University of Technology

Building and Construction

Engineering Department

OPTIMIZATION OF RAFT
FOUNDATION DESIGN

A thesis
Submitted to the Building and Construction Engineering
Department in the University of Technology in partial
Fulfillment of the Requirements for the Degree of Master
of Science in Geotechnical Engineering

By

TAWFIQ AAMIR JAWAD

(B.Sc.1998)

Jumadal-Akhera 1421 August-

2001

Certificate

I certify that the preparation of this thesis was made under my

supervision at the University of Technology in partial fulfillment of the
requirements for the degree of Master of Science in Geotechnical
Engineering.
 Signature:
Advisor: Assist. Professor Dr. Mohammed Majid Al-Ani
Date: / /2001‫‏‏‬

In view of the available recommendation, I forward this thesis for debate is

the Examining Committee.

Signature:
Professor Dr. Namir K.S. Al-Saudi
Chairman of the Branch of
Highway and Bridge Engineering
Date: / / 2001

ACKNOWLEDGMENTS

“In The Name of ALLAH, The Most Gracious, The merciful”

Praise to ALLAH for providing me willingness and strength to

accomplish this work.
I would like to express my sincere appreciation and deepest gratitude
to Dr. Mohammed Majid Al-Ani, whom I had the honor of working under
his supervision for his guidance, valuable assistance, patience and
continuous encouragement throughout this research.
I would like to express also my deepest thanks and gratitude to Eng.
Ra’id Rahman Al-Muhanna for his assistance in this work. Special thank
to the staff of the building and construction department in the University of
Technology and to all those who had participated in any way or another in
accomplishing this work.

Abstract

In all construction works in which the foundations are used especially

the raft foundations, the main goal for designing engineer is to choose the
best foundation taking into account safety and cost.
Optimization technique has been applied for a raft foundation
subjected to the normally occurring loads that transformed by columns to
choose the required design and optimum cost. The problem formulated in
terms of objective function and ten constrains and twelve constrains
represent minimum and maximum values of design variables. Six design
variables represent the dimensions of the raft, effective depth, steel
reinforcement per one meter of raft in two directions and the depth of
footing. The COMPLEX METHOD has been used in the optimum design
problem. A general purpose optimization program has been modified to take
the design parameters into account and generate the initial point for each
design variable to study design parameters effects. These design parameters
are number of floors of the structure resting on raft, distance between
columns in long direction, distance between columns in short direction,
compressive strength of concrete, yielding stress of steel, unit price of
concrete, unit price of steel reinforcement, cohesion of soil under the raft
and angle of internal friction of soil.
The effect and importance of each design parameter on a total cost,
effective depth of raft, steel reinforcement in two directions and depth of
footing has been studied.
Four mathematical models using (2187) trials are presented to obtain
main design variables which are effective depth of raft, steel reinforcement
in both directions and depth of footing for a wider and easier applications for
the designers to obtain the optimal raft foundation without going into
sophisticated optimization details.
‫‏‬
‫‏‬
‫‏‬
‫‏‬
‫‏‬
 CONTENTS

ACKNOWLWDGMENTS
ABSTRACT
CONTENTS
LIST OF GFIGURES
LIST OF SYMBOLES
CHAPTER ONE: INTRODUCTION
1.1 Raft Foundation 1
1.2 Optimization Technique 1
1.3 Optimization Design versus Conventional Design 2
1.4 Scope of Thesis 3
1.5 Layout of Thesis 3
CHAPTER TWO: REVIEW OF LITERATURES
2.1 Raft Foundations 6
2.1.1 Use of raft foundations 6
2.1.2 Common types of raft foundations 7
2.2 Design of Raft Foundation (conventional method) and (ACI) 7
code requirements
2.3 Optimization Theory 15
2.4 Previous Studies 19
CHAPTER THREE: PROBLEM FORMULATION
3.1 Introduction 27
3.2 Problem Formulation Using Optimization Technique 27
3.2.1 Objective function formulation 28
3.2.2 The constraints of problem 29
3.2.3 Optimization problem identification 35
3.3 Solution of Problem 35
3.4 The Complex Method and Programming 35
3.5 Modification of Program and Comments 39
3.6 Numerical Examples 40
CHAPTER FOUR: PARAMETRIC STUDY
4.1 Introduction 44
4.2 Analysis Process 44
4.2.1 Reference example 44
4.2.2 Effect of design parameters on total cost of raft 44
4.2.3 Effect of design parameters on footing volume 51
4.2.4 Effect of design parameters on steel reinforcement 55
 4.2.5 Effect of design parameters on depth of footing 60
CHAPTER FIVE: MODELING
5.1 Introduction 63
5.2 Input Parameters 63
5.3 Dimensional Analysis 65
5.4 Formulation of the Mathematical Models 65
5.4.1 Generalize equation for effective depth 66
5.4.2 Generalize equation for steel area reinforcement in short 67
direction
5.4.3 Generalize equation for steel area reinforcement in long 69
direction
5.4.4 Generalize equation for depth of footing 70
5.5 Comments about the Use of Mathematical Models 71
5.5.1 Numerical example 72
CHAPTER SIX: CONCLUSIONS AND RECOMMENDATIONS
6.1 Introduction 73
6.2 Conclusions 73
6.3 Recommendations for Future Works 74
 LIST OF FIGURES

No. Title page

1.1 Flow chart of thesis layout 5
2.1 Mat versus possible use of spread footings to save labor, 6
forming costs, and negative reinforcing steel
2.2 Common types of raft foundations 8
2.3 Design of rigid mat 11
3.1 Basic problem used in the thesis 28
3.2 Section for diagonal-tension shear 32
3.3 Strips of raft 32
3.4 Flow chart to initiate the complex method 42
3.5 The iterative routine of the complex method 43
4.1 Effect of design parameters on total cost 49
4.2 Design parameters % versus cost 50
4.3 Effect of design parameters on foundation effective 52
depth
4.4 Effect of design parameters on the volume of concrete 53
4.5 Design parameters % versus (L,B,d) % 54
4.6 Effect of design parameters on steel reinforcement in 56
short direction
4.7 Effect of design parameters on steel reinforcement in 57
long direction
4.8 Effect of design parameters on total steel reinforcement 58
4.9 Design parameters % versus (ASB,ASL,Vs) % 59
4.10 Effect of design parameters on depth of footing 61
4.11 Design parameters % versus depth of footing % 62
5.1 Predicted versus observed values of the effective depth 67
of footing
5.2 Predicted versus observed values of the effective depth 68
of footing
5.3 Predicted versus observed values of steel reinforcement 69
in short direction
5.4 Predicted versus observed values of the effective depth 70
of footing
 LIST OF SYMBOLES

A Total area of the mat

AS Steel area
ASL Steel area in long direction per one meter of raft length
ASB Steel area in short direction per one meter of raft width
a Depth of compression block
B Width of footing
B1 Spacing between columns in short direction of the basic
problem used in thesis
b Width of strip of the mat between centers of adjacent bays
bi Scalar constraints
c Cohesion of soil
cc Compression index
cs Swell index
DL Dead load for each floor = 10.6 kN/m2 (includes weights of
slab, tile and partitions)
dc Depth factor for bearing capacity
dq Depth factor for bearing capacity
d Depth factor for bearing capacity
d Effective depth of footing
Ec Modulus of elasticity of concrete
ex,ey Coordinates of the resultant force
eo Void rtio
Fs Factor of safety for bearing capacity
f’c 28-day compression strength of concrete (MPa)
fy Yielding stress of steel (MPa)
f(x) Objective function
gi Constraints
I Moment of inertia of the strip of width b
Ix,Iy Moment of inertia of the mat area with respect to the x and y
axes respectively
kb,ks Coefficient of subgrade reaction
L Length of footing
L1 Spacing between columns in long direction for the basic
problem used in thesis
 l Clear span between two columns in any direction
M Applied bending moment
Mu Ultimate moment from soil pressure on a section at the
column face
Nc Bearing capacity factor
Nq Bearing capacity factor
N Bearing capacity factor
Nf Number of floors of structure
OCR Over consolidation ratio
P Critical load combination
Q Total load on the mat
q Contact soil pressure
qall Allowable soil pressure
qult Ultimate soil pressure
R2 Coefficient of variance
Si Immediate settlement
Sc Consolidation settlement
sc Shape factor for bearing capacity
sq Shape factor for bearing capacity
s Shape factor for bearing capacity
t Thickness of footing
Uc Unit price of concrete
Us Unit price of steel reinforcement
Uex Unit price of excavation
Uf Unit price of fill
Vu,vc Ultimate shear force
Vc Total volume of concrete of the raft
Vs Total quantity of steel reinforcement
w Uniform load of the band
wt Total applied load for settlement
x,y Coordinates of any given point on the mat with respect to the
x and y axes passing through the centroid of the mat area
X Vector of design variables
x1 Length of footing
x2 Width of footing
x3 Effective depth of footing
x4 Steel area per one meter in short direction
x5 Steel area per one meter in long direction
x6 Depth of footing
Z Thickness of sublayer
 Characteristic coefficient
s, Soil density
o Initial average effective overburden pressure for sublayer
 Increase of vertical pressure for sublayer
s Shear reduction factor (0.85) (ACI-9.3)
c Moment reduction factor (0.85) (ACI-9.3)
 Angle of internal friction
 Percentage of steel
min Minimum percentage of steel
max Maximum percentage of steel
b Balanced percentage of steel
s Unit weight of steel =7.85 ton/m3
1,2 Constants whose values indicate the importance of an
objective function to the other
 CHAPTER ONE
INTRODUCTION

1.1 Raft Foundation

A raft or a mat foundation is a combined footing which covers the

entire area beneath a structure and supports all the walls and columns. This
type of foundation is most appropriate and suitable when the allowable soil
pressure is low, or the loading is heavy or for cases when spread footings
would cover more than one half the plan area. Also, when the soil contains
lenses of compressible strata which are likely to cause considerable
differential settlement, a raft foundation is well – suited, since it would tend
to bridge over the erratic spots, by virtue of its rigidity. On occasions, the
principle of floating foundation may be applied best in the case of raft
foundation, in order to minimize settlements (Venkatramaiah, 1993).

1.2 Optimization Technique

Optimization is the act of obtaining the best result under given

circumstances. In design, construction and maintenance of any

engineering system, engineers have to take many technological and

managerial decisions at several stages. The ultimate goal of all such

decisions is to either minimize the effort required or maximize the desired

benefit. Since the effort required or the benefit desired in any practical

situation can be expressed as a function called “Objective Function” of

certain decision variables. Optimization can be defined as the process of

finding the conditions that give the maximum or minimum value of a

 function. There is no single method available for solving all optimization

problems efficiently. Hence a number of optimization methods have been

developed for solving different types of optimization problems. The

optimum seeking methods are also known as mathematical programming

techniques and are generally studied as a part of operations research.

Operations research is a branch of mathematics, which is concerned with

the application of scientific methods and techniques to decision-making

problems and with establishing the best or optimal solutions. The

mathematical programming techniques are useful in finding the minimum

of a function of several variables under a prescribed set of constraints.

1.3 Optimization Design versus Conventional Design

There are two categories which can be used to design a raft foundation:-

1- Conventional (rigid), The conventional method assumes a rigid raft so

that the pressure diagram is uniform. The depth of raft is determined on

the basis of diagonal – tension shear similarly as for a spread footing.

Then the raft is divided into convenient strips to calculate the bending

moment on a strip, this bending moment can be used to calculate

required area of steel in each direction of raft.

2- Flexible Method, If the raft does not meet the rigidity requirements of
conventional method , it should be designed as a flexible plate using:
 a-Approximate flexible procedure
b-Finite differences
c-Finite elements
The conventional design method procedure aim to find an acceptable
or adequate design, which merely satisfies the functional construction and
meet the various code requirements of the problem but not exceedingly over
designed. In general conventional design is step by step procedure and there
will be more than one acceptable design. The purpose of optimization is
raising now to choose the best one out of the many acceptable designs
available (Rao, 1977). Since in the optimization procedure the designs are
carried out by a logical sequence steps which makes a guess of the
appropriate variables to have safe design converging on the lowest cost,
effort or weight or converging on the highest benefits.
Finally optimization method can be of a great value to the designer to
express many decisions in terms of optimization covering safe and economic
conditions, and it should be noted that optimization is a design tool rather
than a new method of design (Al-Ani, 1978).
1.4 Scope of Thesis

It is intended in this thesis to introduce the optimization technique as a

tool for a better design of raft foundation. It is also contemplated in this
thesis to study, the effect of change of a number of floors, spacing between
columns in long direction, spacing between columns in short direction,
strength of concrete, yielding stress of steel, cost of concrete, cost of
reinforcement, cohesion of soil and angle of internal friction of soil which
are named “ Design Parameters” on the design of raft foundation.
Simplifying the application of the optimization technique by presenting
reasonable mathematical models to find optimum effective depth, steel area
in long direction, steel area in short direction of raft and the depth of the
base of footing which are named “Design requirements or Design
variables”.

1.5 Layout of Thesis

The work divided into six chapters. Chapter two presents a review of
literatures concerning the subject which involves a summary about the use,
types and design of raft foundations and introduction to the optimization
technique. The available previous studies on the applications of optimization
techniques on geotechnical engineering are also presented in this chapter.
Chapter three describes the optimization technique formulation for a
raft foundation. Then a general computer program for the application of
optimization technique with flow chart are presented, this chapter also
presents two numerical examples are solved by the program.
In chapter four, parametric study is carried out to demonstrate the
effect of a number of floors, spacing between columns in long direction,
spacing between columns in short direction, strength of concrete, yielding
stress of steel, cost of concrete, cost of reinforcement, cohesion of soil and
angle of internal friction of soil on the design requirements of raft by using
an optimization computer program.
Chapter five presents the mathematical models to determine optimum
design requirements by using STATISTICA package.
Chapter six presents final conclusions of the applications of the
optimization technique in the design of raft foundation. Recommendations
for future works on this subject are also presented in this chapter.
 Chapter Two Introduction to Introduction to
conventional design optimization technique
Review of Literature

Chapter Three
Formulation of deterministic
Problem
Formulation optimization approach

Develop computer program to solve

deterministic approach

Solve two examples using conventional

method and optimization approach

Chapter Four Optimizati

Parametric Study
on analysis

Effect of design parameters

on design variables

Perform four design

Chapter Five equations (depending on
Modeling 2187 trials)

Chapter Six Conclusions of the

Conclusions and study
Recommendations

Recommendations
for future works

Figure 1.1: Flow Chart of the Thesis Layout

 CHAPTER TWO
REVIEW OF LITERATURES

2.1 Raft Foundations

2.1.1 Use of raft foundations

A raft (mat) foundation is usually a large concrete slab supporting a

number of columns or an entire structure. Mats often rest directly on soil or

rock; however, they may be supported on piles as well ( Winterkorn and

Fang, 1975).

A raft or a mat foundation is a combined footing which covers the

entire area beneath a structure and supports all the walls and columns. This

type of foundation is most appropriate and suitable when the allowable soil

pressure is low, or the loading is heavy or for cases when spread footings

would cover more than one half the plan area. Also, when the soil contains

lenses of compressible strata which are likely to cause considerable

differential settlement, a raft foundation is well – suited, since it would tend

to bridge over the erratic spots, by virtue of its rigidity. On occasions, the

principle of floating foundation may be applied best in the case of raft

foundation, in order to minimize settlements (Venkatramaiah, 1993).

2.1.2 Common types of raft foundations

A true mat or raft is a flat concrete slab with uniform thickness

throughout the entire area Fig.2.2 (a), this type is most suitable where the
column loads relatively small. For large column loads a portion ofv the slab
under the column may be thickened, Fig.2.2 (b), to provide sufficient
strength for negative moment and shear (diagonal tension). This may also be
accomplished by providing a pedestal under each column Fig.2.2 (d). if
bending stresses become large because of large column spacing and unequal
column loads, thickened bonds may be used along the column lines in both
directions Fig.2.2 (c). the empty cells are formed by the use of sheet metal or
paper domes. Under extremely heavy column loads, a two – way grid
structure made of cellular construction Fig.2.2 (e) and of intersecting
structural steel trusses has been also used as ribs or deep beams, Fig.2.2 (f)
(Teng, 1983).
2.2 Design of Raft Foundation (Conventional Method) and (ACI) Code
Requirements

In the conventional rigid method the mat is assumed to be infinitely

rigid and the contact pressure is assumed to have planer distribution. The
centroid of the contact pressure coincides with the line of action of the
resultant of all loads acting on the mat. These assumptions can be justified
when the mat is so rigid that its flexural deflection does not alter the
distribution of the contact pressure. For relatively uniform column loads and
relatively uniform column spacing, the mat may be considered rigid when
the column spacing is less than 1.75/, or when the mat is supporting a rigid
superstructure. The characteristic coefficient  is defined as:
 kbb
4 …( 2.1)
4Ec I

where
kb = coefficient of subgrade reaction
b = width of strip of mat between centers of adjacent bays
Ec = modulus of elasticity of concrete
I = moment of inertia of the strip of width b

Figure 2.2: Common types of mat foundations. (a) flat plate;(b) flat
plate thickened under columns; (c) two-way beam and slab; (d) flat
plate with pedestal; (e) cellular construction; (f) basement walls as
rigid frame.
 Based on the above assumption, the mat design can be carried out by
statics. The maximum column and wall loads are computed as a dead
load, the value of live load can be taken depending on the governing
building codes. The centroid of all these loads is determined. The weight
of the mat, however, may not be included in the structural design of the
mat. Every point of the mat is supported directly on the ground as soon as
the fresh concrete is placed. The dead weight of the concrete, therefore,
will not cause any flexural stresses as the concrete hardens, as long as the
contact pressure due to the weight of the mat remains unchanged. For
mats on compressible soils, however, there will be some long-term change
of contact pressure as the supporting soil is consolidated (Teng,1949). The
change of contact pressure due to the weight of mat is relatively small and
hence may be neglected in the structural design (Wintercorn and Fang,
1975).

After the resultant force is located the contact pressure (q) can be
readily determined by statics:

q    Q y  Q x y
Q ey e
…( 2.2 )
A Ix Ix

where
Q = total load on the mat
A = total area of the mat
x, y = coordinates of any given point on the mat with respect to the x and
y axes passing through the centroid of the mat area
ex, ey = coordinates of the resultant force
Ix, Iy = moment of inertia of the mat area with respect to the x and y axes
respectively.
Since the contact pressure is only in compression, the above formula
does not apply when the resultant force is outside the kern. In such cases,
part of the foundation is assumed not to be in contact with the ground.
Charts and formulas are available for determining the contact pressure for
foundations subjected to eccentric load (Teng, 1962).
The mat is analyzed as a whole in each of two perpendicular
directions. Thus, the total shear force acting on any section cutting across the
entire mat is equal to the arithmetic sum of all forces (loads) and reactions
(contact pressure) to the left or right of the section. The total bending
moment acting on such a section is equal to the sum of all moments on one
side of this section.
Although the shear and bending moments on the total section can be
determined by simple statics, the stress distribution along this section is a
highly indeterminate problem. In practice, this is often simplified by
analyzing the mat in a manner similar to that employed in the analysis of
two – way flat slabs. In the cases where the column loads and columns
spacing to do not vary more than about 20 percent from each other, the slab
may be divided into perpendicular bands. In Fig. 2.3 the boundaries of these
bands are shown as centerlines between adjacent column rows, bands A and
B, for example. Each band is assumed to act as an independent beam
subjected to known contact pressures and known column loads. The designer
will discover that often the beam does not satisfy statics; namely, the
resultant of the column loads and the centroid of contact pressure are not
equal and they do not act at the same point. The reason for this is that the
bands do not act independently; there is some shear transfer between the
adjoining bands. In practice, conservative moment coefficients – for
example, wl2/10 – are used, where w is the uniform load of the band and l is
the clear span between two columns in any direction. This approximation is
justified as long as the total bending strength of any section (a-b, c-d, etc.) is
sufficient.
Generally, a rigid mat is of sufficient thickness that there will be no
problem of shear stress, the designer should, however, check the
approximate shear value of each band and, particularly, check the punching
shear under each column (Winterkorn and Fang, 1975).

l
c

b
a
Band
B

d
Band A

Figure 2.3: Design of Rigid Mat

As described before shallow foundation is a structure that is supported

by the soil, laying immediately beneath the structure. Raft foundation is
usually rectangular in plane view. The dimensions of raft usually restricted
with the dimensions of the structure and the available area. The sequential
steps in the design process are:
1. Check the bearing capacity of soil using the given dimensions (L, B)
and the total loads applied on the raft (dead load, DL and live load,
LL) by using the allowable soil pressure

B.L = P / qall …( 2.3)

where
L = length of footing
B = width of footing respectively.
P = critical load combination.
qall = allowable soil pressure
Check the soil bearing capacity for the calculated plane dimensions using
Hansen‟s bearing capacity equation:
qult. = c. Nc . sc . dc + Df . s. Nq. sq. dq + 0.5 B. s. N. s. d … (2.4)
where c = soil cohesion.
s = soil density.
qul t = ultimate bearing capacity
Nc , Nq , N = bearing capacity factors (as specified by Hansen).
sc , sq , s = Hansen‟s shape factors.
dc , dq , d = Hansen‟s depth factors.

Fs = qul t / qall > predetermined value of safety factor.

P  6e L 6e B 
q max  1  …(2.5)
B.L  L B 

qall ≥ qmax …(2.6)

P  6e L 6e B 
q min  1  …(2.7)
B.L  L B 

where eL and eB = Eccentricity of the column load in the long and short
 direction respectively.
qmin ≥ 0 …(2.8)
qult ≥ qall . Fs …(2.9)

2. Check settlement of footing.

The amount of the total settlement, that may be tolerated depends on the
required functional performance of the building and the requirements of
the user, as well as on economic factors, such asset values, insurance cost,
potential production losses, etc. (Whitlow,1995). There are two major
types of settlement that should be accounted for in typical design of
footing.
a. Immediate settlement (si)
wt
si  …(2.10)
ks
where : ks = modulus of subgrade reaction.
Ks = 120.qall …(2.11)
wt = total applied load
This equation is based on reasoning that q all is based on the ultimate soil
pressure divided by safety factor Fs .
b. Consolidation settlement (sc‫)‏‬
For normal consolidated soil (OCR = 1)
cc Z    

sc  log  …(2.12)
1  eo   
 

For over consolidated soil (OCR > 1)

- if  + ∆ ≤ pc

cs Z    

sc  log  …(2.13)
1  eo   
 
 - if o +  > Pc

cs Z  P  c Z    
sc  log c
 c log o  …(2.14)
1  eo   o  1  e  o 
  

where cc = compression index.

Z = thickness of sublayer.
o = initial average effective overburden pressure for sublayer.
 = increase of vertical pressure for sublayer.
eo = void ratio.
cs = swell index.
Pc = pre consolidation pressure.
OCR = over consolidation ratio.
The total settlement must be less than a predefined value.
3. Find the effective depth (d) of raft, which is determined by the two modes
of shear failure (wide beam and punching), largest value should be
chosen. Generally, it is not economic to use shear reinforcement,
therefore, the value of (d) should be capable to sustain the shear stresses.
Effective concrete depth (d) can be obtained as:
- For wide beam (one way) shear action (ACI-11.3)
Vu
d …(2.15)
 f'c 
B.s  
 6 
 
- For punching (two way) shear action (ACI-11.3)
Vu
d …(2.16)
 f'c 
B.s 
3 
 
where d = effective depth of footing
Vu = ultimate shear force
f’c = 28-day compressive strength of concrete (MPa)
sshear reduction factor (0.85)(ACI-9.3)
4. Compute the steel area of bending, using the following equations
Mu
As  
a
c . ( d  ) f y
2
where As = steel area.
Mu = ultimate moment from soil pressure on a section at the
column face.
c= moment reduction factor (0.9) (ACI – 9.3).
fy = tensile strength of steel (Mpa).
a = depth of compression block.
As . f y
a …(2.18)
0.85 f ' c B
Area of steel must be less than the maximum steel area (ACI – 10.3.3)
Asmax = 0.75  b . d . B …(2.19)
where
f'c 600
  0.85 1 ( ) …(2.20)
b fy f y  600

for f’c ≤ 30 MPa

Also area of steel must be greater than the minimum area of steel (for
shrinkage and temperature) (ACI – 10.3.3)
Asmin = 0.002.t.b …(2.21)
where t = thickness of footing
or Asmin = 1.4 / fy …(2.22)
2.3 Optimization Theory

Optimization theory is an act which is concerned with the finding the

set of variables required to achieve the best results from a given situation.
Optimization deals with the optimal solution taking into consideration both
economic and safety conditions. Either maximization or minimization of an
algebraic function of one or more variable can reach this optimum solution.
This algebraic function is named as objective function. The choice of the
objective function is governed by the nature of the problem (Rao, 1977). The
expression of the objective function is written in form of least cost or
maximum or minimum weight or in form of largest improvement for a given
investment.
The choice of values for the variables will in general be restricted by
algebraic equations and/or inequalities called constraints, so that the goal is
not to find the best possible value of the function but rather the best value
permitted by constraints (Simmons, 1975).
To refer to any optimization problem having the following general
form
Find the values of the variable (x1,….., xn) that maximize or minimize
f (x1, ….,xn)
subject to gi (x1,…..,xn) ,i=1,….,m
where f (x) is the objective function
xn is an n-dimensional vector called the design variables
gi is the constraints
bi is scalar constraint
(f ), (gi ) and the variables (xi ) all take on real scalar values
 The number of variables (n) and the number of constraints (m) need
not be related in any way, thus (m) could be less than, equal to, or greater
than (n) in a given mathematical programming problem (Rao, 1977). In
some situations there may be more than one objective function to be
satisfied simultaneously with such multiple objectives there arises possibility
of the actual objective functions as a linear combination of the conflicting
multiple objective functions. Thus if f 1 (x) and f 2 (x) are the two objective
functions then it is possible to construct a new objective function for
optimization as
f (x) = 1 .f 1(x) + 2 .f 2(x) …(2.23)
where 1 and 2 are some constants whose values indicate the
importance of one objective function to the other.
There are many techniques available for the solution of a constrained
nonlinear programming problem. All of these methods can be classified into
two broad categories, namely the direct methods and indirect methods as
shown in table 2.1.

Table (2.1): Methods of Constrained Optimization Techniques

CONSTRAINED OPTIMIZATION TECHNIQUES

 Direct Methods
Indirect Methods
(i)Heuristic search (i) By the transformation
methods- the complex method of variables
(ii)constraint
(ii) penalty function methods
approximation methods – the
(a)interior penalty function method
cutting plane method
(b)exterior penalty function method
(iii)methods of feasible directions
(a) Zoutendij‟s method
(b)Gradient projection method

In the direct methods, the constraints are handled in an explicit

manner whereas in most of the indirect methods, the constrained problem is
solved as a sequence of unconstrained minimization problems.
(i) Heuristic Search Methods. The heuristic search methods are mostly
intuitive and do not have much theoretical support.
(ii) Constraint Approximation Methods. In these methods, the nonlinear
objective function and the constraints are linearized about some point
and using linear programming techniques solves the approximation
linear programming problem. The resulting optimum solution is then
used to construct a new linear approximation, which will again be
solved, by using linear programming techniques. This procedure is
continued until the specified convergence criteria are satisfied. There
are two methods, namely, the cutting plane method and the
approximate programming method, which work on this principle.
(iii) Methods of Feasible Directions. The methods of feasible
directions are those which produce an improving succession of
feasible vectors xi, by moving in a succession of usable feasible
directions. A feasible direction is one along which at least a small step
can be taken without leaving the feasible domain.
A usable feasible direction is a feasible direction along which the
objective function value can be reduced at least by a small amount.
Each iteration consists of two important steps in the methods of feasible
directions. The first step consists of finding a usable feasible direction
at a specified point and the second step consists of determining a proper
step length along the usable feasible direction found in the first step.
The Zoutendijk‟s method of feasible directions and Rosen‟s gradient
projection method can be considered as particular cases of general
methods of feasible directions.
The two basic types of indirect optimization methods can be explained
as follow:
(i) Transformation of Variables. Some of the constrained optimization
problems have their constraints expressed as simple and explicit
functions of the decision variables. In such cases, it may be possible to
make a change of variables such that the constraints are automatically
satisfied. In some other cases, it may be possible to know, in advance,
which constraints will be active at the optimum solution. In these
cases, can be used the particular constraint equation, gj(x) = 0, to
eliminate some of the variables from the problem. Both these
approaches will be discussed under the heading transformation of
variables.
(ii) Penalty Function Methods. There are two types of penalty
function methods- the interior penalty function method the
exterior penalty function method. In both types of methods, the
constrained problem is transformed into a sequence of
unconstrained minimization problems such that the constrained
minimum can be obtained by solving the sequence of
unconstrained minimization problems. In the interior penalty
function methods, the sequence of unconstrained minima lie in
 the feasible region and thus it converges to the constrained
minimum from the interior of the feasible region. In the exterior
methods, the sequence of unconstrained minima lie in the
infeasible region and converges to the desired solution from the
exterior of the feasible region (Rao, 1977).
2.4 Previous Studies

Al – Ani, (1978) used optimization technique as a design tool in

foundation design. Three nonlinear programming methods have been used,
graphical approach, geometric programming and Rosenbrock‟s method to
solve three foundation problems (anchored bulkhead, cantilever retaining
walls, and spread footing). The objective functions of these three problems
were aimed to minimize the total cost of the foundation. The objective
function of each problem subjected to two categories of failure, the
structural failure and soil failure. In the bulkhead problems two designs
variables have been optimized which were the depth of anchor rod and the
embedment depth of the steel pile, afterward, another design variable has
been added which was the depth from ground surface to the bottom of dead
man.
In the cantilever retaining walls five design variables have been
optimized, they were the thickness of the base, length of toe, length of heel,
area of steel of toe and area of steel of heel.
In the spread footing which was subjected to axial load passing
through equal column sides with no eccentricity six design variables have
been optimized which were the length and breadth of footing, effective
depth of footing, depth of excavation and finally the area of steel in both
directions.
Al – Ani concluded that:
1. Optimization technique is more economical and practical tool in the
foundation design than the conventional design procedures.
2. Design using optimization technique has been shown more
economical than the automated design by about 7%.
3. Geometric programming is an efficient method for problem with
variables as large as six variables, while Rosenbrock method efficient
for foundation design problem.
Bhavikatti and Hegde (1979), used sequential linear programming method

to obtain the optimum design of the isolated column footing, they have

considered the footing eccentrically with column so that the soil pressure

is uniform (symmetrical footing). The objective function is aimed to

minimize the material cost and subjected to constraints, which are:

1. The depth from punching shear consideration should be less than the
actual depth.
2. The shear at the critical section (column face + d) should be less than
the permissible shear stress in concrete.
3. Bond stress between steel and concrete of (column face + d) should be
less than the permissible bond stress.
It is assumed that there is one design variable that is the effective
depth of the footing.
Bhavikatti and Hegde found out:
1. The side projection of the footing should be kept the same as the
maximum projection in the direction of moment.
2. Depth is to be kept 16 to 20 percent less than that obtained in the
balanced design with this 8 to 10 percent saving in material cost can
be achieved.
3. The balanced section design is more uneconomical if the bearing
capacity of the soil is low. The improved move limit method of
sequential linear programming is found to be quite satisfactory
technique for the optimum design of isolated column footing.
Namiq (1983), used a graphical approach to obtain the optimal
material cost of anchored sheet pile wall assuming:
1. The sheet pile is driven to and back filled with granular soil.
2. Ground surface is horizontal and there is no surcharge.
3. Water level is below the sheet pile.
A free – end method of static analysis of sheet pile is used in
formulating the problem considering two independent variables (the cross
sectional area of tie rod and area of strip of the wall). A nonlinear and seven
linear inequality constraints define the feasible domain of these variables.
The solution of the problem indicates that by optimal designing of sheet pile
wall, significant cost saving can be obtained. The range of saving depends
on the constraints, which bounds the feasible region.
Desai et al, (1984) carried out cost optimization synthesis for an
isolated sloped square footing resting on dry granular medium. The cost
function and inequality constraints were formulated in terms of width of
footing, depth of excavation, thickness of footing and amount of reinforcing
steel (design variables). The cost function (footing cost) is the summation of
the volume of concrete material and reinforcing steel multiplied by the unit
volume cost of each one. The miscellaneous costs of excavation, refilling the
trench and preparing base support in trench are not accounted for.
Furthermore, the cost of embeded length of column is not calculated in the
cost function. Formulation of constraints based on the assumption that the
isolated sloped footing may be considered rigid compared to the surrounding
supporting medium. Under symmetrical column load, the base pressure may
be assumed uniform over the concrete of footing. The constraints are:
1. The tensile stress developed in reinforcement steel should not exceed
its permissible limit.
2. The shear stress developed in concrete at the critical section should
not exceed the permissible shear stress value.
3. The base pressure developed should not exceed the allowable base
pressure of the supporting medium.
4. Another two sides constraints are:
a. Face of the column length + 0.3 m should be less than the width of
the column (this constraint arises from normal construction
practice).
b. Depth of excavation should be less than the width of footing (arise
from the definition of shallow foundation).
The optimal solution is obtained by the method of Lagrange
multipliers where Lagrange function are solved for stationary points, which
should satisfy Kuhn – Tucker conditions. The variations in width and
thickness of footing, amount of reinforcing steel and cost ratio of footing for
different column loads and relative density of medium were studied. Desai et
al. observed that the self-weight of footing at column face and amount of
steel are greatly influenced by the column loads for footing in loose medium.
However, the influence of column loads on these parameters is small in very
dense medium. Also, they found out that cost of footing is sensitive to
column loads are large in loose medium. However, these variations in dense
medium are significantly small. Finally, they observed that the saving is
large in dense medium when compared to the conventional cost.
 Keskar and Adidam (1989), obtained the optimal cost design for a
simple form of reinforced concrete cantilever retaining wall subjected to the
normally occurring forces. The constrained optimization problems are
formulated in terms of two pre assigned parameter (the total height of the
wall and the top stem thickness), with the material cost as the objective
function. The solution is attempted by the interior penalty function method
using SUMT. The optimal cost per unit length of the wall was given in the
form of charts for different values of top width and height, with five
different types of backfill-retained soils.
Al- Samrrai (1998), used several optimization methods (Hooke and
Jeeves, Rosenbrock‟s and simplex method of Nedler and Mead) in designing
a tied back hinged retaining wall, considering the total cost of retaining wall
as the objective function and its geometric configuration stem thickness,
base thickness, toe width, heel width, depth of tie connection, angle of tie
inclination and tie spacing as the design variables. He made a comparison to
decide the efficiency of each method. He observed that Rosenbrock‟s
method was more powerful among the used methods in locating the
minimum for solved problem.
Al-Jubair (1998-a), used Hooke and Jeeves, simplex method of Nedler
and Mead and Rosenbrock‟s method to evaluate the cost function of a
braced sheet pile system. The objective function consisted of three cost
components; the cost of sheet piling, waling and strutting works. The depth
of struts and strut spacing were considered as design variables. The study
supported the efficiency of optimization methods in manipulating such
decision-making problems. In order to provide the geotechnical engineer
with some useful design curves, an extensive parametric study regarding
excavation height and backfill soil properties were performed utilizing the
more powerful Hooke and Jeeves and simplex methods. The percentage
value of each element cost to the overall structure cost was recommended.
Al-Jubair (1998-b), formulated the problem of designing a circular
cells series cellular cofferdam. The modified Hooke and Jeeves and complex
methods among the constrained methods were used in conjunction with the
Tennessee Valley Authority (TVA) method, to minimize the cost of the
cofferdam which consisted of two cost components; sheet piling works and
filling works. The design variables of the problem were the radii of main and
connecting cells, cells intersection angle and sheet piles embedment depth. It
has been proved by that study, that the constrained optimization methods
were very powerful in selecting the most economical design of the circular
cellular cofferdams.
Abdul Rahman (2000), used optimization to formulate the design of
cantilever retaining wall for optimum cost. The optimum design was
selected to satisfy the ACI-Code requirements. The constraints represent
inequality limitations on flexural strength, shear strength, minimum and
maximum reinforcement and also the stability of the wall. The optimization
problem was formulated for three shapes of the retaining wall, T-shape, T-
shape with key and L-shaped wall with four load cases for each type. The
solution of the problem was attempted using the sequential unconstrained
minimization technique (SUMT). Design figures and tables were presented
for limited values of the cost ratio and the angle of internal friction for the
backfill soil. The accuracy, convergence and the stability of SUMT
algorithm were studied. Overall looping method (OLM) was developed to
study the efficiency of SUMT. It was found that SUMT is a very powerful
method in locating the global minimum with height stability during the
iterative procedure.
 Al – Rawi (2001), applied optimal cost design for a reinforced
concrete spread footing subjected to the normally occurring forces. He used
the nonlinear constrained optimization problem to formulate the objective
function and constraints with the total cost as the objective function for
deterministic and stochastic approaches. The solution is attempted using a
developed computer program, which is based on the interior penalty function
method using (SUMT), which is considered, as one of nonlinear
optimization techniques available from literature. He found that the unit
price of concrete is more important than the other unit prices. Hence, the
total cost of footing consists of 52 % of concrete, 34 % of steel
reinforcement cost, 10 % of excavation cost and 4 % of filling cost. He
found that the bearing capacity, moment, one – way shear action and the
minimum steel ratio are the active constraints and vertical force takes a great
share of importance in the determination of footing cost, volume, steel
reinforcement and depth, especially when it is compared with the horizontal
forces. Also the response against the change of the vertical force shows the
minimum degree of non-symmetrical amongst the design parameters.
Ibrahim (2001), used optimization technique in optimal cost design of
reinforced concrete cantilever retaining wall subjected to the normally
occurring forces. The problem is classified as a nonlinear programming
optimization problem. The problem was solved using the interior penalty
function method (SUMT). The constrained optimization problem was
formulated in terms of properties and unit cost of the materials. Ten design
variables and thirty-one constraints with the total cost of the wall as the
objective function. He concluded that the unit price of concrete has the
largest influence on the total cost, wall dimensions and also the total steel
mass, since the total cost of the retaining wall consists of 51% of concrete
cost, 39% of steel cost, 6% of backfill cost and 4% of excavation cost. He
concluded also that the upper bound on the area of steel max was seen to be
a non-active constraint, so there is no need to consider this restriction in the
optimum solution of the cantilever retaining wall. Ibrahim founded also for
different values of the design parameters considered in the study, the stem
base thickness is governed by the minimum batter requirement for walls
more than 3m in height. For walls less than 3m the stem base thickness is
governed by the moment requirements.
 CHAPTER THREE
PROBLEM FORMULATION

3.1 Introduction

The basic problem shown in Fig. 3.1 is assumed in this study to

represent an actual field problem of a raft foundation supports a
multistory building in AL-NASJAF city. The problem represents a raft
foundation with dimensions (L*B) and a uniform thickness (t) (cover of
concrete is used as 0.1m) is reinforced with a steel reinforcement in
both directions in two layers (ASB and ASL). The raft laying at a depth
(Df) rests on a uniform soil with properties (c,  and where c,  and
are cohesion of soil, angle of internal friction of soil and unit weight of
soil respectively. The raft is subjected to uniform distributed load (dead
load (DL) and live load(LL)) transformed by columns group (3*4) with
a dimensions (0.4m*0.4m) for each column from a number of structure
floors (Nf).

3.2 Problem Formulation Using Optimization Technique

The problem can be formulated with six design variables to be

optimized:
xi={x1, x2, x3, x4, x5, x6} …(3.1)
where
x1= length of foundation (L) in (m)
x2= width of foundation (B) in (m)
x3= effective depth of foundation (d) in (m)
x4= steel area in short direction (ASB) in (mm2)
x5= steel area in long direction (ASL) in (mm2)
x6= depth of foundation (Df) in (m)
 Description of the problem’s formulation can be introduced as
follows:-
3.2.1 Objective function formulation

The objective function is expressed in form of total cost of the raft

f(x)= cost of excavation + cost of concrete + cost of steel + cost of fill
f(x)= x1x2x6Uex + x1x2(x3 + 0.1)Uc + (2*x5*10-6*x2*x1+2* x4*10-6*
x1*x2)*s*Us +x1x2(x6 - x3 + 0.1)Uf …(3.2)
where Uex, Uc, Us and Uf is the unit price of excavation, concrete, steel
and fill respectively.
Uex, Uc, Uf in Unit price/m3
Us in Unit price/Ton

L Free edge

L1 L1 L1

B1
B

B1

Column dimensions
0.4x0.4m

Df
t
q

Figure 3.1: Basic Problem Used in Thesis

 It should be noted that the price values takes into account formwork,
placing of concrete, vibration and includes labor charges and they are
representative according to the rates of construction and may vary from time
to time and place to place.
3.2.2 The constraints of problem

In raft foundation design, failure can be categorized into two general

types of failure that the design should be accounted for. These possible
failures are soil failure an structural failure. Soil failure consists of two main
criteria; general bearing capacity and settlement limitation while structural
failure consists of shear failure and flexural failure.
1- allowable soil pressure (qall ) must be greater than or equal the
maximum applied pressure (qmax )
qall ≥ qmax … (3.3)
qult / Fs ≥ qmax …(3.4)
The ultimate bearing capacity (qult) can be calculated from Hansen‟s
equation (Bowles 1982).
qult = cNcscdc + qNqdq + 0.5BNsd …(3.5)
where:
Nq = etan tan2 (45+/2)
Nc = (Nq-1) cot
= (Nq-1) tan
sc = 1+0.2B/L = 1+0.2x2/ x1
sq = 1+(B/L)tan= 1+(x2/x1) tan
s = 1-0.4 B/L = 1-0.4 x2/x1
dc = 1+0.4 Df/B = 1+ 0.4 x6/x2
dq = 1 + 2 tan (1-sin)2 Df/B = 1 + 2 tan (1-sin)2 x6/x2
d= 1

The maximum applied pressure can be calculated from basic

principle as:
N f (DL  LL)(3L1  0.4)(2B1  0.4)
q max  …(3.6)
x1 x2

q ult
q all 
Fs
…(3.7)

Thus, the constraint can be written as:

 g1 is:
c Nc (1+0.2 x2/x1) (1+0.4 x6/x2) +  x6 etan tan2(45+/2) (1+x2x1 tan) (1 +
(2 tan (1-sin)2 x6/x2) + 0.5  x2 (1.5 (Nq-1) tan) (1-0.4 x2/x1) -
Fs*Nf*(DL+LL) (3L1+0.4)(2B1+0.4) / (x1x2) ≥ 0 …(3.8)
2-Footing immediate settlement must not exceed a permissible limit. To
satisfy this limitation the equation of subgrade reaction is used:
si = wt/ ks = Nf(DL+LL)/si …(3.9)
where ks= modulus of subgrade reaction (kN/m3).
ks = 120 * qall …(3.10)
thus,
 N f (DL  LL)
si  …(3.11)
120(qult /Fs )

If the permissible si = 0.025m

3*qult = Nf *Fs(DL+LL)
Thus, the constraint can be written as:
 g2 is:
3*[c Nc (1+0.2 x2/x1) (1+0.4 x6/x2) +  x6 etan tan2(45+/2) (1+x2x1 tan) (1
+ (2 tan (1-sin)2 x6/x2) + 0.5  x2 (1.5 (Nq-1) tan) (1-0.4 x2/x1)] -
Nf*Fs(DL+LL) ≥ 0 …(3.12)
2- Effective depth of footing must be sufficient so that one way (beam)
shear action does not occur and it must be sufficient so that punching
shear does not occur. If the load is uniformly distributed on the raft (no
eccentricity) the two way shear failure may occur in the interior area of
dimensions (L1*B1) gives the largest load on the column (see Fig. 3.3);
therefore the constraint which is formulated to find the effective depth
developed from Fig. 3.2 as follows:-
∑Fv=0
on diagonal tension zone, and noting this block of footing weight cancels
gives
Pu = 2 d vc (b + d) + 2 d vc (c + d) + (c + d) (c + d) q …(3.13)
Substitution of Pu = Nf.B1.L1.q
for square column c = b = a and obtain
d2 (vc+q/4) + d (vc+q/2) a = (B1L1-a2) q/4 …(3.14)
Equation 3.14 is a general equation used to find effective depth of spread
footing. Simplifying to be applied on the basic problem of the thesis, it can
take the following form:
∑Fv=0 (neglect upward soil pressure)
d *4(0.4+d) vc = (1.4 DL+1.7LL) NfL1B1 …(3.15)
vc = 0.333 s f' c ,s=0.85 (ACI-9.3)

vc = 0.283 f' c …(3.16)

Thus, the constraint can be written as:
 g3 is:
x3*0.283 f' c *1000 (1.6 + 4 x3) - (1.4DL+1.7LL)*L1*B1*Nf ≥ 0…(3.17)

c+d Diagonal
tension
d/2
b+d block
B1
b
c
L1

Figure 3.2: section for diagonal-tension shear

3-Footing must reinforced with sufficient steel area to resist the bending

moment:

Mu = c As (d - a/2)fy …(3.18)
As . f y
a …(3.19)
0.85. f'c .b
then
Mu = f’c As (d- As fy/( 2*0.85 f’c b))fy …(3.20)
M = wl2/10 …(3.21)
where
w= total distributed load on a strip (see Fig 3.3)
l = clear span between two columns in any direction (L1-0.4 or B1-0.4)
L1
L1

B1

B1

Figure 3.3: Strips of Raft

The applied bending moment (M) must be less than allowable bending of
concrete (Mu); thus
w(L1 – 0.4)2/10 ≤ cAs (d-As fy/(2*0.85 f’c b))fy …(3.22)
w(B1 – 0.4)2/10 ≤ cAs (d-As fy/(2*0.85 f’c b))fy …(3.23)
The value of c can be taken 0.9 (ACI-9.3).
Thus; the constraints of bending can be written as:
 g4 is:
0.1(B1-0.4)2 Nf (1.4DL+1.7LL) – 0.9*1000*fy*x3*x4*10-6
+0.5294*1000*x42 *10-10*fy2 f’c-1 ≤ 0 …(3.24)
 g5 is:
0.1(L1-0.4)2 Nf (1.4DL+1.7LL) – 0.9*1000*fy*x3*x4*10-6
+0.5294*1000*x42 *10-10*fy2 f’c-1 ≤ 0 …(3.25)
5- Reinforcement ratio in both directions for bending moment should not be
less than min
min = 1.4/fy …(3.26)
 and
min = 0.002 …(3.27)
Asmin = min .b.d …(3.28)
Asmin = 0.002. b.t for shrinkage and temperature
if b = 1m and the raft designed as a concrete slab
therefore;
min = 0.002 governs usually
thus, the constraints can be written as:
 g6 is:
x4-0.002 (x3+0.1)*1 ≥ 0 …(3.29)
 g7 is:
x5-0.002 (x3+0.1)*1 ≥ 0 …(3.30)
6-Reinforcement ratio in both directions should not exceed the maximum
steel ratio
max = 0.75 b …(3.31)
where b = balanced percentage of steel area
 0.85f' c  600 
 b  1   
 f  600 
…(3.32)
 f y  y 

where 1 = 0.85 for f’c  30 Mpa

thus, the constraints can be written as:
 g8 is:

f'  600 
x4-0.6375 c   x3  0 …(3.33)
f y  f y  600 


 g9 is:

f'  600 
x5-0.6375 c   x3  0 …(3.34)
f y  f y  600 

7- The depth of footing must be larger than the thickness of footing by about
0.4m, then
Df  (d + 0.1) + 0.4 …(3.35)
Df  d + 0.5 …(3.36)
Thus; the constraint can be written as:
 g10 is:
x6 - x3 - 0.5  0 …(3.37)
There are another constraints can be used for the requirements of
using the program illustrate the upper and lower bounds of design variables
(section 3.3), these constraints are:
 3 L1+ 1 ≤ x1 ≤ 3 L1 + 4
 2 B1+ 1 ≤ x2 ≤ 2 B1 + 4
 0.3 ≤ x3 ≤ 1.5
 0.3 * 0.002 * 1 ≤ x4 ≤ 1.5*0.002*1
 0.3 * 0.002 * 1 ≤ x5 ≤ 1.50*0.002*1
 0.8 ≤ x6 ≤ 3
3.2.3 Optimization problem identification
The final form of the optimization mathematical programming problem
can be summarized to be:
Find X which minimize the objective function f(x) when
F(x)= x1x2x6Uex + x1x2(x3 + 0.1)Uc + (x5x2x1 + x4x1x2)2*7.85 Us + x1x2 (x6 -x3
+ 0.1)Uf …(3.38)
subjected to inequality constraints
gj(x)  or  0 j=1, 2, 3,…….,6
where X is the vector of the design variables;
X= [ x1, x2, x3, x4, x5, x6 ]
3.3 Solution of Problem
 The problem formulated in the previous section can be classified as a
constrained optimization problem. The non linear programming can be used
to solve the problem with the benefit of search methods which use function
values. These methods suggested that merely giving the objective function a
very large value (in a minimization problem) whenever the constraints are
violated will suffice.
For each trial point check whether it lies within the constraint region.
If so the objective function can be calculated in the normal way. If not the
objective function should be given a very large value. In this way the search
method will be directed back into the feasible region and hence towards the
minimum point within the feasible region.

3.4 The Complex Method and Programming

The difficulties encountered to implement the search methods
prompted Box in 1964 to devise his own method. Essentially it is a
modification of the Simplex Method of Nedler and Mead so as to take
account of constraints. Box called it the Complex Method.
The problem considered is that of minimizing f(x) = f(x1, x2, …..,xn)
where the x are subjected to the explicit constraints

l j ≤ xj ≤ u j j=1,2,……,n …(3.39)

and also the implicit constraints

gi(x) ≤ bi i=1,2 ,……..,m …(3.40)
If the objective function f(x) is convex and the implicit constraint gi(x) are
convex the problem will have a unique solution. The lj and uj are lower and
upper bounds for the variables are in theory unbounded, assuming „safe‟
bounds which certainly include the optimum will enable the method to be
implemented.
The method is an iterative procedure. It assumes that n and m are
known, the lj and the uj and have an initial point x1 that satisfies all the
constraints (equation 3.39 and 3.40). A set of k points have to be generated
which satisfy the constraints, and evaluate the objective function at those k
points. This set of points is called complex. Box found that k needed to be
larger than n+1, the number of points used in the Simplex Method of Nedler
and Mead. He suggested the value k=2n.
It is assumed that x1, satisfying all the constraints is given. Further points can
be generated which satisfy equation (3.39) by
 Xij = lj + r ( uj – lj ) …(3.41)
For j=1, 2, …….,n and I=2, 3,…..,k where r is a pseudo-random
rectangularly distributed variable on the range (0,1). Such variables are
obtained from the statement Y= RND (X) in BASIC.
Points generated by equation (3.41) for given value of j will
automatically satisfy equation (3.40) they bare accepted as points of the
initial complex. If the points generated by equation (3.41) does not satisfy
equation (3.40) move it half way towards the centroid of already accepted
points, i.e. form
( x i  xc )
x' i  …(3.42)
2
where
i 1
1
xc   xe
i  1 e 1
…(3.43)
If equation (3.42) is still not feasible the procedure can be repeated again
and again until it is. If the gi(x) are convex the constraints will ultimately
satisfy. Of course since x1 is in the constrained region a centroid of already
accepted points will always have. In this way the initial complex of feasible
points are obtained. It is convenient to order these points according to the
magnitude of the corresponding function values.
The steps required to the iterative procedure of the Complex Method in
which the movement towards the minimum point within the constrained
region can be seeked. The steps required are outlined below:
1- find the point with the greatest function value xh and form the centroid
xo of the order ( k-1 ) points.
2- Trying to move away from xh and so form the point xr by reflecting xh
in xo, using a reflection factor (>1). The choice of k=2n and =1.3
are empirical rules suggested by Box.
i.e. xr = (1 + ) xo - xo …(3.44)
3-Next test if xr is feasible.
(i) If not and if lj is violated, setting xrj=lj+10-6; if uj is violated set
xrj=uj-10-6
(ii) If an implicit constraint is violated; move the point xr halfway
towards the centroid xo
i.e. xr(new) = (xr + xo)/2 …(3.45)
Then retest for feasibility and repeat the procedures at step 3 until a
feasible point is obtained.
4-if xr is feasible, evaluate f(xr) and compare it with f(xh) the worst function
value since the function values have been ordered.
If f(xr) > f(xh) i.e. is worse than the worst value obtained so far, move xr
halfway towards xo
i.e. xr(new) = (xr+ xo)/2
and then return to step 3.
5- If f(xr) < f(xh), replace xh by xr and recorder the points and function
values of the complex.
6- Next calculate two quantities that are used to test whether the method has
converged. These are the standard deviation of the k function values, and
the maximum distance dm between two points of complex. The former is
calculated as:
1/2
 k 2

   f(xe )  f /k  …(3.46)
 e1 
where
k
1
f   f ( xe )
k e 1
…(3.47)
but 2 is best calculated from the equivalent formula
k 
 f ( x)
2
 
   f ( xe ) 2 
2
/ k …(3.48)
 e 1 k 
 
7- The test of convergence is made on 2 and dm. If both are sufficiently
small terminate will obtained. Otherwise , return, to step 1 and repeat the
process.
Figures 3.4 and 3.5 shows a flow chart indicating the steps of
programming a previous procedure.
3.5 Modification of Program and Comments
Bunday 1984 built a computer program using a BASIC language to
apply the procedure of previous section. This program has been developed to
perform the design of raft foundation according to optimization technique.
The input file of this approach includes many parameters used in design
which is called parameters of design or (Design Parameters). These
parameters are (L1, B1, f’c, fy, Uc, Us, c, and ). Other parameters which the
program ask about them are considered constant throughout the study, these
are (DL, LL, Uex, Uf, and Fs). Output file gives a design of raft, which is
including (L, B, d, ASB, ASL, and Df). The Complex Method is a useful
procedure, which can be applied to a wide range of constrained optimization
problems. It should not however, be regarded as the best in this field. If the
 objective function is convex and the constrained region is also convex, the
method should succeed, although certain peculiarities of the problem may
call for some modification of the termination criteria. If the objective
function is concave or the constrained region not convex it is easy to see
how the method could fail. Indeed, in the case of a non-convex constrained
region it is not even clear that the centroid of feasible points will also be
feasible. Thus the moving
xr(new) = ( xr + xo ) / 2
are not guaranteed to achieve their objective.
Care also needs to be taken to check that the global minimum rather
than a local minimum ha been found. Box suggests that making more than
one run from different initial points should resolve this difficulty, and this is
the main problem can be made the modification of program to take a
parametric study and the iteration becomes very large as in chapter four. The
random manner in which the initial complex is generated should mean that
initially must generate a good converge on the global minimum.
Convergence of several runs to the same result should clarify the issue.
As in above some modification on the program has been made to
make the program able to take the parametric study in the following chapter
into account. This modification made the program generate the initial point
from the lower bounds lj of the input data, this made the initial point usually
in the feasible region. In this way the program will succeed to continue for
wide range of input data which automatically changed by for-next statement.
Before this modification the program can solve one problem by input the
data which is requested by input statement and the initial point giving as a
value from the designer for each iteration.
3.6 Numerical Examples
Two numerical examples are presented herein. These examples
illustrate the efficiency of the use of optimization technique to solve
engineering problems, one of these problems is a raft foundation design. The
benefit of using this technique is to include the cost of materials of the raft in
a design, that make the design process take into account a very important
factors which do not appear in conventional design.
Example one:
 Design a raft foundation for the following data:
L1 = 4m, B1 = 3m, f’c = 21 MPa, fy = 300 MPa, Uc = 30 Unit price/m3, Us=
200 Unit price/ Ton, Uex = 2 Unit price/m3, Uf = 1 Unit price/m3, Fs=2, Nf=
6, LL= 4.8 kN/m2, DL = 10.6 kN/m2, c=50 kN/m2,  =30ْ
,  = 18 kN/m3
Example two:
Design a raft foundation for the following data:
L1 = 6m, B1 = 4m, f’c = 30 MPa, fy = 400 MPa, Uc = 40 Unit price/m3, Us=
300 Unit price/ Ton, Uex = 2 Unit price/m3, Uf = 1 Unit price/m3, Fs=2, Nf =
9, LL= 4.8 kN/m2, DL = 10.6 kN/m2, c=60 kN/m2,  = 10ْ
,  = 18 kN/m3
The solutions of these examples are shown in Table 3.1.
The results tabulated in Table 3.1 shows that the optimization
technique can be used to design the raft foundation by using the equations of
conventional design, that give a minimum requirements of the conventional
design and the optimum value for Df (x6) which cannot be found from
conventional procedure.

Table (3.1): Results of examples

Results of
Results of Results of
Results of example
example example
Design example one
two by two
variabl one by By using Units
conventio By using
es convention optimizat
n-al optimizatio
-al method ion
method n technique
technique
x1 13.5 13.0018 19.5 19.1717 m
x2 7.5 7 10 9.404 m
x3 0.4 0.3939 0.9 0.8962 m
x4 993.9 993.9 2002.7 2002.7 mm2
x5 1806.3 1806.3 2734 2734 mm2
x6 1 0.898 1.5 1.413 m
 Unit
Cost 2670 2447 12053 11467
price
Functio
n
303 181
evaluat
ions

Read m, n

Read xi,i=1

Read lj,u,j=1,2,…, n
Set xc = x1
Set random
generator

i=i+1

Put xij = lj + r(uj - lj)

is Put
xj feasible? xi=(xi+xc)/2

Update centroid
xc =[(i-1)xc +xI]/i

Find f(xi) = Fi

Is
i<k ?
 xr = (xo +xr)/2
if gi(x)violated

Is xr No xr = (xo +xr)/2
Fr < Fk

Yes

Replace x(k) by xr

Update F1,…, Fk
and points of
complex

Find standard deviation of

function values and maximum
distance between points

Has
No Method
converged ?

Yes

Finish

Figure 3.5: The Iterative Routine of the Complex Method

 CHAPTER FOUR
PARAMETRIC STUDY

4.1 Introduction
In this chapter the effect of design parameters (input parameters) on
the final design of the raft is presented. The effect of design parameters on
the total cost, volume of concrete and quantity of steel are also studied. In
this analysis a “reference” example is chosen to represent actual field
problem, this analysis is made by changing each design parameter to
investigate its effect with the aid of optimization technique.

4.2 Analysis Process

4.2.1 Reference example

The values of design parameters used in this investigation are shown

in Table 4.1. The computer program listed in chapter three using Complex
Method was modified to be able to perform this analysis. A total of 45
trials were executed. Each trial included the chosen set of input parameters
(9) and output design variables (x1…..x6) in addition to the total cost of
footing, total volume of concrete and total quantity of steel. The output
results were organized in tabulated form using Microsoft Excel package as
in Table 4.2. Each trial took about (3-6) minutes in Pentium I computer. To
investigate the effect of each design parameter on design variables, each
parameter was taken separately while the other parameters remained
constant.
4.2.2 Effect of design parameters on total cost of raft
The effect of design parameters on the total cost has been
investigated. Results are illustrated in Fig. 4.1. It can be seen that the total
cost increases with the increase in the applied load, which are represented by
(Nf, L1, B1). It can be seen also that the increase in the strength of concrete
(f’c) and yielding stress of steel (fy) will decrease the total cost with a very
small amount. The effect of increment or decrement of cohesion (c) and
angle of internal friction () is not important because of the large area of raft
which is restricted by the area of structure or available area, therefore, the
bearing capacity always on the safe side for the value of (c) and () used in
analysis.
From Fig. 4.1, one can build an idea about the response of the
footing cost against the change in design parameters under the cover of
the use of optimization technique in the design of the raft. This may give
an indication about the importance of each parameter depending on the
slope of joining any two points. Fig. 4.2 shows the relation between the
design parameters and the total cost as a percentage can be made from
the following relation:

– Xref
%Xdifference = * 100
Xref
where X = investigated design value
Xref = reference value
For example the increase of the value of (Nf) from 6 (reference value) to 8
(% difference = 33.33 %) will increase the total cost with an amount of (20
%). Hence Fig. 4.2, represent a graph of percentage difference of design
parameters as the X – axis and percentage difference of the total cost as Y –
axis and a line is drawn joining these points. From this figure it can be seen
that the higher the slope of the line the more important is the design variable.
Fig. 4.2 shows that the (Nf, L1, B1) are the most important parameters
because they lead to an increase of all design variables which lead to an
increase in the cost. It can be seen also that the unit price of concrete (U c)
and steel unit price (Us) have the same effect on total cost of footing but with
a smaller degree. The strength of concrete (f’c) and yielding stress of steel
(fy) cause a decrease in total cost when they are increased.

Table (4.1): Input Parameters Used in Analysis

N Investigated
Input Parameters Symbol Unit RV*
o. Values
1 Number of Floors Nf No. 6 4, 5, 6, 7, 8,
Spacing Between
2 Columns in Long L1 m 7 5, 6, 7, 8, 9
Direction
Spacing Between
3 Columns in Short B1 m 5 3, 4, 5, 6, 7
Direction
Strength of 22, 24, 26, 28,
4 f’c MPa 26
Concrete 30
Yielding Stress of 300, 325, 350,
5 fy MPa 350
Steel 375, 400
Unit Price of Unit 20, 25, 30, 35,
6 Uc 30
Concrete price 40
Unit 200, 225, 250,
7 Unit Price of Steel Us 250
price 275, 300
30, 40, 50, 60,
8 Cohesion of Soil c kN/m2 50
70
Angle of internal Degre 10, 20, 30, 40,
9  30
friction e 50
RV* = Reference value
Other parameters are assumed constants as DL= 10.6 kN/m2, LL= 4.8 kN/m2
and Fs= 2, =18 kN/m3, Uex=2 unit price/m3, Uf=1 unit price/m3.
 60
NF

40 L1

B1
20
f'c
COST %

0 fy

-5 0 -4 0 -3 0 -2 0 -1 0 0 10 20 30 40 50
Uc

-2 0
Us

-4 0 C

-6 0
D e sig n Pa ra m e trs %

Figure 4.2: Design Parameters % versus Cost %

Many conclusions from Fig. 4.1 and 4.2 can be listed as follows: -
1- Total cost is affected to a large degree by (Nf, L1, B1), which represent
the load applied on the raft.
2- The increase of unit price of concrete (Uc) and unit price of steel (Us)
increase the total cost with a smaller degree than the load.
3- The strength of concrete (f’c) and yielding stress of steel (fy) has a
small effect on the total.
4- There is no effect of cohesion of soil (c) and angle of internal friction
( ) on the total cost.
5- Design parameters follow unsymmetrical behavior around the
reference value. Hence the specific increment and/or decrement for
each design parameter does not meet the same response degree. This
loses of symmetry are different from one parameter to another.
4.2.3 Effect of design parameters on footing volume
The effect of the design parameters on footing volume (Vc) which
represent (x1*x2*(x3+cover) are studied in this section. Fig.4.3 illustrates the
effect of design parameters on effective depth (x3), each parameter was taken
separately to study. Fig. 4.4 shows the relation between design parameters
and volume of concrete (Vc).
Fig. 4.5 shows the effect of design parameters on x1, x2, and x3 (L, B,
d). It is clear that x1 and x2 are not affected by the design parameters. It has
an important effect on the effective depth (x3). This makes a volume of
concrete increase with the increment of (x3) and visa verse.
From Figures 4.3, 4.4, and 4.5, it can be concluded that:
1- The dimensions of the raft are not affected by the design parameters
because of the assumed large area, which make the bearing capacity
always in the allowable limit.
2- The effective depth (x3) is affected by (Nf, L1, B1) which represent
the applied load, therefore, the increase of load lead to the increment
in footing depth.
3- The effective depth decrease when the strength of concrete (f’c) and
unit price of concrete (Uc) increase.
4- The yielding strength of steel (fy) and unit price of steel (Us),
cohesion of soil ( c ) and angle of internal friction (  ) have no effect
on effective depth (x3), (see Fig.4.3 and Fig. 4.4).
5- The volume of concrete (Vc) depends on the value of depth of raft
only because the dimensions of raft are not affect by any design
parameters.
 1 1

0.9 0.9

0.8 0.8
d

d
0.7 0.7

0.6 0.6

0.5 0.5
5 5.5 6 6.5 7 7.5 8 8.5 6 6.5 7 7.5 8 8.5 9 9.5
Nf L1

1 1

0.9 0.9

0.8 0.8
d

d
0.7 0.7

0.6 0.6

0.5 0.5
4 4.5 5 5.5 6 6.5 7 7.5 25 26 27 28 29 30 31
B1 f' c

1 1

0.9 0 .9

0.8 0 .8
d

0.7 0 .7

0.6 0 .6

0.5 0 .5
340 350 360 370 380 390 400 410 20 25 30 35 40 45
fy Uc

1 1

0.9 0.9

0.8 0.8
d

0.7 0.7

0.6 0.6

0.5 0.5
240 250 260 270 280 290 300 310 70 75 80 85 90 95 100 105
Us c

0.9

0.8
d

0.7

0.6

0.5
10 15 20 25 30 35 40 45


Figure 4.3: Effect of Design Parameters on Foundation Effective Depth(m)

 400 400

350 350
Vc 300 300

Vc
250 250

200 200

150 150

100 100
5 5.5 6 6.5 7 7.5 8 8.5 5 6 7 8 9 10
Nf L1

350 400

300 350

300
250
Vc

Vc
250
200
200
150 150

100 100
4 4.5 5 5.5 6 6.5 7 7.5 25 26 27 28 29 30 31
B1 f'c

400 400

350 350

300 300
Vc

Vc

250 250

200 200

150 150

100 100
340 350 360 370 380 390 400 410 20 25 30 35 40 45
fy Uc

400 400

350 350

300 300
Vc

Vc

250 250

200 200

150 150

100 100
240 250 260 270 280 290 300 310 70 75 80 85 90 95 100 105
Us C

400

350

300
Vc

250

200

150

100
10 15 20 25 30 35 40 45


Figure 4.4: Effect of Design parameters on the Volume of Concrete (m3)

 30
NF

20
L1

10 B1

f'c
0
L%

-5 0 -4 0 -3 0 -2 0 -1 0 0 10 20 30 40 50 fy

-1 0
Uc

-2 0 Us

C
-3 0


-4 0
D e sig n Pa ra m e te rs %

50
NF
40
L1
30

B1
20

10 f'c
B%

0 fy

-5 0 -4 0 -3 0 -2 0 -1 0 0 10 20 30 40 50
-1 0 Uc

-2 0 Us

-3 0
C

-4 0


-5 0
D e sig n Pa ra m a te rs %

30
NF

20 L1

B1
10
f'c
d%

0 fy

-5 0 -4 0 -3 0 -2 0 -1 0 0 10 20 30 40 50
Uc

-1 0
Us

-2 0 C

FI

-3 0
D e sig n Pa ra m e te rs %

Figure 4.5: Design Parameters % versus (L,B,d) %

 4.2.4 Effect of design parameters on Total Steel reinforcement

The effect of design parameters on the total quantity of steel

reinforcement (Vs), which is related, with the steel reinforcement in two
directions (ASB and ASL) is presented in this section. Fig. 4.6 was performed
to study the effect of design parameters on steel reinforcement in short
direction (ASB). From Fig. 4.6 it can be seen that the steel reinforcement in
short direction (ASB) increase with the increase of load which is represented
by (Nf, L1, B1) but with a different degree of each one. Hence, if the number
of floors of structure (Nf) increases the steel reinforcement in short direction
(ASB) increases. This can be also seen for the increase of spacing between
columns in two directions (L1,B1) increment. The value of strength of
concrete (f’c) may also affect the quantity of (ASB) but with a smaller degree
than the load. The yielding stress of steel (fy) effect on (ASB) also. The effect
of unit price of concrete (Uc) is very little. It can be seen also that the effect
of unit price of steel (Us), cohesion of soil (c) and angle of internal friction
() can be neglected. In Fig. 4.7 shows the effect of design parameters on
steel reinforcement in long direction (ASL). As in studying (ASB), a steel
reinforcement in long direction (ASL) increases with the (Nf, L1, B1)
increment. The effect of yielding stress of steel (fy) and unit price of steel
(Us) can be seen clearly from Fig.4.7. The effect of design parameters on a
total quantity of steel reinforcement (Vs) is presented in Fig. 4.8.
Fig. 4.9 was presented to show the relation between percent difference
of design parameters % and ASB %, ASL %, Vs % respectively. From this
figure it can be seen the degree of importance of each parameter as a
percentage of increment or decrement. The slope of line joining the
reference values can give a clear idea about this importance.
 3000 3000

2600 2600

2200 2200
ASB

ASB
1800 1800

1400 1400

1000 1000
5 5.5 6 6.5 7 7.5 8 8.5 6 6.5 7 7.5 8 8.5 9 9.5
Nf L1

3000 3000

2600 2600

2200 2200
ASB

ASB
1800 1800

1400 1400

1000 1000
4 4.5 5 5.5 6 6.5 7 7.5 25 26 27 28 29 30 31
B1 f' c

3000 3000

2600 2600

2200 2200
ASB

ASB

1800 1800

1400 1400

1000 1000
340 350 360 370 380 390 400 410 25 27 29 31 33 35 37 39 41
fy Uc

3000 3000

2600 2600

2200 2200
ASB

ASB

1800 1800

1400 1400

1000 1000
240 250 260 270 280 290 300 310 70 75 80 85 90 95 100 105
Us c

3000

2600

2200
ASB

1800

1400

1000
10 15 20 25 30 35 40 45


Figure 4.6: Effect of Design Parameters on Steel Reinforcement in Short

Direction (mm2)
 5000 5000

4500 4500
4000 4000

ASL

ASL
3500 3500
3000 3000
2500
2500
2000
2000
5 5.5 6 6.5 7 7.5 8 8.5
5 6 7 8 9 10
Nf
L1

5000 5000

4500 4500

4000 4000
ASL

ASL
3500 3500

3000 3000

2500 2500

2000 2000
4 4.5 5 5.5 6 6.5 7 7.5 25 26 27 28 29 30 31
B1 f' c

5000 5000

4500 4500

4000 4000
ASL

ASL

3500 3500

3000 3000

2500 2500

2000 2000
340 350 360 370 380 390 400 410 20 25 30 35 40 45
fy Uc

5000 5000

4500 4500

4000 4000
ASL

ASL

3500 3500

3000 3000

2500 2500

2000 2000
240 250 260 270 280 290 300 310 70 75 80 85 90 95 100 105
Us c

5000

4500

4000
ASL

3500

3000

2500

2000
10 15 20 25 30 35 40 45


Figure 4.7: Effect of Design Parameters on Steel Reinforcement in

Long Direction (mm2)
 30 30
28 28
26 26
24 24
22 22
Vs

Vs
20 20
18 18
16 16
14 14
12 12
10 10
5 5.5 6 6.5 7 7.5 8 8.5 6 6.5 7 7.5 8 8.5 9 9.5
Nf L1

30 30
28 28
26 26
24 24
22 22
Vs

Vs
20 20
18 18
16 16
14 14
12 12
10 10
4 4.5 5 5.5 6 6.5 7 7.5 25 26 27 28 29 30 31
B1 f'c

30 30
28 28
26 26
24 24
22 22
Vs
Vs

20 20
18 18
16 16
14 14
12 12
10 10
340 350 360 370 380 390 400 410 20 25 30 35 40 45
fy Uc

30 30
28 28
26 26
24 24
22 22
Vs

Vs

20 20
18 18
16 16
14 14
12 12
10 10
240 250 260 270 280 290 300 310 60 70 80 90 100 110
Us C

30
28
26
24
22
Vs

20
18
16
14
12
10
10 15 20 25 30 35 40 45


Figure 4.8: Effect of Design Parameters on Total Steel Reinforcement (Ton)

From Figures 4.6, 4.7, 4.8 and 4.9 it can be concluded that:
1- The quantity of steel reinforcement in short and long direction (A SB,
ASL) increases if the number of floors (Nf) and spacing between
columns (L1, B1) increase and visa verse.
2- Steel reinforcement in short direction are not affected by the variation
of yielding stress of steel (fy) and the unit price of steel ( Us ) because
it is related with the value of minimum area of steel which depends
on the value of footing thickness (t).
3- The same effect of yielding stress of steel (fy) and unit price of steel
(Us) on the value of steel reinforcement in long direction (A SL) is
observed as well as the load applied from the structure on raft.
4.2.5 Effect of design parameters on depth of footing (Df)

Fig. 4.10 shows the effect of design parameters on the depth of

footing (Df). Each design parameters was taken separately. It can be seen
that the depth of footing (Df) increases with the increment of load with a
large degree. The depth of footing has the same effect of design parameters
in effective depth of footing (d or x 3). This means that the cohesion of soil
(c) and angle of internal friction ( (soil properties) do not affect the depth
of footing (Df) in the raft foundation design.
Fig. 4.11 shows the effect of design parameters on the depth of
footing as a percent difference. From this figure it can be seen if the design
parameter increase as (10 %), what amount of increment in (Df), that give a
clear idea about the effect of each parameter. 
 50
NF

40
L1

30 B1

20 f'c
ASB %

10 fy

Uc
0
-60 -40 -20 0 20 40 60 Us
-10
C
-20


-30
D esign P aram eters %

60
NF

50
L1
40
B1
30
f'c
20
ASL %

10 fy

0 Uc

-60 -40 -20 -10 0 20 40 60 Us

-20 C

-30


-40
D esign P aram eters %

100
NF

80
L1

60
B1

40 f'c
Vs %

20 fy

Uc
0
-5 0 -4 0 -3 0 -2 0 -1 0 0 10 20 30 40 50
Us
-2 0
C
-4 0


-6 0
D e sig n Pa ra m e te rs %

Figure 4.9: Design Parameters % versus (ASB, ASL, Vs) %

 20
NF

15
L1

10 B1

5 f'c
Df %

0 fy

-60 -40 -20 0 20 40 60 Uc

-5
Us
-10
C

-15


-20
D esign P aram eters %

Figure 4.11: Design Parameters % versus Depth of Footing %

4.3 Effect of Geometry

The meaning of the geometry is the layout of columns located entire

area of raft. The geometry of the raft foundation is not affected on the
structural design of raft. That can be seen from the formulation of the
problem. The bearing capacity constraint illustrated that the area of raft have
no effect on the design, The shear constraint is affected by the load and
spacing between columns in two directions. The constraints of the bending
moment, which are used to find area of steel in two directions, depend on the
load, spacing between columns. All of these constraints do not depend on
the geometry of raft; therefore this effect are not studied in this thesis.
 2 2
1.9 1.9
1.8 1.8
1.7 1.7
1.6 1.6
Df

Df
1.5 1.5
1.4 1.4
1.3 1.3
1.2 1.2
1.1 1.1
1 1
5 5.5 6 6.5 7 7.5 8 8.5 6 6.5 7 7.5 8 8.5 9 9.5
Nf L1

2 2
1.9 1.9
1.8 1.8
1.7 1.7
1.6 1.6

Df
Df

1.5 1.5
1.4 1.4
1.3 1.3
1.2 1.2
1.1 1.1
1 1
4 4.5 5 5.5 6 6.5 7 7.5 25 26 27 28 29 30 31
B1 f'c

2 2
1.9 1.9
1.8 1.8
1.7 1.7
1.6 1.6
Df

Df

1.5 1.5
1.4 1.4
1.3 1.3
1.2 1.2
1.1 1.1
1 1
340 350 360 370 380 390 400 410 20 25 30 35 40 45
fy Uc

2 2
1.9 1.9
1.8 1.8
1.7 1.7
1.6 1.6
Df

Df

1.5 1.5
1.4 1.4
1.3 1.3
1.2 1.2
1.1 1.1
1 1
240 250 260 270 280 290 300 310 60 70 80 90 100 110
Us C

2
1.9
1.8
1.7
1.6
Df

1.5
1.4
1.3
1.2
CHAPTER FIVE
1.1
1
MODELING
5.1 Introduction 10 15 20 25 30 35 40 45


Figure 4.10: Effect of Design Parameters on Depth of Footing (Df)

 Non-linear optimization approach in general is not easy to deal with,
since the application of such approach is sophisticated. This is quietly
coinciding with the optimum design of raft foundation. It requires a good
background in optimization detail (a user possessing the qualification to
solve many equations as a systematically using iteration procedure with the
aids of computers as has been illustrated in chapter three).
The usability of this approach necessitates different procedures for
different raft foundations in the same site. Hence any unskilled engineer in
non-linear optimization will not be motivated to use this method. The goal of
this chapter is to simplify the problem through the use of mathematical
models to change hard, complicated approach to more simple equations,
which takes into account design requirements in a similar form to the
conventional (classical) approach.
5.2 Input Parameters

In a previous chapter the effect of design parameters (Nf, L1, B1, f’c, fy,
Uc, Us, c and has been studied. From this study the effect of cohesion of
soil (c) and angle of internal friction (can be neglected in the design
because of the large area of raft that make their effects negligible. Therefore
the input parameters which can be used in the design of raft are (Nf, L1, B1,
f’c, fy, Uc and Us). The values of cohesion of soil (c) and angle of internal
friction (are taken as 50 kN/m2 and 20° respectively.
Table (5.1) shows the values of input parameters used in this analysis.
The computer program (Complex Method) was modified to be able to
perform this analysis. A total 0f 2187 trials were executed. Each trial
includes the chosen set of input parameter (7). The output data are design
variables (x1….x6) and the total cost of the footing as well as the volume of
footing and total quantity of steel reinforcement.

Table (5.1): Input Parameters Used in Analysis of

Modeling
No Investigated
Input Parameters Symbol Unit
. Values
1 Number of Floors Nf No. 5,6,7
Spacing Between Columns
2 L1 m 6, 7, 8
in Long Direction
Spacing Between Columns
3 B1 m 4, 5, 6
in Short Direction
4 Strength of Concrete f’c MPa 23,26,29
300, 350,
5 Yielding Stress of Steel fy MPa
400
Unit
6 Unit Price of Concrete Uc 20,30, 40
price
Unit
7 Unit Price of Steel Us 200,250,300
price

The output design variables (d, ASB, ASL, Df) can be calculated from
four equations, which are performed in the following sections of this
chapter.
Performing of mathematical models does not take the dimensions of
raft (L, B) into account because they do not exhibit any effect by design
parameters. The dimensions therefore are restricted with the total area of
structure.
5.3 Dimensional Analysis
Dimensional analysis is simply a mathematical tool. It is not
necessary to be a mathematician, however, to be able to use and to
understand dimensional analysis. Knowledge of it will enable the engineer to
save considerable time in planning process and in correlating results of a
process or using correlation prepared by others.
In order to apply dimensional analysis to a situation, the engineer need
to know only the variables believed to be involved and their dimensions.
Also user of dimensional analysis needs to use only these variables along
with fundamental algebra to save time, trouble and expense in investigation
or correlation of the problem. From dimensional analysis, it can be
determined if suspected variable is really involved in practical problem or
not.
In all cases dimensional analysis will reduce the number of
experimental variables to be correlated, and often it will point out the best
experimental approach to the problem, however; experiment must still be
relied upon for that purpose.
More complex deterministic models may be encountered when
economic consideration entire into the design making process. Instead of
proportioning an area to make the benefit in its maximum rate, the problem
assumes the form of mathematical models in which the objective function is
expressed in monetary terms and technological requirement takes the form
of constrains.
5.4 Formulation of the Mathematical Models

General equations to determine the main design variables of the raft

foundation have been established for deterministic optimization approach.
The computer program, which has been used to make this analysis, is the
same for all equations (described in chapter three). Different input
parameters have been used in this analysis. Number of points for each model
is taken as (243 points). Table (5.1) shows the input parameters and their
different values.

5.4.1 Generalize equation for effective depth of footing (d)

General equation to determine the effective depth of the raft

foundation (d) has been established. The equation includes five parameters,
number of floors of the structure (Nf), spacing between columns in long
direction (L1), spacing between columns in short direction (B1), compressive
strength of concrete (f’c) and unit price of concrete (Uc).

The equation that has been proposed for footing effective depth (d) is:

0.172 N f 0.609 L10.732 B10.572

d … (5.1)
f'c 0.586 U c 0.021

A computer package (STATISTICA) has been used to make the

analysis for the equation through a nonlinear regression analysis. The values
of the constants as a power of each parameter explain the importance of each
design parameter as well as the figures in chapter four. It has been found that
(R2) of the proposed equation is about 92.91%, which means that the
equation possesses a good agreement.
 A comparison between predicted and the observed values are shown
in Fig. 5.1.

0
.85

R2 =92.91%
0
.75

0
.65

0
.55
ObservedValues

0
.45

0
.35

0
.25
0.2
5 0
.35 0
.45 0
.55 0
.65 0
.75 0
.85
P
re
dic
tedV
alu
es

Figure 5.1: Predicted versus Observed Values of the Effective Depth (d)

5.4.2 Generalize equation for steel reinforcement in short direction (ASB)

A general equation has been proposed to determine area of steel
required per one meter length of footing in short direction (ASB), the
equation include five parameters which are, number of floors of the structure
(Nf), spacing between columns in long direction (L1), spacing between
columns in short direction (B1), yielding stress of steel (fy) and the unit price
of steel (Us).
 The equation has been proposed for steel reinforcement in short
direction (ASB) is:

0.509 0.533 0.461 0.058 0.073

A SB  57.24Nf L1 B1 fy Us …(5.2)

From the power of each parameter can be shown that number of floors
(Nf), the spacing between columns in long direction (L1) and spacing
between columns in short direction (B1) have more importance on the value
of steel reinforcement in short direction (ASB) than the other parameters. It
has been found that the value of (R2) of the proposed equation is about 94.96
%, which mean the equation can be used with a good agreement.
The relation between predicted and observed values shown in Fig.5.2.

2
000

1
800 R2= 94.69%

1
600

1
400
ObservedValues

1
200

1
000

8
00
80
0 1
000 1
200 1
400 1
600 1
800 2
000
P
re
dic
tedV
alu
es

Figure 5.2: Predicted versus Observed Values of The Steel

Reinforcement in Short Direction (ASB)
 5.4.3 Generalize equation for steel reinforcement in long direction (ASL)
A general equation has been proposed to determine area of steel
required for one meter length of footing in long direction (ASL), the equation
include five parameters, number of floors of the structure (Nf), spacing
between columns in long direction (L1), spacing between columns in short
direction (B1), yielding stress of steel (fy) and the unit price of steel (Us).
The equation has been proposed for steel reinforcement in long
direction (ASL) is:

0.334 1.339
Nf L1
ASL  30444 …(5.3)
0.564 0.817 0.035
B1 Us fy

From the power of each parameter can be shown the spacing between
columns in long direction (L1) and unit price of steel (Us) have more
importance on the value of steel reinforcement in long direction (ASL) than
the other parameters. It has been found that the value of (R2) of the proposed
 equation is about 92.65 %, which means the equation can be used with a
good agreement.
The relation between predicted and observed values is shown in Fig.
5.3

4
600
R2 =92.65%
4
000

3
400

2
800
ObservedValues

2
200

1
600

1
000
1
000 1
600 2
200 2
800 3
400 4
000 4
600
P
re
dic
tedV
alu
es

Figure 5.3: Predicted versus Observed Values of the Steel

Reinforcement in Long Direction (ASL)

5.4.4 Generalize equation for Depth of Footing (Df)

A general equation has been proposed to determine the depth of

footing (Df). The equation includes four parameters which are, number of
floors of the structure (Nf), spacing between columns in long direction (L1),
spacing between columns in short direction (B1) and compressive strength of
concrete (f’c).

The equation has been proposed for depth of footing (Df) is:

0.386 0.434 0.351

0.214Nf L1 B1
Df  …(5.4)
0.097
f'c
From the power of each parameter it can be shown that the number of
floors of structure (Nf), the spacing between columns in long direction (L1)
and the spacing between columns in short direction (B1) have more
importance on the value of depth of footing (Df) than the other parameters. It
has been found that the value of (R2) of the proposed equation is about
90.33%, which mean the equation can be used with a good agreement.
The relation between predicted and observed values is shown in Fig.
5.4

1
.35
R2= 90.33%
1
.25

1
.15

1
.05
ObservedValues

0
.95

0
.85

0
.75
0.7 0
.8 0
.9 1
.0 1
.1 1
.2 1
.3 1
.4
P
re
dic
tedV
alu
es

Figure 5.4: Predicted versus Observed Values of The Depth of

Footing (Df)
5.5 Comments About the Use of Mathematical Models

As shown in previous sections, it is easy to use the design equations

with a little effort. It is required to know what the unit of each design
parameter used in these equations and the units of the results of these
equations too.
The design parameters used in mathematical models have following
units:
Nf as a number, L1 in (m), B1 in (m), f’c in (MPa), fy in (MPa),
Uc in (Unit price/m3) and Us in ( Unit price/Ton)
Depending on above units for the design parameters, the results of
design variables which are gotten from design equations as follows:
d in (m)
ASL in (mm2) per meter length of footing
ASB in (mm2) per meter length of footing
Df in (m)

5.1.5 Numerical example

Find the effective depth, area steel in short direction, area steel in long
direction and depth of footing for the following data:
Nf = 6, L1 = 6 m, B1 = 4 m, f’c = 26 MPa, fy = 350 MPa, Uc = 30 Unit
price/m3, Us = 250 Unit price/Ton.
Solution: 1) by applying equations 5.1, 5.2, 5.3, and 5.4 respectively, the
results are:
d = 0.579 m
ASB = 1476 mm2/m length
ASL = 2500 mm2/m length
Df = 1.1 m
2) by applying the computer program
d = 0.576 m
ASB= 1460 mm2/m length
ASL= 2490 mm2/m length
Df = 1.13 m
 CHAPTER SIX
CONCLUSIONS AND RECOMMENDATIONS

6.1 Introduction

This thesis presents the application of optimization technique on the

design of raft foundation. The formulation of problem depends on the
requirements of the ACI-Code. Two examples are solved by the steps of
conventional method with the dimensions of basic problem used in thesis
and assumed depth of footing. These two examples are solved also by
optimization technique. A parametric study for optimization technique has
been tried. Four mathematical models for the application of the design have
been established.

6.2 Conclusions

From the theoretical study of this thesis and result obtained, the
following conclusions are drawn:
1. Unit price of concrete is more important than the other unit prices in
the design of raft foundation. The total cost of raft consists of 53% of
concrete cost, 39% of steel reinforcement cost, 7% of excavation cost and
1% of fill cost.
2. Change of load which represented by (Nf, L1,B1) have a larger amount
of influence on the total cost than the change of unit prices.
3. The change of properties of concrete and steel has a small amount of
influence on a total cost of raft.
4. The dimensions of the raft take the same dimensions of the structure
within the limits of the allowable bearing capacity.
5. The volume of concrete required depends on the effective thickness
(d) of raft.
6. The volume of concrete required affected by (Nf, L1, B1) larger than
the unit price of concrete.
7. Total quantity of steel reinforcement affected by (N f, L1, B1) larger
than the unit prices and properties of concrete and steel.
8. Using of optimization technique in the design of raft give a good
results for the ACI-Code requirements with the comparison by the
conventional design procedure.
9. The saving in total cost by the use of optimization technique by about
(6% - 7%) with the comparison of conventional design procedure.

6.3 Recommendations for Future Works

The following points are recommended to be considered in future studies:

1. Use the optimization technique in the design of raft foundation
installed in layered soil.
2. Applying the optimization technique for a raft subjected to eccentric
load.
3. Applying the four mathematical models for more sites to check their
adequacy.
4. Applying the optimization technique on the design of raft subjected to
different types of load such as dynamic load.
 ‫الخالصة‬

‫ذذذذإنذاهلدفذالرئقديذلؾؿفـدسذادلصؿمذيفذمجقعذاألعؿالذاإلنشائقةذاليتذوتمذادتخدامذاألدسذفقفاذوخاصةذاألدسذ‬
‫ذهوذاحلصولذعؾىذتصؿقمذوأخذذبـظرذاالعتبارذالؽؾػةذواالمان‪(Raft Foundation).‬احلصريوةذ‬
‫ذذذذمتذتطبققذنظروةذاحلؾولذادلثؾىذعؾىذأداسذحصرييذحمؿلذباألمحالذاالعتقادوةذاليتذتـؼؾفاذاألعؿدةذلؾحصولذعؾىذ‬
‫التصؿقمذادلطؾوبذوالؽؾػةذادلثؾىذ(الدنقا)‪.‬‬
‫صقغتذادلدألةذبافرتاضذدتةذمتغرياتذتصؿقؿقهذمتثلذأبعادذاألداسذومسؽهذالػعالذوكؿقاتذحدودذالتدؾقحذادلطؾوبةذ‬
‫باالجتاهنيذالطوليذوالعرضيذوعؿقذاألداس‪,‬ذحقثذاعتربتذكؾػةذاألداسذكدالةذهدف‪.‬ذمتذحلذمدألةذالتصؿقمذاألمثلذ‬
‫)‪ (COMPLEX METHOD‬بادتخدامذالطروؼةذادلعؼدةذ‬ ‫تطؾبتذهذهذالطروؼةذعشرةذحمدداتذتصؿقؿقهذ‬
‫واثـاذعشرذحمددذمتثلذالؼقمذالعؾقاذوالدنقاذلؾؿتغرياتذالتصؿقؿقه‪.‬ذمتذإجراءذبعضذالتعدوالتذعؾىذ)‪(Constraints‬‬
‫برنامجذعامذلؾحؾولذادلثؾىذلقأخذذبـظرذاالعتبارذمتطؾباتذهذاذالتصؿقمذوخؾقذنؼطةذالبداوةذلؽلذمتغريذتصؿقؿيذلؾتؿؽنذ‬
‫منذدرادةذالعواملذادلؤثرةذعؾىذالتصؿقمذ‪,‬هذهذالعواملذهيذعددذطوابقذادلـشأذالذيذودتـدذعؾىذاألداسذاحلصرييذ‬
‫وادلدافةذبنيذاألعؿدةذباالجتاهذالطوليذوادلدافةذبنيذاألعؿدةذباالجتاهذالعرضيذومؼاومةذاإلنضغاطذلؾؽونؽروتذوإجفادذ‬
‫‪(c and ).‬اخلضوعذحلدودذالتدؾقحذوكؾػةذالؽونؽروتذوكؾػةذالتدؾقحذباالضافةذاىلذخواصذالرتبةذاليتذمتثلذ‬
‫ذذذذمتذعؿلذدرادةذحولذتأثريذوأهؿقةذكلذعاملذمنذعواملذالتصؿقمذعؾىذالؽؾػةذالؽؾقةذوعؾىذأبعادذاألداسذوعؾىذمسؽهذ‬
‫الػعالذوبالتاليذعؾىذكؿقةذالؽونؽروتذالؽؾقةذوعؾىذكؿقةذحدودذالتدؾقحذباالجتاهذالطوليذوالعرضيذوبالتاليذالؽؿقةذ‬
‫الؽؾقةذحلدودذالتدؾقحذوعؾىذعؿقذاألداسذأوضاً‪.‬ذ‬
‫ذذذذمتذتؼدومذأربعةذمناذجذرواضقةذمنذخاللذادتخدامذ)‪( 2187‬ذحماولةذحبقثذوتدـىذلؾؿصؿؿنيذمنذاحلصولذعؾىذ‬
‫التصؿقمذادلطؾوبذدونذاخلوضذيفذالتػاصقلذادلعؼدةذلطروؼةذاحلؾولذادلثؾى‪.‬ذذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
 ‫ذ‬

Table (4.2): Input Parameters and Output Computer Results Used in the
Parametric Study

Input parameters Output computer results

L1 B1 Nf f’c fy Uc Us c  L B d ASB ASL Df

5 5 6 26 35 30 25 50 20 16.069 11.077 0.5910 1625.3 1809.5 1.0932
0 0 34 2 73 99 49 85
6 5 6 26 35 30 25 50 20 19.001 11.027 0.8921 2005.3 2031.5 1.4013
0 0 75 8 14 98 83 74
7 5 6 26 35 30 25 50 20 22.013 11 0.7462 1728.5 2644.9 1.2465
0 0 63 49 9 41 67
8 5 6 26 35 30 25 50 20 25.648 11.105 0.8259 1855.7 3174.9 2.1069
0 0 84 47 23 68 51 29
9 5 6 26 35 30 25 50 20 28.101 11.010 0.8473 1940.3 3980.4 1.3485
0 0 53 94 25 21 59 24
7 3 6 26 35 30 25 50 20 22.878 7.0018 0.5495 1300.2 3671.9 1.0515
0 0 4 5 75 9 35 82
7 4 6 26 35 30 25 50 20 22.038 9.0153 0.6366 1475.1 3140.3 1.3263
0 0 78 14 56 39 59 01
7 5 6 26 35 30 25 50 20 22.013 11 0.7462 1728.5 2644.9 1.2465
0 0 63 49 9 41 67
7 6 6 26 35 30 25 50 20 22.010 13.001 0.8141 1829.6 2401.9 1.7573
0 0 38 31 31 5 45 87
7 7 6 26 35 30 25 50 20 22.004 15.013 0.8999 2318.8 2161.5 1.4122
0 0 63 85 91 62 71 62
7 5 4 26 35 30 25 50 20 22.058 11.027 0.5664 1333.6 2321.1 1.0744
0 0 4 38 93 31 66 54
7 5 5 26 35 30 25 50 20 22.075 11.074 0.6513 1534.5 2518.6 1.3150
0 0 95 38 69 38 47 2
7 5 6 26 35 30 25 50 20 22.013 11 0.7462 1728.5 2644.9 1.2465
0 0 63 49 9 41 67
7 5 7 26 35 30 25 50 20 22.059 12.950 0.8007 1824.7 2862.4 1.6160
0 0 71 51 59 85 58 17
7 5 8 26 35 30 25 50 20 22.001 11.168 0.8642 1941.8 3029.9 1.3647
 0 0 73 7 79 28 76 45
7 5 6 22 35 30 25 50 20 22.437 12.942 0.7749 1796.4 2543.4 1.2840
0 0 85 37 78 28 01 34
7 5 6 24 35 30 25 50 20 22.082 11.209 0.7450 1737.3 2651.3 1.6193
0 0 92 87 96 21 12
7 5 6 26 35 30 25 50 20 22.013 11 0.7462 1728.5 2644.9 1.2465
0 0 63 49 9 41 67
7 5 6 28 35 30 25 50 20 23.029 11.001 0.7143 1628.6 2760.7 1.2151
0 0 83 25 43 87 62 38
7 5 6 30 35 30 25 50 20 22.240 11.401 0.6960 1604.9 2848.1 1.2076
0 0 4 17 41 71 69 94
7 5 6 26 30 30 25 50 20 22.041 11.051 0.7939 1788.5 2915.1 1.3030
0 0 02 85 19 31 88 61
7 5 6 26 32 30 25 50 20 22.034 11.062 0.7286 1667.5 2910.7 1.2292
5 0 55 34 5 93 77 76
7 5 6 26 35 30 25 50 20 22.013 11 0.7462 1728.5 2644.9 1.2465
0 0 63 49 9 41 67
7 5 6 26 37 30 25 50 20 23.971 11.067 0.7271 1655.1 2527.8 1.2280
5 0 95 2 4 96 67 94
7 5 6 26 40 30 25 50 20 22.032 11.018 0.7295 1660.5 2368.7 1.2303
0 0 91 76 26 6 85 33
7 5 6 26 35 20 25 50 20 22.009 11.018 0.8060 1812.6 2427.2 1.3063
0 0 91 36 34 63 59 46
7 5 6 26 35 25 25 50 20 22.072 11.091 0.7275 1655.6 2707.7 1.5052
0 0 88 14 26 7 84 68
7 5 6 26 35 30 25 50 20 22.013 11 0.7462 1728.5 2644.9 1.2465
0 0 63 49 9 41 67
7 5 6 26 35 35 25 50 20 22 11.027 0.7275 1657.9 2703.6 1.2288
0 0 01 18 38 75 54
7 5 6 26 35 40 25 50 20 22.011 11.025 0.7271 1676.1 2706.8 1.2333
0 0 73 12 91 43 49 18
7 5 6 26 35 30 20 50 20 22.023 11.019 0.7270 1672.9 2710.9 1.2284
0 0 74 83 55 91 47 47
7 5 6 26 35 30 22 50 20 22.232 11.009 0.7288 1794.6 2698.1 1.2355
0 5 86 52 19 01 98 38
7 5 6 26 35 30 25 50 20 22.013 11 0.7462 1728.5 2644.9 1.2465
0 0 63 49 9 41 67
7 5 6 26 35 30 27 50 20 22.328 11.024 0.8385 1878.0 2415.7 1.3421
 ‫‪0‬‬ ‫‪5‬‬ ‫‪82‬‬ ‫‪41‬‬ ‫‪68‬‬ ‫‪87‬‬ ‫‪29‬‬ ‫‪71‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪30‬‬ ‫‪50 20‬‬ ‫‪22.010‬‬ ‫‪11.003‬‬ ‫‪0.7541‬‬ ‫‪1788.4‬‬ ‫‪2604.2 1.2579‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪33‬‬ ‫‪3‬‬ ‫‪06‬‬ ‫‪91‬‬ ‫‪61‬‬ ‫‪19‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪30 20‬‬ ‫‪22.025‬‬ ‫‪11.001‬‬ ‫‪0.7654‬‬ ‫‪1732.2‬‬ ‫‪2564.6 1.2834‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪94‬‬ ‫‪95‬‬ ‫‪42‬‬ ‫‪52‬‬ ‫‪72‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪40 20‬‬ ‫‪22.018‬‬ ‫‪11.000‬‬ ‫‪0.7273‬‬ ‫‪1724.8‬‬ ‫‪2708.6 1.2294‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪19‬‬ ‫‪26‬‬ ‫‪16‬‬ ‫‪25‬‬ ‫‪42‬‬ ‫‪89‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪50 20‬‬ ‫‪22.013‬‬ ‫‪11‬‬ ‫‪0.7462‬‬ ‫‪1728.5‬‬ ‫‪2644.9 1.2465‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪63‬‬ ‫‪49‬‬ ‫‪9‬‬ ‫‪41‬‬ ‫‪67‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪60 20‬‬ ‫‪22.046‬‬ ‫‪11.085‬‬ ‫‪0.7272‬‬ ‫‪1668.7‬‬ ‫‪2712.2 1.2314‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪94‬‬ ‫‪22‬‬ ‫‪34‬‬ ‫‪26‬‬ ‫‪91‬‬ ‫‪2‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪70 20‬‬ ‫‪22.004‬‬ ‫‪11.013‬‬ ‫‪0.7270‬‬ ‫‪1658.8‬‬ ‫‪2707.7 1.2383‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪62‬‬ ‫‪83‬‬ ‫‪48‬‬ ‫‪15‬‬ ‫‪35‬‬ ‫‪1‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪50 10‬‬ ‫‪22.431‬‬ ‫‪11.016‬‬ ‫‪0.7627‬‬ ‫‪1731.1‬‬ ‫‪2668.6 3.3212‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪15‬‬ ‫‪97‬‬ ‫‪95‬‬ ‫‪77‬‬ ‫‪72‬‬ ‫‪53‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪50 20‬‬ ‫‪22.018‬‬ ‫‪11.000‬‬ ‫‪0.7273‬‬ ‫‪1724.8‬‬ ‫‪2708.6 1.2294‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪19‬‬ ‫‪26‬‬ ‫‪16‬‬ ‫‪25‬‬ ‫‪42‬‬ ‫‪89‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪50 30‬‬ ‫‪22.013‬‬ ‫‪11‬‬ ‫‪0.7462‬‬ ‫‪1728.5‬‬ ‫‪2644.9 1.2465‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪63‬‬ ‫‪49‬‬ ‫‪9‬‬ ‫‪41‬‬ ‫‪67‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪50 40‬‬ ‫‪22.046‬‬ ‫‪11.085‬‬ ‫‪0.7272‬‬ ‫‪1668.7‬‬ ‫‪2712.2 1.2314‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪94‬‬ ‫‪22‬‬ ‫‪34‬‬ ‫‪26‬‬ ‫‪91‬‬ ‫‪2‬‬
‫‪7‬‬ ‫‪5‬‬ ‫‪6 26‬‬ ‫‪35‬‬ ‫‪30‬‬ ‫‪25‬‬ ‫‪50 50‬‬ ‫‪22.004‬‬ ‫‪11.013‬‬ ‫‪0.7270‬‬ ‫‪1658.8‬‬ ‫‪2707.7 1.2383‬‬
‫‪0‬‬ ‫‪0‬‬ ‫‪62‬‬ ‫‪83‬‬ ‫‪48‬‬ ‫‪15‬‬ ‫‪35‬‬ ‫‪1‬‬

‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
‫ذ‬
 ‫ذ‬
References

1. American Concrete Institute: “ACI Standard Building Code

Requirements for Reinforcement Concrete”, ACI 318-17-1999.
2. Abdul Rahman, S. T., (2000), “Optimum Cost Design of Cantilever
Retaining Wall”, M. Sc. Thesis, University of Baghdad, College of
Engineering.
3. American Concrete Institute: “Suggested Design Procedures for
Combined Footings and mats”, Manual of Concrete Practice,
ACI336.2R-1966
4. Al-Ani M. M., (1978), “Optimization of Foundation Design”, M.Sc.
Thesis, University of Baghdad, College of Engineering, Civil
Engineering Department.
5. Al-Ani, M. M., (1998), “Further Contribution to the Analysis of the
Axially Loaded Piles”, Ph.D. Thesis, University of Technology,
Building and Construction Departement.
6. Al-Rawi O. F. A., (2001), “Optimum Design of Spread Footing
Deterministic and Stochastic approach”, M.Sc. Thesis, University of
Technology, Building and Construction Department.
7. Al-Jubair, H. S., (1998-a), “Optimization Design of Braced Sheet
Piling”, The Scientific Journal of Tikrit University, Vol. 5, No.2,
pp.31-44.
8. Al-Jubair, H. S., (1998-b), “Cellular Cofferdam Design
Mathematical programming Approach”, The Scientific Journal of
Tikrit University, Vol. 5, No.1, pp.39-57.
9. Al-Samarrai, Q. A., (1998), “Optimum Design of Tied Bach Hinged
Retaining Walls”, M. Sc. Thesis, University of Tikrit, College of
Engineering, Civil Departement.
10.Beightler C. S., (1979), “Foundations of Optimization”, Second
Edition, Prentice-Hall Inc., N. J. U.S.A.
11.Bowles, J. E., (1984), “Foundation Analysis and Design” Third
Edition, McGraw-Hill Book Company.
12.Bowles, J. E., (1984), “Foundation Analysis and Design” Fifth
Edition, McGraw-Hill Book Company.
13.Bunday, B. D., (1984),”Basic Optimization Methods”, School of
Mathematical Sciences, University of Bradford, Edward Arnold
Publishers.
14.Desai, I.D., Desai, G.N. and Desai, T.B. (1984), “Cost Optimization
of Isolated Sloped Footing in Granular Medium”, First Conference
on the Application of Computers in Civil Engineering, New Delhi.
15.Dieter, G., (1983), “Engineering Design”, McGraw-Hill, Japan.
16.Hynes, M. E, Vanmarcke, E. H, (1975), “Subjective Reliability
Assessment in geotechnical Engineering (Embankment Case
Study)”Massachusetts Institute of technology Dept. of Civil
Engineering.
17.Ibrahim, B. I., (2001), “Optimum Cantilever Retaining Wall Design
Deterministic and Stochastic approaches”, M.Sc. Thesis, University
of Technology, Building and Construction Department.
18.Kapur, K. C., Lamberson, L. R., (1977), “Reliability in Engineering
Design”, wiley and sons, New York.
 19.Keskar, A. V. and adidam, S. R., (1989), “Minimum Cost Design of a
Cantilever Retaining Wall”, The Indian Concrete Journal, Vol.63,
No. 8,pp. 401-405.
20.Morris, A. J., (1982), “Foundations of Structural Optimization: A
unified Approach”, Macmillian India Ltd. Bagalore.
21.Namiq, L. I. And Al-Ani, M. M.,(1983),”Optimization Design of
Footing ”, Iraqi Conference on Engineering, pp.8-12.
22.Nilson, A. H. and Winter, G.,(1986),”Design of Concrete structures”,
Third Edition, McGraw-Hill, Singapore.
23.Peck, R. B., Hanson, W. E., Thornburn, T. H., (1974), “Foundation
Engineering”, Second Eddition, John Wiley and Sons.
24.Rao, S. S.,(1977),”Optimization Theory and applications”, Second
Edition, Wiley Eastern Limited.
25.Simmons, D. M., (1975),” Nonlinear Programming for Operations
Research”, Prentice-Hall International, Inc.
26.Teng, W. C., (1983),”Foundation Design”, Prentice Hall of India
Private Limited, New Delhi.
27.Venkatramaiah, C., (1993),”Geotechnical Engineering”, Wiley
Eastern Limited, New Delhi.
28.Wang, C. K. and Salmon, C. G., (1992), “Reinforced Concrete
Design”, Fifth Eddition, Harper Collins.
29.Wintercorn, H. F. and Fang H. Y., (1975),” Foundation Engineering
Hand Book”, Van Nostrand Reinhold Company Inc. U.S.A.

View publication stats