CURVE FITTING AND OPTIMIZATION

216
minimization) of the objective function z
The optimization (maximization or AX Bis the mathematical formulation ofa
subject to the constraints AXsB or
LPP.

Aset of real values X= ( y ) which satisfies the constraint AX S() B

is called solution.
values x, which satisfies the constraints and also satisfy non negativity
Aset of real
constraints x, > 0 is called feasible solution.
along with non
Aset of realval1 es x, which satisfies the constraints optimal solution.
negativity rest1 ictions and optimizes the objective function is called
function is called
In otherwords feasible solution optimizing the objective
optimal solution.
Note :

1. An LPP can have many optimal solutions.

2. If the optimal value of the objective function is infinity then the LPP is said tohave
unbounded solution. Also an LPP may not possess any feasible solution.
|4.32 Graphical method of solving an LPP
LPP involved with only two decision variables can be solved in this method.
The method is illustrated step wise when the problem is mathematically formulated.
The constraints are considered in the form ofequalities. Obviously these represent
straight lines since there are only two decision variables.

3 The equations are put in the form +=1

b which graphically represents
straight line passingthrough the points (a, 0) and (0, b).
These ines along with the co-ordinate axes forms the boundary of the region
known as the feasible region and the figure so formed by the vertices is called the
convex polygon.
The value of the objective function Z is found at all these vertices.
The optimum/extremevalues of Z(maximum or minimum ) among these values
corresponding to the values of the decision variables is the required optimal
solution of the LPP.

WORKED PROBLEMS
16. Maximize z =x+15y given x 20, y 2 0 subject to the constraints :
x+ 2y s 160, 3x +y s 240 by graphical method.
>> Let us consider the equations,
 oPTIM

3 Optimization
4.3
Optimizationis atechnique of obtaining the best results under the
ces.Optimization means maximization or minimization.
circumstances. prevailing/given
Programmingisa mathematicaltechnique to determine theoptimum use of the limited
available resources.
Linear.rncramming
programmingisis,a decision making technique under the given constraints on the
Ghat the relationship among the variables involved is linear. A general
condition that
relationship among the variables involved is called objective function. The variables
involvedare called I decision variables.
Slution by linear programming methods aims at optimization.
We first present the mathematical formulation of aLinear Programming Problem (LPP)
andproceed to discussthe solution of LPP by Graphical method and Simplex method
431 Mathematical formulation of a linear programming problem
Alinear programming problem is mathematically formulated by first identifying aset
of variables X Xy x,, which are subject to certain linear conditions known as
constraints written in the form of inequalities.

"m1 t+amn y S(2) b,

t m2 X
where the coefficients a,;, b, (1 sis m, 1 j s n) are constants and
x, 20, x, > 0 x20.
or AX 2 B where
The set of inequalities can be put in the matrix form A Xs B

A = X = B=

"m1 "m2"mn b
m

x, X, along with the give

Theobjective function involving the variables
of the for
 217
1()

the straight line AR jolning the

veY the straight
C joining C(80, 0pointn A(
line CD 10,0),
), D(0, 120 ) B0,) And
and CDinthe XOY plarne.
1204D
904
604 02
30+

40 80 120 160

Shaded portion is the feasible region and O CEB is the convex polygon. The point E
sthe point of intersection of the lines AB and CD. We can obtainthís point by solving
Simultaneously the system of equations
+ =l60 and 3r +2y = 240 On solving we obtain E(x,y) (40, 60)
The value of the objective function at the cornersof the convex polygon 0CEBare
abulated.
Corner Value of Z = x+ 1.5y
O(0,0) 0

80
C(80, 0 )
130
E(40, 60 )
120
B(0, 80 ) maximum value equal to 130.
(40, 60 ) & has the
nus Zis maximumn at E (x,y)=(
subject to the constraints: x+2y 2 10, x+y2 8,
4y
Minimize Z = 5x + by graphical method.
0, y z0
2r +y 12, x
12, equations: 12
; 2x + y
Let us considerthe (3)
I+2y = 10 .. (2) ; 6
+
(1) ; 8 12
 CURVE FITTING AND

218
OPTIMIZATION
straight lines AB, CD, EF where
we have,
representthe
Let (1), (2), (3) reespectively =(0, 8) ; E = (6,0 ), F = (0
1os
(0,5) ; C= (8, 0), D
A =(10, 0), B =
plane.
We draw these lines in the XOY

9
9Tp
D

6B

0 2 4 6E 8C 10 X

Shaded portion is the feasible region. Gis the point of intersection of the lines AB &
& EF.
CDand H is the point of intersection of the lines CD
By solving x+ 2y = 10 and x+y = 8 we get G = (6, 2 ).
Also bysolving +y = 8, 2x +y = 12 we get H = (4,4).
Here the feasible region is unbounded and we tabulate the value of the objective
function corresponding to the points A, G, H, F.
Corner Value of
Z = 5x + 4y
A(10, 0) 50

G(6,2) 38

H(4,4) 36

F(0, 12 ) 48
Thus Z is minimum for x= 4, y =4 and the minimum value is 36.

18. Use the graphical method to maximize z = 3x+ 4y subject to the constraints,
2x +y S 40, 2x + 5y S 180, x 2 0, y0.
>> Let us consider the equations :
2x +y = 40 ; 2x +5y = 180
X 2 =1 ...(1) + ...(2)
20 40 90 36 = 1
PMATRW.GRAPHICALMEETHOD
straightlines 219
la(Wand )representthe
(2). AB and CD
respectively
A=(20, 0), B = (0,40 ) ; C-= (90,0), D =(0,36 ) where we have
theselinesin XOY plane.
We
draw

40 B
DAE
36
30

20

101
A C
0 20 40 60 80 90 100 X

Shadedportionis the feasible region and OAED is the convex polygon. The point E
ins the point of intersection of lines ABand CD is obtained by solving the equations:
+y= 40, 2x +5y = 180. E (x, y) = (2.5, 35 )
Te value of the objective function at the corners of the convex polygon OAED are
tabulated.

Corner Value of Z = 3x+ 4y

O(0,0)
A(20, 0) 60

147.5
E(2.5, 35)
D (0,36 ) 144
Thus Zis maximum at the 35) & has the maximum value equal to 147.5
vertex(2.5,
constraints,
subject to the
oe the graphical method to minimize z = 20 x +10 22
Ih+2x, s 40, 3x, t Xg 2 30, 4x, +3x, 260, x, 20, x, 2 0.
»
Let uS consider the
equations = 60

I+2r, =40 i 3x, +x, = 30 .(3)

= 1
...(2) i 1520
40 20 = 1 ...(1) +
30
10
 respectively where weOPTIhave
MIZATION
Let (1), (2), (3) represent the
A = (40, 0), B = (0, 20
straight lines
AB, CD, EF
); C = (10, 0), D =
We draw these lines in the (0,30); E = (15,0), F =
XOY plane. (0, 20 )
40
30D

20 B
H

10?

10 15 20 30 40
Shaded portion is the feasible region andE AHGis the
G is the point of
intersection the lines CD and EF. H isassociated
the lines AB and CD. of convex
the point of polygon.
intersection of
On solving 3x, +X, = 30 & 41, +
3x, = 60 we get G(x, x,)= (6,
On solving 3x, + X, = 30 & x 12)
+2x, = 40 we get H(x, x, ) = (4, 18)
The value of the objective
tabulated. function at the corners of the convex
polygon EAHG are
Corner Value of Z = 20x, + 10x,
E(15, 0 ) 300
A (40, 0) 800
H(4,18) 260
G(6, 12) 240
Thus Z is minimum at the
vertex (6, 12 ) and the
minimum value is 240.
0. Minimize Z = 30x + 20y
subject to the constraints, x-y S 1, X+y > 3,y S4
and x > 0, y 2 0 by graphical
method.
Let us consider the equations
X-y = 1 i X+y = 3 ;y= 4
1
+ y 1 ...(1); 3 1
-1
...(2): y = 4
 GRAPHICAL METHOD
(PTMIZATION.

Let(1)and represent thestraight lines, AB and

(2)
221
A= (1,0), B (0, -1);
CDwhere we have,
lines,
:C-(3,0), D(0,3y
is a straight line parallel to
the X-axis,
Wedrawthese

B
0 1
3

-1fA
Shaded portion is the feasible region and
Eis the point of EDFG is the associated convex polygon.
we get E(x, y) intersection
of AB and CD. On solving x -y = 1 x
=(2, 1). Also F(x, y) = (0, and +y =3
4)
Gis obtained by solving y = 4 and x
-y= 1. That is G(x, y) = (5, 4)
Ihe value of the objective function Zat the
corners of the convexpolygon are tabulated.
Corner Value of Z = 30x + 20y
E(2, 1) 80

D(0, 3) 60
80
F(0, 4)
230
G(5, 4)
Thus Zis minimum at the vertex ( 0, 3) and the minimum value is 60.
21. Use 3aphical method to solve the following LPP, Minimize Z 8r+ 5y subject to
s 10 &x 2 0, y> 0.
6x +y 2 21,x+3 3y > 12,x+y
S> Let us
consider the equations I+3y = 12
i+y 10
6 +y = 21 . .(2) ; + = 1
10 10 ...(3)
(1) ; 12 4
35 +=
+
21
1
 TIMIZATION
Let (1) (2) and (3) respectively represent the straight lines AB, CD and EF
we have, where
A= (3.5, 0), B= (0, 21) ; C = (12, 0), D= (0, 4):E =(10, 0), F = (0. 101
We draw these lines in the XOY plane.
21fB
20

151
F
10[

0 3.5 5 E 10 12

Shaded portion is the feasible region and G, H, I are the vertices of the convex polygon
which is a triangle.
On solving: x+3y = 12 and x+y= 10 we get G(x, y) = (9, 1)
6x +y 21 and x+3y = 12 we get H(x, y) = (3, 3)
*+y = 10 and 6x +y = 21 we get
I(x, y) = (2.2, 7.8)
The value of the objective function at these points are
tabulated.
Corner Valueof Z = 8x + 5y
G(9, 1) 77
H(3,3) 39
I(2.2, 7.8 ) 56.6
Thus the minimum value of Z is 39 at the
vertex (3, 3).
2. Show that the following LPP does
not have any feasible solution.
Dbjective function for maximization : Z = 20x + 30v.
Constraints :3x +4y S 24, 7x +9y 2 63, x2 0, y 20 (use
Let us consider the equations graphical method )
3x + 4y = 24
7x+9y =63
1 ..(1)
8
+
6 G+-1
9 (2)
 )respectively represent
(2).
and the straight 223
Let(1) 0), B=
A= (8, (0, 6) ; lines AB and CD
straightlines are drawn in the
;C= (9,0), D= (0, 7) where we have
The XOY plane.

B6

2 4 6 A8 10 X

Itisevident that there is no feasible region.

Thus we conclude that the LPP does not have any feasible solution.

23. Show that the following LPP has unbounded optimal solution. Objective function to be
Maximized :Z = 8x + 5y Constraints : x-2y s 12, x +2y 2 20, x 2 0, y20.
>» Let us consider the equations.
x-2y = 12 x+2y = 20
+ = 1 ..(1) 20 10 ] (2)
12 -6
)and (2) respectively represent the straight lines ABand CD where we have,
A= (12, 0), B= (0, -6) ; C = (20, 0), D = (0, 10)
Let us draw these straight
lines.
10
5

0 20 25
5 10 12 15
54
6 B
10 unbounded.
regionis
this
Shaded
Thus weportion
evidentthatsolution.
is the feasible region anditis optimal
the LPP has unbounded
 224 CURVE FITTING AND OPTIMIZATION
24. Solve the following LPP graphically.
Maximize Z=3x +5x, subject to x, +2x, s 2000, * t , s 1500, X, s600,
>> Let us consider the equations
X +2x, =2000 i tx, = 1500 i X= 600

2000 + 1000 1 ...(1) ;

1500
+
1500 =1 .. (2) x, = 600
Let (1) and (2) respectively represent the lines AB and CD where we have,
A = (2000,0 ), B = (0,1000 ) ; C (1500,0), D =
(0,1500 )
x, = 600 is a line parallel to the x, axis.
Let us draw these straight lines.

2000+

15004D

1000 B
600
X, = 600
500

500 1000 1500 2000 2500

On solving
X} +2r, = 2000, X +x, =
1500 we get E =
= 1500, X, =600 (1000,, 500 )
we get F =(900, 600 )
+2r, =2000, X = 600
Also we have C = we get G= (800, )
(1500,0 ), H = (0,600 ) 600
The value of the objective
convex polygon. function at these vertices are
tabulated. OCEFGH is the
GPTIMZATIN.
.GRAPH
225
Corner Value of Z = 3x, + 5x,
O(0, 0)
C( 1500, 0) 4500
E(1000, 500 ) 5500
F(900, 600 ) 5700
G(800, 600 ) 5400

Thus (ZMar 5700 when x= 900, X, = 600

constraints .
Moximize z = 50 x +60 X Subject to the 2x +3*, s 1500
3r,+ 2r, s 1500, 0 S x, S 400, 0 S x, S 400, x, >0, x, 20
method)
(usegraphical
equations
>Let us consider the
= 1500 ; 3x, +2x, = 1500, x, = 400, x, = 400

+ = 1 ...(1) ; + = 1 (2)
750 500 500 750

Let (1) and (2) respectively represent the straight lines AB and CD where we have,
A = (750, 0), B =(0, 500 ) ; C= (500, 0 ), D =(0, 750)
Let us draw these straight lines.
8001X2
750%D

600+
5004B
400t X, =400
> i = 400
200
A
750 1000
200 400 500 600
h (400, 150)
solving 3x, +2x4 = 1500 and x, = 400
we get F=

G= (300, 300)
we get
2x, +3x, == 1500 and 3x, + 2x, = 1500 = ( 150, 400)
weget H
2x +3x2 =1500 and x, =400
 Also E = (400, 0) ; I= (0, 400 )
of the objective function at these vertices are tabulated. OEFGHI ia
The value
convex polygon.
Corner Value ofZ = 50x + 60x,
O(0, 0) 0

E (400, 0) 20,000
F(400, 150 ) 29,000
G(300, 300 ) 33,000
H(150, 400) 31,500
I(0, 400 ) 24,000
(Z)Mr =33,000 at x, = 300 and x, = 300

throughtw0
26. Tuo spare parts X and Yare to be produced in a batch. Each one has togo available
processes A and B. The time required in hours per unit and the total time are
given.
X Y Total hours available
Process A 4 48
Process B 9 4 72

Profits per unit of Xand Yare rupees 5 and 6 respectively. Find how many number of spare
parts of X
and Yare to be produced in this batch to maximize the profit. (Use graphical method)
>> The mathematical formulation of the LPP is to maximize Z = 5X+6Y subject to
the constraints 3X +4Y s 48, 9X+4Y s 72.
Let us consider the equations
3X +4Y = 48 9X+4Y =72

XY
16+ 15= 1 X Y
= 1
12 818
8

Let ABand CD represent the straight lines where we have

A = (16, 0), B = (0, 12): C = (8, 0), DD= (0, 18)
The straight lines are drawn in the XOY plane.
PMATN.GRAPHICAL METHOD
227

24
D
18

12%B
E

69

A
4 8 12 16 20 X
Ohsolving3X.+4Y = 48 and 9X +4Y = 72 we get E ( X, Y) = (4, 9). The value of
theobjective function at the vertices of the convex polygon OCEB are tabulated.
Corner Value of Z = 5X + 6Y
O(0, 0 ) 0

C(8, 0) 40
E(4, 9) 74

B (0, 12 ) 72

= 74 at (X, Y)= (4, 9)

(Z)Max
Thus we can say that 4 Nos. of X and9 Nos. of Y spare parts are to be produced to
get the maximum profit of Rs.74

21. Achemist
vishes to provide for his customers, at the least cost the minimum daily
requirements
The amount of of each
three vitamins 1, 2and 3by using a mixture of two products Mand N.
and the vitamin in one, gram of each product, the cost per gram of each product
minimum daily requirements are given below.
Number of units of each vitamin contained in a gram Cost per gram
of each product ofeach
Vitamin 1 Vitamin 2 Vitamin 3 product
Product M 6 20 Paise
Product N 2
2 4
16 Paise

reeachqMiuirnevitamin
immeumnt of 2 12

12 8 24
Find the
Uitamims east e
Cxpensire
(use the combination which provides the minimum requircments of these
8raphical
 CURVE FITTING AND
>> We shallfirst mathematically formulate the OPTIMIZATION
Let
problem :
x= Number of grams of product M in the mixture.
y= Number of grams of product N in the mixture.
From the data the constraints of the L.P.P are
6x + 2y 12, 2x +2y 8, 4x +12y > 24.
The problem is to minimize the
objective function Z = 20x + 16y.
Let us consider the equations,
6x + 2y = 12, 2+2y = 8, 4x +
12y = 24
2=1 ..1) +-11
2
+
6
4 4 (2)
6T2Y 1 . .(3)
Let (1),(2), (3) represent the
straight lines AB, CD, EF respectively where we
A = (2, 0), B =
(0,6), C= (4,0); D = (0,4), E = have
We now draw these lines in (6, 0), F = (0,2)
XOY plane.

4 D

0
2 4
6

Shaded portion is the feasible

intersection of the lines CDandregion
AB.
which is unbounded. The
EF. The point H is the point Gis the point of
The point G(x, intersection of the lines CDana
y)= (3, 1) is obtained
The point H(x, y) =(1, 3) is by solving 2r + 2y = 8
obtained by and 4x+ 12 y = 24.
The value of the objective function atat solving 2r +2y = 8 and 6
: = 12..
the corners E,
G, H,
x+2y
Bare tabulated.
 GAAPHICAL METHOD
TION.
oPTIMIZA
229
Points Value of Z = 20x + 16y
120
E (6, 0 )
G(3, 1 ) 76

H(1,3) 68

B(0,6) 96
Minimumcostis 68 paise. Thus the least expensive combination is 1 part of the
poduct Mand 3parts of the product Nat the minimum cost of 68 paise.

the production
Inrequired of two type of watches afactory uses three machines A, B, C. The time
for each: watch on each machine and the maximum time available on each machine
isgiven below.
Maximum
Time required time available (In
Machine Watch I Watch II hours)

8 380
A 6
300
8
404
4
12
Rs.30. Find what combination should be
and on watch Il is graphical method).
proton watchIis Rs.50 is the maximum profit(use
produced for the maximum profit. What
formulate the problem.
>> We shall first mathematically and y= number of type l
watches
be produced
Let x = number of type I watchesto
to be produced. following constraints has to be
subjectto the
+ 30y,
Ihe objective function Z = 50x
4yS 404
maximized. + 4y s 300, 12r +
380, 8x
6x + 8y equations 300 12x t 4y = 404
4y =
Let us consider the ; 8r t
=1.. (2) : + =1 . .(3)
6x + 8y = 380 ...(1) ; 37.5
75 33.7 101

+
= 1
straightlines AB, CD, EF espectively where we have
47.5 the
63.3 represent (37.5, (0) E= ( 33.7, 0 )
C=
Let (1), (2), (3) (63.3, 0) 75) F= (0, 101 )
D=(0,
A = 47.5)
(0,
B =
 230

lines in the XOYplane.

We draw these Y
10
100
90
80 D
70/
60
50 B
40
30
20
10
10 20 30EC40 50 60A 70 X

associated convex polygon is O E GHB.

the feasible region and the
Shaded portion isintersection His that of the
intersection of
of the lines CD and EF &
Gis the point of
the lines AB and CD.
obtained by solving 8x + 4y = 300 and 12x + 4y = 404
G(x, y) = (26, 23 ) is 6x + 8y = 380.
solving + 4y = 300 and
8x
H(x,y) = (22, 31 ) is obtained by
function at the corners of the convex polygon OEGHB are
The value of the objective
tabulated.

Corner Value of Z = 50x + 30y

O(0,0)
E (33.7, 0) 1685

G( 26, 23 ) 1990

H(22, 31) 2030

B(0,47.5 ) 1425

7 is maximumat the vertex H(22, 31) and the maximum value of Z is 2030.
It can be seen from the figure that His the farthest point from the origin.
Thus the factory has to produce 22 watches of type land 31 watches of type II to
obtain the maximum profitof Rs.2030.
232 CURVE FITTING AND OPTIMIZATION

The value of Z = 2x + 3y at x = 4/9, y = 5/9 is 23/9. It can be easily seen that

value of Z at A(1,0) is 2and the value of Zat D (0,5) is 15.
Zis minimum at A(1,0) and the minimum value is 2.
Thus one unit of food stuff F, and no unit of F, is the least expensive
combination and the cost is Rs.2.

30. Afactory uses 3 types of machines to produce two types of electronicgadgets. The first
gadget requires in hours 12, 4 and 2 respectively on the three types of machines. The second
gadget requires in hours 6, 10 and 3on the machines respectively. The total available time in
hours respectively on the machines are 6000, 4000, 1800. If the two types of gadgetsrespectively
fetches aprofit of rupees 400and 1000 find the number of gadgets of each type to be produced
for getting the maximum profit.
>> Let r be the number of gadget of type-l and ybe the number of gadget of type-2.
By data the constraintsare as follows.
12x +6y s 6000, 4x +10y s 4000, 2x +3y < 1800
The objective function need to be maximized is Z = 400x + 1000y
Let us consider the equations.
12x + 6y = 6000 ; 4x +10y = 4000; 2x +3y = 1800; x 0, y 0

500 1000 = 1 ; 1000 400 =1; 900 600o =

= 1

Let AB, CD, EF respectively represent these three lines where we have,
A = (500, 0) C = (100, 0) E = (900, 0)
B = (0, 1000 ) D= (0, 400 ) F = (0, 600 )

1000Y
B

8001

600

D
400n

200

200 400 500 600 800 E

1000 X
 helinesare.drawn in the XOY plane.
the point of intersectionof 233
6000
the lines AB and
Onsolving
12x +6y m and 4x +10y CD,
Ihe value the objective function at the
of 4000 we get G(x, y)
tabulated. corners of the convex (375, 250 ),
Corner
polygen OAGD are
O(0,0) Value of Z 400x +1000y
(0)
A(500, 0)
G(375, 250 ) 2,00,000
D(0, 400 ) 4,00,000
4,00,000
han be seen that the L.P.P has two optimal
solutions,
Thus we conclude that the factory has to
gadgets of type-2 or 400 gadgets of type-2produce 375 gadgets of type-l and 250
only to get the maximum profit.
A13 Preamble for the Simplex Method
Simplex metthod is an effecient algebraic nethod to solve a L.PP by
procedure and hence an algorithm can be evolved called the simplex algorithm. systematic
hthis method it is necessary that allthe constraints in the inequality fornm is converted
Mo cquality form thus arriving at a system of algebraic equations and we are tamiliar
Wih the vairoustypes of solution of a system of algebraic equations.
|Refer lUnit-VIl in Volume 1book |
Ithe constraint isinvolved with& we add anon zero variables, (say)2 0totheLHS
to make it
an equality and the same varlable is called slak vuriable.
x:7+4 <122 7+4**, - 12, being 1|
Mthe constraint tis involved with weNubtract anonzero variable 8,(sy)2 0 nthe
LMS to make it an equality and the:same varlable is calleodsurpluserable
H,
ntandardfom of L.P!
hall constrainta called a
being equalittesiNNubject to the constrants , mnaximize
v
maximize 7 - ax hy L ito
standard torm of
VA,z
nubject to :
0, v0, he