obtained solution.
However, those conditions can only be used for a few classes of optimization
problems. In a generic problem, this question is answered in a more practical
way. In many engineering design problems, a good solution is usually known
either from the previous studies or from experience. After formulating the
optimal problem and applying the optimization algorithm if a better solution
than the known solution is obtained, the new solution becomes the current
best solution. The optimality of the obtained solution is usually confirmed by
applying the optimization algorithms a number of times from different initial
solutions.
Having discussed different types of optimization algorithms that exist to
solve different kinds of optimization problems, we suggest that for practical
applications there is always a need for customizing an optimization algorithm
for solving a particular problem. The reasons are many. First, there may
not exist a known efficient optimization algorithm for a particular problem at
hand. Second, the size and complexity of solving a problem using an algorithm
may be computationally too expensive for it to be pragmatic. Third, in
many scenarios, the user may not be interested in finding the exact optimal
solution, rather a reasonably good (or better than an existing solution)
may be adequate. In such scenarios, knowledge (or problem information)
about the current problem may be used to design a customized optimization
algorithm that is appropriate for the current problem (Deb, Reddy and Singh,
2003). Such optimization methods are often known as heuristic optimization
40 Optimization for Engineering Design: Algorithms and Examples

methods. This requires that an algorithm is flexible to be updated with

problem information. Mathematical optimization methods are difficult to
update with heuristic information; however, nontraditional methods, such as
genetic algorithms, simulated annealing, and others, are flexible in this regard
and are getting increasingly applied in practice for this singular reason.

1.4 Summary
In order to use optimization algorithms in engineering design activities, the
first task is to formulate the optimization problem. The formulation process
begins with identifying the important design variables that can be changed
in a design. The other design parameters are usually kept fixed. Thereafter,
constraints associated with the design are formulated. The constraints may
arise due to resource limitations such as deflection limitations, strength
limitations, frequency limitations, and others. Constraints may also arise due
to codal restrictions that govern the design. The next task is to formulate
the objective function which the designer is interested in minimizing or
maximizing. The final task of the formulation phase is to identify some
bounding limits for the design variables.
The formulation of an optimization problem can be more difficult than
solving the optimization problem. Unfortunately, every optimization problem
requires different considerations for formulating objectives, constraints, and
variable bounds. Thus, it is not possible to describe all considerations in a
single book. However, many of these considerations require some knowledge
about the problem, which is usually available with the experienced designers
due to their involvement with similar other design problems.
The rest of the book assumes that the formulation of an optimization
problem is available. Chapters 2 to 6 describe a number of different
optimization algorithms—traditional and nontraditional—in step-by-step
format. To demonstrate the working of each algorithm, hand-simulations on
a numerical example problem are illustrated. Sample computer codes for a
number of optimization algorithms are also appended to demonstrate the ease
of conversion of other algorithms into similar computer codes.

REFERENCES

Brucker, P. (2010). Scheduling algorithms. Heidelberg, Germany: Springer.

Chakroborty, P., Deb, K., and Subrahmanyam, P. (1995). Optimal scheduling
of urban transit systems using genetic algorithms. ASCE Journal of
Transportation Engineering, 121(6), 544–553.
Chankong, V. and Haimes, Y. Y. (1983). Multiobjective Decision Making
Theory and Methodology. New York: North-Holland.
Introduction 41

Deb, K. and Datta, R. (2012). Hybrid evolutionary multi-objective opti-

mization and analysis of machining operations. Engineering Optimization,
44(6), 685–706.
Deb, K. and Goldberg, D. E. (1989). An investigation of niche and species
formation in genetic function optimization, In J. D. Schaffer (Ed.),
Proceedings of the Third International Conference on Genetic Algorithms,
(Washington DC, USA), 42–50.
Deb, K. and Gupta, N. (2004). Optimal operating conditions for overhead
crane maneuvering using multi-objective evolutionary algorithms. Pro-
ceedings of the Genetic and Evolutionary Computation Conference, GECCO-
2004 (Seattle, USA). 1042–1053. (Also Lecture Notes in Computer Science
(LNCS) 3102).
Deb, K., Khan, N. and Jindal, S. (2000). Optimal truss-structure design
for multiple objectives. Tenth National Seminar on Aerospace Structures,
168–180.
Deb, K., Rao, U. B., and Karthik, S. (2007). Dynamic multi-objective
optimization and decision-making using modified NSGA-II: A case study
on hydro-thermal power scheduling bi-objective optimization problems.
Proceedings of the Fourth International Conference on Evol. Multi-
Criterion Optimization (EMO-2007), (LNCS, Springer) 803–817.
Deb, K., Reddy, A. R., and Singh, G. (2003). Optimal scheduling of
casting sequence using genetic algorithms. Journal of Materials and
Manufacturing Processes, 18(3), 409–432.
Deb, K. and Saxena, V. (1997). Car suspension design for comfort using
genetic algorithms. In Thomas Back (ed.), Proceedings of the Seventh
International Conference on Genetic Algorithms (ICGA-97), 553–560.
Deb, K. and Sinha, A. (2010). An efficient and accurate solution
methodology for bilevel multi-objective programming problems using a
hybrid evolutionary-local-search algorithm. Evolutionary Computation
Journal, 18(3), 403–449.
Dissanayake, G. and Fang, G. (2001). Minimum-time trajectory for reducing
load sway in quay-cranes. Engineering Optimization, 33, 643–662.
Haykin, S. S. (1999). Neural Networks: A Comprehensive Foundation. New
Delhi: Prentice-Hall of India.
Khare, S., Razdan, M., Munshi, P., Sekhar, B. V., and Balasubramaniam,
K. (2007). Defect detection in carbon-fiber composites using Lamb-wave
tomographic methods. RNDE, 18, 101–119.
Kodali, S. P., Bandaru, S., Deb, K., Munshi, P., and Kishore, N. N. (2008).
Applicability of Genetic Algorithms to Reconstruction of Projected Data
from Ultrasonic Tomography. Proceedings of Genetic and Evolutionary
Computation Conference (GECCO-2008), (Atlanta, USA), 1705–1706.
Mendel, J. M. (2000). Uncertain Rule-Based Fuzzy Logic Systems:
Introduction and New Directions, New Delhi: Prentice-Hall of India.
42 Optimization for Engineering Design: Algorithms and Examples

Murase, A., Roberts, H. L., and Converse, A. O. (1970). Optimal thermal

design of an autothermal ammonia synthesis reactor. Ind. Eng. Chem.
Process Des. Develop., 9, 503–513.
Pinedo, M. (2008). Scheduling: Theory, Algorithms, and Systems. New York:
Springer.
Pratihar, D., Deb, K., and Ghosh, A. (1999). Fuzzy-genetic algorithms and
time-optimal obstacle-free path generation for mobile robots. Engineering
Optimization, 32, 117–142.
Quiza Sardiñas, R., Rivas Santana, M., and Alfonso Brindis, E.
(2006). Genetic algorithm-based multi-objective optimization of cutting
parameters in turning processes. Engineering Applications of Artificial
Intelligence, 19(2), 127–133.
Rao, S. S. (1984). Optimization Theory and Applications. New Delhi: Wiley
Eastern.
Rose, J. L. (2002). A baseline and vision of ultrasonic guided wave inspection
potential. Journal of Pressure Vessel Technology, 124, 273–282.
Rumelhart, D. E. and McClelland, J. L. (1986). Parallel Distributed
Processing: Explorations in the Microstructure of Cognition. Volume I.
Cambridge, MA: MIT Press.
Shigley, J. E. (1986). Mechanical Engineering Design. New York: McGraw-
Hill.
Timoshenko, S. (1986). Strength of Materials, Part 1: Elementary Theory
and Problems. Delhi: CBS Publishers.
Upreti, S. and Deb, K. (1997). Optimal design of an ammonia synthesis
reactor using genetic algorithms. Computers & Chemical Engineering,
21(1), 87–92.